Taint: reject tainted list-separator change

[exim.git] / doc / doc-docbook / spec.xfpt

. /////////////////////////////////////////////////////////////////////////////
. This is the primary source of the Exim Manual. It is an xfpt document that is
. converted into DocBook XML for subsequent conversion into printable and online
. formats. The markup used herein is "standard" xfpt markup, with some extras.
. The markup is summarized in a file called Markup.txt.
.
. WARNING: When you use the .new macro, make sure it appears *before* any
. adjacent index items; otherwise you get an empty "paragraph" which causes
. unwanted vertical space.
. /////////////////////////////////////////////////////////////////////////////

.include stdflags
.include stdmacs

. /////////////////////////////////////////////////////////////////////////////
. This outputs the standard DocBook boilerplate.
. /////////////////////////////////////////////////////////////////////////////

.docbook

. /////////////////////////////////////////////////////////////////////////////
. These lines are processing instructions for the Simple DocBook Processor that
. Philip Hazel has developed as a less cumbersome way of making PostScript and
. PDFs than using xmlto and fop. They will be ignored by all other XML
. processors.
. /////////////////////////////////////////////////////////////////////////////

.literal xml
<?sdop
  foot_right_recto="&chaptertitle; (&chapternumber;)"
  foot_right_verso="&chaptertitle; (&chapternumber;)"
  toc_chapter_blanks="yes,yes"
  table_warn_overflow="overprint"
?>
.literal off

. /////////////////////////////////////////////////////////////////////////////
. This generates the outermost <book> element that wraps the entire document.
. /////////////////////////////////////////////////////////////////////////////

.book

. /////////////////////////////////////////////////////////////////////////////
. These definitions set some parameters and save some typing.
. Update the Copyright year (only) when changing content.
. /////////////////////////////////////////////////////////////////////////////

.set previousversion "4.98"
.include ./local_params

.set ACL "access control lists (ACLs)"
.set I   "&nbsp;&nbsp;&nbsp;&nbsp;"

.set drivernamemax "64"

.macro copyyear
2024
.endmacro

. /////////////////////////////////////////////////////////////////////////////
. Additional xfpt markup used by this document, over and above the default
. provided in the xfpt library.
. /////////////////////////////////////////////////////////////////////////////

. --- Override the &$ flag to automatically insert a $ with the variable name.

.flag &$  $&   "<varname>$"  "</varname>"

. --- Short flags for daggers in option headings. They will always be inside
. --- an italic string, but we want the daggers to be in Roman.

.flag &!!      "</emphasis>&dagger;<emphasis>"
.flag &!?      "</emphasis>&Dagger;<emphasis>"

. --- A macro for an Exim option definition heading, generating a one-line
. --- table with four columns. For cases when the option name is given with
. --- a space, so that it can be split, a fifth argument is used for the
. --- index entry.
. --- Also one for multiple option def headings be grouped in a single
. --- table (but without the split capability).

.macro otable
.itable all 0 0 4 8* left 6* center 6* center 6* right
.endmacro

.macro orow
.row "&%$1%&" "Use: &'$2'&" "Type: &'$3'&" "Default: &'$4'&"
.endmacro

.macro option
.arg 5
.oindex "&%$5%&"
.endarg
.arg -5
.oindex "&%$1%&"
.endarg
.otable
.orow "$1" "$2" "$3" "$4"
.endtable
.endmacro

.macro options
.eacharg
.oindex "&%$+1%&"
.endeach 4
.otable
.eacharg
.orow "$+1" "$+2" "$+3" "$+4"
.endeach 4
.endtable
.endmacro

. --- A macro for the common 2-column tables. The width of the first column
. --- is suitable for the many tables at the start of the main options chapter;
. --- a small number of other 2-column tables override it.

.macro table2 196pt 254pt
.itable none 0 0 2 $1 left $2 left
.endmacro


. --- A macro for a plain variable, including the .vitem and .vindex
.macro var
.vitem $1
.vindex $1
.endmacro

. --- A macro for a "tainted" marker, done as a one-element table
.macro tmark
.itable none 0 0 1 10pt left
.row &'Tainted'&
.endtable
.endmacro

. --- A macro for a tainted variable, adding a taint-marker
.macro tvar
.var $1
.tmark
.endmacro

. --- A macro for a cmdline option, including a .oindex
. --- 1st arg is the option name, undecorated (we do that here).
. --- 2nd arg, optional, text (decorated as needed) to be appended to the name
.macro cmdopt
.vitem  &%$1%&$=2+&~$2+
.oindex &%$1%&
.endmacro

. --- A macro that generates .row, but puts &I; at the start of the first
. --- argument, thus indenting it. Assume a minimum of two arguments, and
. --- allow up to four arguments, which is as many as we'll ever need.

.macro irow
.arg 4
.row "&I;$1" "$2" "$3" "$4"
.endarg
.arg -4
.arg 3
.row "&I;$1" "$2" "$3"
.endarg
.arg -3
.row "&I;$1" "$2"
.endarg
.endarg
.endmacro

. --- Macros for option, variable, and concept index entries. For a "range"
. --- style of entry, use .scindex for the start and .ecindex for the end. The
. --- first argument of .scindex and the only argument of .ecindex must be the
. --- ID that ties them together.
. --- The index entry points to the most-recent chapter head, section or subsection
. --- head, or list-item.

.macro cindex
&<indexterm role="concept">&
&<primary>&$1&</primary>&
.arg 2
&<secondary>&$2&</secondary>&
.endarg
&</indexterm>&
.endmacro

.macro scindex
&<indexterm role="concept" id="$1" class="startofrange">&
&<primary>&$2&</primary>&
.arg 3
&<secondary>&$3&</secondary>&
.endarg
&</indexterm>&
.endmacro

.macro ecindex
&<indexterm role="concept" startref="$1" class="endofrange"/>&
.endmacro

.macro oindex
&<indexterm role="option">&
&<primary>&$1&</primary>&
.arg 2
&<secondary>&$2&</secondary>&
.endarg
&</indexterm>&
.endmacro

. --- The index entry points to the most-recent chapter head, section or subsection
. --- head, or varlist item.

.macro vindex
&<indexterm role="variable">&
&<primary>&$1&</primary>&
.arg 2
&<secondary>&$2&</secondary>&
.endarg
&</indexterm>&
.endmacro

.macro index
.echo "** Don't use .index; use .cindex or .oindex or .vindex"
.endmacro


. use this for a concept-index entry for a header line
.macro chindex
.cindex "&'$1'& header line"
.cindex "header lines" $1
.endmacro
. ////////////////////////////////////////////////////////////////////////////


. ////////////////////////////////////////////////////////////////////////////
. The <bookinfo> element is removed from the XML before processing for ASCII
. output formats.
. ////////////////////////////////////////////////////////////////////////////

.literal xml
<bookinfo>
<title>Specification of the Exim Mail Transfer Agent</title>
<titleabbrev>The Exim MTA</titleabbrev>
<date>
.fulldate
</date>
<author><firstname>Exim</firstname><surname>Maintainers</surname></author>
<authorinitials>EM</authorinitials>
<revhistory><revision>
.versiondatexml
  <authorinitials>EM</authorinitials>
</revision></revhistory>
<copyright><year>
.copyyear
           </year><holder>The Exim Maintainers</holder></copyright>
</bookinfo>
.literal off


. /////////////////////////////////////////////////////////////////////////////
. These implement index entries of the form "x, see y" and "x, see also y".
. However, the DocBook DTD doesn't allow <indexterm> entries
. at the top level, so we have to put the .chapter directive first.
. /////////////////////////////////////////////////////////////////////////////

.chapter "Introduction" "CHID1"

.macro seeother
.literal xml
<indexterm role="$2">
  <primary>$3</primary>
.arg 5
  <secondary>$5</secondary>
.endarg
  <$1><emphasis>$4</emphasis></$1>
</indexterm>
.literal off
.endmacro

. NB: for the 4-arg variant the ordering is awkward
.macro see
.seeother see "$1" "$2" "$3" "$4"
.endmacro
.macro seealso
.seeother seealso "$1" "$2" "$3" "$4"
.endmacro

.see     variable "<emphasis>$1</emphasis>, <emphasis>$2</emphasis>, etc." "numerical variables"
.see     concept address        rewriting                       rewriting
.see     concept "Bounce Address Tag Validation"        BATV
.see     concept "Client SMTP Authorization"            CSA
.see     concept "CR character" "carriage return"
.see     concept CRL            "certificate revocation list"
.seealso concept de-tainting    "tainted data"
.see     concept delivery       "bounce message"                "failure report"
.see     concept dialup         "intermittently connected hosts"
.see     concept exiscan        "content scanning"
.see     concept fallover       fallback
.see     concept filter         "Sieve filter"                  Sieve
.see     concept headers        "header lines"
.see     concept ident          "RFC 1413"
.see     concept "LF character" "linefeed"
.seealso concept maximum        limit
.see     concept monitor        "Exim monitor"
.see     concept "no_<emphasis>xxx</emphasis>"          "entry for xxx"
.see     concept NUL            "binary zero"
.see     concept "passwd file"  "/etc/passwd"
.see     concept "process id"   pid
.see     concept RBL            "DNS list"
.see     concept redirection    "address redirection"
.see     concept "return path"  "envelope sender"
.see     concept scanning       "content scanning"
.see     concept SSL            TLS
.see     concept string         expansion expansion
.see     concept "top bit"      "8-bit characters"
.see     concept variables      "expansion, variables"
.see     concept "zero, binary" "binary zero"


. /////////////////////////////////////////////////////////////////////////////
. This is the real start of the first chapter. See the comment above as to why
. we can't have the .chapter line here.
. chapter "Introduction"
. /////////////////////////////////////////////////////////////////////////////

Exim is a mail transfer agent (MTA) for hosts that are running Unix or
Unix-like operating systems. It was designed on the assumption that it would be
run on hosts that are permanently connected to the Internet. However, it can be
used on intermittently connected hosts with suitable configuration adjustments.

Configuration files currently exist for the following operating systems: AIX,
BSD/OS (aka BSDI), Darwin (Mac OS X), DGUX, Dragonfly, FreeBSD, GNU/Hurd,
GNU/Linux, HI-OSF (Hitachi), HI-UX, HP-UX, IRIX, MIPS RISCOS, NetBSD, OpenBSD,
OpenUNIX, QNX, SCO, SCO SVR4.2 (aka UNIX-SV), Solaris (aka SunOS5), SunOS4,
Tru64-Unix (formerly Digital UNIX, formerly DEC-OSF1), Ultrix, and UnixWare.
Some of these operating systems are no longer current and cannot easily be
tested, so the configuration files may no longer work in practice.

There are also configuration files for compiling Exim in the Cygwin environment
that can be installed on systems running Windows. However, this document does
not contain any information about running Exim in the Cygwin environment.

The terms and conditions for the use and distribution of Exim are contained in
the file &_NOTICE_&. Exim is distributed under the terms of the GNU General
Public Licence, a copy of which may be found in the file &_LICENCE_&.

The use, supply, or promotion of Exim for the purpose of sending bulk,
unsolicited electronic mail is incompatible with the basic aims of Exim,
which revolve around the free provision of a service that enhances the quality
of personal communications. The author of Exim regards indiscriminate
mass-mailing as an antisocial, irresponsible abuse of the Internet.

Exim owes a great deal to Smail 3 and its author, Ron Karr. Without the
experience of running and working on the Smail 3 code, I could never have
contemplated starting to write a new MTA. Many of the ideas and user interfaces
were originally taken from Smail 3, though the actual code of Exim is entirely
new, and has developed far beyond the initial concept.

Many people, both in Cambridge and around the world, have contributed to the
development and the testing of Exim, and to porting it to various operating
systems. I am grateful to them all. The distribution now contains a file called
&_ACKNOWLEDGMENTS_&, in which I have started recording the names of
contributors.


.section "Exim documentation" "SECID1"
. Keep this example change bar when updating the documentation!

.new
.cindex "documentation"
This edition of the Exim specification applies to version &version() of Exim.
Substantive changes from the &previousversion; edition are marked in some
renditions of this document; this paragraph is so marked if the rendition is
capable of showing a change indicator.
.wen

This document is very much a reference manual; it is not a tutorial. The reader
is expected to have some familiarity with the SMTP mail transfer protocol and
with general Unix system administration. Although there are some discussions
and examples in places, the information is mostly organized in a way that makes
it easy to look up, rather than in a natural order for sequential reading.
Furthermore, this manual aims to cover every aspect of Exim in detail, including
a number of rarely-used, special-purpose features that are unlikely to be of
very wide interest.

.cindex "books about Exim"
An &"easier"& discussion of Exim which provides more in-depth explanatory,
introductory, and tutorial material can be found in a book entitled &'The Exim
SMTP Mail Server'& (second edition, 2007), published by UIT Cambridge
(&url(https://www.uit.co.uk/exim-book/)).

The book also contains a chapter that gives a general introduction to SMTP and
Internet mail. Inevitably, however, the book is unlikely to be fully up-to-date
with the latest release of Exim. (Note that the earlier book about Exim,
published by O'Reilly, covers Exim 3, and many things have changed in Exim 4.)

.cindex "Debian" "information sources"
If you are using a Debian distribution of Exim, you will find information about
Debian-specific features in the file
&_/usr/share/doc/exim4-base/README.Debian_&.
The command &(man update-exim.conf)& is another source of Debian-specific
information.

.cindex "&_doc/NewStuff_&"
.cindex "&_doc/ChangeLog_&"
.cindex "change log"
As Exim develops, there may be features in newer versions that have not
yet made it into this document, which is updated only when the most significant
digit of the fractional part of the version number changes. Specifications of
new features that are not yet in this manual are placed in the file
&_doc/NewStuff_& in the Exim distribution.

Some features may be classified as &"experimental"&. These may change
incompatibly while they are developing, or even be withdrawn. For this reason,
they are not documented in this manual. Information about experimental features
can be found in the file &_doc/experimental.txt_&.

All changes to Exim (whether new features, bug fixes, or other kinds of
change) are noted briefly in the file called &_doc/ChangeLog_&.

.cindex "&_doc/spec.txt_&"
This specification itself is available as an ASCII file in &_doc/spec.txt_& so
that it can easily be searched with a text editor. Other files in the &_doc_&
directory are:

.table2 100pt
.row &_OptionLists.txt_&     "list of all options in alphabetical order"
.row &_dbm.discuss.txt_&     "discussion about DBM libraries"
.row &_exim.8_&              "a man page of Exim's command line options"
.row &_experimental.txt_&    "documentation of experimental features"
.row &_filter.txt_&          "specification of the filter language"
.row &_Exim3.upgrade_&       "upgrade notes from release 2 to release 3"
.row &_Exim4.upgrade_&       "upgrade notes from release 3 to release 4"
.row &_openssl.txt_&         "installing a current OpenSSL release"
.endtable

The main specification and the specification of the filtering language are also
available in other formats (HTML, PostScript, PDF, and Texinfo). Section
&<<SECTavail>>& below tells you how to get hold of these.



.section "FTP site and websites" "SECID2"
.cindex "website"
.cindex "FTP site"
The primary site for Exim source distributions is the &%exim.org%& FTP site,
available over HTTPS, HTTP and FTP.  These services, and the &%exim.org%&
website, are hosted at the University of Cambridge.

.cindex "wiki"
.cindex "FAQ"
As well as Exim distribution tar files, the Exim website contains a number of
differently formatted versions of the documentation. A recent addition to the
online information is &url(https://wiki.exim.org,the Exim wiki),
which contains what used to be a separate FAQ, as well as various other
examples, tips, and know-how that have been contributed by Exim users.
The wiki site should always redirect to the correct place, which is currently
provided by GitHub, and is open to editing by anyone with a GitHub account.

.cindex Bugzilla
An Exim Bugzilla exists at &url(https://bugs.exim.org). You can use
this to report bugs, and also to add items to the wish list. Please search
first to check that you are not duplicating a previous entry.
Please do not ask for configuration help in the bug-tracker.


.section "Mailing lists" "SECID3"
.cindex "mailing lists" "for Exim users"
The following Exim mailing lists exist:

.table2 140pt
.row &'exim-announce@lists.exim.org'&   "Moderated, low volume announcements list"
.row &'exim-users@lists.exim.org'&      "General discussion list"
.row &'exim-users-de@lists.exim.org'&   "General discussion list in German language"
.row &'exim-dev@lists.exim.org'&        "Discussion of bugs, enhancements, etc."
.row &'exim-cvs@lists.exim.org'&        "Automated commit messages from the VCS"
.endtable

You can subscribe to these lists, change your existing subscriptions, and view
or search the archives via the mailing lists link on the Exim home page.
.cindex "Debian" "mailing list for"
If you are using a Debian distribution of Exim, you may wish to subscribe to
the Debian-specific mailing list &'pkg-exim4-users@lists.alioth.debian.org'&
via this web page:
.display
&url(https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/pkg-exim4-users)
.endd
Please ask Debian-specific questions on that list and not on the general Exim
lists.

.section "Bug reports" "SECID5"
.cindex "bug reports"
.cindex "reporting bugs"
Reports of obvious bugs can be emailed to &'bugs@exim.org'& or reported
via &url(https://bugs.exim.org,the Bugzilla). However, if you are unsure
whether some behaviour is a bug or not, the best thing to do is to post a
message to the &'exim-dev'& mailing list and have it discussed.



.section "Where to find the Exim distribution" "SECTavail"
.cindex "FTP site"
.cindex "HTTPS download site"
.cindex "distribution" "FTP site"
.cindex "distribution" "https site"
The master distribution site for the Exim distribution is
.display
&url(https://downloads.exim.org/)
.endd
The service is available over HTTPS, HTTP and FTP.
We encourage people to migrate to HTTPS.

The content served at &url(https://downloads.exim.org/) is identical to the
content served at &url(https://ftp.exim.org/pub/exim) and
&url(ftp://ftp.exim.org/pub/exim).

If accessing via a hostname containing &'ftp'&, then the file references that
follow are relative to the &_exim_& directories at these sites.
If accessing via the hostname &'downloads'& then the subdirectories described
here are top-level directories.

There are now quite a number of independent mirror sites around
the world. Those that I know about are listed in the file called &_Mirrors_&.

Within the top exim directory there are subdirectories called &_exim3_& (for
previous Exim 3 distributions), &_exim4_& (for the latest Exim 4
distributions), and &_Testing_& for testing versions. In the &_exim4_&
subdirectory, the current release can always be found in files called
.display
&_exim-n.nn.tar.xz_&
&_exim-n.nn.tar.gz_&
&_exim-n.nn.tar.bz2_&
.endd
where &'n.nn'& is the highest such version number in the directory. The three
files contain identical data; the only difference is the type of compression.
The &_.xz_& file is usually the smallest, while the &_.gz_& file is the
most portable to old systems.

.cindex "distribution" "signing details"
.cindex "distribution" "public key"
.cindex "public key for signed distribution"
The distributions will be PGP signed by an individual key of the Release
Coordinator.  This key will have a uid containing an email address in the
&'exim.org'& domain and will have signatures from other people, including
other Exim maintainers.  We expect that the key will be in the "strong set" of
PGP keys.  There should be a trust path to that key from the Exim Maintainer's
PGP keys, a version of which can be found in the release directory in the file
&_Exim-Maintainers-Keyring.asc_&.  All keys used will be available in public keyserver pools,
such as &'pool.sks-keyservers.net'&.

At the time of the last update, releases were being made by Jeremy Harris and signed
with key &'0xBCE58C8CE41F32DF'&.  Other recent keys used for signing are those
of Heiko Schlittermann, &'0x26101B62F69376CE'&,
and of Phil Pennock, &'0x4D1E900E14C1CC04'&.

The signatures for the tar bundles are in:
.display
&_exim-n.nn.tar.xz.asc_&
&_exim-n.nn.tar.gz.asc_&
&_exim-n.nn.tar.bz2.asc_&
.endd
For each released version, the log of changes is made available in a
separate file in the directory &_ChangeLogs_& so that it is possible to
find out what has changed without having to download the entire distribution.

.cindex "documentation" "available formats"
The main distribution contains ASCII versions of this specification and other
documentation; other formats of the documents are available in separate files
inside the &_exim4_& directory of the FTP site:
.display
&_exim-html-n.nn.tar.gz_&
&_exim-pdf-n.nn.tar.gz_&
&_exim-postscript-n.nn.tar.gz_&
&_exim-texinfo-n.nn.tar.gz_&
.endd
These tar files contain only the &_doc_& directory, not the complete
distribution, and are also available in &_.bz2_& and &_.xz_& forms.


.section "Limitations" "SECID6"
.ilist
.cindex "limitations of Exim"
.cindex "bang paths" "not handled by Exim"
Exim is designed for use as an Internet MTA, and therefore handles addresses in
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
domain format only. It cannot handle UUCP &"bang paths"&, though
simple two-component bang paths can be converted by a straightforward rewriting
configuration. This restriction does not prevent Exim from being interfaced to
UUCP as a transport mechanism, provided that domain addresses are used.
.next
.cindex "domainless addresses"
.cindex "address" "without domain"
Exim insists that every address it handles has a domain attached. For incoming
local messages, domainless addresses are automatically qualified with a
configured domain value. Configuration options specify from which remote
systems unqualified addresses are acceptable. These are then qualified on
arrival.
.next
.cindex "transport" "external"
.cindex "external transports"
The only external transport mechanisms that are currently implemented are SMTP
and LMTP over a TCP/IP network (including support for IPv6). However, a pipe
transport is available, and there are facilities for writing messages to files
and pipes, optionally in &'batched SMTP'& format; these facilities can be used
to send messages to other transport mechanisms such as UUCP, provided they can
handle domain-style addresses. Batched SMTP input is also catered for.
.next
Exim is not designed for storing mail for dial-in hosts. When the volumes of
such mail are large, it is better to get the messages &"delivered"& into files
(that is, off Exim's queue) and subsequently passed on to the dial-in hosts by
other means.
.next
Although Exim does have basic facilities for scanning incoming messages, these
are not comprehensive enough to do full virus or spam scanning. Such operations
are best carried out using additional specialized software packages. If you
compile Exim with the content-scanning extension, straightforward interfaces to
a number of common scanners are provided.
.endlist


.section "Runtime configuration" "SECID7"
Exim's runtime configuration is held in a single text file that is divided
into a number of sections. The entries in this file consist of keywords and
values, in the style of Smail 3 configuration files. A default configuration
file which is suitable for simple online installations is provided in the
distribution, and is described in chapter &<<CHAPdefconfil>>& below.


.section "Calling interface" "SECID8"
.cindex "Sendmail compatibility" "command line interface"
Like many MTAs, Exim has adopted the Sendmail command line interface so that it
can be a straight replacement for &_/usr/lib/sendmail_& or
&_/usr/sbin/sendmail_& when sending mail, but you do not need to know anything
about Sendmail in order to run Exim. For actions other than sending messages,
Sendmail-compatible options also exist, but those that produce output (for
example, &%-bp%&, which lists the messages in the queue) do so in Exim's own
format. There are also some additional options that are compatible with Smail
3, and some further options that are new to Exim. Chapter &<<CHAPcommandline>>&
documents all Exim's command line options. This information is automatically
made into the man page that forms part of the Exim distribution.

Control of messages in the queue can be done via certain privileged command
line options. There is also an optional monitor program called &'eximon'&,
which displays current information in an X window, and which contains a menu
interface to Exim's command line administration options.



.section "Terminology" "SECID9"
.cindex "terminology definitions"
.cindex "body of message" "definition of"
The &'body'& of a message is the actual data that the sender wants to transmit.
It is the last part of a message and is separated from the &'header'& (see
below) by a blank line.

.cindex "bounce message" "definition of"
When a message cannot be delivered, it is normally returned to the sender in a
delivery failure message or a &"non-delivery report"& (NDR). The term
&'bounce'& is commonly used for this action, and the error reports are often
called &'bounce messages'&. This is a convenient shorthand for &"delivery
failure error report"&. Such messages have an empty sender address in the
message's &'envelope'& (see below) to ensure that they cannot themselves give
rise to further bounce messages.

The term &'default'& appears frequently in this manual. It is used to qualify a
value which is used in the absence of any setting in the configuration. It may
also qualify an action which is taken unless a configuration setting specifies
otherwise.

The term &'defer'& is used when the delivery of a message to a specific
destination cannot immediately take place for some reason (a remote host may be
down, or a user's local mailbox may be full). Such deliveries are &'deferred'&
until a later time.

The word &'domain'& is sometimes used to mean all but the first component of a
host's name. It is &'not'& used in that sense here, where it normally refers to
the part of an email address following the @ sign.

.cindex "envelope, definition of"
.cindex "sender" "definition of"
A message in transit has an associated &'envelope'&, as well as a header and a
body. The envelope contains a sender address (to which bounce messages should
be delivered), and any number of recipient addresses. References to the
sender or the recipients of a message usually mean the addresses in the
envelope. An MTA uses these addresses for delivery, and for returning bounce
messages, not the addresses that appear in the header lines.

.cindex "message" "header, definition of"
.cindex "header section" "definition of"
The &'header'& of a message is the first part of a message's text, consisting
of a number of lines, each of which has a name such as &'From:'&, &'To:'&,
&'Subject:'&, etc. Long header lines can be split over several text lines by
indenting the continuations. The header is separated from the body by a blank
line.

.cindex "local part" "definition of"
.cindex "domain" "definition of"
The term &'local part'&, which is taken from
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822), is used to refer to the
part of an email address that precedes the @ sign. The part that follows the
@ sign is called the &'domain'& or &'mail domain'&.

.cindex "local delivery" "definition of"
.cindex "remote delivery, definition of"
The terms &'local delivery'& and &'remote delivery'& are used to distinguish
delivery to a file or a pipe on the local host from delivery by SMTP over
TCP/IP to another host. As far as Exim is concerned, all hosts other than the
host it is running on are &'remote'&.

.cindex "return path" "definition of"
&'Return path'& is another name that is used for the sender address in a
message's envelope.

.cindex "queue" "definition of"
The term &'queue'& is used to refer to the set of messages awaiting delivery
because this term is in widespread use in the context of MTAs. However, in
Exim's case, the reality is more like a pool than a queue, because there is
normally no ordering of waiting messages.

.cindex "queue runner" "definition of"
The term &'queue runner'& is used to describe a process that scans the queue
and attempts to deliver those messages whose retry times have come. This term
is used by other MTAs and also relates to the command &%runq%&, but in Exim
the waiting messages are normally processed in an unpredictable order.

.cindex "spool directory" "definition of"
The term &'spool directory'& is used for a directory in which Exim keeps the
messages in its queue &-- that is, those that it is in the process of
delivering. This should not be confused with the directory in which local
mailboxes are stored, which is called a &"spool directory"& by some people. In
the Exim documentation, &"spool"& is always used in the first sense.






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Incorporated code" "CHID2"
.cindex "incorporated code"
.cindex "regular expressions" "library"
.cindex "PCRE2"
.cindex "OpenDMARC"
A number of pieces of external code are included in the Exim distribution.

.ilist
Regular expressions are supported in the main Exim program and in the
Exim monitor using the freely-distributable PCRE2 library, copyright
&copy; University of Cambridge. The source to PCRE2 is not longer shipped with
Exim, so you will need to use the version of PCRE2 shipped with your system,
or obtain and install the full version of the library from
&url(https://github.com/PhilipHazel/pcre2/releases).
.next
.cindex "cdb" "acknowledgment"
Support for the cdb (Constant DataBase) lookup method is provided by code
contributed by Nigel Metheringham of (at the time he contributed it) Planet
Online Ltd. The implementation is completely contained within the code of Exim.
It does not link against an external cdb library. The code contains the
following statements:

.blockquote
Copyright &copy; 1998 Nigel Metheringham, Planet Online Ltd

This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation; either version 2 of the License, or (at your option) any later
version.
This code implements Dan Bernstein's Constant DataBase (cdb) spec. Information,
the spec and sample code for cdb can be obtained from
&url(https://cr.yp.to/cdb.html). This implementation borrows
some code from Dan Bernstein's implementation (which has no license
restrictions applied to it).
.endblockquote
.next
.cindex "SPA authentication"
.cindex "Samba project"
.cindex "Microsoft Secure Password Authentication"
Client support for Microsoft's &'Secure Password Authentication'& is provided
by code contributed by Marc Prud'hommeaux. Server support was contributed by
Tom Kistner. This includes code taken from the Samba project, which is released
under the Gnu GPL.
.next
.cindex "Cyrus"
.cindex "&'pwcheck'& daemon"
.cindex "&'pwauthd'& daemon"
Support for calling the Cyrus &'pwcheck'& and &'saslauthd'& daemons is provided
by code taken from the Cyrus-SASL library and adapted by Alexander S.
Sabourenkov. The permission notice appears below, in accordance with the
conditions expressed therein.

.blockquote
Copyright &copy; 2001 Carnegie Mellon University.  All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:

.olist
Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
.next
Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in
the documentation and/or other materials provided with the
distribution.
.next
The name &"Carnegie Mellon University"& must not be used to
endorse or promote products derived from this software without
prior written permission. For permission or any other legal
details, please contact
.display
              Office of Technology Transfer
              Carnegie Mellon University
              5000 Forbes Avenue
              Pittsburgh, PA  15213-3890
              (412) 268-4387, fax: (412) 268-7395
              tech-transfer@andrew.cmu.edu
.endd
.next
Redistributions of any form whatsoever must retain the following
acknowledgment:

&"This product includes software developed by Computing Services
at Carnegie Mellon University (&url(https://www.cmu.edu/computing/)."&

CARNEGIE MELLON UNIVERSITY DISCLAIMS ALL WARRANTIES WITH REGARD TO
THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY
AND FITNESS, IN NO EVENT SHALL CARNEGIE MELLON UNIVERSITY BE LIABLE
FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN
AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING
OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
.endlist
.endblockquote

.next
.cindex "Exim monitor" "acknowledgment"
.cindex "X-windows"
.cindex "Athena"
The Exim Monitor program, which is an X-Window application, includes
modified versions of the Athena StripChart and TextPop widgets.
This code is copyright by DEC and MIT, and their permission notice appears
below, in accordance with the conditions expressed therein.

.blockquote
Copyright 1987, 1988 by Digital Equipment Corporation, Maynard, Massachusetts,
and the Massachusetts Institute of Technology, Cambridge, Massachusetts.

All Rights Reserved

Permission to use, copy, modify, and distribute this software and its
documentation for any purpose and without fee is hereby granted,
provided that the above copyright notice appear in all copies and that
both that copyright notice and this permission notice appear in
supporting documentation, and that the names of Digital or MIT not be
used in advertising or publicity pertaining to distribution of the
software without specific, written prior permission.

DIGITAL DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE, INCLUDING
ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO EVENT SHALL
DIGITAL BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR
ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS
SOFTWARE.
.endblockquote

.next
.cindex "opendmarc" "acknowledgment"
The DMARC implementation uses the OpenDMARC library which is Copyrighted by
The Trusted Domain Project. Portions of Exim source which use OpenDMARC
derived code are indicated in the respective source files. The full OpenDMARC
license is provided in the LICENSE.opendmarc file contained in the distributed
source code.

.next
Many people have contributed code fragments, some large, some small, that were
not covered by any specific license requirements. It is assumed that the
contributors are happy to see their code incorporated into Exim under the GPL.
.endlist





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "How Exim receives and delivers mail" "CHID11" &&&
         "Receiving and delivering mail"


.section "Overall philosophy" "SECID10"
.cindex "design philosophy"
Exim is designed to work efficiently on systems that are permanently connected
to the Internet and are handling a general mix of mail. In such circumstances,
most messages can be delivered immediately. Consequently, Exim does not
maintain independent queues of messages for specific domains or hosts, though
it does try to send several messages in a single SMTP connection after a host
has been down, and it also maintains per-host retry information.


.section "Policy control" "SECID11"
.cindex "policy control" "overview"
Policy controls are now an important feature of MTAs that are connected to the
Internet. Perhaps their most important job is to stop MTAs from being abused as
&"open relays"& by misguided individuals who send out vast amounts of
unsolicited junk and want to disguise its source. Exim provides flexible
facilities for specifying policy controls on incoming mail:

.ilist
.cindex "&ACL;" "introduction"
Exim 4 (unlike previous versions of Exim) implements policy controls on
incoming mail by means of &'Access Control Lists'& (ACLs). Each list is a
series of statements that may either grant or deny access. ACLs can be used at
several places in the SMTP dialogue while receiving a message from a remote
host. However, the most common places are after each RCPT command, and at the
very end of the message. The sysadmin can specify conditions for accepting or
rejecting individual recipients or the entire message, respectively, at these
two points (see chapter &<<CHAPACL>>&). Denial of access results in an SMTP
error code.
.next
An ACL is also available for locally generated, non-SMTP messages. In this
case, the only available actions are to accept or deny the entire message.
.next
When Exim is compiled with the content-scanning extension, facilities are
provided in the ACL mechanism for passing the message to external virus and/or
spam scanning software. The result of such a scan is passed back to the ACL,
which can then use it to decide what to do with the message.
.next
When a message has been received, either from a remote host or from the local
host, but before the final acknowledgment has been sent, a locally supplied C
function called &[local_scan()]& can be run to inspect the message and decide
whether to accept it or not (see chapter &<<CHAPlocalscan>>&). If the message
is accepted, the list of recipients can be modified by the function.
.next
Using the &[local_scan()]& mechanism is another way of calling external scanner
software. The &%SA-Exim%& add-on package works this way. It does not require
Exim to be compiled with the content-scanning extension.
.next
After a message has been accepted, a further checking mechanism is available in
the form of the &'system filter'& (see chapter &<<CHAPsystemfilter>>&). This
runs at the start of every delivery process.
.endlist



.section "User filters" "SECID12"
.cindex "filter" "introduction"
.cindex "Sieve filter"
In a conventional Exim configuration, users are able to run private filters by
setting up appropriate &_.forward_& files in their home directories. See
chapter &<<CHAPredirect>>& (about the &(redirect)& router) for the
configuration needed to support this, and the separate document entitled
&'Exim's interfaces to mail filtering'& for user details. Two different kinds
of filtering are available:

.ilist
Sieve filters are written in the standard filtering language that is defined
by &url(https://www.rfc-editor.org/rfc/rfc3028.html,RFC 3028).
.next
Exim filters are written in a syntax that is unique to Exim, but which is more
powerful than Sieve, which it pre-dates.
.endlist

User filters are run as part of the routing process, described below.



.section "Message identification" "SECTmessiden"
.cindex "message ids" "details of format"
.cindex "format" "of message id"
.cindex "id of message"
.cindex "base62"
.cindex "base36"
.cindex "Darwin"
.cindex "Cygwin"
.cindex "exim_msgdate"
Every message handled by Exim is given a &'message id'& which is 23
characters long. It is divided into three parts, separated by hyphens, for
example &`16VDhn-000000001bo-D342`&. Each part is a sequence of letters and digits,
normally encoding numbers in base 62. However, in the Darwin operating
system (Mac OS X) and when Exim is compiled to run under Cygwin, base 36
(avoiding the use of lower case letters) is used instead, because the message
id is used to construct filenames, and the names of files in those systems are
not always case-sensitive.

.cindex "pid (process id)" "re-use of"
The detail of the contents of the message id have changed as Exim has evolved.
Earlier versions relied on the operating system not re-using a process id (pid)
within one second. On modern operating systems, this assumption can no longer
be made, so the algorithm had to be changed. To retain backward compatibility,
the format of the message id was retained, which is why the following rules are
somewhat eccentric:

.ilist
The first six characters of the message id are the time at which the message
started to be received, to a granularity of one second. That is, this field
contains the number of seconds since the start of the epoch (the normal Unix
way of representing the date and time of day).
.next
After the first hyphen, the next
eleven
characters are the id of the process that received the message.
.next
There are two different possibilities for the final four characters:
.olist
.oindex "&%localhost_number%&"
If &%localhost_number%& is not set, this value is the fractional part of the
time of reception, normally in units of
microseconds.
but for systems
that must use base 36 instead of base 62 (because of case-insensitive file
systems), the units are
2 us.
.next
If &%localhost_number%& is set, it is multiplied by
500000 (250000) and added to
the fractional part of the time, which in this case is in units of 2 us (4 us).
.endlist
.endlist

After a message has been received, Exim waits for the clock to tick at the
appropriate resolution before proceeding, so that if another message is
received by the same process, or by another process with the same (re-used)
pid, it is guaranteed that the time will be different. In most cases, the clock
will already have ticked while the message was being received.

The exim_msgdate utility (see section &<<SECTexim_msgdate>>&) can be
used to display the date, and optionally the process id, of an Exim
Message ID.


.section "Receiving mail" "SECID13"
.cindex "receiving mail"
.cindex "message" "reception"
The only way Exim can receive mail from another host is using SMTP over
TCP/IP, in which case the sender and recipient addresses are transferred using
SMTP commands. However, from a locally running process (such as a user's MUA),
there are several possibilities:

.ilist
If the process runs Exim with the &%-bm%& option, the message is read
non-interactively (usually via a pipe), with the recipients taken from the
command line, or from the body of the message if &%-t%& is also used.
.next
If the process runs Exim with the &%-bS%& option, the message is also read
non-interactively, but in this case the recipients are listed at the start of
the message in a series of SMTP RCPT commands, terminated by a DATA
command. This is called &"batch SMTP"& format,
but it isn't really SMTP. The SMTP commands are just another way of passing
envelope addresses in a non-interactive submission.
.next
If the process runs Exim with the &%-bs%& option, the message is read
interactively, using the SMTP protocol. A two-way pipe is normally used for
passing data between the local process and the Exim process.
This is &"real"& SMTP and is handled in the same way as SMTP over TCP/IP. For
example, the ACLs for SMTP commands are used for this form of submission.
.next
A local process may also make a TCP/IP call to the host's loopback address
(127.0.0.1) or any other of its IP addresses. When receiving messages, Exim
does not treat the loopback address specially. It treats all such connections
in the same way as connections from other hosts.
.endlist


.cindex "message sender, constructed by Exim"
.cindex "sender" "constructed by Exim"
In the three cases that do not involve TCP/IP, the sender address is
constructed from the login name of the user that called Exim and a default
qualification domain (which can be set by the &%qualify_domain%& configuration
option). For local or batch SMTP, a sender address that is passed using the
SMTP MAIL command is ignored. However, the system administrator may allow
certain users (&"trusted users"&) to specify a different sender addresses
unconditionally, or all users to specify certain forms of different sender
address. The &%-f%& option or the SMTP MAIL command is used to specify these
different addresses. See section &<<SECTtrustedadmin>>& for details of trusted
users, and the &%untrusted_set_sender%& option for a way of allowing untrusted
users to change sender addresses.

Messages received by either of the non-interactive mechanisms are subject to
checking by the non-SMTP ACL if one is defined. Messages received using SMTP
(either over TCP/IP or interacting with a local process) can be checked by a
number of ACLs that operate at different times during the SMTP session. Either
individual recipients or the entire message can be rejected if local policy
requirements are not met. The &[local_scan()]& function (see chapter
&<<CHAPlocalscan>>&) is run for all incoming messages.

Exim can be configured not to start a delivery process when a message is
received; this can be unconditional, or depend on the number of incoming SMTP
connections or the system load. In these situations, new messages wait on the
queue until a queue runner process picks them up. However, in standard
configurations under normal conditions, delivery is started as soon as a
message is received.





.section "Handling an incoming message" "SECID14"
.cindex "spool directory" "files that hold a message"
.cindex "file" "how a message is held"
When Exim accepts a message, it writes two files in its spool directory. The
first contains the envelope information, the current status of the message, and
the header lines, and the second contains the body of the message. The names of
the two spool files consist of the message id, followed by &`-H`& for the
file containing the envelope and header, and &`-D`& for the data file.

.cindex "spool directory" "&_input_& sub-directory"
By default, all these message files are held in a single directory called
&_input_& inside the general Exim spool directory. Some operating systems do
not perform very well if the number of files in a directory gets large; to
improve performance in such cases, the &%split_spool_directory%& option can be
used. This causes Exim to split up the input files into 62 sub-directories
whose names are single letters or digits. When this is done, the queue is
processed one sub-directory at a time instead of all at once, which can improve
overall performance even when there are not enough files in each directory to
affect file system performance.

The envelope information consists of the address of the message's sender and
the addresses of the recipients. This information is entirely separate from
any addresses contained in the header lines. The status of the message includes
a list of recipients who have already received the message. The format of the
first spool file is described in chapter &<<CHAPspool>>&.

.cindex "rewriting" "addresses"
Address rewriting that is specified in the rewrite section of the configuration
(see chapter &<<CHAPrewrite>>&) is done once and for all on incoming addresses,
both in the header lines and the envelope, at the time the message is accepted.
If during the course of delivery additional addresses are generated (for
example, via aliasing), these new addresses are rewritten as soon as they are
generated. At the time a message is actually delivered (transported) further
rewriting can take place; because this is a transport option, it can be
different for different forms of delivery. It is also possible to specify the
addition or removal of certain header lines at the time the message is
delivered (see chapters &<<CHAProutergeneric>>& and
&<<CHAPtransportgeneric>>&).



.section "Life of a message" "SECID15"
.cindex "message" "life of"
.cindex "message" "frozen"
A message remains in the spool directory until it is completely delivered to
its recipients or to an error address, or until it is deleted by an
administrator or by the user who originally created it. In cases when delivery
cannot proceed &-- for example when a message can neither be delivered to its
recipients nor returned to its sender, the message is marked &"frozen"& on the
spool, and no more deliveries are attempted.

.cindex "frozen messages" "thawing"
.cindex "message" "thawing frozen"
An administrator can &"thaw"& such messages when the problem has been
corrected, and can also freeze individual messages by hand if necessary. In
addition, an administrator can force a delivery error, causing a bounce message
to be sent.

.oindex "&%timeout_frozen_after%&"
.oindex "&%ignore_bounce_errors_after%&"
There are options called &%ignore_bounce_errors_after%& and
&%timeout_frozen_after%&, which discard frozen messages after a certain time.
The first applies only to frozen bounces, the second to all frozen messages.

.cindex "message" "log file for"
.cindex "log" "file for each message"
While Exim is working on a message, it writes information about each delivery
attempt to its main log file. This includes successful, unsuccessful, and
delayed deliveries for each recipient (see chapter &<<CHAPlog>>&). The log
lines are also written to a separate &'message log'& file for each message.
These logs are solely for the benefit of the administrator and are normally
deleted along with the spool files when processing of a message is complete.
The use of individual message logs can be disabled by setting
&%no_message_logs%&; this might give an improvement in performance on very busy
systems.

.cindex "journal file"
.cindex "file" "journal"
All the information Exim itself needs to set up a delivery is kept in the first
spool file, along with the header lines. When a successful delivery occurs, the
address is immediately written at the end of a journal file, whose name is the
message id followed by &`-J`&. At the end of a delivery run, if there are some
addresses left to be tried again later, the first spool file (the &`-H`& file)
is updated to indicate which these are, and the journal file is then deleted.
Updating the spool file is done by writing a new file and renaming it, to
minimize the possibility of data loss.

Should the system or Exim crash after a successful delivery but before
the spool file has been updated, the journal is left lying around. The next
time Exim attempts to deliver the message, it reads the journal file and
updates the spool file before proceeding. This minimizes the chances of double
deliveries caused by crashes.



.section "Processing an address for delivery" "SECTprocaddress"
.cindex "drivers" "definition of"
.cindex "router" "definition of"
.cindex "transport" "definition of"
The main delivery processing elements of Exim are called &'routers'& and
&'transports'&, and collectively these are known as &'drivers'&. Code for a
number of them is provided in the source distribution, and compile-time options
specify which ones are included in the binary. Runtime options specify which
ones are actually used for delivering messages.

.cindex "drivers" "instance definition"
Each driver that is specified in the runtime configuration is an &'instance'&
of that particular driver type. Multiple instances are allowed; for example,
you can set up several different &(smtp)& transports, each with different
option values that might specify different ports or different timeouts. Each
instance has its own identifying name. In what follows we will normally use the
instance name when discussing one particular instance (that is, one specific
configuration of the driver), and the generic driver name when discussing
the driver's features in general.

A &'router'& is a driver that operates on an address, either determining how
its delivery should happen, by assigning it to a specific transport, or
converting the address into one or more new addresses (for example, via an
alias file). A router may also explicitly choose to fail an address, causing it
to be bounced.

A &'transport'& is a driver that transmits a copy of the message from Exim's
spool to some destination. There are two kinds of transport: for a &'local'&
transport, the destination is a file or a pipe on the local host, whereas for a
&'remote'& transport the destination is some other host. A message is passed
to a specific transport as a result of successful routing. If a message has
several recipients, it may be passed to a number of different transports.

.cindex "preconditions" "definition of"
An address is processed by passing it to each configured router instance in
turn, subject to certain preconditions, until a router accepts the address or
specifies that it should be bounced. We will describe this process in more
detail shortly. First, as a simple example, we consider how each recipient
address in a message is processed in a small configuration of three routers.

To make this a more concrete example, it is described in terms of some actual
routers, but remember, this is only an example. You can configure Exim's
routers in many different ways, and there may be any number of routers in a
configuration.

The first router that is specified in a configuration is often one that handles
addresses in domains that are not recognized specifically by the local host.
Typically these are addresses for arbitrary domains on the Internet. A precondition
is set up which looks for the special domains known to the host (for example,
its own domain name), and the router is run for addresses that do &'not'&
match. Typically, this is a router that looks up domains in the DNS in order to
find the hosts to which this address routes. If it succeeds, the address is
assigned to a suitable SMTP transport; if it does not succeed, the router is
configured to fail the address.

The second router is reached only when the domain is recognized as one that
&"belongs"& to the local host. This router does redirection &-- also known as
aliasing and forwarding. When it generates one or more new addresses from the
original, each of them is routed independently from the start. Otherwise, the
router may cause an address to fail, or it may simply decline to handle the
address, in which case the address is passed to the next router.

The final router in many configurations is one that checks to see if the
address belongs to a local mailbox. The precondition may involve a check to
see if the local part is the name of a login account, or it may look up the
local part in a file or a database. If its preconditions are not met, or if
the router declines, we have reached the end of the routers. When this happens,
the address is bounced.



.section "Processing an address for verification" "SECID16"
.cindex "router" "for verification"
.cindex "verifying address" "overview"
As well as being used to decide how to deliver to an address, Exim's routers
are also used for &'address verification'&. Verification can be requested as
one of the checks to be performed in an ACL for incoming messages, on both
sender and recipient addresses, and it can be tested using the &%-bv%& and
&%-bvs%& command line options.

When an address is being verified, the routers are run in &"verify mode"&. This
does not affect the way the routers work, but it is a state that can be
detected. By this means, a router can be skipped or made to behave differently
when verifying. A common example is a configuration in which the first router
sends all messages to a message-scanning program unless they have been
previously scanned. Thus, the first router accepts all addresses without any
checking, making it useless for verifying. Normally, the &%no_verify%& option
would be set for such a router, causing it to be skipped in verify mode.




.section "Running an individual router" "SECTrunindrou"
.cindex "router" "running details"
.cindex "preconditions" "checking"
.cindex "router" "result of running"
As explained in the example above, a number of preconditions are checked before
running a router. If any are not met, the router is skipped, and the address is
passed to the next router. When all the preconditions on a router &'are'& met,
the router is run. What happens next depends on the outcome, which is one of
the following:

.ilist
&'accept'&: The router accepts the address, and either assigns it to a
transport or generates one or more &"child"& addresses. Processing the
original address ceases
.oindex "&%unseen%&"
unless the &%unseen%& option is set on the router. This option
can be used to set up multiple deliveries with different routing (for example,
for keeping archive copies of messages). When &%unseen%& is set, the address is
passed to the next router. Normally, however, an &'accept'& return marks the
end of routing.

Any child addresses generated by the router are processed independently,
starting with the first router by default. It is possible to change this by
setting the &%redirect_router%& option to specify which router to start at for
child addresses. Unlike &%pass_router%& (see below) the router specified by
&%redirect_router%& may be anywhere in the router configuration.
.next
&'pass'&: The router recognizes the address, but cannot handle it itself. It
requests that the address be passed to another router. By default, the address
is passed to the next router, but this can be changed by setting the
&%pass_router%& option. However, (unlike &%redirect_router%&) the named router
must be below the current router (to avoid loops).
.next
&'decline'&: The router declines to accept the address because it does not
recognize it at all. By default, the address is passed to the next router, but
this can be prevented by setting the &%no_more%& option. When &%no_more%& is
set, all the remaining routers are skipped. In effect, &%no_more%& converts
&'decline'& into &'fail'&.
.next
&'fail'&: The router determines that the address should fail, and queues it for
the generation of a bounce message. There is no further processing of the
original address unless &%unseen%& is set on the router.
.next
&'defer'&: The router cannot handle the address at the present time. (A
database may be offline, or a DNS lookup may have timed out.) No further
processing of the address happens in this delivery attempt. It is tried again
next time the message is considered for delivery.
.next
&'error'&: There is some error in the router (for example, a syntax error in
its configuration). The action is as for defer.
.endlist

If an address reaches the end of the routers without having been accepted by
any of them, it is bounced as unrouteable. The default error message in this
situation is &"unrouteable address"&, but you can set your own message by
making use of the &%cannot_route_message%& option. This can be set for any
router; the value from the last router that &"saw"& the address is used.

Sometimes while routing you want to fail a delivery when some conditions are
met but others are not, instead of passing the address on for further routing.
You can do this by having a second router that explicitly fails the delivery
when the relevant conditions are met. The &(redirect)& router has a &"fail"&
facility for this purpose.


.section "Duplicate addresses" "SECID17"
.cindex "case of local parts"
.cindex "address duplicate, discarding"
.cindex "duplicate addresses"
Once routing is complete, Exim scans the addresses that are assigned to local
and remote transports and discards any duplicates that it finds. During this
check, local parts are treated case-sensitively. This happens only when
actually delivering a message; when testing routers with &%-bt%&, all the
routed addresses are shown.



.section "Router preconditions" "SECTrouprecon"
.cindex "router" "preconditions, order of processing"
.cindex "preconditions" "order of processing"
The preconditions that are tested for each router are listed below, in the
order in which they are tested. The individual configuration options are
described in more detail in chapter &<<CHAProutergeneric>>&.

.olist
.cindex affix "router precondition"
The &%local_part_prefix%& and &%local_part_suffix%& options can specify that
the local parts handled by the router may or must have certain prefixes and/or
suffixes. If a mandatory affix (prefix or suffix) is not present, the router is
skipped. These conditions are tested first. When an affix is present, it is
removed from the local part before further processing, including the evaluation
of any other conditions.
.next
Routers can be designated for use only when not verifying an address, that is,
only when routing it for delivery (or testing its delivery routing). If the
&%verify%& option is set false, the router is skipped when Exim is verifying an
address.
Setting the &%verify%& option actually sets two options, &%verify_sender%& and
&%verify_recipient%&, which independently control the use of the router for
sender and recipient verification. You can set these options directly if
you want a router to be used for only one type of verification.
Note that cutthrough delivery is classed as a recipient verification for this purpose.
.next
If the &%address_test%& option is set false, the router is skipped when Exim is
run with the &%-bt%& option to test an address routing. This can be helpful
when the first router sends all new messages to a scanner of some sort; it
makes it possible to use &%-bt%& to test subsequent delivery routing without
having to simulate the effect of the scanner.
.next
Routers can be designated for use only when verifying an address, as
opposed to routing it for delivery. The &%verify_only%& option controls this.
Again, cutthrough delivery counts as a verification.
.next
Individual routers can be explicitly skipped when running the routers to
check an address given in the SMTP EXPN command (see the &%expn%& option).

.next
If the &%domains%& option is set, the domain of the address must be in the set
of domains that it defines.
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using router domains option"
A match verifies the variable &$domain$& (which carries tainted data)
and assigns an untainted value to the &$domain_data$& variable.
Such an untainted value is often needed in the transport.
For specifics of the matching operation and the resulting untainted value,
refer to section &<<SECTdomainlist>>&.

When an untainted value is wanted, use this option
rather than the generic &%condition%& option.

.next
.vindex "&$local_part_prefix$&"
.vindex "&$local_part_prefix_v$&"
.vindex "&$local_part$&"
.vindex "&$local_part_suffix$&"
.vindex "&$local_part_suffix_v$&"
.cindex affix "router precondition"
If the &%local_parts%& option is set, the local part of the address must be in
the set of local parts that it defines.
A match verifies the variable &$local_part$& (which carries tainted data)
and assigns an untainted value to the &$local_part_data$& variable.
Such an untainted value is often needed in the transport.
For specifics of the matching operation and the resulting untainted value,
refer to section &<<SECTlocparlis>>&.

When an untainted value is wanted, use this option
rather than the generic &%condition%& option.

If &%local_part_prefix%& or
&%local_part_suffix%& is in use, the prefix or suffix is removed from the local
part before this check. If you want to do precondition tests on local parts
that include affixes, you can do so by using a &%condition%& option (see below)
that uses the variables &$local_part$&, &$local_part_prefix$&,
&$local_part_prefix_v$&, &$local_part_suffix$&
and &$local_part_suffix_v$& as necessary.

.next
.vindex "&$local_user_uid$&"
.vindex "&$local_user_gid$&"
.vindex "&$home$&"
If the &%check_local_user%& option is set, the local part must be the name of
an account on the local host. If this check succeeds, the uid and gid of the
local user are placed in &$local_user_uid$& and &$local_user_gid$& and the
user's home directory is placed in &$home$&; these values can be used in the
remaining preconditions.

.next
If the &%router_home_directory%& option is set, it is expanded at this point,
because it overrides the value of &$home$&. If this expansion were left till
later, the value of &$home$& as set by &%check_local_user%& would be used in
subsequent tests. Having two different values of &$home$& in the same router
could lead to confusion.

.next
If the &%senders%& option is set, the envelope sender address must be in the
set of addresses that it defines.

.next
If the &%require_files%& option is set, the existence or non-existence of
specified files is tested.

.next
.cindex "customizing" "precondition"
If the &%condition%& option is set, it is evaluated and tested. This option
uses an expanded string to allow you to set up your own custom preconditions.
Expanded strings are described in chapter &<<CHAPexpand>>&.

Note that while using
this option for address matching technically works,
it does not set any de-tainted values.
Such values are often needed, either for router-specific options or
for transport options.
Using the &%domains%& and &%local_parts%& options is usually the most
convenient way to obtain them.
.endlist


Note that &%require_files%& comes near the end of the list, so you cannot use
it to check for the existence of a file in which to lookup up a domain, local
part, or sender. However, as these options are all expanded, you can use the
&%exists%& expansion condition to make such tests within each condition. The
&%require_files%& option is intended for checking files that the router may be
going to use internally, or which are needed by a specific transport (for
example, &_.procmailrc_&).



.section "Delivery in detail" "SECID18"
.cindex "delivery" "in detail"
When a message is to be delivered, the sequence of events is as follows:

.olist
If a system-wide filter file is specified, the message is passed to it. The
filter may add recipients to the message, replace the recipients, discard the
message, cause a new message to be generated, or cause the message delivery to
fail. The format of the system filter file is the same as for Exim user filter
files, described in the separate document entitled &'Exim's interfaces to mail
filtering'&.
.cindex "Sieve filter" "not available for system filter"
(&*Note*&: Sieve cannot be used for system filter files.)

Some additional features are available in system filters &-- see chapter
&<<CHAPsystemfilter>>& for details. Note that a message is passed to the system
filter only once per delivery attempt, however many recipients it has. However,
if there are several delivery attempts because one or more addresses could not
be immediately delivered, the system filter is run each time. The filter
condition &%first_delivery%& can be used to detect the first run of the system
filter.
.next
Each recipient address is offered to each configured router, in turn, subject to
its preconditions, until one is able to handle it. If no router can handle the
address, that is, if they all decline, the address is failed. Because routers
can be targeted at particular domains, several locally handled domains can be
processed entirely independently of each other.
.next
.cindex "routing" "loops in"
.cindex "loop" "while routing"
A router that accepts an address may assign it to a local or a remote
transport. However, the transport is not run at this time. Instead, the address
is placed on a list for the particular transport, which will be run later.
Alternatively, the router may generate one or more new addresses (typically
from alias, forward, or filter files). New addresses are fed back into this
process from the top, but in order to avoid loops, a router ignores any address
which has an identically-named ancestor that was processed by itself.
.next
When all the routing has been done, addresses that have been successfully
handled are passed to their assigned transports. When local transports are
doing real local deliveries, they handle only one address at a time, but if a
local transport is being used as a pseudo-remote transport (for example, to
collect batched SMTP messages for transmission by some other means) multiple
addresses can be handled. Remote transports can always handle more than one
address at a time, but can be configured not to do so, or to restrict multiple
addresses to the same domain.
.next
Each local delivery to a file or a pipe runs in a separate process under a
non-privileged uid, and these deliveries are run one at a time. Remote
deliveries also run in separate processes, normally under a uid that is private
to Exim (&"the Exim user"&), but in this case, several remote deliveries can be
run in parallel. The maximum number of simultaneous remote deliveries for any
one message is set by the &%remote_max_parallel%& option.
The order in which deliveries are done is not defined, except that all local
deliveries happen before any remote deliveries.
.next
.cindex "queue runner"
When it encounters a local delivery during a queue run, Exim checks its retry
database to see if there has been a previous temporary delivery failure for the
address before running the local transport. If there was a previous failure,
Exim does not attempt a new delivery until the retry time for the address is
reached. However, this happens only for delivery attempts that are part of a
queue run. Local deliveries are always attempted when delivery immediately
follows message reception, even if retry times are set for them. This makes for
better behaviour if one particular message is causing problems (for example,
causing quota overflow, or provoking an error in a filter file).
.next
.cindex "delivery" "retry in remote transports"
Remote transports do their own retry handling, since an address may be
deliverable to one of a number of hosts, each of which may have a different
retry time. If there have been previous temporary failures and no host has
reached its retry time, no delivery is attempted, whether in a queue run or
not. See chapter &<<CHAPretry>>& for details of retry strategies.
.next
If there were any permanent errors, a bounce message is returned to an
appropriate address (the sender in the common case), with details of the error
for each failing address. Exim can be configured to send copies of bounce
messages to other addresses.
.next
.cindex "delivery" "deferral"
If one or more addresses suffered a temporary failure, the message is left on
the queue, to be tried again later. Delivery of these addresses is said to be
&'deferred'&.
.next
When all the recipient addresses have either been delivered or bounced,
handling of the message is complete. The spool files and message log are
deleted, though the message log can optionally be preserved if required.
.endlist




.section "Retry mechanism" "SECID19"
.cindex "delivery" "retry mechanism"
.cindex "retry" "description of mechanism"
.cindex "queue runner"
Exim's mechanism for retrying messages that fail to get delivered at the first
attempt is the queue runner process. You must either run an Exim daemon that
uses the &%-q%& option with a time interval to start queue runners at regular
intervals or use some other means (such as &'cron'&) to start them. If you do
not arrange for queue runners to be run, messages that fail temporarily at the
first attempt will remain in your queue forever. A queue runner process works
its way through the queue, one message at a time, trying each delivery that has
passed its retry time.
You can run several queue runners at once.

Exim uses a set of configured rules to determine when next to retry the failing
address (see chapter &<<CHAPretry>>&). These rules also specify when Exim
should give up trying to deliver to the address, at which point it generates a
bounce message. If no retry rules are set for a particular host, address, and
error combination, no retries are attempted, and temporary errors are treated
as permanent.



.subsection "Temporary delivery failure" SECID20
.cindex "delivery" "temporary failure"
There are many reasons why a message may not be immediately deliverable to a
particular address. Failure to connect to a remote machine (because it, or the
connection to it, is down) is one of the most common. Temporary failures may be
detected during routing as well as during the transport stage of delivery.
Local deliveries may be delayed if NFS files are unavailable, or if a mailbox
is on a file system where the user is over quota. Exim can be configured to
impose its own quotas on local mailboxes; where system quotas are set they will
also apply.

If a host is unreachable for a period of time, a number of messages may be
waiting for it by the time it recovers, and sending them in a single SMTP
connection is clearly beneficial. Whenever a delivery to a remote host is
deferred,
.cindex "hints database" "deferred deliveries"
Exim makes a note in its hints database, and whenever a successful
SMTP delivery has happened, it looks to see if any other messages are waiting
for the same host. If any are found, they are sent over the same SMTP
connection, subject to a configuration limit as to the maximum number in any
one connection.



.subsection "Permanent delivery failure" SECID21
.cindex "delivery" "permanent failure"
.cindex "bounce message" "when generated"
When a message cannot be delivered to some or all of its intended recipients, a
bounce message is generated. Temporary delivery failures turn into permanent
errors when their timeout expires. All the addresses that fail in a given
delivery attempt are listed in a single message. If the original message has
many recipients, it is possible for some addresses to fail in one delivery
attempt and others to fail subsequently, giving rise to more than one bounce
message. The wording of bounce messages can be customized by the administrator.
See chapter &<<CHAPemsgcust>>& for details.

.cindex "&'X-Failed-Recipients:'& header line"
Bounce messages contain an &'X-Failed-Recipients:'& header line that lists the
failed addresses, for the benefit of programs that try to analyse such messages
automatically.

.cindex "bounce message" "recipient of"
A bounce message is normally sent to the sender of the original message, as
obtained from the message's envelope. For incoming SMTP messages, this is the
address given in the MAIL command. However, when an address is expanded via a
forward or alias file, an alternative address can be specified for delivery
failures of the generated addresses. For a mailing list expansion (see section
&<<SECTmailinglists>>&) it is common to direct bounce messages to the manager
of the list.



.subsection "Failures to deliver bounce messages" SECID22
.cindex "bounce message" "failure to deliver"
If a bounce message (either locally generated or received from a remote host)
itself suffers a permanent delivery failure, the message is left in the queue,
but it is frozen, awaiting the attention of an administrator. There are options
that can be used to make Exim discard such failed messages, or to keep them
for only a short time (see &%timeout_frozen_after%& and
&%ignore_bounce_errors_after%&).





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Building and installing Exim" "CHID3"
.scindex IIDbuex "building Exim"

.section "Unpacking" "SECID23"
Exim is distributed as a gzipped or bzipped tar file which, when unpacked,
creates a directory with the name of the current release (for example,
&_exim-&version()_&) into which the following files are placed:

.table2 140pt
.irow &_ACKNOWLEDGMENTS_& "contains some acknowledgments"
.irow &_CHANGES_&         "contains a reference to where changes are &&&
  documented"
.irow &_LICENCE_&         "the GNU General Public Licence"
.irow &_Makefile_&        "top-level make file"
.irow &_NOTICE_&          "conditions for the use of Exim"
.irow &_README_&          "list of files, directories and simple build &&&
  instructions"
.endtable

Other files whose names begin with &_README_& may also be present. The
following subdirectories are created:

.table2 140pt
.irow &_Local_&           "an empty directory for local configuration files"
.irow &_OS_&              "OS-specific files"
.irow &_doc_&             "documentation files"
.irow &_exim_monitor_&    "source files for the Exim monitor"
.irow &_scripts_&         "scripts used in the build process"
.irow &_src_&             "remaining source files"
.irow &_util_&            "independent utilities"
.endtable

The main utility programs are contained in the &_src_& directory and are built
with the Exim binary. The &_util_& directory contains a few optional scripts
that may be useful to some sites.


.section "Multiple machine architectures and operating systems" "SECID24"
.cindex "building Exim" "multiple OS/architectures"
The building process for Exim is arranged to make it easy to build binaries for
a number of different architectures and operating systems from the same set of
source files. Compilation does not take place in the &_src_& directory.
Instead, a &'build directory'& is created for each architecture and operating
system.
.cindex "symbolic link" "to build directory"
Symbolic links to the sources are installed in this directory, which is where
the actual building takes place. In most cases, Exim can discover the machine
architecture and operating system for itself, but the defaults can be
overridden if necessary.
.cindex compiler requirements
.cindex compiler version
A C99-capable compiler will be required for the build.


.section "PCRE2 library" "SECTpcre"
.cindex "PCRE2 library"
Exim no longer has an embedded regular-expression library as the vast majority of
modern systems include PCRE2 as a system library, although you may need to
install the PCRE2 package or the PCRE2 development package for your operating
system. If your system has a normal PCRE2 installation the Exim build
process will need no further configuration. If the library or the
headers are in an unusual location you will need to either set the PCRE2_LIBS
and INCLUDE directives appropriately,
or set PCRE2_CONFIG=yes to use the installed &(pcre-config)& command.
If your operating system has no
PCRE2 support then you will need to obtain and build the current PCRE2
from &url(https://github.com/PhilipHazel/pcre2/releases).
More information on PCRE2 is available at &url(https://www.pcre.org/).

.section "DBM libraries" "SECTdb"
.cindex "DBM libraries" "discussion of"
.cindex "hints database" "DBM files used for"
Even if you do not use any DBM files in your configuration, Exim still needs a
DBM library in order to operate, because it uses indexed files for its hints
databases. Unfortunately, there are a number of DBM libraries in existence, and
different operating systems often have different ones installed.

.cindex "Solaris" "DBM library for"
.cindex "IRIX, DBM library for"
.cindex "BSD, DBM library for"
.cindex "Linux, DBM library for"
If you are using Solaris, IRIX, one of the modern BSD systems, or a modern
Linux distribution, the DBM configuration should happen automatically, and you
may be able to ignore this section. Otherwise, you may have to learn more than
you would like about DBM libraries from what follows.

.cindex "&'ndbm'& DBM library"
Licensed versions of Unix normally contain a library of DBM functions operating
via the &'ndbm'& interface, and this is what Exim expects by default. Free
versions of Unix seem to vary in what they contain as standard. In particular,
some early versions of Linux have no default DBM library, and different
distributors have chosen to bundle different libraries with their packaged
versions. However, the more recent releases seem to have standardized on the
Berkeley DB library.

.new
Ownership of the Berkeley DB library has moved to a major corporation;
development seems to have stalled and documentation is not freely available.
This is probably not tenable for the long term use by Exim.
.wen

Different DBM libraries have different conventions for naming the files they
use. When a program opens a file called &_dbmfile_&, there are several
possibilities:

.olist
A traditional &'ndbm'& implementation, such as that supplied as part of
Solaris, operates on two files called &_dbmfile.dir_& and &_dbmfile.pag_&.
.next
.cindex "&'gdbm'& DBM library"
The GNU library, &'gdbm'&, operates on a single file. If used via its &'ndbm'&
compatibility interface it makes two different hard links to it with names
&_dbmfile.dir_& and &_dbmfile.pag_&, but if used via its native interface, the
filename is used unmodified.
.next
.cindex "Berkeley DB library"
The Berkeley DB package, if called via its &'ndbm'& compatibility interface,
operates on a single file called &_dbmfile.db_&, but otherwise looks to the
programmer exactly the same as the traditional &'ndbm'& implementation.
.next
If the Berkeley package is used in its native mode, it operates on a single
file called &_dbmfile_&; the programmer's interface is somewhat different to
the traditional &'ndbm'& interface.
.next
To complicate things further, there are several very different versions of the
Berkeley DB package. Version 1.85 was stable for a very long time, releases
2.&'x'& and 3.&'x'& were current for a while,
but the latest versions when Exim last revamped support were numbered 5.&'x'&.
Maintenance of some of the earlier releases has ceased,
and Exim no longer supports versions before 3.&'x'&.
All versions of Berkeley DB could be obtained from
&url(http://www.sleepycat.com/), which is now a redirect to their new owner's
page with far newer versions listed.
It is probably wise to plan to move your storage configurations away from
Berkeley DB format, as today there are smaller and simpler alternatives more
suited to Exim's usage model.
.next
.cindex "&'tdb'& DBM library"
Yet another DBM library, called &'tdb'&, is available from
&url(https://sourceforge.net/projects/tdb/files/). It has its own interface, and also
operates on a single file.
.next
It is possible to use &url(https://www.sqlite.org/index.html,sqlite3)
for the DBM library.
.endlist

.cindex "USE_DB"
.cindex "DBM libraries" "configuration for building"
Exim and its utilities can be compiled to use any of these interfaces. In order
to use any version of the Berkeley DB package in native mode, you must set
USE_DB in an appropriate configuration file (typically
&_Local/Makefile_&). For example:
.code
USE_DB=yes
.endd
Similarly, for gdbm you set USE_GDBM, for tdb you set USE_TDB,
and for sqlite3 you set USE_SQLITE.
An error is diagnosed if you set more than one of these.
You can set USE_NDBM if needed to override an operating system default.

At the lowest level, the build-time configuration sets none of these options,
thereby assuming an interface of type (1). However, some operating system
configuration files (for example, those for the BSD operating systems and
Linux) assume type (4) by setting USE_DB as their default, and the
configuration files for Cygwin set USE_GDBM. Anything you set in
&_Local/Makefile_&, however, overrides these system defaults.

As well as setting USE_DB, USE_GDBM, or USE_TDB, it may also be
necessary to set DBMLIB, to cause inclusion of the appropriate library, as
in one of these lines:
.code
DBMLIB = -ldb
DBMLIB = -ltdb
DBMLIB = -lsqlite3
DBMLIB = -lgdbm -lgdbm_compat
.endd
The last of those was for a Linux having GDBM provide emulated NDBM facilities.
Settings like that will work if the DBM library is installed in the standard
place. Sometimes it is not, and the library's header file may also not be in
the default path. You may need to set INCLUDE to specify where the header
file is, and to specify the path to the library more fully in DBMLIB, as in
this example:
.code
INCLUDE=-I/usr/local/include/db-4.1
DBMLIB=/usr/local/lib/db-4.1/libdb.a
.endd
There is further detailed discussion about the various DBM libraries in the
file &_doc/dbm.discuss.txt_& in the Exim distribution.

.new
When moving from one DBM library to another,
for the hints databases it suffices to just remove all the files in the
directory named &"db/"& under the spool directory.
This is because hints are only for optimisation and will be rebuilt
during normal operations.
Non-hints DBM databases (used by &"dbm"& lookups in the configuration)
will need individual rebuilds for the new DBM library.
This is not done automatically
.wen



.section "Pre-building configuration" "SECID25"
.cindex "building Exim" "pre-building configuration"
.cindex "configuration for building Exim"
.cindex "&_Local/Makefile_&"
.cindex "&_src/EDITME_&"
Before building Exim, a local configuration file that specifies options
independent of any operating system has to be created with the name
&_Local/Makefile_&. A template for this file is supplied as the file
&_src/EDITME_&, and it contains full descriptions of all the option settings
therein. These descriptions are therefore not repeated here. If you are
building Exim for the first time, the simplest thing to do is to copy
&_src/EDITME_& to &_Local/Makefile_&, then read it and edit it appropriately.

There are three settings that you must supply, because Exim will not build
without them. They are the location of the runtime configuration file
(CONFIGURE_FILE), the directory in which Exim binaries will be installed
(BIN_DIRECTORY), and the identity of the Exim user (EXIM_USER and
maybe EXIM_GROUP as well). The value of CONFIGURE_FILE can in fact be
a colon-separated list of filenames; Exim uses the first of them that exists.

There are a few other parameters that can be specified either at build time or
at runtime, to enable the same binary to be used on a number of different
machines. However, if the locations of Exim's spool directory and log file
directory (if not within the spool directory) are fixed, it is recommended that
you specify them in &_Local/Makefile_& instead of at runtime, so that errors
detected early in Exim's execution (such as a malformed configuration file) can
be logged.

.cindex "content scanning" "specifying at build time"
Exim's interfaces for calling virus and spam scanning software directly from
access control lists are not compiled by default. If you want to include these
facilities, you need to set
.code
WITH_CONTENT_SCAN=yes
.endd
in your &_Local/Makefile_&. For details of the facilities themselves, see
chapter &<<CHAPexiscan>>&.


.cindex "&_Local/eximon.conf_&"
.cindex "&_exim_monitor/EDITME_&"
If you are going to build the Exim monitor, a similar configuration process is
required. The file &_exim_monitor/EDITME_& must be edited appropriately for
your installation and saved under the name &_Local/eximon.conf_&. If you are
happy with the default settings described in &_exim_monitor/EDITME_&,
&_Local/eximon.conf_& can be empty, but it must exist.

This is all the configuration that is needed in straightforward cases for known
operating systems. However, the building process is set up so that it is easy
to override options that are set by default or by operating-system-specific
configuration files, for example, to change the C compiler, which
defaults to &%gcc%&. See section &<<SECToverride>>& below for details of how to
do this.



.section "Support for iconv()" "SECID26"
.cindex "&[iconv()]& support"
.cindex "RFC 2047"
The contents of header lines in messages may be encoded according to the rules
described in &url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047).
This makes it possible to transmit characters that are not
in the ASCII character set, and to label them as being in a particular
character set. When Exim is inspecting header lines by means of the &%$h_%&
mechanism, it decodes them, and translates them into a specified character set
(default is set at build time). The translation is possible only if the operating system
supports the &[iconv()]& function.

However, some of the operating systems that supply &[iconv()]& do not support
very many conversions. The GNU &%libiconv%& library (available from
&url(https://www.gnu.org/software/libiconv/)) can be installed on such
systems to remedy this deficiency, as well as on systems that do not supply
&[iconv()]& at all. After installing &%libiconv%&, you should add
.code
HAVE_ICONV=yes
.endd
to your &_Local/Makefile_& and rebuild Exim.



.section "Including TLS/SSL encryption support" "SECTinctlsssl"
.cindex "TLS" "including support for TLS"
.cindex "encryption" "including support for"
.cindex "OpenSSL" "building Exim with"
.cindex "GnuTLS" "building Exim with"
Exim is usually built to support encrypted SMTP connections, using the STARTTLS
command as per &url(https://www.rfc-editor.org/rfc/rfc2487,RFC 2487).
It can also support clients that expect to
start a TLS session immediately on connection to a non-standard port (see the
&%tls_on_connect_ports%& runtime option and the &%-tls-on-connect%& command
line option).

If you want to build Exim with TLS support, you must first install either the
OpenSSL or GnuTLS library. There is no cryptographic code in Exim itself for
implementing SSL.

If you do not want TLS support you should set
.code
DISABLE_TLS=yes
.endd
in &_Local/Makefile_&.

If OpenSSL is installed, you should set
.code
USE_OPENSL=yes
TLS_LIBS=-lssl -lcrypto
.endd
in &_Local/Makefile_&. You may also need to specify the locations of the
OpenSSL library and include files. For example:
.code
USE_OPENSSL=yes
TLS_LIBS=-L/usr/local/openssl/lib -lssl -lcrypto
TLS_INCLUDE=-I/usr/local/openssl/include/
.endd
.cindex "pkg-config" "OpenSSL"
If you have &'pkg-config'& available, then instead you can just use:
.code
USE_OPENSSL=yes
USE_OPENSSL_PC=openssl
.endd
.cindex "USE_GNUTLS"
If GnuTLS is installed, you should set
.code
USE_GNUTLS=yes
TLS_LIBS=-lgnutls -ltasn1 -lgcrypt
.endd
in &_Local/Makefile_&, and again you may need to specify the locations of the
library and include files. For example:
.code
USE_GNUTLS=yes
TLS_LIBS=-L/usr/gnu/lib -lgnutls -ltasn1 -lgcrypt
TLS_INCLUDE=-I/usr/gnu/include
.endd
.cindex "pkg-config" "GnuTLS"
If you have &'pkg-config'& available, then instead you can just use:
.code
USE_GNUTLS=yes
USE_GNUTLS_PC=gnutls
.endd

You do not need to set TLS_INCLUDE if the relevant directory is already
specified in INCLUDE. Details of how to configure Exim to make use of TLS are
given in chapter &<<CHAPTLS>>&.




.section "Use of tcpwrappers" "SECID27"

.cindex "tcpwrappers, building Exim to support"
.cindex "USE_TCP_WRAPPERS"
.cindex "TCP_WRAPPERS_DAEMON_NAME"
.cindex "tcp_wrappers_daemon_name"
Exim can be linked with the &'tcpwrappers'& library in order to check incoming
SMTP calls using the &'tcpwrappers'& control files. This may be a convenient
alternative to Exim's own checking facilities for installations that are
already making use of &'tcpwrappers'& for other purposes. To do this, you
should set USE_TCP_WRAPPERS in &_Local/Makefile_&, arrange for the file
&_tcpd.h_& to be available at compile time, and also ensure that the library
&_libwrap.a_& is available at link time, typically by including &%-lwrap%& in
EXTRALIBS_EXIM. For example, if &'tcpwrappers'& is installed in &_/usr/local_&,
you might have
.code
USE_TCP_WRAPPERS=yes
CFLAGS=-O -I/usr/local/include
EXTRALIBS_EXIM=-L/usr/local/lib -lwrap
.endd
in &_Local/Makefile_&. The daemon name to use in the &'tcpwrappers'& control
files is &"exim"&. For example, the line
.code
exim : LOCAL  192.168.1.  .friendly.domain.example
.endd
in your &_/etc/hosts.allow_& file allows connections from the local host, from
the subnet 192.168.1.0/24, and from all hosts in &'friendly.domain.example'&.
All other connections are denied. The daemon name used by &'tcpwrappers'&
can be changed at build time by setting TCP_WRAPPERS_DAEMON_NAME in
&_Local/Makefile_&, or by setting tcp_wrappers_daemon_name in the
configure file. Consult the &'tcpwrappers'& documentation for
further details.


.section "Including support for IPv6" "SECID28"
.cindex "IPv6" "including support for"
Exim contains code for use on systems that have IPv6 support. Setting
&`HAVE_IPV6=YES`& in &_Local/Makefile_& causes the IPv6 code to be included;
it may also be necessary to set IPV6_INCLUDE and IPV6_LIBS on systems
where the IPv6 support is not fully integrated into the normal include and
library files.

Two different types of DNS record for handling IPv6 addresses have been
defined. AAAA records (analogous to A records for IPv4) are in use, and are
currently seen as the mainstream. Another record type called A6 was proposed
as better than AAAA because it had more flexibility. However, it was felt to be
over-complex, and its status was reduced to &"experimental"&.
Exim used to
have a compile option for including A6 record support but this has now been
withdrawn.



.section "Dynamically loaded module support" "SECTdynamicmodules"
.cindex "lookup modules"
.cindex "router modules"
.cindex "transport modules"
.cindex "authenticator modules"
.cindex "dynamic modules"
.cindex ".so building"
On some platforms, Exim supports not compiling all lookup types directly into
the main binary, instead putting some into external modules which can be loaded
on demand.
This permits packagers to build Exim with support for lookups with extensive
library dependencies without requiring all systems to install all of those
dependencies.
.new
Any combination of lookup types can be built this way.
Lookup types that provide several variants will be loaded as
Exim starts.
Types that provide only one method are not loaded until used by
the runtime configuration.
.wen

For building
set &`LOOKUP_MODULE_DIR`& to the directory into which the modules will be
installed; Exim will only load modules from that directory, as a security
measure.  You will need to set &`CFLAGS_DYNAMIC`& if not already defined
for your OS; see &_OS/Makefile-Linux_& for an example.
Some other requirements for adjusting &`EXTRALIBS`& may also be necessary,
see &_src/EDITME_& for details.

Then, for each module to be loaded dynamically, define the relevant
&`LOOKUP_`&<&'lookup_type'&> flags to have the value "2" instead of "yes".
For example, this will build in lsearch but load sqlite and mysql support
only if each is installed:
.code
LOOKUP_LSEARCH=yes
LOOKUP_SQLITE=2
LOOKUP_MYSQL=2
.endd
Set also &`LOOKUP_`&<&'lookup_type'&>&` INCLUDE`& and
&`LOOKUP_`&<&'lookup_type'&>`_LIBS if needed for each lookup type,
ensuring that duplicates are not present in more global values.

.new
Similarly, authenticator, router and transport drivers can be built
as external modules.
Modules will be searched for as demanded by the runtime configuration,
permitting a smaller Exim binary.

For building, as above but using
&`AUTH_*`&, &`ROUTER_*`& and &`TRANSPORT_*`& instead of &`LOOKUP_*`&,
.wen


.section "The building process" "SECID29"
.cindex "build directory"
Once &_Local/Makefile_& (and &_Local/eximon.conf_&, if required) have been
created, run &'make'& at the top level. It determines the architecture and
operating system types, and creates a build directory if one does not exist.
For example, on a Sun system running Solaris 8, the directory
&_build-SunOS5-5.8-sparc_& is created.
.cindex "symbolic link" "to source files"
Symbolic links to relevant source files are installed in the build directory.

If this is the first time &'make'& has been run, it calls a script that builds
a make file inside the build directory, using the configuration files from the
&_Local_& directory. The new make file is then passed to another instance of
&'make'&. This does the real work, building a number of utility scripts, and
then compiling and linking the binaries for the Exim monitor (if configured), a
number of utility programs, and finally Exim itself. The command &`make
makefile`& can be used to force a rebuild of the make file in the build
directory, should this ever be necessary.

If you have problems building Exim, check for any comments there may be in the
&_README_& file concerning your operating system, and also take a look at the
FAQ, where some common problems are covered.



.section 'Output from &"make"&' "SECID283"
The output produced by the &'make'& process for compile lines is often very
unreadable, because these lines can be very long. For this reason, the normal
output is suppressed by default, and instead output similar to that which
appears when compiling the 2.6 Linux kernel is generated: just a short line for
each module that is being compiled or linked. However, it is still possible to
get the full output, by calling &'make'& like this:
.code
FULLECHO='' make -e
.endd
The value of FULLECHO defaults to &"@"&, the flag character that suppresses
command reflection in &'make'&. When you ask for the full output, it is
given in addition to the short output.



.section "Overriding build-time options for Exim" "SECToverride"
.cindex "build-time options, overriding"
The main make file that is created at the beginning of the building process
consists of the concatenation of a number of files which set configuration
values, followed by a fixed set of &'make'& instructions. If a value is set
more than once, the last setting overrides any previous ones. This provides a
convenient way of overriding defaults. The files that are concatenated are, in
order:
.display
&_OS/Makefile-Default_&
&_OS/Makefile-_&<&'ostype'&>
&_Local/Makefile_&
&_Local/Makefile-_&<&'ostype'&>
&_Local/Makefile-_&<&'archtype'&>
&_Local/Makefile-_&<&'ostype'&>-<&'archtype'&>
&_OS/Makefile-Base_&
.endd
.cindex "&_Local/Makefile_&"
.cindex "building Exim" "operating system type"
.cindex "building Exim" "architecture type"
where <&'ostype'&> is the operating system type and <&'archtype'&> is the
architecture type. &_Local/Makefile_& is required to exist, and the building
process fails if it is absent. The other three &_Local_& files are optional,
and are often not needed.

The values used for <&'ostype'&> and <&'archtype'&> are obtained from scripts
called &_scripts/os-type_& and &_scripts/arch-type_& respectively. If either of
the environment variables EXIM_OSTYPE or EXIM_ARCHTYPE is set, their
values are used, thereby providing a means of forcing particular settings.
Otherwise, the scripts try to get values from the &%uname%& command. If this
fails, the shell variables OSTYPE and ARCHTYPE are inspected. A number
of &'ad hoc'& transformations are then applied, to produce the standard names
that Exim expects. You can run these scripts directly from the shell in order
to find out what values are being used on your system.


&_OS/Makefile-Default_& contains comments about the variables that are set
therein. Some (but not all) are mentioned below. If there is something that
needs changing, review the contents of this file and the contents of the make
file for your operating system (&_OS/Makefile-<ostype>_&) to see what the
default values are.


.cindex "building Exim" "overriding default settings"
If you need to change any of the values that are set in &_OS/Makefile-Default_&
or in &_OS/Makefile-<ostype>_&, or to add any new definitions, you do not
need to change the original files. Instead, you should make the changes by
putting the new values in an appropriate &_Local_& file. For example,
.cindex "Tru64-Unix build-time settings"
when building Exim in many releases of the Tru64-Unix (formerly Digital UNIX,
formerly DEC-OSF1) operating system, it is necessary to specify that the C
compiler is called &'cc'& rather than &'gcc'&. Also, the compiler must be
called with the option &%-std1%&, to make it recognize some of the features of
Standard C that Exim uses. (Most other compilers recognize Standard C by
default.) To do this, you should create a file called &_Local/Makefile-OSF1_&
containing the lines
.code
CC=cc
CFLAGS=-std1
.endd
If you are compiling for just one operating system, it may be easier to put
these lines directly into &_Local/Makefile_&.

Keeping all your local configuration settings separate from the distributed
files makes it easy to transfer them to new versions of Exim simply by copying
the contents of the &_Local_& directory.


.cindex "NIS lookup type" "including support for"
.cindex "NIS+ lookup type" "including support for"
.cindex "LDAP" "including support for"
.cindex "lookup" "inclusion in binary"
Exim contains support for doing LDAP, NIS, NIS+, and other kinds of file
lookup, but not all systems have these components installed, so the default is
not to include the relevant code in the binary. All the different kinds of file
and database lookup that Exim supports are implemented as separate code modules
which are included only if the relevant compile-time options are set. In the
case of LDAP, NIS, and NIS+, the settings for &_Local/Makefile_& are:
.code
LOOKUP_LDAP=yes
LOOKUP_NIS=yes
LOOKUP_NISPLUS=yes
.endd
and similar settings apply to the other lookup types. They are all listed in
&_src/EDITME_&. In many cases the relevant include files and interface
libraries need to be installed before compiling Exim.
.cindex "cdb" "including support for"
However, there are some optional lookup types (such as cdb) for which
the code is entirely contained within Exim, and no external include
files or libraries are required. When a lookup type is not included in the
binary, attempts to configure Exim to use it cause runtime configuration
errors.

.cindex "pkg-config" "lookups"
.cindex "pkg-config" "authenticators"
Many systems now use a tool called &'pkg-config'& to encapsulate information
about how to compile against a library; Exim has some initial support for
being able to use pkg-config for lookups and authenticators.  For any given
makefile variable which starts &`LOOKUP_`& or &`AUTH_`&, you can add a new
variable with the &`_PC`& suffix in the name and assign as the value the
name of the package to be queried.  The results of querying via the
&'pkg-config'& command will be added to the appropriate Makefile variables
with &`+=`& directives, so your version of &'make'& will need to support that
syntax.  For instance:
.code
LOOKUP_SQLITE=yes
LOOKUP_SQLITE_PC=sqlite3
AUTH_GSASL=yes
AUTH_GSASL_PC=libgsasl
AUTH_HEIMDAL_GSSAPI=yes
AUTH_HEIMDAL_GSSAPI_PC=heimdal-gssapi
.endd

.cindex "Perl" "including support for"
Exim can be linked with an embedded Perl interpreter, allowing Perl
subroutines to be called during string expansion. To enable this facility,
.code
EXIM_PERL=perl.o
.endd
must be defined in &_Local/Makefile_&. Details of this facility are given in
chapter &<<CHAPperl>>&.

.cindex "X11 libraries, location of"
The location of the X11 libraries is something that varies a lot between
operating systems, and there may be different versions of X11 to cope
with. Exim itself makes no use of X11, but if you are compiling the Exim
monitor, the X11 libraries must be available.
The following three variables are set in &_OS/Makefile-Default_&:
.code
X11=/usr/X11R6
XINCLUDE=-I$(X11)/include
XLFLAGS=-L$(X11)/lib
.endd
These are overridden in some of the operating-system configuration files. For
example, in &_OS/Makefile-SunOS5_& there is
.code
X11=/usr/openwin
XINCLUDE=-I$(X11)/include
XLFLAGS=-L$(X11)/lib -R$(X11)/lib
.endd
If you need to override the default setting for your operating system, place a
definition of all three of these variables into your
&_Local/Makefile-<ostype>_& file.

.cindex "EXTRALIBS"
If you need to add any extra libraries to the link steps, these can be put in a
variable called EXTRALIBS, which appears in all the link commands, but by
default is not defined. In contrast, EXTRALIBS_EXIM is used only on the
command for linking the main Exim binary, and not for any associated utilities.

.cindex "DBM libraries" "configuration for building"
There is also DBMLIB, which appears in the link commands for binaries that
use DBM functions (see also section &<<SECTdb>>&). Finally, there is
EXTRALIBS_EXIMON, which appears only in the link step for the Exim monitor
binary, and which can be used, for example, to include additional X11
libraries.

.cindex "configuration file" "editing"
The make file copes with rebuilding Exim correctly if any of the configuration
files are edited. However, if an optional configuration file is deleted, it is
necessary to touch the associated non-optional file (that is,
&_Local/Makefile_& or &_Local/eximon.conf_&) before rebuilding.


.section "OS-specific header files" "SECID30"
.cindex "&_os.h_&"
.cindex "building Exim" "OS-specific C header files"
The &_OS_& directory contains a number of files with names of the form
&_os.h-<ostype>_&. These are system-specific C header files that should not
normally need to be changed. There is a list of macro settings that are
recognized in the file &_OS/os.configuring_&, which should be consulted if you
are porting Exim to a new operating system.



.section "Overriding build-time options for the monitor" "SECID31"
.cindex "building Eximon"
A similar process is used for overriding things when building the Exim monitor,
where the files that are involved are
.display
&_OS/eximon.conf-Default_&
&_OS/eximon.conf-_&<&'ostype'&>
&_Local/eximon.conf_&
&_Local/eximon.conf-_&<&'ostype'&>
&_Local/eximon.conf-_&<&'archtype'&>
&_Local/eximon.conf-_&<&'ostype'&>-<&'archtype'&>
.endd
.cindex "&_Local/eximon.conf_&"
As with Exim itself, the final three files need not exist, and in this case the
&_OS/eximon.conf-<ostype>_& file is also optional. The default values in
&_OS/eximon.conf-Default_& can be overridden dynamically by setting environment
variables of the same name, preceded by EXIMON_. For example, setting
EXIMON_LOG_DEPTH in the environment overrides the value of
LOG_DEPTH at runtime.
.ecindex IIDbuex


.section "Installing Exim binaries and scripts" "SECID32"
.cindex "installing Exim"
.cindex "BIN_DIRECTORY"
The command &`make install`& runs the &(exim_install)& script with no
arguments. The script copies binaries and utility scripts into the directory
whose name is specified by the BIN_DIRECTORY setting in &_Local/Makefile_&.
.cindex "setuid" "installing Exim with"
The install script copies files only if they are newer than the files they are
going to replace. The Exim binary is required to be owned by root and have the
&'setuid'& bit set, for normal configurations. Therefore, you must run &`make
install`& as root so that it can set up the Exim binary in this way. However, in
some special situations (for example, if a host is doing no local deliveries)
it may be possible to run Exim without making the binary setuid root (see
chapter &<<CHAPsecurity>>& for details).

.cindex "CONFIGURE_FILE"
Exim's runtime configuration file is named by the CONFIGURE_FILE setting
in &_Local/Makefile_&. If this names a single file, and the file does not
exist, the default configuration file &_src/configure.default_& is copied there
by the installation script. If a runtime configuration file already exists, it
is left alone. If CONFIGURE_FILE is a colon-separated list, naming several
alternative files, no default is installed.

.cindex "system aliases file"
.cindex "&_/etc/aliases_&"
One change is made to the default configuration file when it is installed: the
default configuration contains a router that references a system aliases file.
The path to this file is set to the value specified by
SYSTEM_ALIASES_FILE in &_Local/Makefile_& (&_/etc/aliases_& by default).
If the system aliases file does not exist, the installation script creates it,
and outputs a comment to the user.

The created file contains no aliases, but it does contain comments about the
aliases a site should normally have. Mail aliases have traditionally been
kept in &_/etc/aliases_&. However, some operating systems are now using
&_/etc/mail/aliases_&. You should check if yours is one of these, and change
Exim's configuration if necessary.

The default configuration uses the local host's name as the only local domain,
and is set up to do local deliveries into the shared directory &_/var/mail_&,
running as the local user. System aliases and &_.forward_& files in users' home
directories are supported, but no NIS or NIS+ support is configured. Domains
other than the name of the local host are routed using the DNS, with delivery
over SMTP.

It is possible to install Exim for special purposes (such as building a binary
distribution) in a private part of the file system. You can do this by a
command such as
.code
make DESTDIR=/some/directory/ install
.endd
This has the effect of pre-pending the specified directory to all the file
paths, except the name of the system aliases file that appears in the default
configuration. (If a default alias file is created, its name &'is'& modified.)
For backwards compatibility, ROOT is used if DESTDIR is not set,
but this usage is deprecated.

.cindex "installing Exim" "what is not installed"
Running &'make install'& does not copy the Exim 4 conversion script
&'convert4r4'&. You will probably run this only once if you are
upgrading from Exim 3. None of the documentation files in the &_doc_&
directory are copied, except for the info files when you have set
INFO_DIRECTORY, as described in section &<<SECTinsinfdoc>>& below.

For the utility programs, old versions are renamed by adding the suffix &_.O_&
to their names. The Exim binary itself, however, is handled differently. It is
installed under a name that includes the version number and the compile number,
for example, &_exim-&version()-1_&. The script then arranges for a symbolic link
called &_exim_& to point to the binary. If you are updating a previous version
of Exim, the script takes care to ensure that the name &_exim_& is never absent
from the directory (as seen by other processes).

.cindex "installing Exim" "testing the script"
If you want to see what the &'make install'& will do before running it for
real, you can pass the &%-n%& option to the installation script by this
command:
.code
make INSTALL_ARG=-n install
.endd
The contents of the variable INSTALL_ARG are passed to the installation
script. You do not need to be root to run this test. Alternatively, you can run
the installation script directly, but this must be from within the build
directory. For example, from the top-level Exim directory you could use this
command:
.code
(cd build-SunOS5-5.5.1-sparc; ../scripts/exim_install -n)
.endd
.cindex "installing Exim" "install script options"
There are two other options that can be supplied to the installation script.

.ilist
&%-no_chown%& bypasses the call to change the owner of the installed binary
to root, and the call to make it a setuid binary.
.next
&%-no_symlink%& bypasses the setting up of the symbolic link &_exim_& to the
installed binary.
.endlist

INSTALL_ARG can be used to pass these options to the script. For example:
.code
make INSTALL_ARG=-no_symlink install
.endd
The installation script can also be given arguments specifying which files are
to be copied. For example, to install just the Exim binary, and nothing else,
without creating the symbolic link, you could use:
.code
make INSTALL_ARG='-no_symlink exim' install
.endd



.section "Installing info documentation" "SECTinsinfdoc"
.cindex "installing Exim" "&'info'& documentation"
Not all systems use the GNU &'info'& system for documentation, and for this
reason, the Texinfo source of Exim's documentation is not included in the main
distribution. Instead it is available separately from the FTP site (see section
&<<SECTavail>>&).

If you have defined INFO_DIRECTORY in &_Local/Makefile_& and the Texinfo
source of the documentation is found in the source tree, running &`make
install`& automatically builds the info files and installs them.



.section "Setting up the spool directory" "SECID33"
.cindex "spool directory" "creating"
When it starts up, Exim tries to create its spool directory if it does not
exist. The Exim uid and gid are used for the owner and group of the spool
directory. Sub-directories are automatically created in the spool directory as
necessary.




.section "Testing" "SECID34"
.cindex "testing" "installation"
Having installed Exim, you can check that the runtime configuration file is
syntactically valid by running the following command, which assumes that the
Exim binary directory is within your PATH environment variable:
.code
exim -bV
.endd
If there are any errors in the configuration file, Exim outputs error messages.
Otherwise it outputs the version number and build date,
the DBM library that is being used, and information about which drivers and
other optional code modules are included in the binary.
Some simple routing tests can be done by using the address testing option. For
example,
.display
&`exim -bt`& <&'local username'&>
.endd
should verify that it recognizes a local mailbox, and
.display
&`exim -bt`& <&'remote address'&>
.endd
a remote one. Then try getting it to deliver mail, both locally and remotely.
This can be done by passing messages directly to Exim, without going through a
user agent. For example:
.code
exim -v postmaster@your.domain.example
From: user@your.domain.example
To: postmaster@your.domain.example
Subject: Testing Exim

This is a test message.
^D
.endd
The &%-v%& option causes Exim to output some verification of what it is doing.
In this case you should see copies of three log lines, one for the message's
arrival, one for its delivery, and one containing &"Completed"&.

.cindex "delivery" "problems with"
If you encounter problems, look at Exim's log files (&'mainlog'& and
&'paniclog'&) to see if there is any relevant information there. Another source
of information is running Exim with debugging turned on, by specifying the
&%-d%& option. If a message is stuck on Exim's spool, you can force a delivery
with debugging turned on by a command of the form
.display
&`exim -d -M`& <&'exim-message-id'&>
.endd
You must be root or an &"admin user"& in order to do this. The &%-d%& option
produces rather a lot of output, but you can cut this down to specific areas.
For example, if you use &%-d-all+route%& only the debugging information
relevant to routing is included. (See the &%-d%& option in chapter
&<<CHAPcommandline>>& for more details.)

.cindex '&"sticky"& bit'
.cindex "lock files"
One specific problem that has shown up on some sites is the inability to do
local deliveries into a shared mailbox directory, because it does not have the
&"sticky bit"& set on it. By default, Exim tries to create a lock file before
writing to a mailbox file, and if it cannot create the lock file, the delivery
is deferred. You can get round this either by setting the &"sticky bit"& on the
directory, or by setting a specific group for local deliveries and allowing
that group to create files in the directory (see the comments above the
&(local_delivery)& transport in the default configuration file). Another
approach is to configure Exim not to use lock files, but just to rely on
&[fcntl()]& locking instead. However, you should do this only if all user
agents also use &[fcntl()]& locking. For further discussion of locking issues,
see chapter &<<CHAPappendfile>>&.

One thing that cannot be tested on a system that is already running an MTA is
the receipt of incoming SMTP mail on the standard SMTP port. However, the
&%-oX%& option can be used to run an Exim daemon that listens on some other
port, or &'inetd'& can be used to do this. The &%-bh%& option and the
&'exim_checkaccess'& utility can be used to check out policy controls on
incoming SMTP mail.

Testing a new version on a system that is already running Exim can most easily
be done by building a binary with a different CONFIGURE_FILE setting. From
within the runtime configuration, all other file and directory names
that Exim uses can be altered, in order to keep it entirely clear of the
production version.


.section "Replacing another MTA with Exim" "SECID35"
.cindex "replacing another MTA"
Building and installing Exim for the first time does not of itself put it in
general use. The name by which the system's MTA is called by mail user agents
is either &_/usr/sbin/sendmail_&, or &_/usr/lib/sendmail_& (depending on the
operating system), and it is necessary to make this name point to the &'exim'&
binary in order to get the user agents to pass messages to Exim. This is
normally done by renaming any existing file and making &_/usr/sbin/sendmail_&
or &_/usr/lib/sendmail_&
.cindex "symbolic link" "to &'exim'& binary"
a symbolic link to the &'exim'& binary. It is a good idea to remove any setuid
privilege and executable status from the old MTA. It is then necessary to stop
and restart the mailer daemon, if one is running.

.cindex "FreeBSD, MTA indirection"
.cindex "&_/etc/mail/mailer.conf_&"
Some operating systems have introduced alternative ways of switching MTAs. For
example, if you are running FreeBSD, you need to edit the file
&_/etc/mail/mailer.conf_& instead of setting up a symbolic link as just
described. A typical example of the contents of this file for running Exim is
as follows:
.code
sendmail            /usr/exim/bin/exim
send-mail           /usr/exim/bin/exim
mailq               /usr/exim/bin/exim -bp
newaliases          /usr/bin/true
.endd
Once you have set up the symbolic link, or edited &_/etc/mail/mailer.conf_&,
your Exim installation is &"live"&. Check it by sending a message from your
favourite user agent.

You should consider what to tell your users about the change of MTA. Exim may
have different capabilities to what was previously running, and there are
various operational differences such as the text of messages produced by
command line options and in bounce messages. If you allow your users to make
use of Exim's filtering capabilities, you should make the document entitled
&'Exim's interface to mail filtering'& available to them.



.section "Running the daemon" SECTdaemonLaunch
The most common command line for launching the Exim daemon looks like
.code
exim -bd -q5m
.endd
This starts a daemon which
.ilist
listens for incoming smtp connections, launching handler processes for
each new one
.next
starts a queue-runner process every five minutes, to inspect queued messages
and run delivery attempts on any that have arrived at their retry time
.endlist
Should a queue run take longer than the time between queue-runner starts,
they will run in parallel.
Numbers of jobs of the various types are subject to policy controls
defined in the configuration.


.section "Upgrading Exim" "SECID36"
.cindex "upgrading Exim"
If you are already running Exim on your host, building and installing a new
version automatically makes it available to MUAs, or any other programs that
call the MTA directly. However, if you are running an Exim daemon, you do need
.cindex restart "on HUP signal"
.cindex signal "HUP, to restart"
to send it a HUP signal, to make it re-execute itself, and thereby pick up the
new binary. You do not need to stop processing mail in order to install a new
version of Exim. The install script does not modify an existing runtime
configuration file.




.section "Stopping the Exim daemon on Solaris" "SECID37"
.cindex "Solaris" "stopping Exim on"
The standard command for stopping the mailer daemon on Solaris is
.code
/etc/init.d/sendmail stop
.endd
If &_/usr/lib/sendmail_& has been turned into a symbolic link, this script
fails to stop Exim because it uses the command &'ps -e'& and greps the output
for the text &"sendmail"&; this is not present because the actual program name
(that is, &"exim"&) is given by the &'ps'& command with these options. A
solution is to replace the line that finds the process id with something like
.code
pid=`cat /var/spool/exim/exim-daemon.pid`
.endd
to obtain the daemon's pid directly from the file that Exim saves it in.

Note, however, that stopping the daemon does not &"stop Exim"&. Messages can
still be received from local processes, and if automatic delivery is configured
(the normal case), deliveries will still occur.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The Exim command line" "CHAPcommandline"
.scindex IIDclo1 "command line" "options"
.scindex IIDclo2 "options" "command line"
Exim's command line takes the standard Unix form of a sequence of options,
each starting with a hyphen character, followed by a number of arguments. The
options are compatible with the main options of Sendmail, and there are also
some additional options, some of which are compatible with Smail 3. Certain
combinations of options do not make sense, and provoke an error if used.
The form of the arguments depends on which options are set.


.section "Setting options by program name" "SECID38"
.cindex "&'mailq'&"
If Exim is called under the name &'mailq'&, it behaves as if the option &%-bp%&
were present before any other options.
The &%-bp%& option requests a listing of the contents of the mail queue on the
standard output.
This feature is for compatibility with some systems that contain a command of
that name in one of the standard libraries, symbolically linked to
&_/usr/sbin/sendmail_& or &_/usr/lib/sendmail_&.

.cindex "&'rsmtp'&"
If Exim is called under the name &'rsmtp'& it behaves as if the option &%-bS%&
were present before any other options, for compatibility with Smail. The
&%-bS%& option is used for reading in a number of messages in batched SMTP
format.

.cindex "&'rmail'&"
If Exim is called under the name &'rmail'& it behaves as if the &%-i%& and
&%-oee%& options were present before any other options, for compatibility with
Smail. The name &'rmail'& is used as an interface by some UUCP systems.

.cindex "&'runq'&"
.cindex "queue runner"
If Exim is called under the name &'runq'& it behaves as if the option &%-q%&
were present before any other options, for compatibility with Smail. The &%-q%&
option causes a single queue runner process to be started.

.cindex "&'newaliases'&"
.cindex "alias file" "building"
.cindex "Sendmail compatibility" "calling Exim as &'newaliases'&"
If Exim is called under the name &'newaliases'& it behaves as if the option
&%-bi%& were present before any other options, for compatibility with Sendmail.
This option is used for rebuilding Sendmail's alias file. Exim does not have
the concept of a single alias file, but can be configured to run a given
command if called with the &%-bi%& option.


.section "Trusted and admin users" "SECTtrustedadmin"
Some Exim options are available only to &'trusted users'& and others are
available only to &'admin users'&. In the description below, the phrases &"Exim
user"& and &"Exim group"& mean the user and group defined by EXIM_USER and
EXIM_GROUP in &_Local/Makefile_& or set by the &%exim_user%& and
&%exim_group%& options. These do not necessarily have to use the name &"exim"&.

.ilist
.cindex "trusted users" "definition of"
.cindex "user" "trusted definition of"
The trusted users are root, the Exim user, any user listed in the
&%trusted_users%& configuration option, and any user whose current group or any
supplementary group is one of those listed in the &%trusted_groups%&
configuration option. Note that the Exim group is not automatically trusted.

.cindex '&"From"& line'
.cindex "envelope from"
.cindex "envelope sender"
Trusted users are always permitted to use the &%-f%& option or a leading
&"From&~"& line to specify the envelope sender of a message that is passed to
Exim through the local interface (see the &%-bm%& and &%-f%& options below).
See the &%untrusted_set_sender%& option for a way of permitting non-trusted
users to set envelope senders.

.chindex From:
.chindex Sender:
For a trusted user, there is never any check on the contents of the &'From:'&
header line, and a &'Sender:'& line is never added. Furthermore, any existing
&'Sender:'& line in incoming local (non-TCP/IP) messages is not removed.

Trusted users may also specify a host name, host address, interface address,
protocol name, ident value, and authentication data when submitting a message
locally. Thus, they are able to insert messages into Exim's queue locally that
have the characteristics of messages received from a remote host. Untrusted
users may in some circumstances use &%-f%&, but can never set the other values
that are available to trusted users.
.next
.cindex "user" "admin definition of"
.cindex "admin user" "definition of"
The admin users are root, the Exim user, and any user that is a member of the
Exim group or of any group listed in the &%admin_groups%& configuration option.
The current group does not have to be one of these groups.

Admin users are permitted to list the queue, and to carry out certain
operations on messages, for example, to force delivery failures. It is also
necessary to be an admin user in order to see the full information provided by
the Exim monitor, and full debugging output.

By default, the use of the &%-M%&, &%-q%&, &%-R%&, and &%-S%& options to cause
Exim to attempt delivery of messages on its queue is restricted to admin users.
However, this restriction can be relaxed by setting the &%prod_requires_admin%&
option false (that is, specifying &%no_prod_requires_admin%&).

Similarly, the use of the &%-bp%& option to list all the messages in the queue
is restricted to admin users unless &%queue_list_requires_admin%& is set
false.
.endlist


&*Warning*&: If you configure your system so that admin users are able to
edit Exim's configuration file, you are giving those users an easy way of
getting root. There is further discussion of this issue at the start of chapter
&<<CHAPconf>>&.




.section "Command line options" "SECID39"
Exim's command line options are described in alphabetical order below. If none
of the options that specifies a specific action (such as starting the daemon or
a queue runner, or testing an address, or receiving a message in a specific
format, or listing the queue) are present, and there is at least one argument
on the command line, &%-bm%& (accept a local message on the standard input,
with the arguments specifying the recipients) is assumed. Otherwise, Exim
outputs a brief message about itself and exits.

. ////////////////////////////////////////////////////////////////////////////
. Insert a stylized XML comment here, to identify the start of the command line
. options. This is for the benefit of the Perl script that automatically
. creates a man page for the options.
. ////////////////////////////////////////////////////////////////////////////

.literal xml
<!-- === Start of command line options === -->
.literal off


.vlist
.cmdopt "--" "--"
.cindex "options" "command line; terminating"
This is a pseudo-option whose only purpose is to terminate the options and
therefore to cause subsequent command line items to be treated as arguments
rather than options, even if they begin with hyphens.

.cmdopt --help
This option causes Exim to output a few sentences stating what it is.
The same output is generated if the Exim binary is called with no options and
no arguments.

.cmdopt --version
This option is an alias for &%-bV%& and causes version information to be
displayed.

.vitem &%-Ac%& &&&
       &%-Am%&
.oindex "&%-Ac%&"
.oindex "&%-Am%&"
These options are used by Sendmail for selecting configuration files and are
ignored by Exim.

.cmdopt -B <&'type'&>
.oindex "&%-B%&"
.cindex "8-bit characters"
.cindex "Sendmail compatibility" "8-bit characters"
This is a Sendmail option for selecting 7 or 8 bit processing. Exim is 8-bit
clean; it ignores this option.

.cmdopt -bd
.cindex "daemon"
.cindex "SMTP" "listener"
.cindex "queue runner"
This option runs Exim as a daemon, awaiting incoming SMTP connections. Usually
the &%-bd%& option is combined with the &%-q%&<&'time'&> option, to specify
that the daemon should also initiate periodic queue runs.

The &%-bd%& option can be used only by an admin user. If either of the &%-d%&
(debugging) or &%-v%& (verifying) options are set, the daemon does not
disconnect from the controlling terminal. When running this way, it can be
stopped by pressing ctrl-C.

By default, Exim listens for incoming connections to the standard SMTP port on
all the host's running interfaces. However, it is possible to listen on other
ports, on multiple ports, and only on specific interfaces. Chapter
&<<CHAPinterfaces>>& contains a description of the options that control this.

When a listening daemon
.cindex "daemon" "process id (pid)"
.cindex "pid (process id)" "of daemon"
is started without the use of &%-oX%& (that is, without overriding the normal
configuration), it writes its process id to a file called &_exim-daemon.pid_&
in Exim's spool directory. This location can be overridden by setting
PID_FILE_PATH in &_Local/Makefile_&. The file is written while Exim is still
running as root.

When &%-oX%& is used on the command line to start a listening daemon, the
process id is not written to the normal pid file path. However, &%-oP%& can be
used to specify a path on the command line if a pid file is required.

The SIGHUP signal
.cindex "SIGHUP"
.cindex restart "on HUP signal"
.cindex signal "HUP, to restart"
.cindex "daemon" "restarting"
.cindex signal "to reload configuration"
.cindex daemon "reload configuration"
.cindex reload configuration
can be used to cause the daemon to re-execute itself. This should be done
whenever Exim's configuration file, or any file that is incorporated into it by
means of the &%.include%& facility, is changed, and also whenever a new version
of Exim is installed. It is not necessary to do this when other files that are
referenced from the configuration (for example, alias files) are changed,
because these are reread each time they are used.

Either a SIGTERM or a SIGINT signal should be used to cause the daemon
to cleanly shut down.
Subprocesses handling recceiving or delivering messages,
or for scanning the queue,
will not be affected by the termination of the daemon process.

.cmdopt -bdf
This option has the same effect as &%-bd%& except that it never disconnects
from the controlling terminal, even when no debugging is specified.

.cmdopt -be
.cindex "testing" "string expansion"
.cindex "expansion" "testing"
Run Exim in expansion testing mode. Exim discards its root privilege, to
prevent ordinary users from using this mode to read otherwise inaccessible
files. If no arguments are given, Exim runs interactively, prompting for lines
of data. Otherwise, it processes each argument in turn.

If Exim was built with USE_READLINE=yes in &_Local/Makefile_&, it tries
to load the &%libreadline%& library dynamically whenever the &%-be%& option is
used without command line arguments. If successful, it uses the &[readline()]&
function, which provides extensive line-editing facilities, for reading the
test data. A line history is supported.

Long expansion expressions can be split over several lines by using backslash
continuations. As in Exim's runtime configuration, white space at the start of
continuation lines is ignored. Each argument or data line is passed through the
string expansion mechanism, and the result is output. Variable values from the
configuration file (for example, &$qualify_domain$&) are available, but no
message-specific values (such as &$message_exim_id$&) are set, because no message
is being processed (but see &%-bem%& and &%-Mset%&).

&*Note*&: If you use this mechanism to test lookups, and you change the data
files or databases you are using, you must exit and restart Exim before trying
the same lookup again. Otherwise, because each Exim process caches the results
of lookups, you will just get the same result as before.

Macro processing is done on lines before string-expansion: new macros can be
defined and macros will be expanded.
Because macros in the config file are often used for secrets, those are only
available to admin users.

The word &"set"& at the start of a line, followed by a single space,
is recognised specially as defining a value for a variable.
.cindex "tainted data" "expansion testing"
If the sequence &",t"& is inserted before the space,
the value is marked as tainted.
The syntax is otherwise the same as the ACL modifier &"set ="&.

.cmdopt -bem <&'filename'&>
.cindex "testing" "string expansion"
.cindex "expansion" "testing"
This option operates like &%-be%& except that it must be followed by the name
of a file. For example:
.code
exim -bem /tmp/testmessage
.endd
The file is read as a message (as if receiving a locally-submitted non-SMTP
message) before any of the test expansions are done. Thus, message-specific
variables such as &$message_size$& and &$header_from:$& are available. However,
no &'Received:'& header is added to the message. If the &%-t%& option is set,
recipients are read from the headers in the normal way, and are shown in the
&$recipients$& variable. Note that recipients cannot be given on the command
line, because further arguments are taken as strings to expand (just like
&%-be%&).

.cmdopt -bF <&'filename'&>
.cindex "system filter" "testing"
.cindex "testing" "system filter"
This option is the same as &%-bf%& except that it assumes that the filter being
tested is a system filter. The additional commands that are available only in
system filters are recognized.

.cmdopt -bf <&'filename'&>
.cindex "filter" "testing"
.cindex "testing" "filter file"
.cindex "forward file" "testing"
.cindex "testing" "forward file"
.cindex "Sieve filter" "testing"
This option runs Exim in user filter testing mode; the file is the filter file
to be tested, and a test message must be supplied on the standard input. If
there are no message-dependent tests in the filter, an empty file can be
supplied.

If you want to test a system filter file, use &%-bF%& instead of &%-bf%&. You
can use both &%-bF%& and &%-bf%& on the same command, in order to test a system
filter and a user filter in the same run. For example:
.code
exim -bF /system/filter -bf /user/filter </test/message
.endd
This is helpful when the system filter adds header lines or sets filter
variables that are used by the user filter.

If the test filter file does not begin with one of the special lines
.code
# Exim filter
# Sieve filter
.endd
it is taken to be a normal &_.forward_& file, and is tested for validity under
that interpretation. See sections &<<SECTitenonfilred>>& to
&<<SECTspecitredli>>& for a description of the possible contents of non-filter
redirection lists.

The result of an Exim command that uses &%-bf%&, provided no errors are
detected, is a list of the actions that Exim would try to take if presented
with the message for real. More details of filter testing are given in the
separate document entitled &'Exim's interfaces to mail filtering'&.

When testing a filter file,
.cindex "&""From""& line"
.cindex "envelope from"
.cindex "envelope sender"
.oindex "&%-f%&" "for filter testing"
the envelope sender can be set by the &%-f%& option,
or by a &"From&~"& line at the start of the test message. Various parameters
that would normally be taken from the envelope recipient address of the message
can be set by means of additional command line options (see the next four
options).

.cmdopt -bfd <&'domain'&>
.vindex "&$qualify_domain$&"
This sets the domain of the recipient address when a filter file is being
tested by means of the &%-bf%& option. The default is the value of
&$qualify_domain$&.

.cmdopt -bfl <&'local&~part'&>
This sets the local part of the recipient address when a filter file is being
tested by means of the &%-bf%& option. The default is the username of the
process that calls Exim. A local part should be specified with any prefix or
suffix stripped, because that is how it appears to the filter when a message is
actually being delivered.

.cmdopt -bfp <&'prefix'&>
.cindex affix "filter testing"
This sets the prefix of the local part of the recipient address when a filter
file is being tested by means of the &%-bf%& option. The default is an empty
prefix.

.cmdopt -bfs <&'suffix'&>
.cindex affix "filter testing"
This sets the suffix of the local part of the recipient address when a filter
file is being tested by means of the &%-bf%& option. The default is an empty
suffix.

.cmdopt -bh <&'IP&~address'&>
.cindex "testing" "incoming SMTP"
.cindex "SMTP" "testing incoming"
.cindex "testing" "relay control"
.cindex "relaying" "testing configuration"
.cindex "policy control" "testing"
.cindex "debugging" "&%-bh%& option"
This option runs a fake SMTP session as if from the given IP address, using the
standard input and output. The IP address may include a port number at the end,
after a full stop. For example:
.code
exim -bh 10.9.8.7.1234
exim -bh fe80::a00:20ff:fe86:a061.5678
.endd
When an IPv6 address is given, it is converted into canonical form. In the case
of the second example above, the value of &$sender_host_address$& after
conversion to the canonical form is
&`fe80:0000:0000:0a00:20ff:fe86:a061.5678`&.

Comments as to what is going on are written to the standard error file. These
include lines beginning with &"LOG"& for anything that would have been logged.
This facility is provided for testing configuration options for incoming
messages, to make sure they implement the required policy. For example, you can
test your relay controls using &%-bh%&.

&*Warning 1*&:
.cindex "RFC 1413"
You can test features of the configuration that rely on ident
(&url(https://www.rfc-editor.org/rfc/rfc2487,RFC 1413))
information by using the &%-oMt%& option. However, Exim cannot actually perform
an ident callout when testing using &%-bh%& because there is no incoming SMTP
connection.

&*Warning 2*&: Address verification callouts (see section &<<SECTcallver>>&)
are also skipped when testing using &%-bh%&. If you want these callouts to
occur, use &%-bhc%& instead.

Messages supplied during the testing session are discarded, and nothing is
written to any of the real log files. There may be pauses when DNS (and other)
lookups are taking place, and of course these may time out. The &%-oMi%& option
can be used to specify a specific IP interface and port if this is important,
and &%-oMaa%& and &%-oMai%& can be used to set parameters as if the SMTP
session were authenticated.

The &'exim_checkaccess'& utility is a &"packaged"& version of &%-bh%& whose
output just states whether a given recipient address from a given host is
acceptable or not. See section &<<SECTcheckaccess>>&.

Features such as authentication and encryption, where the client input is not
plain text, cannot easily be tested with &%-bh%&. Instead, you should use a
specialized SMTP test program such as
&url(https://www.jetmore.org/john/code/swaks/,swaks,swaks).

.cmdopt -bhc <&'IP&~address'&>
This option operates in the same way as &%-bh%&, except that address
verification callouts are performed if required. This includes consulting and
updating the callout cache database.

.cmdopt -bi
.cindex "alias file" "building"
.cindex "building alias file"
.cindex "Sendmail compatibility" "&%-bi%& option"
Sendmail interprets the &%-bi%& option as a request to rebuild its alias file.
Exim does not have the concept of a single alias file, and so it cannot mimic
this behaviour. However, calls to &_/usr/lib/sendmail_& with the &%-bi%& option
tend to appear in various scripts such as NIS make files, so the option must be
recognized.

If &%-bi%& is encountered, the command specified by the &%bi_command%&
configuration option is run, under the uid and gid of the caller of Exim. If
the &%-oA%& option is used, its value is passed to the command as an argument.
The command set by &%bi_command%& may not contain arguments. The command can
use the &'exim_dbmbuild'& utility, or some other means, to rebuild alias files
if this is required. If the &%bi_command%& option is not set, calling Exim with
&%-bi%& is a no-op.

. // Keep :help first, then the rest in alphabetical order
.cmdopt -bI:help
.cindex "querying exim information"
We shall provide various options starting &`-bI:`& for querying Exim for
information.  The output of many of these will be intended for machine
consumption.  This one is not.  The &%-bI:help%& option asks Exim for a
synopsis of supported options beginning &`-bI:`&.  Use of any of these
options shall cause Exim to exit after producing the requested output.

.cmdopt -bI:dscp
.cindex "DSCP" "values"
This option causes Exim to emit an alphabetically sorted list of all
recognised DSCP names.

.cmdopt -bI:sieve
.cindex "Sieve filter" "capabilities"
This option causes Exim to emit an alphabetically sorted list of all supported
Sieve protocol extensions on stdout, one per line.  This is anticipated to be
useful for ManageSieve
(&url(https://datatracker.ietf.org/doc/html/rfc5804.html,RFC 5804))
implementations, in providing that protocol's
&`SIEVE`& capability response line.  As the precise list may depend upon
compile-time build options, which this option will adapt to, this is the only
way to guarantee a correct response.

.cmdopt -bm
.cindex "local message reception"
This option runs an Exim receiving process that accepts an incoming,
locally-generated message on the standard input. The recipients are given as the
command arguments (except when &%-t%& is also present &-- see below). Each
argument can be a comma-separated list of
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822) addresses.
This is the
default option for selecting the overall action of an Exim call; it is assumed
if no other conflicting option is present.

If any addresses in the message are unqualified (have no domain), they are
qualified by the values of the &%qualify_domain%& or &%qualify_recipient%&
options, as appropriate. The &%-bnq%& option (see below) provides a way of
suppressing this for special cases.

Policy checks on the contents of local messages can be enforced by means of
the non-SMTP ACL. See section &<<SECnonSMTP>>& for details.

.cindex "return code" "for &%-bm%&"
The return code is zero if the message is successfully accepted. Otherwise, the
action is controlled by the &%-oe%&&'x'& option setting &-- see below.

The format
.cindex "message" "format"
.cindex "format" "message"
.cindex "&""From""& line"
.cindex "UUCP" "&""From""& line"
.cindex "Sendmail compatibility" "&""From""& line"
of the message must be as defined in
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822),
except that, for
compatibility with Sendmail and Smail, a line in one of the forms
.code
From sender Fri Jan  5 12:55 GMT 1997
From sender Fri, 5 Jan 97 12:55:01
.endd
(with the weekday optional, and possibly with additional text after the date)
is permitted to appear at the start of the message. There appears to be no
authoritative specification of the format of this line. Exim recognizes it by
matching against the regular expression defined by the &%uucp_from_pattern%&
option, which can be changed if necessary.

.oindex "&%-f%&" "overriding &""From""& line"
The specified sender is treated as if it were given as the argument to the
&%-f%& option, but if a &%-f%& option is also present, its argument is used in
preference to the address taken from the message. The caller of Exim must be a
trusted user for the sender of a message to be set in this way.

.cmdopt -bmalware <&'filename'&>
.cindex "testing", "malware"
.cindex "malware scan test"
This debugging option causes Exim to scan the given file or directory
(depending on the used scanner interface),
using the malware scanning framework.  The option of &%av_scanner%& influences
this option, so if &%av_scanner%&'s value is dependent upon an expansion then
the expansion should have defaults which apply to this invocation.  ACLs are
not invoked, so if &%av_scanner%& references an ACL variable then that variable
will never be populated and &%-bmalware%& will fail.

Exim will have changed working directory before resolving the filename, so
using fully qualified pathnames is advisable.  Exim will be running as the Exim
user when it tries to open the file, rather than as the invoking user.
This option requires admin privileges.

The &%-bmalware%& option will not be extended to be more generally useful,
there are better tools for file-scanning.  This option exists to help
administrators verify their Exim and AV scanner configuration.

.cmdopt -bnq
.cindex "address qualification, suppressing"
By default, Exim automatically qualifies unqualified addresses (those
without domains) that appear in messages that are submitted locally (that
is, not over TCP/IP). This qualification applies both to addresses in
envelopes, and addresses in header lines. Sender addresses are qualified using
&%qualify_domain%&, and recipient addresses using &%qualify_recipient%& (which
defaults to the value of &%qualify_domain%&).

Sometimes, qualification is not wanted. For example, if &%-bS%& (batch SMTP) is
being used to re-submit messages that originally came from remote hosts after
content scanning, you probably do not want to qualify unqualified addresses in
header lines. (Such lines will be present only if you have not enabled a header
syntax check in the appropriate ACL.)

The &%-bnq%& option suppresses all qualification of unqualified addresses in
messages that originate on the local host. When this is used, unqualified
addresses in the envelope provoke errors (causing message rejection) and
unqualified addresses in header lines are left alone.


.cmdopt -bP
.cindex "configuration options" "extracting"
.cindex "options" "configuration &-- extracting"
If this option is given with no arguments, it causes the values of all Exim's
main configuration options to be written to the standard output. The values
of one or more specific options can be requested by giving their names as
arguments, for example:
.code
exim -bP qualify_domain hold_domains
.endd
.cindex "hiding configuration option values"
.cindex "configuration options" "hiding value of"
.cindex "options" "hiding value of"
However, any option setting that is preceded by the word &"hide"& in the
configuration file is not shown in full, except to an admin user. For other
users, the output is as in this example:
.code
mysql_servers = <value not displayable>
.endd
If &%config%& is given as an argument, the config is
output, as it was parsed, any include file resolved, any comment removed.

If &%config_file%& is given as an argument, the name of the runtime
configuration file is output. (&%configure_file%& works too, for
backward compatibility.)
If a list of configuration files was supplied, the value that is output here
is the name of the file that was actually used.

.cindex "options" "hiding name of"
If the &%-n%& flag is given, then for most modes of &%-bP%& operation the
name will not be output.

.cindex "daemon" "process id (pid)"
.cindex "pid (process id)" "of daemon"
If &%log_file_path%& or &%pid_file_path%& are given, the names of the
directories where log files and daemon pid files are written are output,
respectively. If these values are unset, log files are written in a
sub-directory of the spool directory called &%log%&, and the pid file is
written directly into the spool directory.

If &%-bP%& is followed by a name preceded by &`+`&, for example,
.code
exim -bP +local_domains
.endd
it searches for a matching named list of any type (domain, host, address, or
local part) and outputs what it finds.

.cindex "options" "router &-- extracting"
.cindex "options" "transport &-- extracting"
.cindex "options" "authenticator &-- extracting"
If one of the words &%router%&, &%transport%&, or &%authenticator%& is given,
followed by the name of an appropriate driver instance, the option settings for
that driver are output. For example:
.code
exim -bP transport local_delivery
.endd
The generic driver options are output first, followed by the driver's private
options. A list of the names of drivers of a particular type can be obtained by
using one of the words &%router_list%&, &%transport_list%&, or
&%authenticator_list%&, and a complete list of all drivers with their option
settings can be obtained by using &%routers%&, &%transports%&, or
&%authenticators%&.

.cindex "environment"
If &%environment%& is given as an argument, the set of environment
variables is output, line by line. Using the &%-n%& flag suppresses the value of the
variables.

.cindex "options" "macro &-- extracting"
If invoked by an admin user, then &%macro%&, &%macro_list%& and &%macros%&
are available, similarly to the drivers.  Because macros are sometimes used
for storing passwords, this option is restricted.
The output format is one item per line.
For the "-bP macro <name>" form, if no such macro is found
the exit status will be nonzero.

.cmdopt -bp
.cindex "queue" "listing messages in"
.cindex "listing" "messages in the queue"
This option requests a listing of the contents of the mail queue on the
standard output. If the &%-bp%& option is followed by a list of message ids,
just those messages are listed. By default, this option can be used only by an
admin user. However, the &%queue_list_requires_admin%& option can be set false
to allow any user to see the queue.

Each message in the queue is displayed as in the following example:
.code
25m  2.9K 0t5C6f-0000c8-00 <alice@wonderland.fict.example>
          red.king@looking-glass.fict.example
          <other addresses>
.endd
.cindex "message" "size in queue listing"
.cindex "size" "of message"
The first line contains the length of time the message has been in the queue
(in this case 25 minutes), the size of the message (2.9K), the unique local
identifier for the message, and the message sender, as contained in the
envelope. For bounce messages, the sender address is empty, and appears as
&"<>"&. If the message was submitted locally by an untrusted user who overrode
the default sender address, the user's login name is shown in parentheses
before the sender address.

.cindex "frozen messages" "in queue listing"
If the message is frozen (attempts to deliver it are suspended) then the text
&"*** frozen ***"& is displayed at the end of this line.

The recipients of the message (taken from the envelope, not the headers) are
displayed on subsequent lines. Those addresses to which the message has already
been delivered are marked with the letter D. If an original address gets
expanded into several addresses via an alias or forward file, the original is
displayed with a D only when deliveries for all of its child addresses are
complete.


.cmdopt -bpa
This option operates like &%-bp%&, but in addition it shows delivered addresses
that were generated from the original top level address(es) in each message by
alias or forwarding operations. These addresses are flagged with &"+D"& instead
of just &"D"&.


.cmdopt -bpc
.cindex "queue" "count of messages on"
This option counts the number of messages in the queue, and writes the total
to the standard output. It is restricted to admin users, unless
&%queue_list_requires_admin%& is set false.


.cmdopt -bpi
.cindex queue "list of message IDs"
This option operates like &%-bp%&, but only outputs message ids
(one per line).


.cmdopt -bpr
This option operates like &%-bp%&, but the output is not sorted into
chronological order of message arrival. This can speed it up when there are
lots of messages in the queue, and is particularly useful if the output is
going to be post-processed in a way that doesn't need the sorting.

.cmdopt -bpra
This option is a combination of &%-bpr%& and &%-bpa%&.

.cmdopt -bpri
This option is a combination of &%-bpr%& and &%-bpi%&.

.cmdopt -bpru
This option is a combination of &%-bpr%& and &%-bpu%&.


.cmdopt -bpu
This option operates like &%-bp%& but shows only undelivered top-level
addresses for each message displayed. Addresses generated by aliasing or
forwarding are not shown, unless the message was deferred after processing by a
router with the &%one_time%& option set.


.cmdopt -brt
.cindex "testing" "retry configuration"
.cindex "retry" "configuration testing"
This option is for testing retry rules, and it must be followed by up to three
arguments. It causes Exim to look for a retry rule that matches the values
and to write it to the standard output. For example:
.code
exim -brt bach.comp.mus.example
Retry rule: *.comp.mus.example  F,2h,15m; F,4d,30m;
.endd
See chapter &<<CHAPretry>>& for a description of Exim's retry rules. The first
argument, which is required, can be a complete address in the form
&'local_part@domain'&, or it can be just a domain name. If the second argument
contains a dot, it is interpreted as an optional second domain name; if no
retry rule is found for the first argument, the second is tried. This ties in
with Exim's behaviour when looking for retry rules for remote hosts &-- if no
rule is found that matches the host, one that matches the mail domain is
sought. Finally, an argument that is the name of a specific delivery error, as
used in setting up retry rules, can be given. For example:
.code
exim -brt haydn.comp.mus.example quota_3d
Retry rule: *@haydn.comp.mus.example quota_3d  F,1h,15m
.endd

.cmdopt -brw
.cindex "testing" "rewriting"
.cindex "rewriting" "testing"
This option is for testing address rewriting rules, and it must be followed by
a single argument, consisting of either a local part without a domain, or a
complete address with a fully qualified domain. Exim outputs how this address
would be rewritten for each possible place it might appear. See chapter
&<<CHAPrewrite>>& for further details.

.cmdopt -bS
.cindex "SMTP" "batched incoming"
.cindex "batched SMTP input"
This option is used for batched SMTP input, which is an alternative interface
for non-interactive local message submission. A number of messages can be
submitted in a single run. However, despite its name, this is not really SMTP
input. Exim reads each message's envelope from SMTP commands on the standard
input, but generates no responses. If the caller is trusted, or
&%untrusted_set_sender%& is set, the senders in the SMTP MAIL commands are
believed; otherwise the sender is always the caller of Exim.

The message itself is read from the standard input, in SMTP format (leading
dots doubled), terminated by a line containing just a single dot. An error is
provoked if the terminating dot is missing. A further message may then follow.

As for other local message submissions, the contents of incoming batch SMTP
messages can be checked using the non-SMTP ACL (see section &<<SECnonSMTP>>&).
Unqualified addresses are automatically qualified using &%qualify_domain%& and
&%qualify_recipient%&, as appropriate, unless the &%-bnq%& option is used.

Some other SMTP commands are recognized in the input. HELO and EHLO act
as RSET; VRFY, EXPN, ETRN, ATRN, and HELP act as NOOP;
QUIT quits, ignoring the rest of the standard input.

.cindex "return code" "for &%-bS%&"
If any error is encountered, reports are written to the standard output and
error streams, and Exim gives up immediately. The return code is 0 if no error
was detected; it is 1 if one or more messages were accepted before the error
was detected; otherwise it is 2.

More details of input using batched SMTP are given in section
&<<SECTincomingbatchedSMTP>>&.

.cmdopt -bs
.cindex "SMTP" "local input"
.cindex "local SMTP input"
This option causes Exim to accept one or more messages by reading SMTP commands
on the standard input, and producing SMTP replies on the standard output. SMTP
policy controls, as defined in ACLs (see chapter &<<CHAPACL>>&) are applied.
Some user agents use this interface as a way of passing locally-generated
messages to the MTA.

In
.cindex "sender" "source of"
this usage, if the caller of Exim is trusted, or &%untrusted_set_sender%& is
set, the senders of messages are taken from the SMTP MAIL commands.
Otherwise the content of these commands is ignored and the sender is set up as
the calling user. Unqualified addresses are automatically qualified using
&%qualify_domain%& and &%qualify_recipient%&, as appropriate, unless the
&%-bnq%& option is used.

.cindex "inetd"
The
&%-bs%& option is also used to run Exim from &'inetd'&, as an alternative to
using a listening daemon. Exim can distinguish the two cases by checking
whether the standard input is a TCP/IP socket. When Exim is called from
&'inetd'&, the source of the mail is assumed to be remote, and the comments
above concerning senders and qualification do not apply. In this situation,
Exim behaves in exactly the same way as it does when receiving a message via
the listening daemon.

.cmdopt -bt
.cindex "testing" "addresses"
.cindex "address" "testing"
This option runs Exim in address testing mode, in which each argument is taken
as a recipient address to be tested for deliverability. The results are
written to the standard output. If a test fails, and the caller is not an admin
user, no details of the failure are output, because these might contain
sensitive information such as usernames and passwords for database lookups.

If no arguments are given, Exim runs in an interactive manner, prompting with a
right angle bracket for addresses to be tested.

Unlike the &%-be%& test option, you cannot arrange for Exim to use the
&[readline()]& function, because it is running as &'root'& and there are
security issues.

Each address is handled as if it were the recipient address of a message
(compare the &%-bv%& option). It is passed to the routers and the result is
written to the standard output. However, any router that has
&%no_address_test%& set is bypassed. This can make &%-bt%& easier to use for
genuine routing tests if your first router passes everything to a scanner
program.

.cindex "return code" "for &%-bt%&"
The return code is 2 if any address failed outright; it is 1 if no address
failed outright but at least one could not be resolved for some reason. Return
code 0 is given only when all addresses succeed.

.cindex "duplicate addresses"
&*Note*&: When actually delivering a message, Exim removes duplicate recipient
addresses after routing is complete, so that only one delivery takes place.
This does not happen when testing with &%-bt%&; the full results of routing are
always shown.

&*Warning*&: &%-bt%& can only do relatively simple testing. If any of the
routers in the configuration makes any tests on the sender address of a
message,
.oindex "&%-f%&" "for address testing"
you can use the &%-f%& option to set an appropriate sender when running
&%-bt%& tests. Without it, the sender is assumed to be the calling user at the
default qualifying domain. However, if you have set up (for example) routers
whose behaviour depends on the contents of an incoming message, you cannot test
those conditions using &%-bt%&. The &%-N%& option provides a possible way of
doing such tests.

.cmdopt -bV
.cindex "version number of Exim"
This option causes Exim to write the current version number, compilation
number, and compilation date of the &'exim'& binary to the standard output.
It also lists the DBM library that is being used, the optional modules (such as
specific lookup types), the drivers that are included in the binary, and the
name of the runtime configuration file that is in use.

As part of its operation, &%-bV%& causes Exim to read and syntax check its
configuration file. However, this is a static check only. It cannot check
values that are to be expanded. For example, although a misspelt ACL verb is
detected, an error in the verb's arguments is not. You cannot rely on &%-bV%&
alone to discover (for example) all the typos in the configuration; some
realistic testing is needed. The &%-bh%& and &%-N%& options provide more
dynamic testing facilities.

.cmdopt -bv
.cindex "verifying address" "using &%-bv%&"
.cindex "address" "verification"
This option runs Exim in address verification mode, in which each argument is
taken as a recipient address to be verified by the routers. (This does
not involve any verification callouts). During normal operation, verification
happens mostly as a consequence processing a &%verify%& condition in an ACL
(see chapter &<<CHAPACL>>&). If you want to test an entire ACL, possibly
including callouts, see the &%-bh%& and &%-bhc%& options.

If verification fails, and the caller is not an admin user, no details of the
failure are output, because these might contain sensitive information such as
usernames and passwords for database lookups.

If no arguments are given, Exim runs in an interactive manner, prompting with a
right angle bracket for addresses to be verified.

Unlike the &%-be%& test option, you cannot arrange for Exim to use the
&[readline()]& function, because it is running as &'exim'& and there are
security issues.

Verification differs from address testing (the &%-bt%& option) in that routers
that have &%no_verify%& set are skipped, and if the address is accepted by a
router that has &%fail_verify%& set, verification fails. The address is
verified as a recipient if &%-bv%& is used; to test verification for a sender
address, &%-bvs%& should be used.

If the &%-v%& option is not set, the output consists of a single line for each
address, stating whether it was verified or not, and giving a reason in the
latter case. Without &%-v%&, generating more than one address by redirection
causes verification to end successfully, without considering the generated
addresses. However, if just one address is generated, processing continues,
and the generated address must verify successfully for the overall verification
to succeed.

When &%-v%& is set, more details are given of how the address has been handled,
and in the case of address redirection, all the generated addresses are also
considered. Verification may succeed for some and fail for others.

The
.cindex "return code" "for &%-bv%&"
return code is 2 if any address failed outright; it is 1 if no address
failed outright but at least one could not be resolved for some reason. Return
code 0 is given only when all addresses succeed.

If any of the routers in the configuration makes any tests on the sender
address of a message, you should use the &%-f%& option to set an appropriate
sender when running &%-bv%& tests. Without it, the sender is assumed to be the
calling user at the default qualifying domain.

.cmdopt -bvs
This option acts like &%-bv%&, but verifies the address as a sender rather
than a recipient address. This affects any rewriting and qualification that
might happen.

.cmdopt -bw
.cindex "daemon"
.cindex "inetd"
.cindex "inetd" "wait mode"
This option runs Exim as a daemon, awaiting incoming SMTP connections,
similarly to the &%-bd%& option.  All port specifications on the command-line
and in the configuration file are ignored.  Queue-running may not be specified.

In this mode, Exim expects to be passed a socket as fd 0 (stdin) which is
listening for connections.  This permits the system to start up and have
inetd (or equivalent) listen on the SMTP ports, starting an Exim daemon for
each port only when the first connection is received.

If the option is given as &%-bw%&<&'time'&> then the time is a timeout, after
which the daemon will exit, which should cause inetd to listen once more.

.cmdopt -C <&'filelist'&>
.cindex "configuration file" "alternate"
.cindex "CONFIGURE_FILE"
.cindex "alternate configuration file"
This option causes Exim to find the runtime configuration file from the given
list instead of from the list specified by the CONFIGURE_FILE
compile-time setting. Usually, the list will consist of just a single filename,
but it can be a colon-separated list of names. In this case, the first
file that exists is used. Failure to open an existing file stops Exim from
proceeding any further along the list, and an error is generated.

When this option is used by a caller other than root, and the list is different
from the compiled-in list, Exim gives up its root privilege immediately, and
runs with the real and effective uid and gid set to those of the caller.
However, if a TRUSTED_CONFIG_LIST file is defined in &_Local/Makefile_&, that
file contains a list of full pathnames, one per line, for configuration files
which are trusted. Root privilege is retained for any configuration file so
listed, as long as the caller is the Exim user (or the user specified in the
CONFIGURE_OWNER option, if any), and as long as the configuration file is
not writeable by inappropriate users or groups.

Leaving TRUSTED_CONFIG_LIST unset precludes the possibility of testing a
configuration using &%-C%& right through message reception and delivery,
even if the caller is root. The reception works, but by that time, Exim is
running as the Exim user, so when it re-executes to regain privilege for the
delivery, the use of &%-C%& causes privilege to be lost. However, root can
test reception and delivery using two separate commands (one to put a message
in the queue, using &%-odq%&, and another to do the delivery, using &%-M%&).

If ALT_CONFIG_PREFIX is defined &_in Local/Makefile_&, it specifies a
prefix string with which any file named in a &%-C%& command line option
must start. In addition, the filename must not contain the sequence &`/../`&.
However, if the value of the &%-C%& option is identical to the value of
CONFIGURE_FILE in &_Local/Makefile_&, Exim ignores &%-C%& and proceeds as
usual. There is no default setting for ALT_CONFIG_PREFIX; when it is
unset, any filename can be used with &%-C%&.

ALT_CONFIG_PREFIX can be used to confine alternative configuration files
to a directory to which only root has access. This prevents someone who has
broken into the Exim account from running a privileged Exim with an arbitrary
configuration file.

The &%-C%& facility is useful for ensuring that configuration files are
syntactically correct, but cannot be used for test deliveries, unless the
caller is privileged, or unless it is an exotic configuration that does not
require privilege. No check is made on the owner or group of the files
specified by this option.


.vitem &%-D%&<&'macro'&>=<&'value'&>
.oindex "&%-D%&"
.cindex "macro" "setting on command line"
This option can be used to override macro definitions in the configuration file
(see section &<<SECTmacrodefs>>&). However, like &%-C%&, if it is used by an
unprivileged caller, it causes Exim to give up its root privilege.
If DISABLE_D_OPTION is defined in &_Local/Makefile_&, the use of &%-D%& is
completely disabled, and its use causes an immediate error exit.

If WHITELIST_D_MACROS is defined in &_Local/Makefile_& then it should be a
colon-separated list of macros which are considered safe and, if &%-D%& only
supplies macros from this list, and the values are acceptable, then Exim will
not give up root privilege if the caller is root, the Exim run-time user, or
the CONFIGURE_OWNER, if set.  This is a transition mechanism and is expected
to be removed in the future.  Acceptable values for the macros satisfy the
regexp: &`^[A-Za-z0-9_/.-]*$`&

The entire option (including equals sign if present) must all be within one
command line item. &%-D%& can be used to set the value of a macro to the empty
string, in which case the equals sign is optional. These two commands are
synonymous:
.code
exim -DABC  ...
exim -DABC= ...
.endd
To include spaces in a macro definition item, quotes must be used. If you use
quotes, spaces are permitted around the macro name and the equals sign. For
example:
.code
exim '-D ABC = something' ...
.endd
&%-D%& may be repeated up to 10 times on a command line.
Only macro names up to 22 letters long can be set.


.vitem &%-d%&<&'debug&~options'&>
.oindex "&%-d%&"
.cindex "debugging" "list of selectors"
.cindex "debugging" "&%-d%& option"
This option causes debugging information to be written to the standard
error stream. It is restricted to admin users because debugging output may show
database queries that contain password information. Also, the details of users'
filter files should be protected. If a non-admin user uses &%-d%&, Exim
writes an error message to the standard error stream and exits with a non-zero
return code.

When &%-d%& is used, &%-v%& is assumed. If &%-d%& is given on its own, a lot of
standard debugging data is output. This can be reduced, or increased to include
some more rarely needed information, by directly following &%-d%& with a string
made up of names preceded by plus or minus characters. These add or remove sets
of debugging data, respectively. For example, &%-d+filter%& adds filter
debugging, whereas &%-d-all+filter%& selects only filter debugging. Note that
no spaces are allowed in the debug setting. The available debugging categories
are:
.itable none 0 0 2 20* left 80* left
.irow acl            "ACL interpretation"
.irow auth           "authenticators"
.irow deliver        "general delivery logic"
.irow dns            "DNS lookups (see also resolver)"
.irow dnsbl          "DNS black list (aka RBL) code"
.irow exec           "arguments for &[execv()]& calls"
.irow expand         "detailed debugging for string expansions"
.irow filter         "filter handling"
.irow hints_lookup   "hints data lookups"
.irow host_lookup    "all types of name-to-IP address handling"
.irow ident          "ident lookup"
.irow interface      "lists of local interfaces"
.irow lists          "matching things in lists"
.irow load           "system load checks"
.irow local_scan     "can be used by &[local_scan()]& (see chapter &&&
                    &<<CHAPlocalscan>>&)"
.irow lookup         "general lookup code and all lookups"
.irow memory         "memory handling"
.irow noutf8         "modifier: avoid UTF-8 line-drawing"
.irow pid            "modifier: add pid to debug output lines"
.irow process_info   "setting info for the process log"
.irow queue_run      "queue runs"
.irow receive        "general message reception logic"
.irow resolver       "turn on the DNS resolver's debugging output"
.irow retry          "retry handling"
.irow rewrite        "address rewriting""
.irow route          "address routing"
.irow timestamp      "modifier: add timestamp to debug output lines"
.irow tls            "TLS logic"
.irow transport      "transports"
.irow uid            "changes of uid/gid and looking up uid/gid"
.irow verify         "address verification logic"
.irow all            "almost all of the above (see below), and also &%-v%&"
.endtable
The &`all`& option excludes &`memory`& when used as &`+all`&, but includes it
for &`-all`&. The reason for this is that &`+all`& is something that people
tend to use when generating debug output for Exim maintainers. If &`+memory`&
is included, an awful lot of output that is very rarely of interest is
generated, so it now has to be explicitly requested. However, &`-all`& does
turn everything off.

.cindex "resolver, debugging output"
.cindex "DNS resolver, debugging output"
The &`resolver`& option produces output only if the DNS resolver was compiled
with DEBUG enabled. This is not the case in some operating systems. Also,
unfortunately, debugging output from the DNS resolver is written to stdout
rather than stderr.

The default (&%-d%& with no argument) omits &`expand`&, &`filter`&,
&`interface`&, &`load`&, &`memory`&, &`pid`&, &`resolver`&, and &`timestamp`&.
However, the &`pid`& selector is forced when debugging is turned on for a
daemon, which then passes it on to any re-executed Exims. Exim also
automatically adds the pid to debug lines when several remote deliveries are
run in parallel.

The &`timestamp`& selector causes the current time to be inserted at the start
of all debug output lines. This can be useful when trying to track down delays
in processing.

.cindex debugging "UTF-8 in"
.cindex UTF-8 "in debug output"
The &`noutf8`& selector disables the use of
UTF-8 line-drawing characters to group related information.
When disabled. ascii-art is used instead.
Using the &`+all`& option does not set this modifier,

If the &%debug_print%& option is set in any driver, it produces output whenever
any debugging is selected, or if &%-v%& is used.

.vitem &%-dd%&<&'debug&~options'&>
.oindex "&%-dd%&"
This option behaves exactly like &%-d%& except when used on a command that
starts a daemon process. In that case, debugging is turned off for the
subprocesses that the daemon creates. Thus, it is useful for monitoring the
behaviour of the daemon without creating as much output as full debugging does.

.cmdopt -dropcr
This is an obsolete option that is now a no-op. It used to affect the way Exim
handled CR and LF characters in incoming messages. What happens now is
described in section &<<SECTlineendings>>&.

.cmdopt -E
.cindex "bounce message" "generating"
This option specifies that an incoming message is a locally-generated delivery
failure report. It is used internally by Exim when handling delivery failures
and is not intended for external use. Its only effect is to stop Exim
generating certain messages to the postmaster, as otherwise message cascades
could occur in some situations. As part of the same option, a message id may
follow the characters &%-E%&. If it does, the log entry for the receipt of the
new message contains the id, following &"R="&, as a cross-reference.

.vitem &%-e%&&'x'&
.oindex "&%-e%&&'x'&"
There are a number of Sendmail options starting with &%-oe%& which seem to be
called by various programs without the leading &%o%& in the option. For
example, the &%vacation%& program uses &%-eq%&. Exim treats all options of the
form &%-e%&&'x'& as synonymous with the corresponding &%-oe%&&'x'& options.

.cmdopt -F <&'string'&>
.cindex "sender" "name"
.cindex "name" "of sender"
This option sets the sender's full name for use when a locally-generated
message is being accepted. In the absence of this option, the user's &'gecos'&
entry from the password data is used. As users are generally permitted to alter
their &'gecos'& entries, no security considerations are involved. White space
between &%-F%& and the <&'string'&> is optional.

.cmdopt -f <&'address'&>
.cindex "sender" "address"
.cindex "address" "sender"
.cindex "trusted users"
.cindex "envelope from"
.cindex "envelope sender"
.cindex "user" "trusted"
This option sets the address of the envelope sender of a locally-generated
message (also known as the return path). The option can normally be used only
by a trusted user, but &%untrusted_set_sender%& can be set to allow untrusted
users to use it.

Processes running as root or the Exim user are always trusted. Other
trusted users are defined by the &%trusted_users%& or &%trusted_groups%&
options. In the absence of &%-f%&, or if the caller is not trusted, the sender
of a local message is set to the caller's login name at the default qualify
domain.

There is one exception to the restriction on the use of &%-f%&: an empty sender
can be specified by any user, trusted or not, to create a message that can
never provoke a bounce. An empty sender can be specified either as an empty
string, or as a pair of angle brackets with nothing between them, as in these
examples of shell commands:
.code
exim -f '<>' user@domain
exim -f "" user@domain
.endd
In addition, the use of &%-f%& is not restricted when testing a filter file
with &%-bf%& or when testing or verifying addresses using the &%-bt%& or
&%-bv%& options.

Allowing untrusted users to change the sender address does not of itself make
it possible to send anonymous mail. Exim still checks that the &'From:'& header
refers to the local user, and if it does not, it adds a &'Sender:'& header,
though this can be overridden by setting &%no_local_from_check%&.

White
.cindex "&""From""& line"
space between &%-f%& and the <&'address'&> is optional (that is, they can be
given as two arguments or one combined argument). The sender of a
locally-generated message can also be set (when permitted) by an initial
&"From&~"& line in the message &-- see the description of &%-bm%& above &-- but
if &%-f%& is also present, it overrides &"From&~"&.

.cmdopt -G
.cindex "submission fixups, suppressing (command-line)"
This option is equivalent to an ACL applying:
.code
control = suppress_local_fixups
.endd
for every message received.  Note that Sendmail will complain about such
bad formatting, where Exim silently just does not fix it up.  This may change
in future.

As this affects audit information, the caller must be a trusted user to use
this option.

.cmdopt -h <&'number'&>
.cindex "Sendmail compatibility" "&%-h%& option ignored"
This option is accepted for compatibility with Sendmail, but has no effect. (In
Sendmail it overrides the &"hop count"& obtained by counting &'Received:'&
headers.)

.cmdopt -i
.cindex "Solaris" "&'mail'& command"
.cindex "dot" "in incoming non-SMTP message"
This option, which has the same effect as &%-oi%&, specifies that a dot on a
line by itself should not terminate an incoming, non-SMTP message.
Solaris 2.4 (SunOS 5.4) Sendmail has a similar &%-i%& processing option
&url(https://docs.oracle.com/cd/E19457-01/801-6680-1M/801-6680-1M.pdf),
p. 1M-529), and therefore a &%-oi%& command line option, which both are used
by its &'mailx'& command.

.cmdopt -L <&'tag'&>
.cindex "syslog" "process name; set with flag"
This option is equivalent to setting &%syslog_processname%& in the config
file and setting &%log_file_path%& to &`syslog`&.
Its use is restricted to administrators.  The configuration file has to be
read and parsed, to determine access rights, before this is set and takes
effect, so early configuration file errors will not honour this flag.

The tag should not be longer than 32 characters.

.cmdopt -M <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "forcing delivery"
.cindex "delivery" "forcing attempt"
.cindex "frozen messages" "forcing delivery"
This option requests Exim to run a delivery attempt on each message in turn. If
any of the messages are frozen, they are automatically thawed before the
delivery attempt. The settings of &%queue_domains%&, &%queue_smtp_domains%&,
and &%hold_domains%& are ignored.

Retry
.cindex "hints database" "overriding retry hints"
hints for any of the addresses are overridden &-- Exim tries to deliver even if
the normal retry time has not yet been reached. This option requires the caller
to be an admin user. However, there is an option called &%prod_requires_admin%&
which can be set false to relax this restriction (and also the same requirement
for the &%-q%&, &%-R%&, and &%-S%& options).

The deliveries happen synchronously, that is, the original Exim process does
not terminate until all the delivery attempts have finished. No output is
produced unless there is a serious error. If you want to see what is happening,
use the &%-v%& option as well, or inspect Exim's main log.

.cmdopt -Mar <&'message&~id'&>&~<&'address'&>&~<&'address'&>&~...
.cindex "message" "adding recipients"
.cindex "recipient" "adding"
This option requests Exim to add the addresses to the list of recipients of the
message (&"ar"& for &"add recipients"&). The first argument must be a message
id, and the remaining ones must be email addresses. However, if the message is
active (in the middle of a delivery attempt), it is not altered. This option
can be used only by an admin user.

.vitem "&%-MC%&&~<&'transport'&>&~<&'hostname'&>&&&
        &~<&'host&~IP'&>&&&
        &~<&'sequence&~number'&>&&&
        &~<&'message&~id'&>"
.oindex "&%-MC%&"
.cindex "SMTP" "passed connection"
.cindex "SMTP" "multiple deliveries"
.cindex "multiple SMTP deliveries"
This option is not intended for use by external callers. It is used internally
by Exim to invoke another instance of itself to deliver a waiting message using
an existing SMTP connection, which is passed as the standard input. Details are
given in chapter &<<CHAPSMTP>>&. This must be the final option, and the caller
must be root or the Exim user in order to use it.

.cmdopt -MCA
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that the
connection to the remote host has been authenticated.

.cmdopt -MCD
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that the
remote host supports the ESMTP &_DSN_& extension.

.cmdopt -MCd
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-d%& option
to pass on an information string on the purpose of the process.

.cmdopt -MCG <&'queue&~name'&>
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that an
alternate queue is used, named by the following argument.

.cmdopt -MCK
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that a
remote host supports the ESMTP &_CHUNKING_& extension.

.cmdopt -MCL
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that the server to
which Exim is connected advertised limits on numbers of mails, recipients or
recipient domains.
The limits are given by the following three arguments.

.cmdopt -MCP
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that the server to
which Exim is connected supports pipelining.

.cmdopt -MCp
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option. It signifies that the connection
t a remote server is via a SOCKS proxy, using addresses and ports given by
the following four arguments.

.cmdopt -MCQ <&'process&~id'&>&~<&'pipe&~fd'&>
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option when the original delivery was
started by a queue runner. It passes on the process id of the queue runner,
together with the file descriptor number of an open pipe. Closure of the pipe
signals the final completion of the sequence of processes that are passing
messages through the same SMTP connection.

.cmdopt -MCq <&'recipient&~address'&>&~<&'size'&>
This option is not intended for use by external callers. It is used internally
by Exim to implement quota checking for local users.

.cmdopt -MCS
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option, and passes on the fact that the
ESMTP SIZE option should be used on messages delivered down the existing
connection.

.cmdopt -MCT
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option, and passes on the fact that the
host to which Exim is connected supports TLS encryption.

.vitem &%-MCr%&&~<&'SNI'&> &&&
       &%-MCs%&&~<&'SNI'&>
.oindex "&%-MCs%&"
.oindex "&%-MCr%&"
These options are not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MCt%& option, and passes on the fact that
a TLS Server Name Indication was sent as part of the channel establishment.
The argument gives the SNI string.
The "r" variant indicates a DANE-verified connection.

.cmdopt -MCt <&'IP&~address'&>&~<&'port'&>&~<&'cipher'&>
This option is not intended for use by external callers. It is used internally
by Exim in conjunction with the &%-MC%& option, and passes on the fact that the
connection is being proxied by a parent process for handling TLS encryption.
The arguments give the local address and port being proxied, and the TLS cipher.

.cmdopt -Mc <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "hints database" "not overridden by &%-Mc%&"
.cindex "delivery" "manually started &-- not forced"
This option requests Exim to run a delivery attempt on each message, in turn,
but unlike the &%-M%& option, it does check for retry hints, and respects any
that are found. This option is not very useful to external callers. It is
provided mainly for internal use by Exim when it needs to re-invoke itself in
order to regain root privilege for a delivery (see chapter &<<CHAPsecurity>>&).
However, &%-Mc%& can be useful when testing, in order to run a delivery that
respects retry times and other options such as &%hold_domains%& that are
overridden when &%-M%& is used. Such a delivery does not count as a queue run.
If you want to run a specific delivery as if in a queue run, you should use
&%-q%& with a message id argument. A distinction between queue run deliveries
and other deliveries is made in one or two places.

.cmdopt -Mes <&'message&~id'&>&~<&'address'&>
.cindex "message" "changing sender"
.cindex "sender" "changing"
This option requests Exim to change the sender address in the message to the
given address, which must be a fully qualified address or &"<>"& (&"es"& for
&"edit sender"&). There must be exactly two arguments. The first argument must
be a message id, and the second one an email address. However, if the message
is active (in the middle of a delivery attempt), its status is not altered.
This option can be used only by an admin user.

.cmdopt -Mf <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "freezing messages"
.cindex "message" "manually freezing"
This option requests Exim to mark each listed message as &"frozen"&. This
prevents any delivery attempts taking place until the message is &"thawed"&,
either manually or as a result of the &%auto_thaw%& configuration option.
However, if any of the messages are active (in the middle of a delivery
attempt), their status is not altered. This option can be used only by an admin
user.

.cmdopt -Mg <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "giving up on messages"
.cindex "message" "abandoning delivery attempts"
.cindex "delivery" "abandoning further attempts"
This option requests Exim to give up trying to deliver the listed messages,
including any that are frozen. However, if any of the messages are active,
their status is not altered. For non-bounce messages, a delivery error message
is sent to the sender.
Bounce messages are just discarded. This option can be used only by an admin
user.

.cmdopt -MG <&'queue&~name'&>&~<&'message&~id'&>&~<&'message&~id'&>&~...
.cindex queue named
.cindex "named queues" "moving messages"
.cindex "queue" "moving messages"
This option requests that each listed message be moved from its current
queue to the given named queue.
The destination queue name argument is required, but can be an empty
string to define the default queue.
If the messages are not currently located in the default queue,
a &%-qG<name>%& option will be required to define the source queue.

.cmdopt -Mmad <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "delivery" "cancelling all"
This option requests Exim to mark all the recipient addresses in the messages
as already delivered (&"mad"& for &"mark all delivered"&). However, if any
message is active (in the middle of a delivery attempt), its status is not
altered. This option can be used only by an admin user.

.cmdopt -Mmd <&'message&~id'&>&~<&'address'&>&~<&'address'&>&~...
.cindex "delivery" "cancelling by address"
.cindex "recipient" "removing"
.cindex "removing recipients"
This option requests Exim to mark the given addresses as already delivered
(&"md"& for &"mark delivered"&). The first argument must be a message id, and
the remaining ones must be email addresses. These are matched to recipient
addresses in the message in a case-sensitive manner. If the message is active
(in the middle of a delivery attempt), its status is not altered. This option
can be used only by an admin user.

.cmdopt -Mrm <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "removing messages"
.cindex "abandoning mail"
.cindex "message" "manually discarding"
This option requests Exim to remove the given messages from the queue. No
bounce messages are sent; each message is simply forgotten. However, if any of
the messages are active, their status is not altered. This option can be used
only by an admin user or by the user who originally caused the message to be
placed in the queue.

. .new
. .vitem &%-MS%&
. .oindex "&%-MS%&"
. .cindex REQUIRETLS
. This option is used to request REQUIRETLS processing on the message.
. It is used internally by Exim in conjunction with -E when generating
. a bounce message.
. .wen

.cmdopt -Mset <&'message&~id'&>
.cindex "testing" "string expansion"
.cindex "expansion" "testing"
This option is useful only in conjunction with &%-be%& (that is, when testing
string expansions). Exim loads the given message from its spool before doing
the test expansions, thus setting message-specific variables such as
&$message_size$& and the header variables. The &$recipients$& variable is made
available. This feature is provided to make it easier to test expansions that
make use of these variables. However, this option can be used only by an admin
user. See also &%-bem%&.

.cmdopt -Mt <&'message&~id'&>&~<&'message&~id'&>&~...
.cindex "thawing messages"
.cindex "unfreezing messages"
.cindex "frozen messages" "thawing"
.cindex "message" "thawing frozen"
This option requests Exim to &"thaw"& any of the listed messages that are
&"frozen"&, so that delivery attempts can resume. However, if any of the
messages are active, their status is not altered. This option can be used only
by an admin user.

.cmdopt -Mvb <&'message&~id'&>
.cindex "listing" "message body"
.cindex "message" "listing body of"
This option causes the contents of the message body (-D) spool file to be
written to the standard output. This option can be used only by an admin user.

.cmdopt -Mvc <&'message&~id'&>
.cindex "message" "listing in RFC 2822 format"
.cindex "listing" "message in RFC 2822 format"
This option causes a copy of the complete message (header lines plus body) to
be written to the standard output in
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
format. This option can be used only by an admin user.

.cmdopt -Mvh <&'message&~id'&>
.cindex "listing" "message headers"
.cindex "header lines" "listing"
.cindex "message" "listing header lines"
This option causes the contents of the message headers (-H) spool file to be
written to the standard output. This option can be used only by an admin user.

.cmdopt -Mvl <&'message&~id'&>
.cindex "listing" "message log"
.cindex "message" "listing message log"
This option causes the contents of the message log spool file to be written to
the standard output. This option can be used only by an admin user.

.cmdopt -m
This is a synonym for &%-om%& that is accepted by Sendmail
(&url(https://docs.oracle.com/cd/E19457-01/801-6680-1M/801-6680-1M.pdf)
p. 1M-258), so Exim treats it that way too.

.cmdopt -N
.cindex "debugging" "&%-N%& option"
.cindex "debugging" "suppressing delivery"
This is a debugging option that inhibits delivery of a message at the transport
level. It implies &%-v%&. Exim goes through many of the motions of delivery &--
it just doesn't actually transport the message, but instead behaves as if it
had successfully done so. However, it does not make any updates to the retry
database, and the log entries for deliveries are flagged with &"*>"& rather
than &"=>"&.

Because &%-N%& discards any message to which it applies, only root or the Exim
user are allowed to use it with &%-bd%&, &%-q%&, &%-R%& or &%-M%&. In other
words, an ordinary user can use it only when supplying an incoming message to
which it will apply. Although transportation never fails when &%-N%& is set, an
address may be deferred because of a configuration problem on a transport, or a
routing problem. Once &%-N%& has been used for a delivery attempt, it sticks to
the message, and applies to any subsequent delivery attempts that may happen
for that message.

.cmdopt -n
This option is interpreted by Sendmail to mean &"no aliasing"&.
For normal modes of operation, it is ignored by Exim.
When combined with &%-bP%& it makes the output more terse (suppresses
option names, environment values and config pretty printing).

.cmdopt -O <&'data'&>
This option is interpreted by Sendmail to mean &`set option`&. It is ignored by
Exim.

.cmdopt -oA <&'file&~name'&>
.cindex "Sendmail compatibility" "&%-oA%& option"
This option is used by Sendmail in conjunction with &%-bi%& to specify an
alternative alias filename. Exim handles &%-bi%& differently; see the
description above.

.cmdopt -oB <&'n'&>
.cindex "SMTP" "passed connection"
.cindex "SMTP" "multiple deliveries"
.cindex "multiple SMTP deliveries"
This is a debugging option which limits the maximum number of messages that can
be delivered down one SMTP connection, overriding the value set in any &(smtp)&
transport. If <&'n'&> is omitted, the limit is set to 1.

.cmdopt -odb
.cindex "background delivery"
.cindex "delivery" "in the background"
This option applies to all modes in which Exim accepts incoming messages,
including the listening daemon. It requests &"background"& delivery of such
messages, which means that the accepting process automatically starts a
delivery process for each message received, but does not wait for the delivery
processes to finish.

When all the messages have been received, the reception process exits,
leaving the delivery processes to finish in their own time. The standard output
and error streams are closed at the start of each delivery process.
This is the default action if none of the &%-od%& options are present.

If one of the queueing options in the configuration file
(&%queue_only%& or &%queue_only_file%&, for example) is in effect, &%-odb%&
overrides it if &%queue_only_override%& is set true, which is the default
setting. If &%queue_only_override%& is set false, &%-odb%& has no effect.

.cmdopt -odf
.cindex "foreground delivery"
.cindex "delivery" "in the foreground"
This option requests &"foreground"& (synchronous) delivery when Exim has
accepted a locally-generated message. (For the daemon it is exactly the same as
&%-odb%&.) A delivery process is automatically started to deliver the message,
and Exim waits for it to complete before proceeding.

The original Exim reception process does not finish until the delivery
process for the final message has ended. The standard error stream is left open
during deliveries.

However, like &%-odb%&, this option has no effect if &%queue_only_override%& is
false and one of the queueing options in the configuration file is in effect.

If there is a temporary delivery error during foreground delivery, the
message is left in the queue for later delivery, and the original reception
process exits. See chapter &<<CHAPnonqueueing>>& for a way of setting up a
restricted configuration that never queues messages.


.cmdopt -odi
This option is synonymous with &%-odf%&. It is provided for compatibility with
Sendmail.

.cmdopt -odq
.cindex "non-immediate delivery"
.cindex "delivery" "suppressing immediate"
.cindex "queueing incoming messages"
This option applies to all modes in which Exim accepts incoming messages,
including the listening daemon. It specifies that the accepting process should
not automatically start a delivery process for each message received. Messages
are placed in the queue, and remain there until a subsequent queue runner
process encounters them. There are several configuration options (such as
&%queue_only%&) that can be used to queue incoming messages under certain
conditions. This option overrides all of them and also &%-odqs%&. It always
forces queueing.

.cmdopt -odqs
.cindex "SMTP" "delaying delivery"
.cindex "first pass routing"
This option is a hybrid between &%-odb%&/&%-odi%& and &%-odq%&.
However, like &%-odb%& and &%-odi%&, this option has no effect if
&%queue_only_override%& is false and one of the queueing options in the
configuration file is in effect.

When &%-odqs%& does operate, a delivery process is started for each incoming
message, in the background by default, but in the foreground if &%-odi%& is
also present. The recipient addresses are routed, and local deliveries are done
in the normal way. However, if any SMTP deliveries are required, they are not
done at this time, so the message remains in the queue until a subsequent queue
runner process encounters it. Because routing was done, Exim knows which
messages are waiting for which hosts, and so a number of messages for the same
host can be sent in a single SMTP connection. The &%queue_smtp_domains%&
configuration option has the same effect for specific domains. See also the
&%-qq%& option.

.cmdopt -oee
.cindex "error" "reporting"
If an error is detected while a non-SMTP message is being received (for
example, a malformed address), the error is reported to the sender in a mail
message.

.cindex "return code" "for &%-oee%&"
Provided
this error message is successfully sent, the Exim receiving process
exits with a return code of zero. If not, the return code is 2 if the problem
is that the original message has no recipients, or 1 for any other error.
This is the default &%-oe%&&'x'& option if Exim is called as &'rmail'&.

.cmdopt -oem
.cindex "error" "reporting"
.cindex "return code" "for &%-oem%&"
This is the same as &%-oee%&, except that Exim always exits with a non-zero
return code, whether or not the error message was successfully sent.
This is the default &%-oe%&&'x'& option, unless Exim is called as &'rmail'&.

.cmdopt -oep
.cindex "error" "reporting"
If an error is detected while a non-SMTP message is being received, the
error is reported by writing a message to the standard error file (stderr).
.cindex "return code" "for &%-oep%&"
The return code is 1 for all errors.

.cmdopt -oeq
.cindex "error" "reporting"
This option is supported for compatibility with Sendmail, but has the same
effect as &%-oep%&.

.cmdopt -oew
.cindex "error" "reporting"
This option is supported for compatibility with Sendmail, but has the same
effect as &%-oem%&.

.cmdopt -oi
.cindex "dot" "in incoming non-SMTP message"
This option, which has the same effect as &%-i%&, specifies that a dot on a
line by itself should not terminate an incoming, non-SMTP message. Otherwise, a
single dot does terminate, though Exim does no special processing for other
lines that start with a dot. This option is set by default if Exim is called as
&'rmail'&. See also &%-ti%&.

.cmdopt -oitrue
This option is treated as synonymous with &%-oi%&.

.cmdopt -oMa <&'host&~address'&>
.cindex "sender" "host address, specifying for local message"
A number of options starting with &%-oM%& can be used to set values associated
with remote hosts on locally-submitted messages (that is, messages not received
over TCP/IP). These options can be used by any caller in conjunction with the
&%-bh%&, &%-be%&, &%-bf%&, &%-bF%&, &%-bt%&, or &%-bv%& testing options. In
other circumstances, they are ignored unless the caller is trusted.

The &%-oMa%& option sets the sender host address. This may include a port
number at the end, after a full stop (period). For example:
.code
exim -bs -oMa 10.9.8.7.1234
.endd
An alternative syntax is to enclose the IP address in square brackets,
followed by a colon and the port number:
.code
exim -bs -oMa [10.9.8.7]:1234
.endd
The IP address is placed in the &$sender_host_address$& variable, and the
port, if present, in &$sender_host_port$&. If both &%-oMa%& and &%-bh%&
are present on the command line, the sender host IP address is taken from
whichever one is last.

.cmdopt -oMaa <&'name'&>
.cindex "authentication" "name, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMaa%&
option sets the value of &$sender_host_authenticated$& (the authenticator
name). See chapter &<<CHAPSMTPAUTH>>& for a discussion of SMTP authentication.
This option can be used with &%-bh%& and &%-bs%& to set up an
authenticated SMTP session without actually using the SMTP AUTH command.

.cmdopt -oMai <&'string'&>
.cindex "authentication" "id, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMai%&
option sets the value of &$authenticated_id$& (the id that was authenticated).
This overrides the default value (the caller's login id, except with &%-bh%&,
where there is no default) for messages from local sources. See chapter
&<<CHAPSMTPAUTH>>& for a discussion of authenticated ids.

.cmdopt -oMas <&'address'&>
.cindex "authentication" "sender, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMas%&
option sets the authenticated sender value in &$authenticated_sender$&. It
overrides the sender address that is created from the caller's login id for
messages from local sources, except when &%-bh%& is used, when there is no
default. For both &%-bh%& and &%-bs%&, an authenticated sender that is
specified on a MAIL command overrides this value. See chapter
&<<CHAPSMTPAUTH>>& for a discussion of authenticated senders.

.cmdopt -oMi <&'interface&~address'&>
.cindex "interface" "address, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMi%&
option sets the IP interface address value. A port number may be included,
using the same syntax as for &%-oMa%&. The interface address is placed in
&$received_ip_address$& and the port number, if present, in &$received_port$&.

.cmdopt -oMm <&'message&~reference'&>
.cindex "message reference" "message reference, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMm%&
option sets the message reference, e.g. message-id, and is logged during
delivery. This is useful when some kind of audit trail is required to tie
messages together. The format of the message reference is checked and will
abort if the format is invalid. The option will only be accepted if exim is
running in trusted mode, not as any regular user.

The best example of a message reference is when Exim sends a bounce message.
The message reference is the message-id of the original message for which Exim
is sending the bounce.

.cmdopt -oMr <&'protocol&~name'&>
.cindex "protocol, specifying for local message"
.vindex "&$received_protocol$&"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMr%&
option sets the received protocol value that is stored in
&$received_protocol$&. However, it does not apply (and is ignored) when &%-bh%&
or &%-bs%& is used. For &%-bh%&, the protocol is forced to one of the standard
SMTP protocol names (see the description of &$received_protocol$& in section
&<<SECTexpvar>>&). For &%-bs%&, the protocol is always &"local-"& followed by
one of those same names. For &%-bS%& (batched SMTP) however, the protocol can
be set by &%-oMr%&. Repeated use of this option is not supported.

.cmdopt -oMs <&'host&~name'&>
.cindex "sender" "host name, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMs%&
option sets the sender host name in &$sender_host_name$&. When this option is
present, Exim does not attempt to look up a host name from an IP address; it
uses the name it is given.

.cmdopt -oMt <&'ident&~string'&>
.cindex "sender" "ident string, specifying for local message"
See &%-oMa%& above for general remarks about the &%-oM%& options. The &%-oMt%&
option sets the sender ident value in &$sender_ident$&. The default setting for
local callers is the login id of the calling process, except when &%-bh%& is
used, when there is no default.

.cmdopt -om
.cindex "Sendmail compatibility" "&%-om%& option ignored"
In Sendmail, this option means &"me too"&, indicating that the sender of a
message should receive a copy of the message if the sender appears in an alias
expansion. Exim always does this, so the option does nothing.

.cmdopt -oo
.cindex "Sendmail compatibility" "&%-oo%& option ignored"
This option is ignored. In Sendmail it specifies &"old style headers"&,
whatever that means.

.cmdopt -oP <&'path'&>
.cindex "pid (process id)" "of daemon"
.cindex "daemon" "process id (pid)"
This option is useful only in conjunction with &%-bd%& or &%-q%& with a time
value. The option specifies the file to which the process id of the daemon is
written. When &%-oX%& is used with &%-bd%&, or when &%-q%& with a time is used
without &%-bd%&, this is the only way of causing Exim to write a pid file,
because in those cases, the normal pid file is not used.

.cmdopt -oPX
.cindex "pid (process id)" "of daemon"
.cindex "daemon" "process id (pid)"
This option is not intended for general use.
The daemon uses it when terminating due to a SIGTEM, possibly in
combination with &%-oP%&&~<&'path'&>.
It causes the pid file to be removed.

.cmdopt -or <&'time'&>
.cindex "timeout" "for non-SMTP input"
This option sets a timeout value for incoming non-SMTP messages. If it is not
set, Exim will wait forever for the standard input. The value can also be set
by the &%receive_timeout%& option. The format used for specifying times is
described in section &<<SECTtimeformat>>&.

.cmdopt -os <&'time'&>
.cindex "timeout" "for SMTP input"
.cindex "SMTP" "input timeout"
This option sets a timeout value for incoming SMTP messages. The timeout
applies to each SMTP command and block of data. The value can also be set by
the &%smtp_receive_timeout%& option; it defaults to 5 minutes. The format used
for specifying times is described in section &<<SECTtimeformat>>&.

.cmdopt -ov
This option has exactly the same effect as &%-v%&.

.cmdopt -oX <&'number&~or&~string'&>
.cindex "TCP/IP" "setting listening ports"
.cindex "TCP/IP" "setting listening interfaces"
.cindex "port" "receiving TCP/IP"
This option is relevant only when the &%-bd%& (start listening daemon) option
is also given. It controls which ports and interfaces the daemon uses. Details
of the syntax, and how it interacts with configuration file options, are given
in chapter &<<CHAPinterfaces>>&. When &%-oX%& is used to start a daemon, no pid
file is written unless &%-oP%& is also present to specify a pid filename.

.cmdopt -oY
.cindex "daemon notifier socket"
This option controls the creation of an inter-process communications endpoint
by the Exim daemon.
It is only relevant when the &%-bd%& (start listening daemon) option is also
given.
Normally the daemon creates this socket, unless a &%-oX%& and &*no*& &%-oP%&
option is also present.
If this option is given then the socket will not be created.  This is required
if the system is running multiple daemons, in which case it should
be used on all.
The features supported by the socket will not be available in such cases.

The socket is currently used for
.ilist
fast ramp-up of queue runner processes
.next
caching compiled regexes
.next
obtaining a current queue size
.endlist

.cmdopt -pd
.cindex "Perl" "starting the interpreter"
This option applies when an embedded Perl interpreter is linked with Exim (see
chapter &<<CHAPperl>>&). It overrides the setting of the &%perl_at_start%&
option, forcing the starting of the interpreter to be delayed until it is
needed.

.cmdopt -ps
.cindex "Perl" "starting the interpreter"
This option applies when an embedded Perl interpreter is linked with Exim (see
chapter &<<CHAPperl>>&). It overrides the setting of the &%perl_at_start%&
option, forcing the starting of the interpreter to occur as soon as Exim is
started.

.vitem &%-p%&<&'rval'&>:<&'sval'&>
.oindex "&%-p%&"
For compatibility with Sendmail, this option is equivalent to
.display
&`-oMr`& <&'rval'&> &`-oMs`& <&'sval'&>
.endd
It sets the incoming protocol and host name (for trusted callers). The
host name and its colon can be omitted when only the protocol is to be set.
Note the Exim already has two private options, &%-pd%& and &%-ps%&, that refer
to embedded Perl. It is therefore impossible to set a protocol value of &`d`&
or &`s`& using this option (but that does not seem a real limitation).
Repeated use of this option is not supported.

.cmdopt -q
.cindex "queue runner" "starting manually"
This option is normally restricted to admin users. However, there is a
configuration option called &%prod_requires_admin%& which can be set false to
relax this restriction (and also the same requirement for the &%-M%&, &%-R%&,
and &%-S%& options).

.cindex "queue runner" "description of operation"
If other commandline options do not specify an action,
the &%-q%& option starts one queue runner process. This scans the queue of
waiting messages, and runs a delivery process for each one in turn. It waits
for each delivery process to finish before starting the next one. A delivery
process may not actually do any deliveries if the retry times for the addresses
have not been reached. Use &%-qf%& (see below) if you want to override this.

If
.cindex "SMTP" "passed connection"
.cindex "SMTP" "multiple deliveries"
.cindex "multiple SMTP deliveries"
the delivery process spawns other processes to deliver other messages down
passed SMTP connections, the queue runner waits for these to finish before
proceeding.

When all the queued messages have been considered, the original queue runner
process terminates. In other words, a single pass is made over the waiting
mail, one message at a time. Use &%-q%& with a time (see below) if you want
this to be repeated periodically.

Exim processes the waiting messages in an unpredictable order. It isn't very
random, but it is likely to be different each time, which is all that matters.
If one particular message screws up a remote MTA, other messages to the same
MTA have a chance of getting through if they get tried first.

It is possible to cause the messages to be processed in lexical message id
order, which is essentially the order in which they arrived, by setting the
&%queue_run_in_order%& option, but this is not recommended for normal use.

.vitem &%-q%&<&'qflags'&>
The &%-q%& option may be followed by one or more flag letters that change its
behaviour. They are all optional, but if more than one is present, they must
appear in the correct order. Each flag is described in a separate item below.

.vitem &%-qq...%&
.oindex "&%-qq%&"
.cindex "queue" "double scanning"
.cindex "queue" "routing"
.cindex "routing" "whole queue before delivery"
.cindex "first pass routing"
.cindex "queue runner" "two phase"
An option starting with &%-qq%& requests a two-stage queue run. In the first
stage, the queue is scanned as if the &%queue_smtp_domains%& option matched
every domain. Addresses are routed, local deliveries happen, but no remote
transports are run.

Performance will be best if the &%queue_run_in_order%& option is false.
If that is so and
the &%queue_fast_ramp%& option is true
and a daemon-notifier socket is available
then in the first phase of the run,
once a threshold number of messages are routed for a given host,
a delivery process is forked in parallel with the rest of the scan.

.cindex "hints database" "remembering routing"
The hints database that remembers which messages are waiting for specific hosts
is updated, as if delivery to those hosts had been deferred.

After the first queue scan complete,
a second, normal queue scan is done, with routing and delivery taking
place as normal.
Messages that are routed to the same host should mostly be
delivered down a single SMTP
.cindex "SMTP" "passed connection"
.cindex "SMTP" "multiple deliveries"
.cindex "multiple SMTP deliveries"
connection because of the hints that were set up during the first queue scan.

Two-phase queue runs should be used on systems which, even intermittently,
have a large queue (such as mailing-list operators).
They may also be useful for hosts that are connected to the Internet
intermittently.

.vitem &%-q[q]i...%&
.oindex "&%-qi%&"
.cindex "queue" "initial delivery"
If the &'i'& flag is present, the queue runner runs delivery processes only for
those messages that haven't previously been tried. (&'i'& stands for &"initial
delivery"&.) This can be helpful if you are putting messages in the queue using
&%-odq%& and want a queue runner just to process the new messages.

.vitem &%-q[q][i]f...%&
.oindex "&%-qf%&"
.cindex "queue" "forcing delivery"
.cindex "delivery" "forcing in queue run"
If one &'f'& flag is present, a delivery attempt is forced for each non-frozen
message, whereas without &'f'& only those non-frozen addresses that have passed
their retry times are tried.

.vitem &%-q[q][i]ff...%&
.oindex "&%-qff%&"
.cindex "frozen messages" "forcing delivery"
If &'ff'& is present, a delivery attempt is forced for every message, whether
frozen or not.

.vitem &%-q[q][i][f[f]]l%&
.oindex "&%-ql%&"
.cindex "queue" "local deliveries only"
The &'l'& (the letter &"ell"&) flag specifies that only local deliveries are to
be done. If a message requires any remote deliveries, it remains in the queue
for later delivery.

.vitem &%-q[q][i][f[f]][l][G<name>[/<time>]]]%&
.oindex "&%-qG%&"
.cindex queue named
.cindex "named queues"  "deliver from"
.cindex "queue" "delivering specific messages"
If the &'G'& flag and a name is present, the queue runner operates on the
queue with the given name rather than the default queue.
The name should not contain a &'/'& character.
For a periodic queue run (see below)
append to the name a slash and a time value.

If other commandline options specify an action, a &'-qG<name>'& option
will specify a queue to operate on.
For example:
.code
exim -bp -qGquarantine
mailq -qGquarantine
exim -qGoffpeak -Rf @special.domain.example
.endd

.vitem &%-q%&<&'qflags'&>&~<&'start&~id'&>&~<&'end&~id'&>
When scanning the queue, Exim can be made to skip over messages whose ids are
lexically less than a given value by following the &%-q%& option with a
starting message id. For example:
.code
exim -q 0t5C6f-0000c8-00
.endd
Messages that arrived earlier than &`0t5C6f-0000c8-00`& are not inspected. If a
second message id is given, messages whose ids are lexically greater than it
are also skipped. If the same id is given twice, for example,
.code
exim -q 0t5C6f-0000c8-00 0t5C6f-0000c8-00
.endd
just one delivery process is started, for that message. This differs from
&%-M%& in that retry data is respected, and it also differs from &%-Mc%& in
that it counts as a delivery from a queue run. Note that the selection
mechanism does not affect the order in which the messages are scanned. There
are also other ways of selecting specific sets of messages for delivery in a
queue run &-- see &%-R%& and &%-S%&.

.vitem &%-q%&<&'qflags'&><&'time'&>
.cindex "queue runner" "starting periodically"
.cindex "periodic queue running"
When a time value is present, the &%-q%& option causes Exim to run as a daemon,
starting a queue runner process at intervals specified by the given time value
(whose format is described in section &<<SECTtimeformat>>&). This form of the
&%-q%& option is commonly combined with the &%-bd%& option, in which case a
single daemon process handles both functions. A common way of starting up a
combined daemon at system boot time is to use a command such as
.code
/usr/exim/bin/exim -bd -q30m
.endd
Such a daemon listens for incoming SMTP calls, and also starts a queue runner
process every 30 minutes.

.cindex "named queues" "queue runners"
It is possible to set up runners for multiple named queues within one daemon,
For example:
.code
exim -qGhipri/2m -q10m -qqGmailinglist/1h
.endd

When a daemon is started by &%-q%& with a time value, but without &%-bd%&, no
pid file is written unless one is explicitly requested by the &%-oP%& option.

.vitem &%-qR%&<&'rsflags'&>&~<&'string'&>
.oindex "&%-qR%&"
This option is synonymous with &%-R%&. It is provided for Sendmail
compatibility.

.vitem &%-qS%&<&'rsflags'&>&~<&'string'&>
.oindex "&%-qS%&"
This option is synonymous with &%-S%&.

.vitem &%-R%&<&'rsflags'&>&~<&'string'&>
.oindex "&%-R%&"
.cindex "queue runner" "for specific recipients"
.cindex "delivery" "to given domain"
.cindex "domain" "delivery to"
The <&'rsflags'&> may be empty, in which case the white space before the string
is optional, unless the string is &'f'&, &'ff'&, &'r'&, &'rf'&, or &'rff'&,
which are the possible values for <&'rsflags'&>. White space is required if
<&'rsflags'&> is not empty.

This option is similar to &%-q%& with no time value, that is, it causes Exim to
perform a single queue run, except that, when scanning the messages on the
queue, Exim processes only those that have at least one undelivered recipient
address containing the given string, which is checked in a case-independent
way. If the <&'rsflags'&> start with &'r'&, <&'string'&> is interpreted as a
regular expression; otherwise it is a literal string.

If you want to do periodic queue runs for messages with specific recipients,
you can combine &%-R%& with &%-q%& and a time value. For example:
.code
exim -q25m -R @special.domain.example
.endd
This example does a queue run for messages with recipients in the given domain
every 25 minutes. Any additional flags that are specified with &%-q%& are
applied to each queue run.

Once a message is selected for delivery by this mechanism, all its addresses
are processed. For the first selected message, Exim overrides any retry
information and forces a delivery attempt for each undelivered address. This
means that if delivery of any address in the first message is successful, any
existing retry information is deleted, and so delivery attempts for that
address in subsequently selected messages (which are processed without forcing)
will run. However, if delivery of any address does not succeed, the retry
information is updated, and in subsequently selected messages, the failing
address will be skipped.

.cindex "frozen messages" "forcing delivery"
If the <&'rsflags'&> contain &'f'& or &'ff'&, the delivery forcing applies to
all selected messages, not just the first; frozen messages are included when
&'ff'& is present.

The &%-R%& option makes it straightforward to initiate delivery of all messages
to a given domain after a host has been down for some time. When the SMTP
command ETRN is accepted by its ACL (see section &<<SECTETRN>>&), its default
effect is to run Exim with the &%-R%& option, but it can be configured to run
an arbitrary command instead.

.cmdopt -r
This is a documented (for Sendmail) obsolete alternative name for &%-f%&.

.vitem &%-S%&<&'rsflags'&>&~<&'string'&>
.oindex "&%-S%&"
.cindex "delivery" "from given sender"
.cindex "queue runner" "for specific senders"
This option acts like &%-R%& except that it checks the string against each
message's sender instead of against the recipients. If &%-R%& is also set, both
conditions must be met for a message to be selected. If either of the options
has &'f'& or &'ff'& in its flags, the associated action is taken.

.cmdopt -Tqt <&'times'&>
This is an option that is exclusively for use by the Exim testing suite. It is not
recognized when Exim is run normally. It allows for the setting up of explicit
&"queue times"& so that various warning/retry features can be tested.

.cmdopt -t
.cindex "recipient" "extracting from header lines"
.chindex Bcc:
.chindex Cc:
.chindex To:
When Exim is receiving a locally-generated, non-SMTP message on its standard
input, the &%-t%& option causes the recipients of the message to be obtained
from the &'To:'&, &'Cc:'&, and &'Bcc:'& header lines in the message instead of
from the command arguments. The addresses are extracted before any rewriting
takes place and the &'Bcc:'& header line, if present, is then removed.

.cindex "Sendmail compatibility" "&%-t%& option"
If the command has any arguments, they specify addresses to which the message
is &'not'& to be delivered. That is, the argument addresses are removed from
the recipients list obtained from the headers. This is compatible with Smail 3
and in accordance with the documented behaviour of several versions of
Sendmail, as described in man pages on a number of operating systems (e.g.
Solaris 8, IRIX 6.5, HP-UX 11). However, some versions of Sendmail &'add'&
argument addresses to those obtained from the headers, and the O'Reilly
Sendmail book documents it that way. Exim can be made to add argument addresses
instead of subtracting them by setting the option
&%extract_addresses_remove_arguments%& false.

.cindex "&%Resent-%& header lines" "with &%-t%&"
If there are any &%Resent-%& header lines in the message, Exim extracts
recipients from all &'Resent-To:'&, &'Resent-Cc:'&, and &'Resent-Bcc:'& header
lines instead of from &'To:'&, &'Cc:'&, and &'Bcc:'&. This is for compatibility
with Sendmail and other MTAs. (Prior to release 4.20, Exim gave an error if
&%-t%& was used in conjunction with &%Resent-%& header lines.)

&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
talks about different sets of &%Resent-%& header lines (for when a
message is resent several times). The RFC also specifies that they should be
added at the front of the message, and separated by &'Received:'& lines. It is
not at all clear how &%-t%& should operate in the present of multiple sets,
nor indeed exactly what constitutes a &"set"&.
In practice, it seems that MUAs do not follow the RFC. The &%Resent-%& lines
are often added at the end of the header, and if a message is resent more than
once, it is common for the original set of &%Resent-%& headers to be renamed as
&%X-Resent-%& when a new set is added. This removes any possible ambiguity.

.cmdopt -ti
This option is exactly equivalent to &%-t%& &%-i%&. It is provided for
compatibility with Sendmail.

.cmdopt -tls-on-connect
.cindex "TLS" "use without STARTTLS"
.cindex "TLS" "automatic start"
This option is available when Exim is compiled with TLS support. It forces all
incoming SMTP connections to behave as if the incoming port is listed in the
&%tls_on_connect_ports%& option. See section &<<SECTsupobssmt>>& and chapter
&<<CHAPTLS>>& for further details.


.cmdopt -U
.cindex "Sendmail compatibility" "&%-U%& option ignored"
Sendmail uses this option for &"initial message submission"&, and its
documentation states that in future releases, it may complain about
syntactically invalid messages rather than fixing them when this flag is not
set. Exim ignores this option.

.cmdopt -v
This option causes Exim to write information to the standard error stream,
describing what it is doing. In particular, it shows the log lines for
receiving and delivering a message, and if an SMTP connection is made, the SMTP
dialogue is shown. Some of the log lines shown may not actually be written to
the log if the setting of &%log_selector%& discards them. Any relevant
selectors are shown with each log line. If none are shown, the logging is
unconditional.

.cmdopt -x
AIX uses &%-x%& for a private purpose (&"mail from a local mail program has
National Language Support extended characters in the body of the mail item"&).
It sets &%-x%& when calling the MTA from its &%mail%& command. Exim ignores
this option.

.cmdopt -X <&'logfile'&>
This option is interpreted by Sendmail to cause debug information to be sent
to the named file.  It is ignored by Exim.

.cmdopt -z <&'log-line'&>
This option writes its argument to Exim's logfile.
Use is restricted to administrators; the intent is for operational notes.
Quotes should be used to maintain a multi-word item as a single argument,
under most shells.
.endlist

.ecindex IIDclo1
.ecindex IIDclo2


. ////////////////////////////////////////////////////////////////////////////
. Insert a stylized DocBook comment here, to identify the end of the command
. line options. This is for the benefit of the Perl script that automatically
. creates a man page for the options.
. ////////////////////////////////////////////////////////////////////////////

.literal xml
<!-- === End of command line options === -->
.literal off





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////


.chapter "The Exim runtime configuration file" "CHAPconf" &&&
         "The runtime configuration file"

.cindex "runtime configuration"
.cindex "configuration file" "general description"
.cindex "CONFIGURE_FILE"
.cindex "configuration file" "errors in"
.cindex "error" "in configuration file"
.cindex "return code" "for bad configuration"
Exim uses a single runtime configuration file that is read whenever an Exim
binary is executed. Note that in normal operation, this happens frequently,
because Exim is designed to operate in a distributed manner, without central
control.

If a syntax error is detected while reading the configuration file, Exim
writes a message on the standard error, and exits with a non-zero return code.
The message is also written to the panic log. &*Note*&: Only simple syntax
errors can be detected at this time. The values of any expanded options are
not checked until the expansion happens, even when the expansion does not
actually alter the string.

The name of the configuration file is compiled into the binary for security
reasons, and is specified by the CONFIGURE_FILE compilation option. In
most configurations, this specifies a single file. However, it is permitted to
give a colon-separated list of filenames, in which case Exim uses the first
existing file in the list.

.cindex "EXIM_USER"
.cindex "EXIM_GROUP"
.cindex "CONFIGURE_OWNER"
.cindex "CONFIGURE_GROUP"
.cindex "configuration file" "ownership"
.cindex "ownership" "configuration file"
The runtime configuration file must be owned by root or by the user that is
specified at compile time by the CONFIGURE_OWNER option (if set). The
configuration file must not be world-writeable, or group-writeable unless its
group is the root group or the one specified at compile time by the
CONFIGURE_GROUP option.

&*Warning*&: In a conventional configuration, where the Exim binary is setuid
to root, anybody who is able to edit the runtime configuration file has an
easy way to run commands as root. If you specify a user or group in the
CONFIGURE_OWNER or CONFIGURE_GROUP options, then that user and/or any users
who are members of that group will trivially be able to obtain root privileges.

Up to Exim version 4.72, the runtime configuration file was also permitted to
be writeable by the Exim user and/or group. That has been changed in Exim 4.73
since it offered a simple privilege escalation for any attacker who managed to
compromise the Exim user account.

A default configuration file, which will work correctly in simple situations,
is provided in the file &_src/configure.default_&. If CONFIGURE_FILE
defines just one filename, the installation process copies the default
configuration to a new file of that name if it did not previously exist. If
CONFIGURE_FILE is a list, no default is automatically installed. Chapter
&<<CHAPdefconfil>>& is a &"walk-through"& discussion of the default
configuration.



.section "Using a different configuration file" "SECID40"
.cindex "configuration file" "alternate"
A one-off alternate configuration can be specified by the &%-C%& command line
option, which may specify a single file or a list of files. However, when
&%-C%& is used, Exim gives up its root privilege, unless called by root (or
unless the argument for &%-C%& is identical to the built-in value from
CONFIGURE_FILE), or is listed in the TRUSTED_CONFIG_LIST file and the caller
is the Exim user or the user specified in the CONFIGURE_OWNER setting. &%-C%&
is useful mainly for checking the syntax of configuration files before
installing them. No owner or group checks are done on a configuration file
specified by &%-C%&, if root privilege has been dropped.

Even the Exim user is not trusted to specify an arbitrary configuration file
with the &%-C%& option to be used with root privileges, unless that file is
listed in the TRUSTED_CONFIG_LIST file. This locks out the possibility of
testing a configuration using &%-C%& right through message reception and
delivery, even if the caller is root. The reception works, but by that time,
Exim is running as the Exim user, so when it re-execs to regain privilege for
the delivery, the use of &%-C%& causes privilege to be lost. However, root
can test reception and delivery using two separate commands (one to put a
message in the queue, using &%-odq%&, and another to do the delivery, using
&%-M%&).

If ALT_CONFIG_PREFIX is defined &_in Local/Makefile_&, it specifies a
prefix string with which any file named in a &%-C%& command line option must
start. In addition, the filename must not contain the sequence &"&`/../`&"&.
There is no default setting for ALT_CONFIG_PREFIX; when it is unset, any
filename can be used with &%-C%&.

One-off changes to a configuration can be specified by the &%-D%& command line
option, which defines and overrides values for macros used inside the
configuration file. However, like &%-C%&, the use of this option by a
non-privileged user causes Exim to discard its root privilege.
If DISABLE_D_OPTION is defined in &_Local/Makefile_&, the use of &%-D%& is
completely disabled, and its use causes an immediate error exit.

The WHITELIST_D_MACROS option in &_Local/Makefile_& permits the binary builder
to declare certain macro names trusted, such that root privilege will not
necessarily be discarded.
WHITELIST_D_MACROS defines a colon-separated list of macros which are
considered safe and, if &%-D%& only supplies macros from this list, and the
values are acceptable, then Exim will not give up root privilege if the caller
is root, the Exim run-time user, or the CONFIGURE_OWNER, if set.  This is a
transition mechanism and is expected to be removed in the future.  Acceptable
values for the macros satisfy the regexp: &`^[A-Za-z0-9_/.-]*$`&

Some sites may wish to use the same Exim binary on different machines that
share a file system, but to use different configuration files on each machine.
If CONFIGURE_FILE_USE_NODE is defined in &_Local/Makefile_&, Exim first
looks for a file whose name is the configuration filename followed by a dot
and the machine's node name, as obtained from the &[uname()]& function. If this
file does not exist, the standard name is tried. This processing occurs for
each filename in the list given by CONFIGURE_FILE or &%-C%&.

In some esoteric situations different versions of Exim may be run under
different effective uids and the CONFIGURE_FILE_USE_EUID is defined to
help with this. See the comments in &_src/EDITME_& for details.



.section "Configuration file format" "SECTconffilfor"
.cindex "configuration file" "format of"
.cindex "format" "configuration file"
Exim's configuration file is divided into a number of different parts. General
option settings must always appear at the start of the file. The other parts
are all optional, and may appear in any order. Each part other than the first
is introduced by the word &"begin"& followed by at least one literal
space, and the name of the part. The optional parts are:

.ilist
&'ACL'&: Access control lists for controlling incoming SMTP mail (see chapter
&<<CHAPACL>>&).
.next
.cindex "AUTH" "configuration"
&'authenticators'&: Configuration settings for the authenticator drivers. These
are concerned with the SMTP AUTH command (see chapter &<<CHAPSMTPAUTH>>&).
.next
&'routers'&: Configuration settings for the router drivers. Routers process
addresses and determine how the message is to be delivered (see chapters
&<<CHAProutergeneric>>&&--&<<CHAPredirect>>&).
.next
&'transports'&: Configuration settings for the transport drivers. Transports
define mechanisms for copying messages to destinations (see chapters
&<<CHAPtransportgeneric>>&&--&<<CHAPsmtptrans>>&).
.next
&'retry'&: Retry rules, for use when a message cannot be delivered immediately.
If there is no retry section, or if it is empty (that is, no retry rules are
defined), Exim will not retry deliveries. In this situation, temporary errors
are treated the same as permanent errors. Retry rules are discussed in chapter
&<<CHAPretry>>&.
.next
&'rewrite'&: Global address rewriting rules, for use when a message arrives and
when new addresses are generated during delivery. Rewriting is discussed in
chapter &<<CHAPrewrite>>&.
.next
&'local_scan'&: Private options for the &[local_scan()]& function. If you
want to use this feature, you must set
.code
LOCAL_SCAN_HAS_OPTIONS=yes
.endd
in &_Local/Makefile_& before building Exim. Details of the &[local_scan()]&
facility are given in chapter &<<CHAPlocalscan>>&.
.endlist

.cindex "configuration file" "leading white space in"
.cindex "configuration file" "trailing white space in"
.cindex "white space" "in configuration file"
Leading and trailing white space in configuration lines is always ignored.

Blank lines in the file, and lines starting with a # character (ignoring
leading white space) are treated as comments and are ignored. &*Note*&: A
# character other than at the beginning of a line is not treated specially,
and does not introduce a comment.

Any non-comment line can be continued by ending it with a backslash. Note that
the general rule for white space means that trailing white space after the
backslash and leading white space at the start of continuation
lines is ignored. Comment lines beginning with # (but not empty lines) may
appear in the middle of a sequence of continuation lines.

A convenient way to create a configuration file is to start from the
default, which is supplied in &_src/configure.default_&, and add, delete, or
change settings as required.

The ACLs, retry rules, and rewriting rules have their own syntax which is
described in chapters &<<CHAPACL>>&, &<<CHAPretry>>&, and &<<CHAPrewrite>>&,
respectively. The other parts of the configuration file have some syntactic
items in common, and these are described below, from section &<<SECTcos>>&
onwards. Before that, the inclusion, macro, and conditional facilities are
described.



.section "File inclusions in the configuration file" "SECID41"
.cindex "inclusions in configuration file"
.cindex "configuration file" "including other files"
.cindex "&`.include`& in configuration file"
.cindex "&`.include_if_exists`& in configuration file"
You can include other files inside Exim's runtime configuration file by
using this syntax:
.display
&`.include`& <&'filename'&>
&`.include_if_exists`& <&'filename'&>
.endd
on a line by itself. Double quotes round the filename are optional. If you use
the first form, a configuration error occurs if the file does not exist; the
second form does nothing for non-existent files.
The first form allows a relative name. It is resolved relative to
the directory of the including file. For the second form an absolute filename
is required.

Includes may be nested to any depth, but remember that Exim reads its
configuration file often, so it is a good idea to keep them to a minimum.
If you change the contents of an included file, you must HUP the daemon,
because an included file is read only when the configuration itself is read.

The processing of inclusions happens early, at a physical line level, so, like
comment lines, an inclusion can be used in the middle of an option setting,
for example:
.code
hosts_lookup = a.b.c \
               .include /some/file
.endd
Include processing happens after macro processing (see below). Its effect is to
process the lines of the included file as if they occurred inline where the
inclusion appears.



.section "Macros in the configuration file" "SECTmacrodefs"
.cindex "macro" "description of"
.cindex "configuration file" "macros"
If a line in the main part of the configuration (that is, before the first
&"begin"& line) begins with an upper case letter, it is taken as a macro
definition, and must be of the form
.display
<&'name'&> = <&'rest of line'&>
.endd
The name must consist of letters, digits, and underscores, and need not all be
in upper case, though that is recommended. The rest of the line, including any
continuations, is the replacement text, and has leading and trailing white
space removed. Quotes are not removed. The replacement text can never end with
a backslash character, but this doesn't seem to be a serious limitation.

Macros may also be defined between router, transport, authenticator, or ACL
definitions. They may not, however, be defined within an individual driver or
ACL, or in the &%local_scan%&, retry, or rewrite sections of the configuration.

.section "Macro substitution" "SECID42"
Once a macro is defined, all subsequent lines in the file (and any included
files) are scanned for the macro name; if there are several macros, the line is
scanned for each, in turn, in the order in which the macros are defined. The
replacement text is not re-scanned for the current macro, though it is scanned
for subsequently defined macros. For this reason, a macro name may not contain
the name of a previously defined macro as a substring. You could, for example,
define
.display
&`ABCD_XYZ = `&<&'something'&>
&`ABCD = `&<&'something else'&>
.endd
but putting the definitions in the opposite order would provoke a configuration
error. Macro expansion is applied to individual physical lines from the file,
before checking for line continuation or file inclusion (see above). If a line
consists solely of a macro name, and the expansion of the macro is empty, the
line is ignored. A macro at the start of a line may turn the line into a
comment line or a &`.include`& line.


.section "Redefining macros" "SECID43"
Once defined, the value of a macro can be redefined later in the configuration
(or in an included file). Redefinition is specified by using &'=='& instead of
&'='&. For example:
.code
MAC =  initial value
...
MAC == updated value
.endd
Redefinition does not alter the order in which the macros are applied to the
subsequent lines of the configuration file. It is still the same order in which
the macros were originally defined. All that changes is the macro's value.
Redefinition makes it possible to accumulate values. For example:
.code
MAC =  initial value
...
MAC == MAC and something added
.endd
This can be helpful in situations where the configuration file is built
from a number of other files.

.section "Overriding macro values" "SECID44"
The values set for macros in the configuration file can be overridden by the
&%-D%& command line option, but Exim gives up its root privilege when &%-D%& is
used, unless called by root or the Exim user. A definition on the command line
using the &%-D%& option causes all definitions and redefinitions within the
file to be ignored.



.section "Example of macro usage" "SECID45"
As an example of macro usage, consider a configuration where aliases are looked
up in a MySQL database. It helps to keep the file less cluttered if long
strings such as SQL statements are defined separately as macros, for example:
.code
ALIAS_QUERY = select mailbox from user where \
              login='${quote_mysql:$local_part}';
.endd
This can then be used in a &(redirect)& router setting like this:
.code
data = ${lookup mysql{ALIAS_QUERY}}
.endd
In earlier versions of Exim macros were sometimes used for domain, host, or
address lists. In Exim 4 these are handled better by named lists &-- see
section &<<SECTnamedlists>>&.


.section "Builtin macros" "SECTbuiltinmacros"
Exim defines some macros depending on facilities available, which may
differ due to build-time definitions and from one release to another.
All of these macros start with an underscore.
They can be used to conditionally include parts of a configuration
(see below).

The following classes of macros are defined:
.display
&` _HAVE_*                    `&  build-time defines
&` _DRIVER_ROUTER_*           `&  router drivers
&` _DRIVER_TRANSPORT_*        `&  transport drivers
&` _DRIVER_AUTHENTICATOR_*    `&  authenticator drivers
&` _EXP_COND_*                `&  expansion conditions
&` _EXP_ITEM_*                `&  expansion items
&` _EXP_OP_*                  `&  expansion operators
&` _EXP_VAR_*                 `&  expansion variables
&` _LOG_*                     `&  log_selector values
&` _OPT_MAIN_*                `&  main config options
&` _OPT_ROUTERS_*             `&  generic router options
&` _OPT_TRANSPORTS_*          `&  generic transport options
&` _OPT_AUTHENTICATORS_*      `&  generic authenticator options
&` _OPT_ROUTER_*_*            `&  private router options
&` _OPT_TRANSPORT_*_*         `&  private transport options
&` _OPT_AUTHENTICATOR_*_*     `&  private authenticator options
.endd

Use an &"exim -bP macros"& command to get the list of macros.


.section "Conditional skips in the configuration file" "SECID46"
.cindex "configuration file" "conditional skips"
.cindex "&`.ifdef`&"
You can use the directives &`.ifdef`&, &`.ifndef`&, &`.elifdef`&,
&`.elifndef`&, &`.else`&, and &`.endif`& to dynamically include or exclude
portions of the configuration file. The processing happens whenever the file is
read (that is, when an Exim binary starts to run).

The implementation is very simple. Instances of the first four directives must
be followed by text that includes the names of one or macros. The condition
that is tested is whether or not any macro substitution has taken place in the
line. Thus:
.code
.ifdef AAA
message_size_limit = 50M
.else
message_size_limit = 100M
.endif
.endd
sets a message size limit of 50M if the macro &`AAA`& is defined
(or &`A`& or &`AA`&), and 100M
otherwise. If there is more than one macro named on the line, the condition
is true if any of them are defined. That is, it is an &"or"& condition. To
obtain an &"and"& condition, you need to use nested &`.ifdef`&s.

Although you can use a macro expansion to generate one of these directives,
it is not very useful, because the condition &"there was a macro substitution
in this line"& will always be true.

Text following &`.else`& and &`.endif`& is ignored, and can be used as comment
to clarify complicated nestings.



.section "Common option syntax" "SECTcos"
.cindex "common option syntax"
.cindex "syntax of common options"
.cindex "configuration file" "common option syntax"
For the main set of options, driver options, and &[local_scan()]& options,
each setting is on a line by itself, and starts with a name consisting of
lower-case letters and underscores. Many options require a data value, and in
these cases the name must be followed by an equals sign (with optional white
space) and then the value. For example:
.code
qualify_domain = mydomain.example.com
.endd
.cindex "hiding configuration option values"
.cindex "configuration options" "hiding value of"
.cindex "options" "hiding value of"
Some option settings may contain sensitive data, for example, passwords for
accessing databases. To stop non-admin users from using the &%-bP%& command
line option to read these values, you can precede the option settings with the
word &"hide"&. For example:
.code
hide mysql_servers = localhost/users/admin/secret-password
.endd
For non-admin users, such options are displayed like this:
.code
mysql_servers = <value not displayable>
.endd
If &"hide"& is used on a driver option, it hides the value of that option on
all instances of the same driver.

The following sections describe the syntax used for the different data types
that are found in option settings.


.section "Boolean options" "SECID47"
.cindex "format" "boolean"
.cindex "boolean configuration values"
.oindex "&%no_%&&'xxx'&"
.oindex "&%not_%&&'xxx'&"
Options whose type is given as boolean are on/off switches. There are two
different ways of specifying such options: with and without a data value. If
the option name is specified on its own without data, the switch is turned on;
if it is preceded by &"no_"& or &"not_"& the switch is turned off. However,
boolean options may be followed by an equals sign and one of the words
&"true"&, &"false"&, &"yes"&, or &"no"&, as an alternative syntax. For example,
the following two settings have exactly the same effect:
.code
queue_only
queue_only = true
.endd
The following two lines also have the same (opposite) effect:
.code
no_queue_only
queue_only = false
.endd
You can use whichever syntax you prefer.




.section "Integer values" "SECID48"
.cindex "integer configuration values"
.cindex "format" "integer"
If an option's type is given as &"integer"&, the value can be given in decimal,
hexadecimal, or octal. If it starts with a digit greater than zero, a decimal
number is assumed. Otherwise, it is treated as an octal number unless it starts
with the characters &"0x"&, in which case the remainder is interpreted as a
hexadecimal number.

If an integer value is followed by the letter K, it is multiplied by 1024; if
it is followed by the letter M, it is multiplied by 1024x1024;
if by the letter G, 1024x1024x1024.
When the values
of integer option settings are output, values which are an exact multiple of
1024 or 1024x1024 are sometimes, but not always, printed using the letters K
and M. The printing style is independent of the actual input format that was
used.


.section "Octal integer values" "SECID49"
.cindex "integer format"
.cindex "format" "octal integer"
If an option's type is given as &"octal integer"&, its value is always
interpreted as an octal number, whether or not it starts with the digit zero.
Such options are always output in octal.


.section "Fixed point numbers" "SECID50"
.cindex "fixed point configuration values"
.cindex "format" "fixed point"
If an option's type is given as &"fixed-point"&, its value must be a decimal
integer, optionally followed by a decimal point and up to three further digits.



.section "Time intervals" "SECTtimeformat"
.cindex "time interval" "specifying in configuration"
.cindex "format" "time interval"
A time interval is specified as a sequence of numbers, each followed by one of
the following letters, with no intervening white space:

.table2 30pt
.irow &%s%& seconds
.irow &%m%& minutes
.irow &%h%& hours
.irow &%d%& days
.irow &%w%& weeks
.endtable

For example, &"3h50m"& specifies 3 hours and 50 minutes. The values of time
intervals are output in the same format. Exim does not restrict the values; it
is perfectly acceptable, for example, to specify &"90m"& instead of &"1h30m"&.



.section "String values" "SECTstrings"
.cindex "string" "format of configuration values"
.cindex "format" "string"
If an option's type is specified as &"string"&, the value can be specified with
or without double-quotes. If it does not start with a double-quote, the value
consists of the remainder of the line plus any continuation lines, starting at
the first character after any leading white space, with trailing white space
removed, and with no interpretation of the characters in the string. Because
Exim removes comment lines (those beginning with #) at an early stage, they can
appear in the middle of a multi-line string. The following two settings are
therefore equivalent:
.code
trusted_users = uucp:mail
trusted_users = uucp:\
                # This comment line is ignored
                mail
.endd
.cindex "string" "quoted"
.cindex "escape characters in quoted strings"
If a string does start with a double-quote, it must end with a closing
double-quote, and any backslash characters other than those used for line
continuation are interpreted as escape characters, as follows:

.table2 100pt
.irow &`\\`&                     "single backslash"
.irow &`\n`&                     "newline"
.irow &`\r`&                     "carriage return"
.irow &`\t`&                     "tab"
.irow "&`\`&<&'octal digits'&>"  "up to 3 octal digits specify one character"
.irow "&`\x`&<&'hex digits'&>"   "up to 2 hexadecimal digits specify one &&&
                                   character"
.endtable

If a backslash is followed by some other character, including a double-quote
character, that character replaces the pair.

Quoting is necessary only if you want to make use of the backslash escapes to
insert special characters, or if you need to specify a value with leading or
trailing spaces. These cases are rare, so quoting is almost never needed in
current versions of Exim. In versions of Exim before 3.14, quoting was required
in order to continue lines, so you may come across older configuration files
and examples that apparently quote unnecessarily.


.section "Expanded strings" "SECID51"
.cindex "expansion" "definition of"
Some strings in the configuration file are subjected to &'string expansion'&,
by which means various parts of the string may be changed according to the
circumstances (see chapter &<<CHAPexpand>>&). The input syntax for such strings
is as just described; in particular, the handling of backslashes in quoted
strings is done as part of the input process, before expansion takes place.
However, backslash is also an escape character for the expander, so any
backslashes that are required for that reason must be doubled if they are
within a quoted configuration string.


.section "User and group names" "SECID52"
.cindex "user name" "format of"
.cindex "format" "user name"
.cindex "groups" "name format"
.cindex "format" "group name"
User and group names are specified as strings, using the syntax described
above, but the strings are interpreted specially. A user or group name must
either consist entirely of digits, or be a name that can be looked up using the
&[getpwnam()]& or &[getgrnam()]& function, as appropriate.


.section "List construction" "SECTlistconstruct"
.cindex "list" "syntax of in configuration"
.cindex "format" "list item in configuration"
.cindex "string" "list, definition of"
The data for some configuration options is a list of items, with colon as the
default separator. Many of these options are shown with type &"string list"& in
the descriptions later in this document. Others are listed as &"domain list"&,
&"host list"&, &"address list"&, or &"local part list"&. Syntactically, they
are all the same; however, those other than &"string list"& are subject to
particular kinds of interpretation, as described in chapter
&<<CHAPdomhosaddlists>>&.

In all these cases, the entire list is treated as a single string as far as the
input syntax is concerned. The &%trusted_users%& setting in section
&<<SECTstrings>>& above is an example. If a colon is actually needed in an item
in a list, it must be entered as two colons. Leading and trailing white space
on each item in a list is ignored. This makes it possible to include items that
start with a colon, and in particular, certain forms of IPv6 address. For
example, the list
.code
local_interfaces = 127.0.0.1 : ::::1
.endd
contains two IP addresses, the IPv4 address 127.0.0.1 and the IPv6 address ::1.

&*Note*&: Although leading and trailing white space is ignored in individual
list items, it is not ignored when parsing the list. The spaces around the first
colon in the example above are necessary. If they were not there, the list would
be interpreted as the two items 127.0.0.1:: and 1.

.subsection "Changing list separators" "SECTlistsepchange"
.cindex "list separator" "changing"
.cindex "IPv6" "addresses in lists"
Doubling colons in IPv6 addresses is an unwelcome chore, so a mechanism was
introduced to allow the separator character to be changed. If a list begins
with a left angle bracket, followed by any punctuation character, that
character is used instead of colon as the list separator. For example, the list
above can be rewritten to use a semicolon separator like this:
.code
local_interfaces = <; 127.0.0.1 ; ::1
.endd
This facility applies to all lists, with the exception of the list in
&%log_file_path%&. It is recommended that the use of non-colon separators be
confined to circumstances where they really are needed.

.cindex "list separator" "newline as"
.cindex "newline" "as list separator"
It is also possible to use newline and other control characters (those with
code values less than 32, plus DEL) as separators in lists. Such separators
must be provided literally at the time the list is processed. For options that
are string-expanded, you can write the separator using a normal escape
sequence. This will be processed by the expander before the string is
interpreted as a list. For example, if a newline-separated list of domains is
generated by a lookup, you can process it directly by a line such as this:
.code
domains = <\n ${lookup mysql{.....}}
.endd
This avoids having to change the list separator in such data. You are unlikely
to want to use a control character as a separator in an option that is not
expanded, because the value is literal text. However, it can be done by giving
the value in quotes. For example:
.code
local_interfaces = "<\n 127.0.0.1 \n ::1"
.endd
Unlike printing character separators, which can be included in list items by
doubling, it is not possible to include a control character as data when it is
set as the separator. Two such characters in succession are interpreted as
enclosing an empty list item.



.subsection "Empty items in lists" "SECTempitelis"
.cindex "list" "empty item in"
An empty item at the end of a list is always ignored. In other words, trailing
separator characters are ignored. Thus, the list in
.code
senders = user@domain :
.endd
contains only a single item. If you want to include an empty string as one item
in a list, it must not be the last item. For example, this list contains three
items, the second of which is empty:
.code
senders = user1@domain : : user2@domain
.endd
&*Note*&: There must be white space between the two colons, as otherwise they
are interpreted as representing a single colon data character (and the list
would then contain just one item). If you want to specify a list that contains
just one, empty item, you can do it as in this example:
.code
senders = :
.endd
In this case, the first item is empty, and the second is discarded because it
is at the end of the list.




.section "Format of driver configurations" "SECTfordricon"
.cindex "drivers" "configuration format"
There are separate parts in the configuration for defining routers, transports,
and authenticators. In each part, you are defining a number of driver
instances, each with its own set of options. Each driver instance is defined by
a sequence of lines like this:
.display
<&'instance name'&>:
  <&'option'&>
  ...
  <&'option'&>
.endd
In the following example, the instance name is &(localuser)&, and it is
followed by three options settings:
.code
localuser:
  driver = accept
  check_local_user
  transport = local_delivery
.endd
For each driver instance, you specify which Exim code module it uses &-- by the
setting of the &%driver%& option &-- and (optionally) some configuration
settings. For example, in the case of transports, if you want a transport to
deliver with SMTP you would use the &(smtp)& driver; if you want to deliver to
a local file you would use the &(appendfile)& driver. Each of the drivers is
described in detail in its own separate chapter later in this manual.

You can have several routers, transports, or authenticators that are based on
the same underlying driver (each must have a different instance name).

The order in which routers are defined is important, because addresses are
passed to individual routers one by one, in order. The order in which
transports are defined does not matter at all. The order in which
authenticators are defined is used only when Exim, as a client, is searching
them to find one that matches an authentication mechanism offered by the
server.

.cindex "generic options"
.cindex "options" "generic &-- definition of"
Within a driver instance definition, there are two kinds of option: &'generic'&
and &'private'&. The generic options are those that apply to all drivers of the
same type (that is, all routers, all transports or all authenticators). The
&%driver%& option is a generic option that must appear in every definition.
.cindex "private options"
The private options are special for each driver, and none need appear, because
they all have default values.

The options may appear in any order, except that the &%driver%& option must
precede any private options, since these depend on the particular driver. For
this reason, it is recommended that &%driver%& always be the first option.

Driver instance names, which are used for reference in log entries and
elsewhere, can be any sequence of letters, digits, and underscores (starting
with a letter) and must be unique among drivers of the same type. A router and
a transport (for example) can each have the same name, but no two router
instances can have the same name. The name of a driver instance should not be
confused with the name of the underlying driver module. For example, the
configuration lines:
.code
remote_smtp:
  driver = smtp
.endd
create an instance of the &(smtp)& transport driver whose name is
&(remote_smtp)&. The same driver code can be used more than once, with
different instance names and different option settings each time. A second
instance of the &(smtp)& transport, with different options, might be defined
thus:
.code
special_smtp:
  driver = smtp
  port = 1234
  command_timeout = 10s
.endd
The names &(remote_smtp)& and &(special_smtp)& would be used to reference
these transport instances from routers, and these names would appear in log
lines.

Comment lines may be present in the middle of driver specifications. The full
list of option settings for any particular driver instance, including all the
defaulted values, can be extracted by making use of the &%-bP%& command line
option.






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The default configuration file" "CHAPdefconfil"
.scindex IIDconfiwal "configuration file" "default &""walk through""&"
.cindex "default" "configuration file &""walk through""&"
The default configuration file supplied with Exim as &_src/configure.default_&
is sufficient for a host with simple mail requirements. As an introduction to
the way Exim is configured, this chapter &"walks through"& the default
configuration, giving brief explanations of the settings. Detailed descriptions
of the options are given in subsequent chapters. The default configuration file
itself contains extensive comments about ways you might want to modify the
initial settings. However, note that there are many options that are not
mentioned at all in the default configuration.



.section "Macros" "SECTdefconfmacros"
All macros should be defined before any options.

One macro is specified, but commented out, in the default configuration:
.code
# ROUTER_SMARTHOST=MAIL.HOSTNAME.FOR.CENTRAL.SERVER.EXAMPLE
.endd
If all off-site mail is expected to be delivered to a "smarthost", then set the
hostname here and uncomment the macro.  This will affect which router is used
later on.  If this is left commented out, then Exim will perform direct-to-MX
deliveries using a &(dnslookup)& router.

In addition to macros defined here, Exim includes a number of built-in macros
to enable configuration to be guarded by a binary built with support for a
given feature.  See section &<<SECTbuiltinmacros>>& for more details.


.section "Main configuration settings" "SECTdefconfmain"
The main (global) configuration option settings section must always come first
in the file, after the macros.
The first thing you'll see in the file, after some initial comments, is the line
.code
# primary_hostname =
.endd
This is a commented-out setting of the &%primary_hostname%& option. Exim needs
to know the official, fully qualified name of your host, and this is where you
can specify it. However, in most cases you do not need to set this option. When
it is unset, Exim uses the &[uname()]& system function to obtain the host name.

The first three non-comment configuration lines are as follows:
.code
domainlist local_domains    = @
domainlist relay_to_domains =
hostlist   relay_from_hosts = 127.0.0.1
.endd
These are not, in fact, option settings. They are definitions of two named
domain lists and one named host list. Exim allows you to give names to lists of
domains, hosts, and email addresses, in order to make it easier to manage the
configuration file (see section &<<SECTnamedlists>>&).

The first line defines a domain list called &'local_domains'&; this is used
later in the configuration to identify domains that are to be delivered
on the local host.

.cindex "@ in a domain list"
There is just one item in this list, the string &"@"&. This is a special form
of entry which means &"the name of the local host"&. Thus, if the local host is
called &'a.host.example'&, mail to &'any.user@a.host.example'& is expected to
be delivered locally. Because the local host's name is referenced indirectly,
the same configuration file can be used on different hosts.

The second line defines a domain list called &'relay_to_domains'&, but the
list itself is empty. Later in the configuration we will come to the part that
controls mail relaying through the local host; it allows relaying to any
domains in this list. By default, therefore, no relaying on the basis of a mail
domain is permitted.

The third line defines a host list called &'relay_from_hosts'&. This list is
used later in the configuration to permit relaying from any host or IP address
that matches the list. The default contains just the IP address of the IPv4
loopback interface, which means that processes on the local host are able to
submit mail for relaying by sending it over TCP/IP to that interface. No other
hosts are permitted to submit messages for relaying.

Just to be sure there's no misunderstanding: at this point in the configuration
we aren't actually setting up any controls. We are just defining some domains
and hosts that will be used in the controls that are specified later.

The next two configuration lines are genuine option settings:
.code
acl_smtp_rcpt = acl_check_rcpt
acl_smtp_data = acl_check_data
.endd
These options specify &'Access Control Lists'& (ACLs) that are to be used
during an incoming SMTP session for every recipient of a message (every RCPT
command), and after the contents of the message have been received,
respectively. The names of the lists are &'acl_check_rcpt'& and
&'acl_check_data'&, and we will come to their definitions below, in the ACL
section of the configuration. The RCPT ACL controls which recipients are
accepted for an incoming message &-- if a configuration does not provide an ACL
to check recipients, no SMTP mail can be accepted. The DATA ACL allows the
contents of a message to be checked.

Two commented-out option settings are next:
.code
# av_scanner = clamd:/tmp/clamd
# spamd_address = 127.0.0.1 783
.endd
These are example settings that can be used when Exim is compiled with the
content-scanning extension. The first specifies the interface to the virus
scanner, and the second specifies the interface to SpamAssassin. Further
details are given in chapter &<<CHAPexiscan>>&.

Three more commented-out option settings follow:
.code
# tls_advertise_hosts = *
# tls_certificate = /etc/ssl/exim.crt
# tls_privatekey = /etc/ssl/exim.pem
.endd
These are example settings that can be used when Exim is compiled with
support for TLS (aka SSL) as described in section &<<SECTinctlsssl>>&. The
first one specifies the list of clients that are allowed to use TLS when
connecting to this server; in this case, the wildcard means all clients. The
other options specify where Exim should find its TLS certificate and private
key, which together prove the server's identity to any clients that connect.
More details are given in chapter &<<CHAPTLS>>&.

Another two commented-out option settings follow:
.code
# daemon_smtp_ports = 25 : 465 : 587
# tls_on_connect_ports = 465
.endd
.cindex "port" "465 and 587"
.cindex "port" "for message submission"
.cindex "message" "submission, ports for"
.cindex "submissions protocol"
.cindex "smtps protocol"
.cindex "ssmtp protocol"
.cindex "SMTP" "submissions protocol"
.cindex "SMTP" "ssmtp protocol"
.cindex "SMTP" "smtps protocol"
These options provide better support for roaming users who wish to use this
server for message submission. They are not much use unless you have turned on
TLS (as described in the previous paragraph) and authentication (about which
more in section &<<SECTdefconfauth>>&).
Mail submission from mail clients (MUAs) should be separate from inbound mail
to your domain (MX delivery) for various good reasons (eg, ability to impose
much saner TLS protocol and ciphersuite requirements without unintended
consequences).
&url(https://www.rfc-editor.org/rfc/rfc6409,RFC 6409) (previously 4409)
specifies use of port 587 for SMTP Submission,
which uses STARTTLS, so this is the &"submission"& port.
&url(https://www.rfc-editor.org/rfc/rfc8314,RFC 8314)
specifies use of port 465 as the &"submissions"& protocol,
which should be used in preference to 587.
You should also consider deploying SRV records to help clients find
these ports.
Older names for &"submissions"& are &"smtps"& and &"ssmtp"&.

Two more commented-out options settings follow:
.code
# qualify_domain =
# qualify_recipient =
.endd
The first of these specifies a domain that Exim uses when it constructs a
complete email address from a local login name. This is often needed when Exim
receives a message from a local process. If you do not set &%qualify_domain%&,
the value of &%primary_hostname%& is used. If you set both of these options,
you can have different qualification domains for sender and recipient
addresses. If you set only the first one, its value is used in both cases.

.cindex "domain literal" "recognizing format"
The following line must be uncommented if you want Exim to recognize
addresses of the form &'user@[10.11.12.13]'& that is, with a &"domain literal"&
(an IP address within square brackets) instead of a named domain.
.code
# allow_domain_literals
.endd
The RFCs still require this form, but many people think that in the modern
Internet it makes little sense to permit mail to be sent to specific hosts by
quoting their IP addresses. This ancient format has been used by people who
try to abuse hosts by using them for unwanted relaying. However, some
people believe there are circumstances (for example, messages addressed to
&'postmaster'&) where domain literals are still useful.

The next configuration line is a kind of trigger guard:
.code
never_users = root
.endd
It specifies that no delivery must ever be run as the root user. The normal
convention is to set up &'root'& as an alias for the system administrator. This
setting is a guard against slips in the configuration.
The list of users specified by &%never_users%& is not, however, the complete
list; the build-time configuration in &_Local/Makefile_& has an option called
FIXED_NEVER_USERS specifying a list that cannot be overridden. The
contents of &%never_users%& are added to this list. By default
FIXED_NEVER_USERS also specifies root.

When a remote host connects to Exim in order to send mail, the only information
Exim has about the host's identity is its IP address. The next configuration
line,
.code
host_lookup = *
.endd
specifies that Exim should do a reverse DNS lookup on all incoming connections,
in order to get a host name. This improves the quality of the logging
information, but if you feel it is too expensive, you can remove it entirely,
or restrict the lookup to hosts on &"nearby"& networks.
Note that it is not always possible to find a host name from an IP address,
because not all DNS reverse zones are maintained, and sometimes DNS servers are
unreachable.

The next two lines are concerned with &'ident'& callbacks, as defined by
&url(https://www.rfc-editor.org/rfc/rfc1413,RFC 1413)
(hence their names):
.code
rfc1413_hosts = *
rfc1413_query_timeout = 0s
.endd
These settings cause Exim to avoid ident callbacks for all incoming SMTP calls.
Few hosts offer RFC1413 service these days; calls have to be
terminated by a timeout and this needlessly delays the startup
of an incoming SMTP connection.
If you have hosts for which you trust RFC1413 and need this
information, you can change this.

This line enables an efficiency SMTP option.  It is negotiated by clients
and not expected to cause problems but can be disabled if needed.
.code
prdr_enable = true
.endd

When Exim receives messages over SMTP connections, it expects all addresses to
be fully qualified with a domain, as required by the SMTP definition. However,
if you are running a server to which simple clients submit messages, you may
find that they send unqualified addresses. The two commented-out options:
.code
# sender_unqualified_hosts =
# recipient_unqualified_hosts =
.endd
show how you can specify hosts that are permitted to send unqualified sender
and recipient addresses, respectively.

The &%log_selector%& option is used to increase the detail of logging
over the default:
.code
log_selector = +smtp_protocol_error +smtp_syntax_error \
               +tls_certificate_verified
.endd

The &%percent_hack_domains%& option is also commented out:
.code
# percent_hack_domains =
.endd
It provides a list of domains for which the &"percent hack"& is to operate.
This is an almost obsolete form of explicit email routing. If you do not know
anything about it, you can safely ignore this topic.

The next two settings in the main part of the default configuration are
concerned with messages that have been &"frozen"& on Exim's queue. When a
message is frozen, Exim no longer continues to try to deliver it. Freezing
occurs when a bounce message encounters a permanent failure because the sender
address of the original message that caused the bounce is invalid, so the
bounce cannot be delivered. This is probably the most common case, but there
are also other conditions that cause freezing, and frozen messages are not
always bounce messages.
.code
ignore_bounce_errors_after = 2d
timeout_frozen_after = 7d
.endd
The first of these options specifies that failing bounce messages are to be
discarded after 2 days in the queue. The second specifies that any frozen
message (whether a bounce message or not) is to be timed out (and discarded)
after a week. In this configuration, the first setting ensures that no failing
bounce message ever lasts a week.

Exim queues it's messages in a spool directory. If you expect to have
large queues, you may consider using this option. It splits the spool
directory into subdirectories to avoid file system degradation from
many files in a single directory, resulting in better performance.
Manual manipulation of queued messages becomes more complex (though fortunately
not often needed).
.code
# split_spool_directory = true
.endd

In an ideal world everybody follows the standards. For non-ASCII
messages &url(https://www.rfc-editor.org/rfc/rfc2048,RFC 2047)
is a standard, allowing a maximum line length of 76
characters. Exim adheres that standard and won't process messages which
violate this standard. (Even ${rfc2047:...} expansions will fail.)
In particular, the Exim maintainers have had multiple reports of
problems from Russian administrators of issues until they disable this
check, because of some popular, yet buggy, mail composition software.
.code
# check_rfc2047_length = false
.endd

If you need to be strictly RFC compliant you may wish to disable the
8BITMIME advertisement. Use this, if you exchange mails with systems
that are not 8-bit clean.
.code
# accept_8bitmime = false
.endd

Libraries you use may depend on specific environment settings.  This
imposes a security risk (e.g. PATH). There are two lists:
&%keep_environment%& for the variables to import as they are, and
&%add_environment%& for variables we want to set to a fixed value.
Note that TZ is handled separately, by the &%timezone%& runtime
option and by the TIMEZONE_DEFAULT buildtime option.
.code
# keep_environment = ^LDAP
# add_environment = PATH=/usr/bin::/bin
.endd


.section "ACL configuration" "SECID54"
.cindex "default" "ACLs"
.cindex "&ACL;" "default configuration"
In the default configuration, the ACL section follows the main configuration.
It starts with the line
.code
begin acl
.endd
and it contains the definitions of two ACLs, called &'acl_check_rcpt'& and
&'acl_check_data'&, that were referenced in the settings of &%acl_smtp_rcpt%&
and &%acl_smtp_data%& above.

.cindex "RCPT" "ACL for"
The first ACL is used for every RCPT command in an incoming SMTP message. Each
RCPT command specifies one of the message's recipients. The ACL statements
are considered in order, until the recipient address is either accepted or
rejected. The RCPT command is then accepted or rejected, according to the
result of the ACL processing.
.code
acl_check_rcpt:
.endd
This line, consisting of a name terminated by a colon, marks the start of the
ACL, and names it.
.code
accept  hosts = :
.endd
This ACL statement accepts the recipient if the sending host matches the list.
But what does that strange list mean? It doesn't actually contain any host
names or IP addresses. The presence of the colon puts an empty item in the
list; Exim matches this only if the incoming message did not come from a remote
host, because in that case, the remote hostname is empty. The colon is
important. Without it, the list itself is empty, and can never match anything.

What this statement is doing is to accept unconditionally all recipients in
messages that are submitted by SMTP from local processes using the standard
input and output (that is, not using TCP/IP). A number of MUAs operate in this
manner.
.code
deny    domains       = +local_domains
        local_parts   = ^[.] : ^.*[@%!/|]
        message       = Restricted characters in address

deny    domains       = !+local_domains
        local_parts   = ^[./|] : ^.*[@%!] : ^.*/\\.\\./
        message       = Restricted characters in address
.endd
These statements are concerned with local parts that contain any of the
characters &"@"&, &"%"&, &"!"&, &"/"&, &"|"&, or dots in unusual places.
Although these characters are entirely legal in local parts (in the case of
&"@"& and leading dots, only if correctly quoted), they do not commonly occur
in Internet mail addresses.

The first three have in the past been associated with explicitly routed
addresses (percent is still sometimes used &-- see the &%percent_hack_domains%&
option). Addresses containing these characters are regularly tried by spammers
in an attempt to bypass relaying restrictions, and also by open relay testing
programs. Unless you really need them it is safest to reject these characters
at this early stage. This configuration is heavy-handed in rejecting these
characters for all messages it accepts from remote hosts. This is a deliberate
policy of being as safe as possible.

The first rule above is stricter, and is applied to messages that are addressed
to one of the local domains handled by this host. This is implemented by the
first condition, which restricts it to domains that are listed in the
&'local_domains'& domain list. The &"+"& character is used to indicate a
reference to a named list. In this configuration, there is just one domain in
&'local_domains'&, but in general there may be many.

The second condition on the first statement uses two regular expressions to
block local parts that begin with a dot or contain &"@"&, &"%"&, &"!"&, &"/"&,
or &"|"&. If you have local accounts that include these characters, you will
have to modify this rule.

Empty components (two dots in a row) are not valid in
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822),
but Exim
allows them because they have been encountered in practice. (Consider the
common convention of local parts constructed as
&"&'first-initial.second-initial.family-name'&"& when applied to someone like
the author of Exim, who has no second initial.) However, a local part starting
with a dot or containing &"/../"& can cause trouble if it is used as part of a
filename (for example, for a mailing list). This is also true for local parts
that contain slashes. A pipe symbol can also be troublesome if the local part
is incorporated unthinkingly into a shell command line.

The second rule above applies to all other domains, and is less strict. This
allows your own users to send outgoing messages to sites that use slashes
and vertical bars in their local parts. It blocks local parts that begin
with a dot, slash, or vertical bar, but allows these characters within the
local part. However, the sequence &"/../"& is barred. The use of &"@"&, &"%"&,
and &"!"& is blocked, as before. The motivation here is to prevent your users
(or your users' viruses) from mounting certain kinds of attack on remote sites.
.code
accept  local_parts   = postmaster
        domains       = +local_domains
.endd
This statement, which has two conditions, accepts an incoming address if the
local part is &'postmaster'& and the domain is one of those listed in the
&'local_domains'& domain list. The &"+"& character is used to indicate a
reference to a named list. In this configuration, there is just one domain in
&'local_domains'&, but in general there may be many.

The presence of this statement means that mail to postmaster is never blocked
by any of the subsequent tests. This can be helpful while sorting out problems
in cases where the subsequent tests are incorrectly denying access.
.code
require verify        = sender
.endd
This statement requires the sender address to be verified before any subsequent
ACL statement can be used. If verification fails, the incoming recipient
address is refused. Verification consists of trying to route the address, to
see if a bounce message could be delivered to it. In the case of remote
addresses, basic verification checks only the domain, but &'callouts'& can be
used for more verification if required. Section &<<SECTaddressverification>>&
discusses the details of address verification.
.code
accept  hosts         = +relay_from_hosts
        control       = submission
.endd
This statement accepts the address if the message is coming from one of the
hosts that are defined as being allowed to relay through this host. Recipient
verification is omitted here, because in many cases the clients are dumb MUAs
that do not cope well with SMTP error responses. For the same reason, the
second line specifies &"submission mode"& for messages that are accepted. This
is described in detail in section &<<SECTsubmodnon>>&; it causes Exim to fix
messages that are deficient in some way, for example, because they lack a
&'Date:'& header line. If you are actually relaying out from MTAs, you should
probably add recipient verification here, and disable submission mode.
.code
accept  authenticated = *
        control       = submission
.endd
This statement accepts the address if the client host has authenticated itself.
Submission mode is again specified, on the grounds that such messages are most
likely to come from MUAs. The default configuration does not define any
authenticators, though it does include some nearly complete commented-out
examples described in &<<SECTdefconfauth>>&. This means that no client can in
fact authenticate until you complete the authenticator definitions.
.code
require message = relay not permitted
        domains = +local_domains : +relay_to_domains
.endd
This statement rejects the address if its domain is neither a local domain nor
one of the domains for which this host is a relay.
.code
require verify = recipient
.endd
This statement requires the recipient address to be verified; if verification
fails, the address is rejected.
.code
# deny    dnslists    = black.list.example
#         message     = rejected because $sender_host_address \
#                       is in a black list at $dnslist_domain\n\
#                       $dnslist_text
#
# warn    dnslists    = black.list.example
#         add_header  = X-Warning: $sender_host_address is in \
#                       a black list at $dnslist_domain
#         log_message = found in $dnslist_domain
.endd
These commented-out lines are examples of how you could configure Exim to check
sending hosts against a DNS black list. The first statement rejects messages
from blacklisted hosts, whereas the second just inserts a warning header
line.
.code
# require verify = csa
.endd
This commented-out line is an example of how you could turn on client SMTP
authorization (CSA) checking. Such checks do DNS lookups for special SRV
records.
.code
accept
.endd
The final statement in the first ACL unconditionally accepts any recipient
address that has successfully passed all the previous tests.
.code
acl_check_data:
.endd
This line marks the start of the second ACL, and names it. Most of the contents
of this ACL are commented out:
.code
# deny    malware   = *
#         message   = This message contains a virus \
#                     ($malware_name).
.endd
These lines are examples of how to arrange for messages to be scanned for
viruses when Exim has been compiled with the content-scanning extension, and a
suitable virus scanner is installed. If the message is found to contain a
virus, it is rejected with the given custom error message.
.code
# warn    spam      = nobody
#         message   = X-Spam_score: $spam_score\n\
#                     X-Spam_score_int: $spam_score_int\n\
#                     X-Spam_bar: $spam_bar\n\
#                     X-Spam_report: $spam_report
.endd
These lines are an example of how to arrange for messages to be scanned by
SpamAssassin when Exim has been compiled with the content-scanning extension,
and SpamAssassin has been installed. The SpamAssassin check is run with
&`nobody`& as its user parameter, and the results are added to the message as a
series of extra header line. In this case, the message is not rejected,
whatever the spam score.
.code
accept
.endd
This final line in the DATA ACL accepts the message unconditionally.


.section "Router configuration" "SECID55"
.cindex "default" "routers"
.cindex "routers" "default"
The router configuration comes next in the default configuration, introduced
by the line
.code
begin routers
.endd
Routers are the modules in Exim that make decisions about where to send
messages. An address is passed to each router, in turn, until it is either
accepted, or failed. This means that the order in which you define the routers
matters. Each router is fully described in its own chapter later in this
manual. Here we give only brief overviews.
.code
# domain_literal:
#   driver = ipliteral
#   domains = !+local_domains
#   transport = remote_smtp
.endd
.cindex "domain literal" "default router"
This router is commented out because the majority of sites do not want to
support domain literal addresses (those of the form &'user@[10.9.8.7]'&). If
you uncomment this router, you also need to uncomment the setting of
&%allow_domain_literals%& in the main part of the configuration.

Which router is used next depends upon whether or not the ROUTER_SMARTHOST
macro has been defined, per
.code
.ifdef ROUTER_SMARTHOST
smarthost:
#...
.else
dnslookup:
#...
.endif
.endd

If ROUTER_SMARTHOST has been defined, either at the top of the file or on the
command-line, then we route all non-local mail to that smarthost; otherwise, we'll
perform DNS lookups for direct-to-MX lookup.  Any mail which is to a local domain will
skip these routers because of the &%domains%& option.

.code
smarthost:
  driver = manualroute
  domains = ! +local_domains
  transport = smarthost_smtp
  route_data = ROUTER_SMARTHOST
  ignore_target_hosts = <; 0.0.0.0 ; 127.0.0.0/8 ; ::1
  no_more
.endd
This router only handles mail which is not to any local domains; this is
specified by the line
.code
domains = ! +local_domains
.endd
The &%domains%& option lists the domains to which this router applies, but the
exclamation mark is a negation sign, so the router is used only for domains
that are not in the domain list called &'local_domains'& (which was defined at
the start of the configuration). The plus sign before &'local_domains'&
indicates that it is referring to a named list. Addresses in other domains are
passed on to the following routers.

The name of the router driver is &(manualroute)& because we are manually
specifying how mail should be routed onwards, instead of using DNS MX.
While the name of this router instance is arbitrary, the &%driver%& option must
be one of the driver modules that is in the Exim binary.

With no pre-conditions other than &%domains%&, all mail for non-local domains
will be handled by this router, and the &%no_more%& setting will ensure that no
other routers will be used for messages matching the pre-conditions.  See
&<<SECTrouprecon>>& for more on how the pre-conditions apply.  For messages which
are handled by this router, we provide a hostname to deliver to in &%route_data%&
and the macro supplies the value; the address is then queued for the
&(smarthost_smtp)& transport.

.code
dnslookup:
  driver = dnslookup
  domains = ! +local_domains
  transport = remote_smtp
  ignore_target_hosts = 0.0.0.0 : 127.0.0.0/8
  no_more
.endd
The &%domains%& option behaves as per smarthost, above.

The name of the router driver is &(dnslookup)&,
and is specified by the &%driver%& option. Do not be confused by the fact that
the name of this router instance is the same as the name of the driver. The
instance name is arbitrary, but the name set in the &%driver%& option must be
one of the driver modules that is in the Exim binary.

The &(dnslookup)& router routes addresses by looking up their domains in the
DNS in order to obtain a list of hosts to which the address is routed. If the
router succeeds, the address is queued for the &(remote_smtp)& transport, as
specified by the &%transport%& option. If the router does not find the domain
in the DNS, no further routers are tried because of the &%no_more%& setting, so
the address fails and is bounced.

The &%ignore_target_hosts%& option specifies a list of IP addresses that are to
be entirely ignored. This option is present because a number of cases have been
encountered where MX records in the DNS point to host names
whose IP addresses are 0.0.0.0 or are in the 127 subnet (typically 127.0.0.1).
Completely ignoring these IP addresses causes Exim to fail to route the
email address, so it bounces. Otherwise, Exim would log a routing problem, and
continue to try to deliver the message periodically until the address timed
out.
.code
system_aliases:
  driver = redirect
  allow_fail
  allow_defer
  data = ${lookup{$local_part}lsearch{/etc/aliases}}
# user = exim
  file_transport = address_file
  pipe_transport = address_pipe
.endd
Control reaches this and subsequent routers only for addresses in the local
domains. This router checks to see whether the local part is defined as an
alias in the &_/etc/aliases_& file, and if so, redirects it according to the
data that it looks up from that file. If no data is found for the local part,
the value of the &%data%& option is empty, causing the address to be passed to
the next router.

&_/etc/aliases_& is a conventional name for the system aliases file that is
often used. That is why it is referenced by from the default configuration
file. However, you can change this by setting SYSTEM_ALIASES_FILE in
&_Local/Makefile_& before building Exim.
.code
userforward:
  driver = redirect
  check_local_user
# local_part_suffix = +* : -*
# local_part_suffix_optional
  file = $home/.forward
# allow_filter
  no_verify
  no_expn
  check_ancestor
  file_transport = address_file
  pipe_transport = address_pipe
  reply_transport = address_reply
.endd
This is the most complicated router in the default configuration. It is another
redirection router, but this time it is looking for forwarding data set up by
individual users. The &%check_local_user%& setting specifies a check that the
local part of the address is the login name of a local user. If it is not, the
router is skipped. The two commented options that follow &%check_local_user%&,
namely:
.code
# local_part_suffix = +* : -*
# local_part_suffix_optional
.endd
.vindex "&$local_part_suffix$&"
show how you can specify the recognition of local part suffixes. If the first
is uncommented, a suffix beginning with either a plus or a minus sign, followed
by any sequence of characters, is removed from the local part and placed in the
variable &$local_part_suffix$&. The second suffix option specifies that the
presence of a suffix in the local part is optional. When a suffix is present,
the check for a local login uses the local part with the suffix removed.

When a local user account is found, the file called &_.forward_& in the user's
home directory is consulted. If it does not exist, or is empty, the router
declines. Otherwise, the contents of &_.forward_& are interpreted as
redirection data (see chapter &<<CHAPredirect>>& for more details).

.cindex "Sieve filter" "enabling in default router"
Traditional &_.forward_& files contain just a list of addresses, pipes, or
files. Exim supports this by default. However, if &%allow_filter%& is set (it
is commented out by default), the contents of the file are interpreted as a set
of Exim or Sieve filtering instructions, provided the file begins with &"#Exim
filter"& or &"#Sieve filter"&, respectively. User filtering is discussed in the
separate document entitled &'Exim's interfaces to mail filtering'&.

The &%no_verify%& and &%no_expn%& options mean that this router is skipped when
verifying addresses, or when running as a consequence of an SMTP EXPN command.
There are two reasons for doing this:

.olist
Whether or not a local user has a &_.forward_& file is not really relevant when
checking an address for validity; it makes sense not to waste resources doing
unnecessary work.
.next
More importantly, when Exim is verifying addresses or handling an EXPN
command during an SMTP session, it is running as the Exim user, not as root.
The group is the Exim group, and no additional groups are set up.
It may therefore not be possible for Exim to read users' &_.forward_& files at
this time.
.endlist

The setting of &%check_ancestor%& prevents the router from generating a new
address that is the same as any previous address that was redirected. (This
works round a problem concerning a bad interaction between aliasing and
forwarding &-- see section &<<SECTredlocmai>>&).

The final three option settings specify the transports that are to be used when
forwarding generates a direct delivery to a file, or to a pipe, or sets up an
auto-reply, respectively. For example, if a &_.forward_& file contains
.code
a.nother@elsewhere.example, /home/spqr/archive
.endd
the delivery to &_/home/spqr/archive_& is done by running the &%address_file%&
transport.
.code
localuser:
  driver = accept
  check_local_user
# local_part_suffix = +* : -*
# local_part_suffix_optional
  transport = local_delivery
.endd
The final router sets up delivery into local mailboxes, provided that the local
part is the name of a local login, by accepting the address and assigning it to
the &(local_delivery)& transport. Otherwise, we have reached the end of the
routers, so the address is bounced. The commented suffix settings fulfil the
same purpose as they do for the &(userforward)& router.


.section "Transport configuration" "SECID56"
.cindex "default" "transports"
.cindex "transports" "default"
Transports define mechanisms for actually delivering messages. They operate
only when referenced from routers, so the order in which they are defined does
not matter. The transports section of the configuration starts with
.code
begin transports
.endd
Two remote transports and four local transports are defined.
.code
remote_smtp:
  driver = smtp
  message_size_limit = ${if > {$max_received_linelength}{998} {1}{0}}
.ifdef _HAVE_PRDR
  hosts_try_prdr = *
.endif
.endd
This transport is used for delivering messages over SMTP connections.
The list of remote hosts comes from the router.
The &%message_size_limit%& usage is a hack to avoid sending on messages
with over-long lines.

The &%hosts_try_prdr%& option enables an efficiency SMTP option.  It is
negotiated between client and server and not expected to cause problems
but can be disabled if needed.  The built-in macro _HAVE_PRDR guards the
use of the &%hosts_try_prdr%& configuration option.

The other remote transport is used when delivering to a specific smarthost
with whom there must be some kind of existing relationship, instead of the
usual federated system.

.code
smarthost_smtp:
  driver = smtp
  message_size_limit = ${if > {$max_received_linelength}{998} {1}{0}}
  multi_domain
  #
.ifdef _HAVE_TLS
  # Comment out any of these which you have to, then file a Support
  # request with your smarthost provider to get things fixed:
  hosts_require_tls = *
  tls_verify_hosts = *
  # As long as tls_verify_hosts is enabled, this this will have no effect,
  # but if you have to comment it out then this will at least log whether
  # you succeed or not:
  tls_try_verify_hosts = *
  #
  # The SNI name should match the name which we'll expect to verify;
  # many mail systems don't use SNI and this doesn't matter, but if it does,
  # we need to send a name which the remote site will recognize.
  # This _should_ be the name which the smarthost operators specified as
  # the hostname for sending your mail to.
  tls_sni = ROUTER_SMARTHOST
  #
.ifdef _HAVE_OPENSSL
  tls_require_ciphers = HIGH:!aNULL:@STRENGTH
.endif
.ifdef _HAVE_GNUTLS
  tls_require_ciphers = SECURE192:-VERS-SSL3.0:-VERS-TLS1.0:-VERS-TLS1.1
.endif
.endif
.ifdef _HAVE_PRDR
  hosts_try_prdr = *
.endif
.endd
After the same &%message_size_limit%& hack, we then specify that this Transport
can handle messages to multiple domains in one run.  The assumption here is
that you're routing all non-local mail to the same place and that place is
happy to take all messages from you as quickly as possible.
All other options depend upon built-in macros; if Exim was built without TLS support
then no other options are defined.
If TLS is available, then we configure "stronger than default" TLS ciphersuites
and versions using the &%tls_require_ciphers%& option, where the value to be
used depends upon the library providing TLS.
Beyond that, the options adopt the stance that you should have TLS support available
from your smarthost on today's Internet, so we turn on requiring TLS for the
mail to be delivered, and requiring that the certificate be valid, and match
the expected hostname.  The &%tls_sni%& option can be used by service providers
to select an appropriate certificate to present to you and here we re-use the
ROUTER_SMARTHOST macro, because that is unaffected by CNAMEs present in DNS.
You want to specify the hostname which you'll expect to validate for, and that
should not be subject to insecure tampering via DNS results.

For the &%hosts_try_prdr%& option see the previous transport.

All other options are defaulted.
.code
local_delivery:
  driver = appendfile
  file = /var/mail/$local_part_data
  delivery_date_add
  envelope_to_add
  return_path_add
# group = mail
# mode = 0660
.endd
This &(appendfile)& transport is used for local delivery to user mailboxes in
traditional BSD mailbox format.

We prefer to avoid using &$local_part$& directly to define the mailbox filename,
as it is provided by a potential bad actor.
Instead we use &$local_part_data$&,
the result of looking up &$local_part$& in the user database
(done by using &%check_local_user%& in the the router).

By default &(appendfile)& runs under the uid and gid of the
local user, which requires the sticky bit to be set on the &_/var/mail_&
directory. Some systems use the alternative approach of running mail deliveries
under a particular group instead of using the sticky bit. The commented options
show how this can be done.

Exim adds three headers to the message as it delivers it: &'Delivery-date:'&,
&'Envelope-to:'& and &'Return-path:'&. This action is requested by the three
similarly-named options above.
.code
address_pipe:
  driver = pipe
  return_output
.endd
This transport is used for handling deliveries to pipes that are generated by
redirection (aliasing or users' &_.forward_& files). The &%return_output%&
option specifies that any output on stdout or stderr generated by the pipe is to
be returned to the sender.
.code
address_file:
  driver = appendfile
  delivery_date_add
  envelope_to_add
  return_path_add
.endd
This transport is used for handling deliveries to files that are generated by
redirection. The name of the file is not specified in this instance of
&(appendfile)&, because it comes from the &(redirect)& router.
.code
address_reply:
  driver = autoreply
.endd
This transport is used for handling automatic replies generated by users'
filter files.



.section "Default retry rule" "SECID57"
.cindex "retry" "default rule"
.cindex "default" "retry rule"
The retry section of the configuration file contains rules which affect the way
Exim retries deliveries that cannot be completed at the first attempt. It is
introduced by the line
.code
begin retry
.endd
In the default configuration, there is just one rule, which applies to all
errors:
.code
*   *   F,2h,15m; G,16h,1h,1.5; F,4d,6h
.endd
This causes any temporarily failing address to be retried every 15 minutes for
2 hours, then at intervals starting at one hour and increasing by a factor of
1.5 until 16 hours have passed, then every 6 hours up to 4 days. If an address
is not delivered after 4 days of temporary failure, it is bounced. The time is
measured from first failure, not from the time the message was received.

If the retry section is removed from the configuration, or is empty (that is,
if no retry rules are defined), Exim will not retry deliveries. This turns
temporary errors into permanent errors.


.section "Rewriting configuration" "SECID58"
The rewriting section of the configuration, introduced by
.code
begin rewrite
.endd
contains rules for rewriting addresses in messages as they arrive. There are no
rewriting rules in the default configuration file.



.section "Authenticators configuration" "SECTdefconfauth"
.cindex "AUTH" "configuration"
The authenticators section of the configuration, introduced by
.code
begin authenticators
.endd
defines mechanisms for the use of the SMTP AUTH command. The default
configuration file contains two commented-out example authenticators
which support plaintext username/password authentication using the
standard PLAIN mechanism and the traditional but non-standard LOGIN
mechanism, with Exim acting as the server. PLAIN and LOGIN are enough
to support most MUA software.

The example PLAIN authenticator looks like this:
.code
#PLAIN:
#  driver                  = plaintext
#  server_set_id           = $auth2
#  server_prompts          = :
#  server_condition        = Authentication is not yet configured
#  server_advertise_condition = ${if def:tls_in_cipher }
.endd
And the example LOGIN authenticator looks like this:
.code
#LOGIN:
#  driver                  = plaintext
#  server_set_id           = $auth1
#  server_prompts          = <| Username: | Password:
#  server_condition        = Authentication is not yet configured
#  server_advertise_condition = ${if def:tls_in_cipher }
.endd

The &%server_set_id%& option makes Exim remember the authenticated username
in &$authenticated_id$&, which can be used later in ACLs or routers. The
&%server_prompts%& option configures the &(plaintext)& authenticator so
that it implements the details of the specific authentication mechanism,
i.e. PLAIN or LOGIN. The &%server_advertise_condition%& setting controls
when Exim offers authentication to clients; in the examples, this is only
when TLS or SSL has been started, so to enable the authenticators you also
need to add support for TLS as described in section &<<SECTdefconfmain>>&.

The &%server_condition%& setting defines how to verify that the username and
password are correct. In the examples it just produces an error message.
To make the authenticators work, you can use a string expansion
expression like one of the examples in chapter &<<CHAPplaintext>>&.

Beware that the sequence of the parameters to PLAIN and LOGIN differ; the
usercode and password are in different positions.
Chapter &<<CHAPplaintext>>& covers both.

.ecindex IIDconfiwal



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Regular expressions" "CHAPregexp"

.cindex "regular expressions" "library"
.cindex "PCRE2"
Exim supports the use of regular expressions in many of its options. It
uses the PCRE2 regular expression library; this provides regular expression
matching that is compatible with Perl 5. The syntax and semantics of
regular expressions is discussed in
online Perl manpages, in
many Perl reference books, and also in
Jeffrey Friedl's &'Mastering Regular Expressions'&, which is published by
O'Reilly (see &url(http://www.oreilly.com/catalog/regex2/)).
. --- the http: URL here redirects to another page with the ISBN in the URL
. --- where trying to use https: just redirects back to http:, so sticking
. --- to the old URL for now.  2018-09-07.

The documentation for the syntax and semantics of the regular expressions that
are supported by PCRE2 is included in the PCRE2 distribution, and no further
description is included here. The PCRE2 functions are called from Exim using
the default option settings (that is, with no PCRE2 options set), except that
the PCRE2_CASELESS option is set when the matching is required to be
case-insensitive.

In most cases, when a regular expression is required in an Exim configuration,
it has to start with a circumflex, in order to distinguish it from plain text
or an &"ends with"& wildcard. In this example of a configuration setting, the
second item in the colon-separated list is a regular expression.
.code
domains = a.b.c : ^\\d{3} : *.y.z : ...
.endd
The doubling of the backslash is required because of string expansion that
precedes interpretation &-- see section &<<SECTlittext>>& for more discussion
of this issue, and a way of avoiding the need for doubling backslashes. The
regular expression that is eventually used in this example contains just one
backslash. The circumflex is included in the regular expression, and has the
normal effect of &"anchoring"& it to the start of the string that is being
matched.

There are, however, two cases where a circumflex is not required for the
recognition of a regular expression: these are the &%match%& condition in a
string expansion, and the &%matches%& condition in an Exim filter file. In
these cases, the relevant string is always treated as a regular expression; if
it does not start with a circumflex, the expression is not anchored, and can
match anywhere in the subject string.

In all cases, if you want a regular expression to match at the end of a string,
you must code the $ metacharacter to indicate this. For example:
.code
domains = ^\\d{3}\\.example
.endd
matches the domain &'123.example'&, but it also matches &'123.example.com'&.
You need to use:
.code
domains = ^\\d{3}\\.example\$
.endd
if you want &'example'& to be the top-level domain. The backslash before the
$ is needed because string expansion also interprets dollar characters.



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "File and database lookups" "CHAPfdlookup"
.scindex IIDfidalo1 "file" "lookups"
.scindex IIDfidalo2 "database" "lookups"
.cindex "lookup" "description of"
Exim can be configured to look up data in files or databases as it processes
messages. Two different kinds of syntax are used:

.olist
A string that is to be expanded may contain explicit lookup requests. These
cause parts of the string to be replaced by data that is obtained from the
lookup. Lookups of this type are conditional expansion items. Different results
can be defined for the cases of lookup success and failure. See chapter
&<<CHAPexpand>>&, where string expansions are described in detail.
The key for the lookup is &*specified*& as part of the string to be expanded.
.next
Lists of domains, hosts, and email addresses can contain lookup requests as a
way of avoiding excessively long linear lists. In this case, the data that is
returned by the lookup is often (but not always) discarded; whether the lookup
succeeds or fails is what really counts. These kinds of list are described in
chapter &<<CHAPdomhosaddlists>>&.
Depending on the lookup type (see below)
the key for the lookup may need to be &*specified*& as above
or may be &*implicit*&,
given by the context in which the list is being checked.
.endlist

String expansions, lists, and lookups interact with each other in such a way
that there is no order in which to describe any one of them that does not
involve references to the others. Each of these three chapters makes more sense
if you have read the other two first. If you are reading this for the first
time, be aware that some of it will make a lot more sense after you have read
chapters &<<CHAPdomhosaddlists>>& and &<<CHAPexpand>>&.

.section "Examples of different lookup syntax" "SECID60"
It is easy to confuse the two different kinds of lookup, especially as the
lists that may contain the second kind are always expanded before being
processed as lists. Therefore, they may also contain lookups of the first kind.
Be careful to distinguish between the following two examples:
.code
domains = ${lookup{$sender_host_address}lsearch{/some/file}}
domains = lsearch;/some/file
.endd
.ilist
The first uses a string expansion, the result of which must be a domain list.
The key for an expansion-style lookup must be given explicitly.
No strings have been specified for a successful or a failing lookup; the
defaults in this case are the looked-up data and an empty string, respectively.
The expansion takes place before the string is processed as a list, and the
file that is searched could contain lines like this:
.code
192.168.3.4: domain1:domain2:...
192.168.1.9: domain3:domain4:...
.endd
When the lookup succeeds, the result of the expansion is a list of domains (and
possibly other types of item that are allowed in domain lists).
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using a lookup expansion"
The result of the expansion is not tainted.

.next
In the second example, the lookup is a single item in a domain list. It causes
Exim to use a lookup to see if the domain that is being processed can be found
in the file.
The file could contains lines like this:
.code
domain1:
domain2:
.endd
Any data that follows the keys is not relevant when checking that the domain
matches the list item.

The key for a list-style lookup is implicit, from the lookup context, if
the lookup is a single-key type (see below).
For query-style lookup types the query must be given explicitly.
.endlist

It is possible, though no doubt confusing, to use both kinds of lookup at once.
Consider a file containing lines like this:
.code
192.168.5.6: lsearch;/another/file
.endd
If the value of &$sender_host_address$& is 192.168.5.6, expansion of the
first &%domains%& setting above generates the second setting, which therefore
causes a second lookup to occur.

The lookup type may optionally be followed by a comma
and a comma-separated list of options.
Each option is a &"name=value"& pair.
Whether an option is meaningful depends on the lookup type.

All lookups support the option &"cache=no_rd"&.
If this is given then the cache that Exim manages for lookup results
is not checked before doing the lookup.
The result of the lookup is still written to the cache.

The rest of this chapter describes the different lookup types that are
available. Any of them can be used in any part of the configuration where a
lookup is permitted.


.section "Lookup types" "SECID61"
.cindex "lookup" "types of"
.cindex "single-key lookup" "definition of"
Two different types of data lookup are implemented:

.ilist
The &'single-key'& type requires the specification of a file in which to look,
and a single key to search for. The key must be a non-empty string for the
lookup to succeed. The lookup type determines how the file is searched.
.cindex "tainted data" "single-key lookups"
The file string may not be tainted.

.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using a single-key lookup"
All single-key lookups support the option &"ret=key"&.
If this is given and the lookup
(either underlying implementation or cached value)
returns data, the result is replaced with a non-tainted
version of the lookup key.

.next
.cindex "query-style lookup" "definition of"
The &'query-style'& type accepts a generalized database query. No particular
key value is assumed by Exim for query-style lookups. You can use whichever
Exim variables you need to construct the database query.

For the string-expansion kind of lookups, the query is given in the first
bracketed argument of the &${lookup ...}$& expansion.
For the list-argument kind of lookup the query is given by the remainder of the
list item after the first semicolon.

.cindex "tainted data" "quoting for lookups"
If tainted data is used in the query then it should be quoted by
using the &*${quote_*&<&'lookup-type'&>&*:*&<&'string'&>&*}*& expansion operator
appropriate for the lookup.
.endlist

The code for each lookup type is in a separate source file that is included in
the binary of Exim only if the corresponding compile-time option is set. The
default settings in &_src/EDITME_& are:
.code
LOOKUP_DBM=yes
LOOKUP_LSEARCH=yes
.endd
which means that only linear searching and DBM lookups are included by default.
For some types of lookup (e.g. SQL databases), you need to install appropriate
libraries and header files before building Exim.




.section "Single-key lookup types" "SECTsinglekeylookups"
.cindex "lookup" "single-key types"
.cindex "single-key lookup" "list of types"
The following single-key lookup types are implemented:

.subsection cdb
.cindex "cdb" "description of"
.cindex "lookup" "cdb"
.cindex "binary zero" "in lookup key"
The given file is searched as a Constant DataBase file, using the key
string without a terminating binary zero. The cdb format is designed for
indexed files that are read frequently and never updated, except by total
re-creation. As such, it is particularly suitable for large files containing
aliases or other indexed data referenced by an MTA. Information about cdb and
tools for building the files can be found in several places:
.display
&url(https://cr.yp.to/cdb.html)
&url(https://www.corpit.ru/mjt/tinycdb.html)
&url(https://packages.debian.org/stable/utils/freecdb)
&url(https://github.com/philpennock/cdbtools) (in Go)
.endd
A cdb distribution is not needed in order to build Exim with cdb support,
because the code for reading cdb files is included directly in Exim itself.
However, no means of building or testing cdb files is provided with Exim, so
you need to obtain a cdb distribution in order to do this.

.subsection dbm
.cindex "DBM" "lookup type"
.cindex "lookup" "dbm"
.cindex "binary zero" "in lookup key"
Calls to DBM library functions are used to extract data from the given
DBM file by looking up the record with the given key. A terminating binary
zero is included in the key that is passed to the DBM library. See section
&<<SECTdb>>& for a discussion of DBM libraries.

.cindex "Berkeley DB library" "file format"
For all versions of Berkeley DB, Exim uses the DB_HASH style of database
when building DBM files using the &%exim_dbmbuild%& utility. However, when
using Berkeley DB versions 3 or 4, it opens existing databases for reading with
the DB_UNKNOWN option. This enables it to handle any of the types of database
that the library supports, and can be useful for accessing DBM files created by
other applications. (For earlier DB versions, DB_HASH is always used.)

.subsection dbmjz
.cindex "lookup" "dbmjz"
.cindex "lookup" "dbm &-- embedded NULs"
.cindex "sasldb2"
.cindex "dbmjz lookup type"
This is the same as &(dbm)&, except that the lookup key is
interpreted as an Exim list; the elements of the list are joined together with
ASCII NUL characters to form the lookup key.  An example usage would be to
authenticate incoming SMTP calls using the passwords from Cyrus SASL's
&_/etc/sasldb2_& file with the &(gsasl)& authenticator or Exim's own
&(cram_md5)& authenticator.

.subsection dbmnz
.cindex "lookup" "dbmnz"
.cindex "lookup" "dbm &-- terminating zero"
.cindex "binary zero" "in lookup key"
.cindex "Courier"
.cindex "&_/etc/userdbshadow.dat_&"
.cindex "dbmnz lookup type"
This is the same as &(dbm)&, except that a terminating binary zero
is not included in the key that is passed to the DBM library. You may need this
if you want to look up data in files that are created by or shared with some
other application that does not use terminating zeros. For example, you need to
use &(dbmnz)& rather than &(dbm)& if you want to authenticate incoming SMTP
calls using the passwords from Courier's &_/etc/userdbshadow.dat_& file. Exim's
utility program for creating DBM files (&'exim_dbmbuild'&) includes the zeros
by default, but has an option to omit them (see section &<<SECTdbmbuild>>&).

.subsection dsearch
.cindex "lookup" "dsearch"
.cindex "dsearch lookup type"
The given file must be an absolute directory path; this is searched for an entry
whose name is the key by calling the &[lstat()]& function.
Unless the options (below) permit a path,
the key may not contain any forward slash characters.
If &[lstat()]& succeeds then so does the lookup.
.cindex "tainted data" "dsearch result"
The result is regarded as untainted.

Options for the lookup can be given by appending them after the word "dsearch",
separated by a comma.  Options, if present, are a comma-separated list having
each element starting with a tag name and an equals.

Three options are supported, for the return value and for filtering match
candidates.
The "ret" option requests an alternate result value of
the entire path for the entry. Example:
.code
${lookup {passwd} dsearch,ret=full {/etc}}
.endd
The default result is just the requested entry.

The "filter" option requests that only directory entries of a given type
are matched. The match value is one of "file", "dir" or "subdir" (the latter
not matching "." or ".."). Example:
.code
${lookup {passwd} dsearch,filter=file {/etc}}
.endd
The default matching is for any entry type, including directories
and symlinks.

The "key" option relaxes the restriction that only a simple path component can
be searched for, to permit a sequence of path components. Example:
.code
${lookup {foo/bar} dsearch,key=path {/etc}}
.endd
If this option is used, a ".." component in the key is specifically disallowed.
The default operation is that the key may only be a single path component.

An example of how this
lookup can be used to support virtual domains is given in section
&<<SECTvirtualdomains>>&.

.subsection iplsearch
.cindex "lookup" "iplsearch"
.cindex "iplsearch lookup type"
The given file is a text file containing keys and data. A key is
terminated by a colon or white space or the end of the line. The keys in the
file must be IP addresses, or IP addresses with CIDR masks. Keys that involve
IPv6 addresses must be enclosed in quotes to prevent the first internal colon
being interpreted as a key terminator. For example:
.code
1.2.3.4:           data for 1.2.3.4
192.168.0.0/16:    data for 192.168.0.0/16
"abcd::cdab":      data for abcd::cdab
"abcd:abcd::/32"   data for abcd:abcd::/32
.endd
The key for an &(iplsearch)& lookup must be an IP address (without a mask). The
file is searched linearly, using the CIDR masks where present, until a matching
key is found. The first key that matches is used; there is no attempt to find a
&"best"& match. Apart from the way the keys are matched, the processing for
&(iplsearch)& is the same as for &(lsearch)&.

&*Warning 1*&: Unlike most other single-key lookup types, a file of data for
&(iplsearch)& can &'not'& be turned into a DBM or cdb file, because those
lookup types support only literal keys.

&*Warning 2*&: In a host list, you must always use &(net-iplsearch)& so that
the implicit key is the host's IP address rather than its name
(see section &<<SECThoslispatsikey>>&).

&*Warning 3*&: Do not use an IPv4-mapped IPv6 address for a key; use the
IPv4, in dotted-quad form. (Exim converts IPv4-mapped IPv6 addresses to this
notation before executing the lookup.)

One option is supported, "ret=full", to request the return of the entire line
rather than omitting the key portion.
Note however that the key portion will have been de-quoted.

.subsection json
.cindex lookup json
.cindex json "lookup type"
.cindex JSON expansions
The given file is a text file with a JSON structure.
An element of the structure is extracted, defined by the search key.
The key is a list of subelement selectors
(colon-separated by default but changeable in the usual way)
which are applied in turn to select smaller and smaller portions
of the JSON structure.
If a selector is numeric, it must apply to a JSON array; the (zero-based)
nunbered array element is selected.
Otherwise it must apply to a JSON object; the named element is selected.
The final resulting element can be a simple JSON type or a JSON object
or array; for the latter two a string-representation of the JSON
is returned.
For elements of type string, the returned value is de-quoted.


.subsection lmdb
.cindex LMDB
.cindex lookup lmdb
.cindex database lmdb
The given file is an LMDB database.
LMDB is a memory-mapped key-value store,
with API modeled loosely on that of BerkeleyDB.
See &url(https://symas.com/products/lightning-memory-mapped-database/,LMDB)
for the feature set and operation modes.

Exim provides read-only access via the LMDB C library.
The library can be obtained from &url(https://github.com/LMDB/lmdb)
or your operating system package repository.
To enable LMDB support in Exim set LOOKUP_LMDB=yes in &_Local/Makefile_&.

You will need to separately create the LMDB database file,
possibly using the &"mdb_load"& utility.


.subsection lsearch
.cindex "linear search"
.cindex "lookup" "lsearch"
.cindex "lsearch lookup type"
.cindex "case sensitivity" "in lsearch lookup"
The given file is a text file that is searched linearly for a
line beginning with the search key, terminated by a colon or white space or the
end of the line. The search is case-insensitive; that is, upper and lower case
letters are treated as the same. The first occurrence of the key that is found
in the file is used.

White space between the key and the colon is permitted. The remainder of the
line, with leading and trailing white space removed, is the data. This can be
continued onto subsequent lines by starting them with any amount of white
space, but only a single space character is included in the data at such a
junction. If the data begins with a colon, the key must be terminated by a
colon, for example:
.code
baduser:  :fail:
.endd
Empty lines and lines beginning with # are ignored, even if they occur in the
middle of an item. This is the traditional textual format of alias files. Note
that the keys in an &(lsearch)& file are literal strings. There is no
wildcarding of any kind.

.cindex "lookup" "lsearch &-- colons in keys"
.cindex "white space" "in lsearch key"
In most &(lsearch)& files, keys are not required to contain colons or #
characters, or white space. However, if you need this feature, it is available.
If a key begins with a doublequote character, it is terminated only by a
matching quote (or end of line), and the normal escaping rules apply to its
contents (see section &<<SECTstrings>>&). An optional colon is permitted after
quoted keys (exactly as for unquoted keys). There is no special handling of
quotes for the data part of an &(lsearch)& line.

.subsection nis
.cindex "NIS lookup type"
.cindex "lookup" "NIS"
.cindex "binary zero" "in lookup key"
The given file is the name of a NIS map, and a NIS lookup is done with
the given key, without a terminating binary zero. There is a variant called
&(nis0)& which does include the terminating binary zero in the key. This is
reportedly needed for Sun-style alias files. Exim does not recognize NIS
aliases; the full map names must be used.

.subsection (n)wildlsearch
.cindex "wildlsearch lookup type"
.cindex "lookup" "wildlsearch"
.cindex "nwildlsearch lookup type"
.cindex "lookup" "nwildlsearch"
&(wildlsearch)& or &(nwildlsearch)&: These search a file linearly, like
&(lsearch)&, but instead of being interpreted as a literal string, each key in
the file may be wildcarded. The difference between these two lookup types is
that for &(wildlsearch)&, each key in the file is string-expanded before being
used, whereas for &(nwildlsearch)&, no expansion takes place.

.cindex "case sensitivity" "in (n)wildlsearch lookup"
Like &(lsearch)&, the testing is done case-insensitively. However, keys in the
file that are regular expressions can be made case-sensitive by the use of
&`(-i)`& within the pattern. The following forms of wildcard are recognized:

.olist
The string may begin with an asterisk to mean &"ends with"&. For example:
.code
*.a.b.c       data for anything.a.b.c
*fish         data for anythingfish
.endd
.next
The string may begin with a circumflex to indicate a regular expression. For
example, for &(wildlsearch)&:
.code
^\N\d+\.a\.b\N    data for <digits>.a.b
.endd
Note the use of &`\N`& to disable expansion of the contents of the regular
expression. If you are using &(nwildlsearch)&, where the keys are not
string-expanded, the equivalent entry is:
.code
^\d+\.a\.b        data for <digits>.a.b
.endd
The case-insensitive flag is set at the start of compiling the regular
expression, but it can be turned off by using &`(-i)`& at an appropriate point.
For example, to make the entire pattern case-sensitive:
.code
^(?-i)\d+\.a\.b        data for <digits>.a.b
.endd

If the regular expression contains white space or colon characters, you must
either quote it (see &(lsearch)& above), or represent these characters in other
ways. For example, &`\s`& can be used for white space and &`\x3A`& for a
colon. This may be easier than quoting, because if you quote, you have to
escape all the backslashes inside the quotes.

&*Note*&: It is not possible to capture substrings in a regular expression
match for later use, because the results of all lookups are cached. If a lookup
is repeated, the result is taken from the cache, and no actual pattern matching
takes place. The values of all the numeric variables are unset after a
&((n)wildlsearch)& match.

.next
Although I cannot see it being of much use, the general matching function that
is used to implement &((n)wildlsearch)& means that the string may begin with a
lookup name terminated by a semicolon, and followed by lookup data. For
example:
.code
cdb;/some/file  data for keys that match the file
.endd
The data that is obtained from the nested lookup is discarded.
.endlist olist

Keys that do not match any of these patterns are interpreted literally. The
continuation rules for the data are the same as for &(lsearch)&, and keys may
be followed by optional colons.

&*Warning*&: Unlike most other single-key lookup types, a file of data for
&((n)wildlsearch)& can &'not'& be turned into a DBM or cdb file, because those
lookup types support only literal keys.

.subsection spf
.cindex "spf lookup type"
.cindex "lookup" "spf"
If Exim is built with SPF support, manual lookups can be done
(as opposed to the standard ACL condition method).
For details see section &<<SECSPF>>&.


.section "Query-style lookup types" "SECTquerystylelookups"
.cindex "lookup" "query-style types"
.cindex "query-style lookup" "list of types"
The supported query-style lookup types are listed below. Further details about
many of them are given in later sections.

.subsection dnsdb
.cindex "DNS" "as a lookup type"
.cindex "lookup" "DNS"
This does a DNS search for one or more records whose domain names
are given in the supplied query. The resulting data is the contents of the
records. See section &<<SECTdnsdb>>&.

.subsection ibase
.cindex "InterBase lookup type"
.cindex "lookup" "InterBase"
This does a lookup in an InterBase database.

.subsection ldap
.cindex "LDAP" "lookup type"
.cindex "lookup" "LDAP"
This does an LDAP lookup using a query in the form of a URL, and
returns attributes from a single entry. There is a variant called &(ldapm)&
that permits values from multiple entries to be returned. A third variant
called &(ldapdn)& returns the Distinguished Name of a single entry instead of
any attribute values. See section &<<SECTldap>>&.

.subsection mysql
.cindex "MySQL" "lookup type"
.cindex "lookup" "MySQL"
The format of the query is an SQL statement that is passed to a
MySQL database. See section &<<SECTsql>>&.

.subsection nisplus
.cindex "NIS+ lookup type"
.cindex "lookup" "NIS+"
This does a NIS+ lookup using a query that can specify the name of
the field to be returned. See section &<<SECTnisplus>>&.

.subsection oracle
.cindex "Oracle" "lookup type"
.cindex "lookup" "Oracle"
The format of the query is an SQL statement that is passed to an
Oracle database. See section &<<SECTsql>>&.

.subsection passwd
.cindex "lookup" "passwd"
.cindex "passwd lookup type"
.cindex "&_/etc/passwd_&"
This is a query-style lookup with queries that are just user names. The
lookup calls &[getpwnam()]& to interrogate the system password data, and on
success, the result string is the same as you would get from an &(lsearch)&
lookup on a traditional &_/etc/passwd file_&, though with &`*`& for the
password value. For example:
.code
*:42:42:King Rat:/home/kr:/bin/bash
.endd

.subsection pgsql
.cindex "PostgreSQL lookup type"
.cindex "lookup" "PostgreSQL"
The format of the query is an SQL statement that is passed to a
PostgreSQL database. See section &<<SECTsql>>&.

.subsection redis
.cindex "Redis lookup type"
.cindex lookup Redis
The format of the query is either a simple get or simple set,
passed to a Redis database. See section &<<SECTsql>>&.

.subsection sqlite
.cindex "sqlite lookup type"
.cindex "lookup" "sqlite"
The format of the query is
an SQL statement that is passed to an SQLite database. See section &<<SECTsqlite>>&.

.subsection testdb
This is a lookup type that is used for testing Exim. It is
not likely to be useful in normal operation.

.subsection whoson
.cindex "whoson lookup type"
.cindex "lookup" "whoson"
&'Whoson'& (&url(http://whoson.sourceforge.net)) is a protocol that
allows a server to check whether a particular (dynamically allocated) IP
address is currently allocated to a known (trusted) user and, optionally, to
obtain the identity of the said user. For SMTP servers, &'Whoson'& was popular
at one time for &"POP before SMTP"& authentication, but that approach has been
superseded by SMTP authentication. In Exim, &'Whoson'& can be used to implement
&"POP before SMTP"& checking using ACL statements such as
.code
require condition = \
  ${lookup whoson {$sender_host_address}{yes}{no}}
.endd
The query consists of a single IP address. The value returned is the name of
the authenticated user, which is stored in the variable &$value$&. However, in
this example, the data in &$value$& is not used; the result of the lookup is
one of the fixed strings &"yes"& or &"no"&.



.section "Temporary errors in lookups" "SECID63"
.cindex "lookup" "temporary error in"
Lookup functions can return temporary error codes if the lookup cannot be
completed. For example, an SQL or LDAP database might be unavailable. For this
reason, it is not advisable to use a lookup that might do this for critical
options such as a list of local domains.

When a lookup cannot be completed in a router or transport, delivery
of the message (to the relevant address) is deferred, as for any other
temporary error. In other circumstances Exim may assume the lookup has failed,
or may give up altogether.



.section "Default values in single-key lookups" "SECTdefaultvaluelookups"
.cindex "wildcard lookups"
.cindex "lookup" "default values"
.cindex "lookup" "wildcard"
.cindex "lookup" "* added to type"
.cindex "default" "in single-key lookups"
In this context, a &"default value"& is a value specified by the administrator
that is to be used if a lookup fails.

&*Note:*& This section applies only to single-key lookups. For query-style
lookups, the facilities of the query language must be used. An attempt to
specify a default for a query-style lookup provokes an error.

If &"*"& is added to a single-key lookup type (for example, &%lsearch*%&)
and the initial lookup fails, the key &"*"& is looked up in the file to
provide a default value. See also the section on partial matching below.

.cindex "*@ with single-key lookup"
.cindex "lookup" "*@ added to type"
.cindex "alias file" "per-domain default"
Alternatively, if &"*@"& is added to a single-key lookup type (for example
&%dbm*@%&) then, if the initial lookup fails and the key contains an @
character, a second lookup is done with everything before the last @ replaced
by *. This makes it possible to provide per-domain defaults in alias files
that include the domains in the keys. If the second lookup fails (or doesn't
take place because there is no @ in the key), &"*"& is looked up.
For example, a &(redirect)& router might contain:
.code
data = ${lookup{$local_part@$domain}lsearch*@{/etc/mix-aliases}}
.endd
Suppose the address that is being processed is &'jane@eyre.example'&. Exim
looks up these keys, in this order:
.code
jane@eyre.example
*@eyre.example
*
.endd
The data is taken from whichever key it finds first. &*Note*&: In an
&(lsearch)& file, this does not mean the first of these keys in the file. A
complete scan is done for each key, and only if it is not found at all does
Exim move on to try the next key.



.section "Partial matching in single-key lookups" "SECTpartiallookup"
.cindex "partial matching"
.cindex "wildcard lookups"
.cindex "lookup" "partial matching"
.cindex "lookup" "wildcard"
.cindex "asterisk" "in search type"
The normal operation of a single-key lookup is to search the file for an exact
match with the given key. However, in a number of situations where domains are
being looked up, it is useful to be able to do partial matching. In this case,
information in the file that has a key starting with &"*."& is matched by any
domain that ends with the components that follow the full stop. For example, if
a key in a DBM file is
.code
*.dates.fict.example
.endd
then when partial matching is enabled this is matched by (amongst others)
&'2001.dates.fict.example'& and &'1984.dates.fict.example'&. It is also matched
by &'dates.fict.example'&, if that does not appear as a separate key in the
file.

&*Note*&: Partial matching is not available for query-style lookups. It is
also not available for any lookup items in address lists (see section
&<<SECTaddresslist>>&).

Partial matching is implemented by doing a series of separate lookups using
keys constructed by modifying the original subject key. This means that it can
be used with any of the single-key lookup types, provided that
partial matching keys
beginning with a special prefix (default &"*."&) are included in the data file.
Keys in the file that do not begin with the prefix are matched only by
unmodified subject keys when partial matching is in use.

Partial matching is requested by adding the string &"partial-"& to the front of
the name of a single-key lookup type, for example, &%partial-dbm%&. When this
is done, the subject key is first looked up unmodified; if that fails, &"*."&
is added at the start of the subject key, and it is looked up again. If that
fails, further lookups are tried with dot-separated components removed from the
start of the subject key, one-by-one, and &"*."& added on the front of what
remains.

A minimum number of two non-* components are required. This can be adjusted
by including a number before the hyphen in the search type. For example,
&%partial3-lsearch%& specifies a minimum of three non-* components in the
modified keys. Omitting the number is equivalent to &"partial2-"&. If the
subject key is &'2250.dates.fict.example'& then the following keys are looked
up when the minimum number of non-* components is two:
.code
2250.dates.fict.example
*.2250.dates.fict.example
*.dates.fict.example
*.fict.example
.endd
As soon as one key in the sequence is successfully looked up, the lookup
finishes.

.cindex "lookup" "partial matching &-- changing prefix"
.cindex "prefix" "for partial matching"
The use of &"*."& as the partial matching prefix is a default that can be
changed. The motivation for this feature is to allow Exim to operate with file
formats that are used by other MTAs. A different prefix can be supplied in
parentheses instead of the hyphen after &"partial"&. For example:
.code
domains = partial(.)lsearch;/some/file
.endd
In this example, if the domain is &'a.b.c'&, the sequence of lookups is
&`a.b.c`&, &`.a.b.c`&, and &`.b.c`& (the default minimum of 2 non-wild
components is unchanged). The prefix may consist of any punctuation characters
other than a closing parenthesis. It may be empty, for example:
.code
domains = partial1()cdb;/some/file
.endd
For this example, if the domain is &'a.b.c'&, the sequence of lookups is
&`a.b.c`&, &`b.c`&, and &`c`&.

If &"partial0"& is specified, what happens at the end (when the lookup with
just one non-wild component has failed, and the original key is shortened right
down to the null string) depends on the prefix:

.ilist
If the prefix has zero length, the whole lookup fails.
.next
If the prefix has length 1, a lookup for just the prefix is done. For
example, the final lookup for &"partial0(.)"& is for &`.`& alone.
.next
Otherwise, if the prefix ends in a dot, the dot is removed, and the
remainder is looked up. With the default prefix, therefore, the final lookup is
for &"*"& on its own.
.next
Otherwise, the whole prefix is looked up.
.endlist


If the search type ends in &"*"& or &"*@"& (see section
&<<SECTdefaultvaluelookups>>& above), the search for an ultimate default that
this implies happens after all partial lookups have failed. If &"partial0"& is
specified, adding &"*"& to the search type has no effect with the default
prefix, because the &"*"& key is already included in the sequence of partial
lookups. However, there might be a use for lookup types such as
&"partial0(.)lsearch*"&.

The use of &"*"& in lookup partial matching differs from its use as a wildcard
in domain lists and the like. Partial matching works only in terms of
dot-separated components; a key such as &`*fict.example`&
in a database file is useless, because the asterisk in a partial matching
subject key is always followed by a dot.

When the lookup is done from a string-expansion,
the variables &$1$& and &$2$& contain the wild and non-wild parts of the key
during the expansion of the replacement text.
They return to their previous values at the end of the lookup item.




.section "Lookup caching" "SECID64"
.cindex "lookup" "caching"
.cindex "caching" "lookup data"
Exim caches all lookup results in order to avoid needless repetition of
lookups. However, because (apart from the daemon) Exim operates as a collection
of independent, short-lived processes, this caching applies only within a
single Exim process. There is no inter-process lookup caching facility.

If an option &"cache=no_rd"& is used on the lookup then
the cache is only written to, cached data is not used for the operation
and a real lookup is done.

For single-key lookups, Exim keeps the relevant files open in case there is
another lookup that needs them. In some types of configuration this can lead to
many files being kept open for messages with many recipients. To avoid hitting
the operating system limit on the number of simultaneously open files, Exim
closes the least recently used file when it needs to open more files than its
own internal limit, which can be changed via the &%lookup_open_max%& option.

The single-key lookup files are closed and the lookup caches are flushed at
strategic points during delivery &-- for example, after all routing is
complete.




.section "Quoting lookup data" "SECID65"
.cindex "lookup" "quoting"
.cindex "quoting" "in lookups"
When data from an incoming message is included in a query-style lookup, there
is the possibility of special characters in the data messing up the syntax of
the query. For example, a NIS+ query that contains
.code
[name=$local_part]
.endd
will be broken if the local part happens to contain a closing square bracket.
For NIS+, data can be enclosed in double quotes like this:
.code
[name="$local_part"]
.endd
but this still leaves the problem of a double quote in the data. The rule for
NIS+ is that double quotes must be doubled. Other lookup types have different
rules, and to cope with the differing requirements, an expansion operator
of the following form is provided:
.code
${quote_<lookup-type>:<string>}
.endd
For example, the way to write the NIS+ query is
.code
[name="${quote_nisplus:$local_part}"]
.endd
.cindex "tainted data" "in lookups"
&*All*& tainted data used in a query-style lookup must be quoted
using a mechanism appropriate for the lookup type.
See chapter &<<CHAPexpand>>& for full coverage of string expansions. The quote
operator can be used for all lookup types, but has no effect for single-key
lookups, since no quoting is ever needed in their key strings.




.section "More about dnsdb" "SECTdnsdb"
.cindex "dnsdb lookup"
.cindex "lookup" "dnsdb"
.cindex "DNS" "as a lookup type"
The &(dnsdb)& lookup type uses the DNS as its database. A simple query consists
of a record type and a domain name, separated by an equals sign. For example,
an expansion string could contain:
.code
${lookup dnsdb{mx=a.b.example}{$value}fail}
.endd
If the lookup succeeds, the result is placed in &$value$&, which in this case
is used on its own as the result. If the lookup does not succeed, the
&`fail`& keyword causes a &'forced expansion failure'& &-- see section
&<<SECTforexpfai>>& for an explanation of what this means.

The supported DNS record types are A, CNAME, MX, NS, PTR, SOA, SPF, SRV, TLSA
and TXT, and, when Exim is compiled with IPv6 support, AAAA.
If no type is given, TXT is assumed.

For any record type, if multiple records are found, the data is returned as a
concatenation, with newline as the default separator. The order, of course,
depends on the DNS resolver. You can specify a different separator character
between multiple records by putting a right angle-bracket followed immediately
by the new separator at the start of the query. For example:
.code
${lookup dnsdb{>: a=host1.example}}
.endd
It is permitted to specify a space as the separator character. Further
white space is ignored.
For lookup types that return multiple fields per record,
an alternate field separator can be specified using a comma after the main
separator character, followed immediately by the field separator.

.cindex "PTR record" "in &(dnsdb)& lookup"
When the type is PTR,
the data can be an IP address, written as normal; inversion and the addition of
&%in-addr.arpa%& or &%ip6.arpa%& happens automatically. For example:
.code
${lookup dnsdb{ptr=192.168.4.5}{$value}fail}
.endd
If the data for a PTR record is not a syntactically valid IP address, it is not
altered and nothing is added.

.cindex "MX record" "in &(dnsdb)& lookup"
.cindex "SRV record" "in &(dnsdb)& lookup"
For an MX lookup, both the preference value and the host name are returned for
each record, separated by a space. For an SRV lookup, the priority, weight,
port, and host name are returned for each record, separated by spaces.
The field separator can be modified as above.

.cindex "TXT record" "in &(dnsdb)& lookup"
.cindex "SPF record" "in &(dnsdb)& lookup"
For TXT records with multiple items of data, only the first item is returned,
unless a field separator is specified.
To concatenate items without a separator, use a semicolon instead.
For SPF records the
default behaviour is to concatenate multiple items without using a separator.
.code
${lookup dnsdb{>\n,: txt=a.b.example}}
${lookup dnsdb{>\n; txt=a.b.example}}
${lookup dnsdb{spf=example.org}}
.endd
It is permitted to specify a space as the separator character. Further
white space is ignored.

.cindex "SOA record" "in &(dnsdb)& lookup"
For an SOA lookup, while no result is obtained the lookup is redone with
successively more leading components dropped from the given domain.
Only the primary-nameserver field is returned unless a field separator is
specified.
.code
${lookup dnsdb{>:,; soa=a.b.example.com}}
.endd

.subsection "Dnsdb lookup modifiers" SECTdnsdb_mod
.cindex "dnsdb modifiers"
.cindex "modifiers" "dnsdb"
.cindex "options" "dnsdb"
Modifiers for &(dnsdb)& lookups are given by optional keywords,
each followed by a comma,
that may appear before the record type.

The &(dnsdb)& lookup fails only if all the DNS lookups fail. If there is a
temporary DNS error for any of them, the behaviour is controlled by
a defer-option modifier.
The possible keywords are
&"defer_strict"&, &"defer_never"&, and &"defer_lax"&.
With &"strict"& behaviour, any temporary DNS error causes the
whole lookup to defer. With &"never"& behaviour, a temporary DNS error is
ignored, and the behaviour is as if the DNS lookup failed to find anything.
With &"lax"& behaviour, all the queries are attempted, but a temporary DNS
error causes the whole lookup to defer only if none of the other lookups
succeed. The default is &"lax"&, so the following lookups are equivalent:
.code
${lookup dnsdb{defer_lax,a=one.host.com:two.host.com}}
${lookup dnsdb{a=one.host.com:two.host.com}}
.endd
Thus, in the default case, as long as at least one of the DNS lookups
yields some data, the lookup succeeds.

.cindex "DNSSEC" "dns lookup"
Use of &(DNSSEC)& is controlled by a dnssec modifier.
The possible keywords are
&"dnssec_strict"&, &"dnssec_lax"&, and &"dnssec_never"&.
With &"strict"& or &"lax"& DNSSEC information is requested
with the lookup.
With &"strict"& a response from the DNS resolver that
is not labelled as authenticated data
is treated as equivalent to a temporary DNS error.
The default is &"lax"&.

See also the &$lookup_dnssec_authenticated$& variable.

.cindex timeout "dns lookup"
.cindex "DNS" timeout
Timeout for the dnsdb lookup can be controlled by a retrans modifier.
The form is &"retrans_VAL"& where VAL is an Exim time specification
(e.g. &"5s"&).
The default value is set by the main configuration option &%dns_retrans%&.

Retries for the dnsdb lookup can be controlled by a retry modifier.
The form if &"retry_VAL"& where VAL is an integer.
The default count is set by the main configuration option &%dns_retry%&.

.cindex caching "of dns lookup"
.cindex TTL "of dns lookup"
.cindex DNS TTL
Dnsdb lookup results are cached within a single process (and its children).
The cache entry lifetime is limited to the smallest time-to-live (TTL)
value of the set of returned DNS records.


.subsection "Pseudo dnsdb record types" SECID66
.cindex "MX record" "in &(dnsdb)& lookup"
By default, both the preference value and the host name are returned for
each MX record, separated by a space. If you want only host names, you can use
the pseudo-type MXH:
.code
${lookup dnsdb{mxh=a.b.example}}
.endd
In this case, the preference values are omitted, and just the host names are
returned.

.cindex "name server for enclosing domain"
Another pseudo-type is ZNS (for &"zone NS"&). It performs a lookup for NS
records on the given domain, but if none are found, it removes the first
component of the domain name, and tries again. This process continues until NS
records are found or there are no more components left (or there is a DNS
error). In other words, it may return the name servers for a top-level domain,
but it never returns the root name servers. If there are no NS records for the
top-level domain, the lookup fails. Consider these examples:
.code
${lookup dnsdb{zns=xxx.quercite.com}}
${lookup dnsdb{zns=xxx.edu}}
.endd
Assuming that in each case there are no NS records for the full domain name,
the first returns the name servers for &%quercite.com%&, and the second returns
the name servers for &%edu%&.

You should be careful about how you use this lookup because, unless the
top-level domain does not exist, the lookup always returns some host names. The
sort of use to which this might be put is for seeing if the name servers for a
given domain are on a blacklist. You can probably assume that the name servers
for the high-level domains such as &%com%& or &%co.uk%& are not going to be on
such a list.

.cindex "CSA" "in &(dnsdb)& lookup"
A third pseudo-type is CSA (Client SMTP Authorization). This looks up SRV
records according to the CSA rules, which are described in section
&<<SECTverifyCSA>>&. Although &(dnsdb)& supports SRV lookups directly, this is
not sufficient because of the extra parent domain search behaviour of CSA. The
result of a successful lookup such as:
.code
${lookup dnsdb {csa=$sender_helo_name}}
.endd
has two space-separated fields: an authorization code and a target host name.
The authorization code can be &"Y"& for yes, &"N"& for no, &"X"& for explicit
authorization required but absent, or &"?"& for unknown.

.cindex "A+" "in &(dnsdb)& lookup"
The pseudo-type A+ performs an AAAA
and then an A lookup.  All results are returned; defer processing
(see below) is handled separately for each lookup.  Example:
.code
${lookup dnsdb {>; a+=$sender_helo_name}}
.endd


.subsection "Multiple dnsdb lookups" SECID67
In the previous sections, &(dnsdb)& lookups for a single domain are described.
However, you can specify a list of domains or IP addresses in a single
&(dnsdb)& lookup. The list is specified in the normal Exim way, with colon as
the default separator, but with the ability to change this. For example:
.code
${lookup dnsdb{one.domain.com:two.domain.com}}
${lookup dnsdb{a=one.host.com:two.host.com}}
${lookup dnsdb{ptr = <; 1.2.3.4 ; 4.5.6.8}}
.endd
In order to retain backwards compatibility, there is one special case: if
the lookup type is PTR and no change of separator is specified, Exim looks
to see if the rest of the string is precisely one IPv6 address. In this
case, it does not treat it as a list.

The data from each lookup is concatenated, with newline separators by default,
in the same way that multiple DNS records for a single item are handled. A
different separator can be specified, as described above.




.section "More about LDAP" "SECTldap"
.cindex "LDAP" "lookup, more about"
.cindex "lookup" "LDAP"
.cindex "Solaris" "LDAP"
The original LDAP implementation came from the University of Michigan; this has
become &"Open LDAP"&, and there are now two different releases. Another
implementation comes from Netscape, and Solaris 7 and subsequent releases
contain inbuilt LDAP support. Unfortunately, though these are all compatible at
the lookup function level, their error handling is different. For this reason
it is necessary to set a compile-time variable when building Exim with LDAP, to
indicate which LDAP library is in use. One of the following should appear in
your &_Local/Makefile_&:
.code
LDAP_LIB_TYPE=UMICHIGAN
LDAP_LIB_TYPE=OPENLDAP1
LDAP_LIB_TYPE=OPENLDAP2
LDAP_LIB_TYPE=NETSCAPE
LDAP_LIB_TYPE=SOLARIS
.endd
If LDAP_LIB_TYPE is not set, Exim assumes &`OPENLDAP1`&, which has the
same interface as the University of Michigan version.

There are three LDAP lookup types in Exim. These behave slightly differently in
the way they handle the results of a query:

.ilist
&(ldap)& requires the result to contain just one entry; if there are more, it
gives an error.
.next
&(ldapdn)& also requires the result to contain just one entry, but it is the
Distinguished Name that is returned rather than any attribute values.
.next
&(ldapm)& permits the result to contain more than one entry; the attributes
from all of them are returned.
.endlist


For &(ldap)& and &(ldapm)&, if a query finds only entries with no attributes,
Exim behaves as if the entry did not exist, and the lookup fails. The format of
the data returned by a successful lookup is described in the next section.
First we explain how LDAP queries are coded.


.subsection "Format of LDAP queries" SECTforldaque
.cindex "LDAP" "query format"
An LDAP query takes the form of a URL as defined in
&url(https://www.rfc-editor.org/rfc/rfc2255,RFC 2255). For example, in
the configuration of a &(redirect)& router one might have this setting:
.code
data = ${lookup ldap \
  {ldap:///cn=$local_part,o=University%20of%20Cambridge,\
  c=UK?mailbox?base?}}
.endd
.cindex "LDAP" "with TLS"
The URL may begin with &`ldap`& or &`ldaps`& if your LDAP library supports
secure (encrypted) LDAP connections. The second of these ensures that an
encrypted TLS connection is used.

With sufficiently modern LDAP libraries, Exim supports forcing TLS over regular
LDAP connections, rather than the SSL-on-connect &`ldaps`&.
See the &%ldap_start_tls%& option.

Starting with Exim 4.83, the initialization of LDAP with TLS is more tightly
controlled. Every part of the TLS configuration can be configured by settings in
&_exim.conf_&. Depending on the version of the client libraries installed on
your system, some of the initialization may have required setting options in
&_/etc/ldap.conf_& or &_~/.ldaprc_& to get TLS working with self-signed
certificates. This revealed a nuance where the current UID that exim was
running as could affect which config files it read. With Exim 4.83, these
methods become optional, only taking effect if not specifically set in
&_exim.conf_&.


.subsection "LDAP quoting" SECID68
.cindex "LDAP" "quoting"
Two levels of quoting are required in LDAP queries, the first for LDAP itself
and the second because the LDAP query is represented as a URL. Furthermore,
within an LDAP query, two different kinds of quoting are required. For this
reason, there are two different LDAP-specific quoting operators.

The &%quote_ldap%& operator is designed for use on strings that are part of
filter specifications. Conceptually, it first does the following conversions on
the string:
.code
*   =>   \2A
(   =>   \28
)   =>   \29
\   =>   \5C
.endd
in accordance with
&url(https://www.rfc-editor.org/rfc/rfc2254,RFC 2254).
The resulting string is then quoted according
to the rules for URLs, that is, all non-alphanumeric characters except
.code
! $ ' - . _ ( ) * +
.endd
are converted to their hex values, preceded by a percent sign. For example:
.code
${quote_ldap: a(bc)*, a<yz>; }
.endd
yields
.code
%20a%5C28bc%5C29%5C2A%2C%20a%3Cyz%3E%3B%20
.endd
Removing the URL quoting, this is (with a leading and a trailing space):
.code
a\28bc\29\2A, a<yz>;
.endd
The &%quote_ldap_dn%& operator is designed for use on strings that are part of
base DN specifications in queries. Conceptually, it first converts the string
by inserting a backslash in front of any of the following characters:
.code
, + " \ < > ;
.endd
It also inserts a backslash before any leading spaces or # characters, and
before any trailing spaces. (These rules are in
&url(https://www.rfc-editor.org/rfc/rfc2255,RFC 2253).)
The resulting string is then quoted according to the rules for URLs.
For example:
.code
${quote_ldap_dn: a(bc)*, a<yz>; }
.endd
yields
.code
%5C%20a(bc)*%5C%2C%20a%5C%3Cyz%5C%3E%5C%3B%5C%20
.endd
Removing the URL quoting, this is (with a trailing space):
.code
\ a(bc)*\, a\<yz\>\;\
.endd
There are some further comments about quoting in the section on LDAP
authentication below.


.subsection "LDAP connections" SECID69
.cindex "LDAP" "connections"
The connection to an LDAP server may either be over TCP/IP, or, when OpenLDAP
is in use, via a Unix domain socket. The example given above does not specify
an LDAP server. A server that is reached by TCP/IP can be specified in a query
by starting it with
.code
ldap://<hostname>:<port>/...
.endd
If the port (and preceding colon) are omitted, the standard LDAP port (389) is
used. When no server is specified in a query, a list of default servers is
taken from the &%ldap_default_servers%& configuration option. This supplies a
colon-separated list of servers which are tried in turn until one successfully
handles a query, or there is a serious error. Successful handling either
returns the requested data, or indicates that it does not exist. Serious errors
are syntactical, or multiple values when only a single value is expected.
Errors which cause the next server to be tried are connection failures, bind
failures, and timeouts.

For each server name in the list, a port number can be given. The standard way
of specifying a host and port is to use a colon separator
(&url(https://www.rfc-editor.org/rfc/rfc1738,RFC 1738)). Because
&%ldap_default_servers%& is a colon-separated list, such colons have to be
doubled. For example
.code
ldap_default_servers = ldap1.example.com::145:ldap2.example.com
.endd
If &%ldap_default_servers%& is unset, a URL with no server name is passed
to the LDAP library with no server name, and the library's default (normally
the local host) is used.

If you are using the OpenLDAP library, you can connect to an LDAP server using
a Unix domain socket instead of a TCP/IP connection. This is specified by using
&`ldapi`& instead of &`ldap`& in LDAP queries. What follows here applies only
to OpenLDAP. If Exim is compiled with a different LDAP library, this feature is
not available.

For this type of connection, instead of a host name for the server, a pathname
for the socket is required, and the port number is not relevant. The pathname
can be specified either as an item in &%ldap_default_servers%&, or inline in
the query. In the former case, you can have settings such as
.code
ldap_default_servers = /tmp/ldap.sock : backup.ldap.your.domain
.endd
When the pathname is given in the query, you have to escape the slashes as
&`%2F`& to fit in with the LDAP URL syntax. For example:
.code
${lookup ldap {ldapi://%2Ftmp%2Fldap.sock/o=...
.endd
When Exim processes an LDAP lookup and finds that the &"hostname"& is really
a pathname, it uses the Unix domain socket code, even if the query actually
specifies &`ldap`& or &`ldaps`&. In particular, no encryption is used for a
socket connection. This behaviour means that you can use a setting of
&%ldap_default_servers%& such as in the example above with traditional &`ldap`&
or &`ldaps`& queries, and it will work. First, Exim tries a connection via
the Unix domain socket; if that fails, it tries a TCP/IP connection to the
backup host.

If an explicit &`ldapi`& type is given in a query when a host name is
specified, an error is diagnosed. However, if there are more items in
&%ldap_default_servers%&, they are tried. In other words:

.ilist
Using a pathname with &`ldap`& or &`ldaps`& forces the use of the Unix domain
interface.
.next
Using &`ldapi`& with a host name causes an error.
.endlist


Using &`ldapi`& with no host or path in the query, and no setting of
&%ldap_default_servers%&, does whatever the library does by default.



.subsection "LDAP authentication and control information" SECID70
.cindex "LDAP" "authentication"
The LDAP URL syntax provides no way of passing authentication and other control
information to the server. To make this possible, the URL in an LDAP query may
be preceded by any number of <&'name'&>=<&'value'&> settings, separated by
spaces. If a value contains spaces it must be enclosed in double quotes, and
when double quotes are used, backslash is interpreted in the usual way inside
them. The following names are recognized:
.itable none 0 0 2 20* left 80* left
.irow DEREFERENCE "set the dereferencing parameter"
.irow NETTIME     "set a timeout for a network operation"
.irow USER        "set the DN, for authenticating the LDAP bind"
.irow PASS        "set the password, likewise"
.irow REFERRALS   "set the referrals parameter"
.irow SERVERS     "set alternate server list for this query only"
.irow SIZE        "set the limit for the number of entries returned"
.irow TIME        "set the maximum waiting time for a query"
.endtable
The value of the DEREFERENCE parameter must be one of the words &"never"&,
&"searching"&, &"finding"&, or &"always"&. The value of the REFERRALS parameter
must be &"follow"& (the default) or &"nofollow"&. The latter stops the LDAP
library from trying to follow referrals issued by the LDAP server.

.cindex LDAP timeout
.cindex timeout "LDAP lookup"
The name CONNECT is an obsolete name for NETTIME, retained for
backwards compatibility. This timeout (specified as a number of seconds) is
enforced from the client end for operations that can be carried out over a
network. Specifically, it applies to network connections and calls to the
&'ldap_result()'& function. If the value is greater than zero, it is used if
LDAP_OPT_NETWORK_TIMEOUT is defined in the LDAP headers (OpenLDAP), or
if LDAP_X_OPT_CONNECT_TIMEOUT is defined in the LDAP headers (Netscape
SDK 4.1). A value of zero forces an explicit setting of &"no timeout"& for
Netscape SDK; for OpenLDAP no action is taken.

The TIME parameter (also a number of seconds) is passed to the server to
set a server-side limit on the time taken to complete a search.

The SERVERS parameter allows you to specify an alternate list of ldap servers
to use for an individual lookup.  The global &%ldap_default_servers%& option provides a
default list of ldap servers, and a single lookup can specify a single ldap
server to use.  But when you need to do a lookup with a list of servers that is
different than the default list (maybe different order, maybe a completely
different set of servers), the SERVERS parameter allows you to specify this
alternate list (colon-separated).

Here is an example of an LDAP query in an Exim lookup that uses some of these
values. This is a single line, folded to fit on the page:
.code
${lookup ldap
  {user="cn=manager,o=University of Cambridge,c=UK" pass=secret
  ldap:///o=University%20of%20Cambridge,c=UK?sn?sub?(cn=foo)}
  {$value}fail}
.endd
The encoding of spaces as &`%20`& is a URL thing which should not be done for
any of the auxiliary data. Exim configuration settings that include lookups
which contain password information should be preceded by &"hide"& to prevent
non-admin users from using the &%-bP%& option to see their values.

The auxiliary data items may be given in any order. The default is no
connection timeout (the system timeout is used), no user or password, no limit
on the number of entries returned, and no time limit on queries.

When a DN is quoted in the USER= setting for LDAP authentication, Exim
removes any URL quoting that it may contain before passing it to the LDAP library.
Apparently
some libraries do this for themselves, but some do not. Removing the URL
quoting has two advantages:

.ilist
It makes it possible to use the same &%quote_ldap_dn%& expansion for USER=
DNs as with DNs inside actual queries.
.next
It permits spaces inside USER= DNs.
.endlist

For example, a setting such as
.code
USER=cn=${quote_ldap_dn:$1}
.endd
should work even if &$1$& contains spaces.

Expanded data for the PASS= value should be quoted using the &%quote%&
expansion operator, rather than the LDAP quote operators. The only reason this
field needs quoting is to ensure that it conforms to the Exim syntax, which
does not allow unquoted spaces. For example:
.code
PASS=${quote:$3}
.endd
The LDAP authentication mechanism can be used to check passwords as part of
SMTP authentication. See the &%ldapauth%& expansion string condition in chapter
&<<CHAPexpand>>&.



.subsection "Format of data returned by LDAP" SECID71
.cindex "LDAP" "returned data formats"
The &(ldapdn)& lookup type returns the Distinguished Name from a single entry
as a sequence of values, for example
.code
cn=manager,o=University of Cambridge,c=UK
.endd
The &(ldap)& lookup type generates an error if more than one entry matches the
search filter, whereas &(ldapm)& permits this case, and inserts a newline in
the result between the data from different entries. It is possible for multiple
values to be returned for both &(ldap)& and &(ldapm)&, but in the former case
you know that whatever values are returned all came from a single entry in the
directory.

In the common case where you specify a single attribute in your LDAP query, the
result is not quoted, and does not contain the attribute name. If the attribute
has multiple values, they are separated by commas. Any comma that is
part of an attribute's value is doubled.

If you specify multiple attributes, the result contains space-separated, quoted
strings, each preceded by the attribute name and an equals sign. Within the
quotes, the quote character, backslash, and newline are escaped with
backslashes, and commas are used to separate multiple values for the attribute.
Any commas in attribute values are doubled
(permitting treatment of the values as a comma-separated list).
Apart from the escaping, the string within quotes takes the same form as the
output when a single attribute is requested. Specifying no attributes is the
same as specifying all of an entry's attributes.

Here are some examples of the output format. The first line of each pair is an
LDAP query, and the second is the data that is returned. The attribute called
&%attr1%& has two values, one of them with an embedded comma, whereas
&%attr2%& has only one value. Both attributes are derived from &%attr%&
(they have SUP &%attr%& in their schema definitions).

.code
ldap:///o=base?attr1?sub?(uid=fred)
value1.1,value1,,2

ldap:///o=base?attr2?sub?(uid=fred)
value two

ldap:///o=base?attr?sub?(uid=fred)
value1.1,value1,,2,value two

ldap:///o=base?attr1,attr2?sub?(uid=fred)
attr1="value1.1,value1,,2" attr2="value two"

ldap:///o=base??sub?(uid=fred)
objectClass="top" attr1="value1.1,value1,,2" attr2="value two"
.endd
You can
make use of Exim's &%-be%& option to run expansion tests and thereby check the
results of LDAP lookups.
The &%extract%& operator in string expansions can be used to pick out
individual fields from data that consists of &'key'&=&'value'& pairs.
The &%listextract%& operator should be used to pick out individual values
of attributes, even when only a single value is expected.
The doubling of embedded commas allows you to use the returned data as a
comma separated list (using the "<," syntax for changing the input list separator).




.section "More about NIS+" "SECTnisplus"
.cindex "NIS+ lookup type"
.cindex "lookup" "NIS+"
NIS+ queries consist of a NIS+ &'indexed name'& followed by an optional colon
and field name. If this is given, the result of a successful query is the
contents of the named field; otherwise the result consists of a concatenation
of &'field-name=field-value'& pairs, separated by spaces. Empty values and
values containing spaces are quoted. For example, the query
.code
[name=mg1456],passwd.org_dir
.endd
might return the string
.code
name=mg1456 passwd="" uid=999 gid=999 gcos="Martin Guerre"
home=/home/mg1456 shell=/bin/bash shadow=""
.endd
(split over two lines here to fit on the page), whereas
.code
[name=mg1456],passwd.org_dir:gcos
.endd
would just return
.code
Martin Guerre
.endd
with no quotes. A NIS+ lookup fails if NIS+ returns more than one table entry
for the given indexed key. The effect of the &%quote_nisplus%& expansion
operator is to double any quote characters within the text.



.section "SQL lookups" "SECTsql"
.cindex "SQL lookup types"
.cindex "MySQL" "lookup type"
.cindex "PostgreSQL lookup type"
.cindex "lookup" "MySQL"
.cindex "lookup" "PostgreSQL"
.cindex "Oracle" "lookup type"
.cindex "lookup" "Oracle"
.cindex "InterBase lookup type"
.cindex "lookup" "InterBase"
.cindex "Redis lookup type"
.cindex lookup Redis
Exim can support lookups in InterBase, MySQL, Oracle, PostgreSQL, Redis,
and SQLite
databases. Queries for these databases contain SQL statements, so an example
might be
.code
${lookup mysql{select mailbox from users where id='userx'}\
  {$value}fail}
.endd
If the result of the query contains more than one field, the data for each
field in the row is returned, preceded by its name, so the result of
.code
${lookup pgsql{select home,name from users where id='userx'}\
  {$value}}
.endd
might be
.code
home=/home/userx name="Mister X"
.endd
Empty values and values containing spaces are double quoted, with embedded
quotes escaped by a backslash. If the result of the query contains just one
field, the value is passed back verbatim, without a field name, for example:
.code
Mister X
.endd
If the result of the query yields more than one row, it is all concatenated,
with a newline between the data for each row.


.subsection "More about MySQL, PostgreSQL, Oracle, InterBase, and Redis" SECID72
.cindex "MySQL" "lookup type"
.cindex "PostgreSQL lookup type"
.cindex "lookup" "MySQL"
.cindex "lookup" "PostgreSQL"
.cindex "Oracle" "lookup type"
.cindex "lookup" "Oracle"
.cindex "InterBase lookup type"
.cindex "lookup" "InterBase"
.cindex "Redis lookup type"
.cindex lookup Redis
If any MySQL, PostgreSQL, Oracle, InterBase or Redis lookups are used, the
&%mysql_servers%&, &%pgsql_servers%&, &%oracle_servers%&, &%ibase_servers%&,
or &%redis_servers%&
option (as appropriate) must be set to a colon-separated list of server
information.
.oindex &%mysql_servers%&
.oindex &%pgsql_servers%&
.oindex &%oracle_servers%&
.oindex &%ibase_servers%&
.oindex &%redis_servers%&
(For MySQL and PostgreSQL, the global option need not be set if all
queries contain their own server information &-- see section
&<<SECTspeserque>>&.)
For all but Redis
each item in the list is a slash-separated list of four
items: host name, database name, user name, and password. In the case of
Oracle, the host name field is used for the &"service name"&, and the database
name field is not used and should be empty. For example:
.code
hide oracle_servers = oracle.plc.example//userx/abcdwxyz
.endd
Because password data is sensitive, you should always precede the setting with
&"hide"&, to prevent non-admin users from obtaining the setting via the &%-bP%&
option. Here is an example where two MySQL servers are listed:
.code
hide mysql_servers = localhost/users/root/secret:\
                     otherhost/users/root/othersecret
.endd
For MySQL and PostgreSQL, a host may be specified as <&'name'&>:<&'port'&> but
because this is a colon-separated list, the colon has to be doubled. For each
query, these parameter groups are tried in order until a connection is made and
a query is successfully processed. The result of a query may be that no data is
found, but that is still a successful query. In other words, the list of
servers provides a backup facility, not a list of different places to look.

For Redis the global option need not be specified if all queries contain their
own server information &-- see section &<<SECTspeserque>>&.
If specified, the option must be set to a colon-separated list of server
information.
Each item in the list is a slash-separated list of three items:
host, database number, and password.
.olist
The host is required and may be either an IPv4 address and optional
port number (separated by a colon, which needs doubling due to the
higher-level list), or a Unix socket pathname enclosed in parentheses
.next
The database number is optional; if present that number is selected in the backend
.next
The password is optional; if present it is used to authenticate to the backend
.endlist

The &%quote_mysql%&, &%quote_pgsql%&, and &%quote_oracle%& expansion operators
convert newline, tab, carriage return, and backspace to \n, \t, \r, and \b
respectively, and the characters single-quote, double-quote, and backslash
itself are escaped with backslashes.

The &%quote_redis%& expansion operator
escapes whitespace and backslash characters with a backslash.

.subsection "Specifying the server in the query" SECTspeserque
For MySQL, PostgreSQL and Redis lookups (but not currently for Oracle and InterBase),
it is possible to specify a list of servers with an individual query. This is
done by appending a comma-separated option to the query type:
.display
&`,servers=`&&'server1:server2:server3:...'&
.endd
Each item in the list may take one of two forms:
.olist
If it contains no slashes it is assumed to be just a host name. The appropriate
global option (&%mysql_servers%& or &%pgsql_servers%&) is searched for a host
of the same name, and the remaining parameters (database, user, password) are
taken from there.
.next
If it contains any slashes, it is taken as a complete parameter set.
.endlist
The list of servers is used in exactly the same way as the global list.
Once a connection to a server has happened and a query has been
successfully executed, processing of the lookup ceases.

This feature is intended for use in master/slave situations where updates
are occurring and you want to update the master rather than a slave. If the
master is in the list as a backup for reading, you might have a global setting
like this:
.code
mysql_servers = slave1/db/name/pw:\
                slave2/db/name/pw:\
                master/db/name/pw
.endd
In an updating lookup, you could then write:
.code
${lookup mysql,servers=master {UPDATE ...} }
.endd
That query would then be sent only to the master server. If, on the other hand,
the master is not to be used for reading, and so is not present in the global
option, you can still update it by a query of this form:
.code
${lookup pgsql,servers=master/db/name/pw {UPDATE ...} }
.endd

A now-deprecated syntax places the servers specification before the query,
semicolon separated:
.code
${lookup mysql{servers=master; UPDATE ...} }
.endd
The new version avoids issues with tainted
arguments explicitly expanded as part of the query.
The entire string within the braces becomes tainted,
including the server specification - which is not permissible.
If the older sytax is used, a warning message will be logged.
This syntax will be removed in a future release.

&*Note*&: server specifications in list-style lookups are still problematic.


.subsection "Special MySQL features" SECID73
For MySQL, an empty host name or the use of &"localhost"& in &%mysql_servers%&
causes a connection to the server on the local host by means of a Unix domain
socket. An alternate socket can be specified in parentheses.
An option group name for MySQL option files can be specified in square brackets;
the default value is &"exim"&.
The full syntax of each item in &%mysql_servers%& is:
.display
<&'hostspec'&><&'portspec'&>(<&'socket name'&>)[<&'option group'&>]/&&&
   <&'database'&>/<&'user'&>/<&'password'&>
.endd
Any of the four sub-parts of the first field can be omitted. For normal use on
the local host it can be left blank or set to just &"localhost"&.

.new
A &'hostspec'& can be a hostname, an IPv4 address or an IPv6 address.
For the latter, a &'portspec'& is a dot followed by a port number;
for the other two a &'portspec'& is a colon followed by a port number.
.wen

Note that the default list-separator for the list of servers is a colon so
(unless that is changed) all colons in list items must be doubled.

No database need be supplied &-- but if it is absent here, it must be given in
the queries.

If a MySQL query is issued that does not request any data (an insert, update,
or delete command), the result of the lookup is the number of rows affected.

&*Warning*&: This can be misleading. If an update does not actually change
anything (for example, setting a field to the value it already has), the result
is zero because no rows are affected.

To get an encryted connection, use a Mysql option file with the required
parameters for the connection.


.subsection "Special PostgreSQL features" SECID74
.new
The &'hostspec'& for PostgreSQL follows the same rules as for MySQL above.
.wen

PostgreSQL lookups can also use Unix domain socket connections to the database.
This is usually faster and costs less CPU time than a TCP/IP connection.
However it can be used only if the mail server runs on the same machine as the
database server. A configuration line for PostgreSQL via Unix domain sockets
looks like this:
.code
hide pgsql_servers = (/tmp/.s.PGSQL.5432)/db/user/password : ...
.endd
In other words, instead of supplying a host name, a path to the socket is
given. The path name is enclosed in parentheses so that its slashes aren't
visually confused with the delimiters for the other server parameters.

If a PostgreSQL query is issued that does not request any data (an insert,
update, or delete command), the result of the lookup is the number of rows
affected.

.subsection "More about SQLite" SECTsqlite
.cindex "lookup" "SQLite"
.cindex "sqlite lookup type"
SQLite is different to the other SQL lookups because a filename is required in
addition to the SQL query. An SQLite database is a single file, and there is no
daemon as in the other SQL databases.

.oindex &%sqlite_dbfile%&
There are two ways of
specifying the file.
The first is by using the &%sqlite_dbfile%& main option.
The second, which allows separate files for each query,
is to use an option appended, comma-separated, to the &"sqlite"&
lookup type word.  The option is the word &"file"&, then an equals,
then the filename.
The filename in this case cannot contain whitespace or open-brace charachters.

A deprecated method is available, prefixing the query with the filename
separated by white space.
This means that
.cindex "tainted data" "sqlite file"
the query cannot use any tainted values, as that taints
the entire query including the filename - resulting in a refusal to open
the file.

In all the above cases the filename must be an absolute path.

Here is a lookup expansion example:
.code
sqlite_dbfile = /some/thing/sqlitedb
...
${lookup sqlite {select name from aliases where id='userx';}}
.endd
In a list, the syntax is similar. For example:
.code
domainlist relay_to_domains = sqlite;\
   select * from relays where ip='$sender_host_address';
.endd
The only character affected by the &%quote_sqlite%& operator is a single
quote, which it doubles.

.cindex timeout SQLite
.cindex sqlite "lookup timeout"
The SQLite library handles multiple simultaneous accesses to the database
internally. Multiple readers are permitted, but only one process can
update at once. Attempts to access the database while it is being updated
are rejected after a timeout period, during which the SQLite library
waits for the lock to be released. In Exim, the default timeout is set
to 5 seconds, but it can be changed by means of the &%sqlite_lock_timeout%&
option.

.subsection "More about Redis" SECTredis
.cindex "lookup" "Redis"
.cindex "redis lookup type"
Redis is a non-SQL database. Commands are simple get and set.
Examples:
.code
${lookup redis{set keyname ${quote_redis:objvalue plus}}}
${lookup redis{get keyname}}
.endd

As of release 4.91, "lightweight" support for Redis Cluster is available.
Requires &%redis_servers%& list to contain all the servers in the cluster, all
of which must be reachable from the running exim instance. If the cluster has
master/slave replication, the list must contain all the master and slave
servers.

When the Redis Cluster returns a "MOVED" response to a query, Exim does not
immediately follow the redirection but treats the response as a DEFER, moving on
to the next server in the &%redis_servers%& list until the correct server is
reached.

.ecindex IIDfidalo1
.ecindex IIDfidalo2


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Domain, host, address, and local part lists" &&&
         "CHAPdomhosaddlists" &&&
         "Domain, host, and address lists"
.scindex IIDdohoadli "lists of domains; hosts; etc."
A number of Exim configuration options contain lists of domains, hosts,
email addresses, or local parts. For example, the &%hold_domains%& option
contains a list of domains whose delivery is currently suspended. These lists
are also used as data in ACL statements (see chapter &<<CHAPACL>>&), and as
arguments to expansion conditions such as &%match_domain%&.

Each item in one of these lists is a pattern to be matched against a domain,
host, email address, or local part, respectively. In the sections below, the
different types of pattern for each case are described, but first we cover some
general facilities that apply to all four kinds of list.

Note that other parts of Exim use a &'string list'& which does not
support all the complexity available in
domain, host, address and local part lists.



.section "Results of list checking" SECTlistresults
The primary result of doing a list check is a truth value.
In some contexts additional information is stored
about the list element that matched:
.vlist
.vitem hosts
A &%hosts%& ACL condition
will store a result in the &$host_data$& variable.
.vitem local_parts
A &%local_parts%& router option or &%local_parts%& ACL condition
will store a result in the &$local_part_data$& variable.
.vitem domains
A &%domains%& router option or &%domains%& ACL condition
will store a result in the &$domain_data$& variable.
.vitem senders
A &%senders%& router option or &%senders%& ACL condition
will store a result in the &$sender_data$& variable.
.vitem recipients
A &%recipients%& ACL condition
will store a result in the &$recipient_data$& variable.
.endlist

The detail of the additional information depends on the
type of match and is given below as the &*value*& information.




.section "Expansion of lists" "SECTlistexpand"
.cindex "expansion" "of lists"
.new
Each list, after any leading change-of-separator specification
(see &<<SECTlistsepchange>>&) is expanded as a single string,
.cindex "tainted data" tracking
&*Note*&: As a result, if any component was tainted then the
entire expansion result string becomes tainted.

Splitting out a leading explicit change-of-separator permits
one being safely used on a list that has tainted components
while still detecting the use of a tainted setting.
The latter is not permitted.
.wen

&'Exception: the router headers_remove option, where list-item
splitting is done before string-expansion.'&

The result of
expansion must be a list, possibly containing empty items, which is split up
into separate items for matching. By default, colon is the separator character,
but this can be varied if necessary. See sections &<<SECTlistconstruct>>& and
&<<SECTempitelis>>& for details of the list syntax; the second of these
discusses the way to specify empty list items.


If the string expansion is forced to fail, Exim behaves as if the item it is
testing (domain, host, address, or local part) is not in the list. Other
expansion failures cause temporary errors.

If an item in a list is a regular expression, backslashes, dollars and possibly
other special characters in the expression must be protected against
misinterpretation by the string expander. The easiest way to do this is to use
the &`\N`& expansion feature to indicate that the contents of the regular
expression should not be expanded. For example, in an ACL you might have:
.code
deny senders = \N^\d{8}\w@.*\.baddomain\.example$\N : \
               ${lookup{$domain}lsearch{/badsenders/bydomain}}
.endd
The first item is a regular expression that is protected from expansion by
&`\N`&, whereas the second uses the expansion to obtain a list of unwanted
senders based on the receiving domain.




.subsection "Negated items in lists" SECID76
.cindex "list" "negation"
.cindex "negation" "in lists"
Items in a list may be positive or negative. Negative items are indicated by a
leading exclamation mark, which may be followed by optional white space. A list
defines a set of items (domains, etc). When Exim processes one of these lists,
it is trying to find out whether a domain, host, address, or local part
(respectively) is in the set that is defined by the list. It works like this:

The list is scanned from left to right. If a positive item is matched, the
subject that is being checked is in the set; if a negative item is matched, the
subject is not in the set. If the end of the list is reached without the
subject having matched any of the patterns, it is in the set if the last item
was a negative one, but not if it was a positive one. For example, the list in
.code
domainlist relay_to_domains = !a.b.c : *.b.c
.endd
matches any domain ending in &'.b.c'& except for &'a.b.c'&. Domains that match
neither &'a.b.c'& nor &'*.b.c'& do not match, because the last item in the
list is positive. However, if the setting were
.code
domainlist relay_to_domains = !a.b.c
.endd
then all domains other than &'a.b.c'& would match because the last item in the
list is negative. In other words, a list that ends with a negative item behaves
as if it had an extra item &`:*`& on the end.

Another way of thinking about positive and negative items in lists is to read
the connector as &"or"& after a positive item and as &"and"& after a negative
item.



.subsection "File names in lists" SECTfilnamlis
.cindex "list" "filename in"
If an item in a domain, host, address, or local part list is an absolute
filename (beginning with a slash character), each line of the file is read and
processed as if it were an independent item in the list, except that further
filenames are not allowed,
and no expansion of the data from the file takes place.
Empty lines in the file are ignored, and the file may also contain comment
lines:

.ilist
For domain and host lists, if a # character appears anywhere in a line of the
file, it and all following characters are ignored.
.next
Because local parts may legitimately contain # characters, a comment in an
address list or local part list file is recognized only if # is preceded by
white space or the start of the line. For example:
.code
not#comment@x.y.z   # but this is a comment
.endd
.endlist

Putting a filename in a list has the same effect as inserting each line of the
file as an item in the list (blank lines and comments excepted). However, there
is one important difference: the file is read each time the list is processed,
so if its contents vary over time, Exim's behaviour changes.

If a filename is preceded by an exclamation mark, the sense of any match
within the file is inverted. For example, if
.code
hold_domains = !/etc/nohold-domains
.endd
and the file contains the lines
.code
!a.b.c
*.b.c
.endd
then &'a.b.c'& is in the set of domains defined by &%hold_domains%&, whereas
any domain matching &`*.b.c`& is not.



.subsection "An lsearch file is not an out-of-line list" SECID77
As will be described in the sections that follow, lookups can be used in lists
to provide indexed methods of checking list membership. There has been some
confusion about the way &(lsearch)& lookups work in lists. Because
an &(lsearch)& file contains plain text and is scanned sequentially, it is
sometimes thought that it is allowed to contain wild cards and other kinds of
non-constant pattern. This is not the case. The keys in an &(lsearch)& file are
always fixed strings, just as for any other single-key lookup type.

If you want to use a file to contain wild-card patterns that form part of a
list, just give the filename on its own, without a search type, as described
in the previous section. You could also use the &(wildlsearch)& or
&(nwildlsearch)&, but there is no advantage in doing this.




.subsection "Named lists" SECTnamedlists
.cindex "named lists"
.cindex "list" "named"
A list of domains, hosts, email addresses, or local parts can be given a name
which is then used to refer to the list elsewhere in the configuration. This is
particularly convenient if the same list is required in several different
places. It also allows lists to be given meaningful names, which can improve
the readability of the configuration. For example, it is conventional to define
a domain list called &'local_domains'& for all the domains that are handled
locally on a host, using a configuration line such as
.code
domainlist local_domains = localhost:my.dom.example
.endd
Named lists are referenced by giving their name preceded by a plus sign, so,
for example, a router that is intended to handle local domains would be
configured with the line
.code
domains = +local_domains
.endd
The first router in a configuration is often one that handles all domains
except the local ones, using a configuration with a negated item like this:
.code
dnslookup:
  driver = dnslookup
  domains = ! +local_domains
  transport = remote_smtp
  no_more
.endd
The four kinds of named list are created by configuration lines starting with
the words &%domainlist%&, &%hostlist%&, &%addresslist%&, or &%localpartlist%&,
respectively. Then there follows the name that you are defining, followed by an
equals sign and the list itself. For example:
.code
hostlist    relay_from_hosts = 192.168.23.0/24 : my.friend.example
addresslist bad_senders = cdb;/etc/badsenders
.endd
A named list may refer to other named lists:
.code
domainlist  dom1 = first.example : second.example
domainlist  dom2 = +dom1 : third.example
domainlist  dom3 = fourth.example : +dom2 : fifth.example
.endd
&*Warning*&: If the last item in a referenced list is a negative one, the
effect may not be what you intended, because the negation does not propagate
out to the higher level. For example, consider:
.code
domainlist  dom1 = !a.b
domainlist  dom2 = +dom1 : *.b
.endd
The second list specifies &"either in the &%dom1%& list or &'*.b'&"&. The first
list specifies just &"not &'a.b'&"&, so the domain &'x.y'& matches it. That
means it matches the second list as well. The effect is not the same as
.code
domainlist  dom2 = !a.b : *.b
.endd
where &'x.y'& does not match. It's best to avoid negation altogether in
referenced lists if you can.

.cindex "hiding named list values"
.cindex "named lists" "hiding value of"
Some named list definitions may contain sensitive data, for example, passwords for
accessing databases. To stop non-admin users from using the &%-bP%& command
line option to read these values, you can precede the definition with the
word &"hide"&. For example:
.code
hide domainlist filter_for_domains = ldap;PASS=secret ldap::/// ...
.endd


Named lists may have a performance advantage. When Exim is routing an
address or checking an incoming message, it caches the result of tests on named
lists. So, if you have a setting such as
.code
domains = +local_domains
.endd
on several of your routers
or in several ACL statements,
the actual test is done only for the first one. However, the caching works only
if there are no expansions within the list itself or any sublists that it
references. In other words, caching happens only for lists that are known to be
the same each time they are referenced.

By default, there may be up to 16 named lists of each type. This limit can be
extended by changing a compile-time variable. The use of domain and host lists
is recommended for concepts such as local domains, relay domains, and relay
hosts. The default configuration is set up like this.



.subsection "Named lists compared with macros" SECID78
.cindex "list" "named compared with macro"
.cindex "macro" "compared with named list"
At first sight, named lists might seem to be no different from macros in the
configuration file. However, macros are just textual substitutions. If you
write
.code
ALIST = host1 : host2
auth_advertise_hosts = !ALIST
.endd
it probably won't do what you want, because that is exactly the same as
.code
auth_advertise_hosts = !host1 : host2
.endd
Notice that the second host name is not negated. However, if you use a host
list, and write
.code
hostlist alist = host1 : host2
auth_advertise_hosts = ! +alist
.endd
the negation applies to the whole list, and so that is equivalent to
.code
auth_advertise_hosts = !host1 : !host2
.endd


.subsection "Named list caching" SECID79
.cindex "list" "caching of named"
.cindex "caching" "named lists"
While processing a message, Exim caches the result of checking a named list if
it is sure that the list is the same each time. In practice, this means that
the cache operates only if the list contains no $ characters, which guarantees
that it will not change when it is expanded. Sometimes, however, you may have
an expanded list that you know will be the same each time within a given
message. For example:
.code
domainlist special_domains = \
           ${lookup{$sender_host_address}cdb{/some/file}}
.endd
This provides a list of domains that depends only on the sending host's IP
address. If this domain list is referenced a number of times (for example,
in several ACL lines, or in several routers) the result of the check is not
cached by default, because Exim does not know that it is going to be the
same list each time.

By appending &`_cache`& to &`domainlist`& you can tell Exim to go ahead and
cache the result anyway. For example:
.code
domainlist_cache special_domains = ${lookup{...
.endd
If you do this, you should be absolutely sure that caching is going to do
the right thing in all cases. When in doubt, leave it out.



.section "Domain lists" "SECTdomainlist"
.cindex "domain list" "patterns for"
.cindex "list" "domain list"
Domain lists contain patterns that are to be matched against a mail domain.
The following types of item may appear in domain lists:

.ilist
.cindex "primary host name"
.cindex "host name" "matched in domain list"
.oindex "&%primary_hostname%&"
.cindex "domain list" "matching primary host name"
.cindex "@ in a domain list"
If a pattern consists of a single @ character, it matches the local host name,
as set by the &%primary_hostname%& option (or defaulted). This makes it
possible to use the same configuration file on several different hosts that
differ only in their names.

The value for a match will be the primary host name.


.next
.cindex "@[] in a domain list"
.cindex "domain list" "matching local IP interfaces"
.cindex "domain literal"
If a pattern consists of the string &`@[]`& it matches an IP address enclosed
in square brackets (as in an email address that contains a domain literal), but
only if that IP address is recognized as local for email routing purposes. The
&%local_interfaces%& and &%extra_local_interfaces%& options can be used to
control which of a host's several IP addresses are treated as local.
In today's Internet, the use of domain literals is controversial;
see the &%allow_domain_literals%& main option.

The value for a match will be the string &`@[]`&.


.next
.cindex "@mx_any"
.cindex "@mx_primary"
.cindex "@mx_secondary"
.cindex "domain list" "matching MX pointers to local host"
If a pattern consists of the string &`@mx_any`& it matches any domain that
has an MX record pointing to the local host or to any host that is listed in
.oindex "&%hosts_treat_as_local%&"
&%hosts_treat_as_local%&. The items &`@mx_primary`& and &`@mx_secondary`&
are similar, except that the first matches only when a primary MX target is the
local host, and the second only when no primary MX target is the local host,
but a secondary MX target is. &"Primary"& means an MX record with the lowest
preference value &-- there may of course be more than one of them.

The MX lookup that takes place when matching a pattern of this type is
performed with the resolver options for widening names turned off. Thus, for
example, a single-component domain will &'not'& be expanded by adding the
resolver's default domain. See the &%qualify_single%& and &%search_parents%&
options of the &(dnslookup)& router for a discussion of domain widening.

Sometimes you may want to ignore certain IP addresses when using one of these
patterns. You can specify this by following the pattern with &`/ignore=`&<&'ip
list'&>, where <&'ip list'&> is a list of IP addresses. These addresses are
ignored when processing the pattern (compare the &%ignore_target_hosts%& option
on a router). For example:
.code
domains = @mx_any/ignore=127.0.0.1
.endd
This example matches any domain that has an MX record pointing to one of
the local host's IP addresses other than 127.0.0.1.

The list of IP addresses is in fact processed by the same code that processes
host lists, so it may contain CIDR-coded network specifications and it may also
contain negative items.

Because the list of IP addresses is a sublist within a domain list, you have to
be careful about delimiters if there is more than one address. Like any other
list, the default delimiter can be changed. Thus, you might have:
.code
domains = @mx_any/ignore=<;127.0.0.1;0.0.0.0 : \
          an.other.domain : ...
.endd
so that the sublist uses semicolons for delimiters. When IPv6 addresses are
involved, it is easiest to change the delimiter for the main list as well:
.code
domains = <? @mx_any/ignore=<;127.0.0.1;::1 ? \
          an.other.domain ? ...
.endd
The value for a match will be the list element string (starting &`@mx_`&).


.next
.cindex "asterisk" "in domain list"
.cindex "domain list" "asterisk in"
.cindex "domain list" "matching &""ends with""&"
If a pattern starts with an asterisk, the remaining characters of the pattern
are compared with the terminating characters of the domain. The use of &"*"& in
domain lists differs from its use in partial matching lookups. In a domain
list, the character following the asterisk need not be a dot, whereas partial
matching works only in terms of dot-separated components. For example, a domain
list item such as &`*key.ex`& matches &'donkey.ex'& as well as
&'cipher.key.ex'&.

The value for a match will be the list element string (starting with the asterisk).
Additionally, &$0$& will be set to the matched string
and &$1$& to the variable portion which the asterisk matched.

.next
.cindex "regular expressions" "in domain list"
.cindex "domain list" "matching regular expression"
If a pattern starts with a circumflex character, it is treated as a regular
expression, and matched against the domain using a regular expression matching
function. The circumflex is treated as part of the regular expression.
Email domains are case-independent, so this regular expression match is by
default case-independent, but you can make it case-dependent by starting it
with &`(?-i)`&. References to descriptions of the syntax of regular expressions
are given in chapter &<<CHAPregexp>>&.

&*Warning*&: Because domain lists are expanded before being processed, you
must escape any backslash and dollar characters in the regular expression, or
use the special &`\N`& sequence (see chapter &<<CHAPexpand>>&) to specify that
it is not to be expanded (unless you really do want to build a regular
expression by expansion, of course).

The value for a match will be the list element string (starting with the circumflex).
Additionally, &$0$& will be set to the string matching the regular expression,
and &$1$& (onwards) to any submatches identified by parentheses.



.next
.cindex "lookup" "in domain list"
.cindex "domain list" "matching by lookup"
If a pattern starts with the name of a single-key lookup type followed by a
semicolon (for example, &"dbm;"& or &"lsearch;"&), the remainder of the pattern
must be a filename in a suitable format for the lookup type. For example, for
&"cdb;"& it must be an absolute path:
.code
domains = cdb;/etc/mail/local_domains.cdb
.endd
The appropriate type of lookup is done on the file using the domain name as the
key. In most cases, the value resulting from the lookup is not used; Exim is interested
only in whether or not the key is present in the file. However, when a lookup
is used for the &%domains%& option on a router
or a &%domains%& condition in an ACL statement, the value is preserved in the
&$domain_data$& variable and can be referred to in other router options or
other statements in the same ACL.
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using ACL domains condition"
The value will be untainted.

&*Note*&: If the data result of the lookup (as opposed to the key)
is empty, then this empty value is stored in &$domain_data$&.
The option to return the key for the lookup, as the value,
may be what is wanted.


.next
Any of the single-key lookup type names may be preceded by
&`partial`&<&'n'&>&`-`&, where the <&'n'&> is optional, for example,
.code
domains = partial-dbm;/partial/domains
.endd
This causes partial matching logic to be invoked; a description of how this
works is given in section &<<SECTpartiallookup>>&.

.next
.cindex "asterisk" "in lookup type"
Any of the single-key lookup types may be followed by an asterisk. This causes
a default lookup for a key consisting of a single asterisk to be done if the
original lookup fails. This is not a useful feature when using a domain list to
select particular domains (because any domain would match), but it might have
value if the result of the lookup is being used via the &$domain_data$&
expansion variable.

.next
If the pattern starts with the name of a query-style lookup type followed by a
semicolon (for example, &"nisplus;"& or &"ldap;"&), the remainder of the
pattern must be an appropriate query for the lookup type, as described in
chapter &<<CHAPfdlookup>>&. For example:
.code
hold_domains = mysql;select domain from holdlist \
  where domain = '${quote_mysql:$domain}';
.endd
In most cases, the value resulting from the lookup is not used (so for an SQL query, for
example, it doesn't matter what field you select). Exim is interested only in
whether or not the query succeeds. However, when a lookup is used for the
&%domains%& option on a router, the value is preserved in the &$domain_data$&
variable and can be referred to in other options.
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using router domains option"
The value will be untainted.

.next
If the pattern starts with the name of a lookup type
of either kind (single-key or query-style) it may be
followed by a comma and options,
The options are lookup-type specific and consist of a comma-separated list.
Each item starts with a tag and and equals "=" sign.

.next
.cindex "domain list" "matching literal domain name"
If none of the above cases apply, a caseless textual comparison is made
between the pattern and the domain.

The value for a match will be the list element string.
.cindex "tainted data" "de-tainting"
Note that this is commonly untainted
(depending on the way the list was created).
Specifically, explicit text in the configuration file in not tainted.
This is a useful way of obtaining an untainted equivalent to
the domain, for later operations.

However if the list (including one-element lists)
is created by expanding a variable containing tainted data,
it is tainted and so will the match value be.
.endlist


Here is an example that uses several different kinds of pattern:
.code
domainlist funny_domains = \
  @ : \
  lib.unseen.edu : \
  *.foundation.fict.example : \
  \N^[1-2]\d{3}\.fict\.example$\N : \
  partial-dbm;/opt/data/penguin/book : \
  nis;domains.byname : \
  nisplus;[name=$domain,status=local],domains.org_dir
.endd
There are obvious processing trade-offs among the various matching modes. Using
an asterisk is faster than a regular expression, and listing a few names
explicitly probably is too. The use of a file or database lookup is expensive,
but may be the only option if hundreds of names are required. Because the
patterns are tested in order, it makes sense to put the most commonly matched
patterns earlier.



.section "Host lists" "SECThostlist"
.cindex "host list" "patterns in"
.cindex "list" "host list"
Host lists are used to control what remote hosts are allowed to do. For
example, some hosts may be allowed to use the local host as a relay, and some
may be permitted to use the SMTP ETRN command. Hosts can be identified in
two different ways, by name or by IP address. In a host list, some types of
pattern are matched to a host name, and some are matched to an IP address.
You need to be particularly careful with this when single-key lookups are
involved, to ensure that the right value is being used as the key.


.subsection "Special host list patterns" SECID80
.cindex "empty item in hosts list"
.cindex "host list" "empty string in"
If a host list item is the empty string, it matches only when no remote host is
involved. This is the case when a message is being received from a local
process using SMTP on the standard input, that is, when a TCP/IP connection is
not used.

.cindex "asterisk" "in host list"
The special pattern &"*"& in a host list matches any host or no host. Neither
the IP address nor the name is actually inspected.



.subsection "Host list patterns that match by IP address" SECThoslispatip
.cindex "host list" "matching IP addresses"
If an IPv4 host calls an IPv6 host and the call is accepted on an IPv6 socket,
the incoming address actually appears in the IPv6 host as
&`::ffff:`&<&'v4address'&>. When such an address is tested against a host
list, it is converted into a traditional IPv4 address first. (Not all operating
systems accept IPv4 calls on IPv6 sockets, as there have been some security
concerns.)

The following types of pattern in a host list check the remote host by
inspecting its IP address:

.ilist
If the pattern is a plain domain name (not a regular expression, not starting
with *, not a lookup of any kind), Exim calls the operating system function
to find the associated IP address(es). Exim uses the newer
&[getipnodebyname()]& function when available, otherwise &[gethostbyname()]&.
This typically causes a forward DNS lookup of the name. The result is compared
with the IP address of the subject host.

If there is a temporary problem (such as a DNS timeout) with the host name
lookup, a temporary error occurs. For example, if the list is being used in an
ACL condition, the ACL gives a &"defer"& response, usually leading to a
temporary SMTP error code. If no IP address can be found for the host name,
what happens is described in section &<<SECTbehipnot>>& below.

.next
.cindex "@ in a host list"
If the pattern is &"@"&, the primary host name is substituted and used as a
domain name, as just described.

.next
If the pattern is an IP address, it is matched against the IP address of the
subject host. IPv4 addresses are given in the normal &"dotted-quad"& notation.
IPv6 addresses can be given in colon-separated format, but the colons have to
be doubled so as not to be taken as item separators when the default list
separator is used. IPv6 addresses are recognized even when Exim is compiled
without IPv6 support. This means that if they appear in a host list on an
IPv4-only host, Exim will not treat them as host names. They are just addresses
that can never match a client host.

.next
.cindex "@[] in a host list"
If the pattern is &"@[]"&, it matches the IP address of any IP interface on
the local host. For example, if the local host is an IPv4 host with one
interface address 10.45.23.56, these two ACL statements have the same effect:
.code
accept hosts = 127.0.0.1 : 10.45.23.56
accept hosts = @[]
.endd
.next
.cindex "CIDR notation"
If the pattern is an IP address followed by a slash and a mask length, for
example
.code
10.11.42.0/24
.endd
, it is matched against the IP address of the subject
host under the given mask. This allows an entire network of hosts to be
included (or excluded) by a single item. The mask uses CIDR notation; it
specifies the number of address bits that must match, starting from the most
significant end of the address.

&*Note*&: The mask is &'not'& a count of addresses, nor is it the high number
of a range of addresses. It is the number of bits in the network portion of the
address. The above example specifies a 24-bit netmask, so it matches all 256
addresses in the 10.11.42.0 network. An item such as
.code
192.168.23.236/31
.endd
matches just two addresses, 192.168.23.236 and 192.168.23.237. A mask value of
32 for an IPv4 address is the same as no mask at all; just a single address
matches.

Here is another example which shows an IPv4 and an IPv6 network:
.code
recipient_unqualified_hosts = 192.168.0.0/16: \
                              3ffe::ffff::836f::::/48
.endd
The doubling of list separator characters applies only when these items
appear inline in a host list. It is not required when indirecting via a file.
For example:
.code
recipient_unqualified_hosts = /opt/exim/unqualnets
.endd
could make use of a file containing
.code
172.16.0.0/12
3ffe:ffff:836f::/48
.endd
to have exactly the same effect as the previous example. When listing IPv6
addresses inline, it is usually more convenient to use the facility for
changing separator characters. This list contains the same two networks:
.code
recipient_unqualified_hosts = <; 172.16.0.0/12; \
                                 3ffe:ffff:836f::/48
.endd
The separator is changed to semicolon by the leading &"<;"& at the start of the
list.
.endlist



.subsection "Host list patterns for single-key lookups by host address" &&&
         SECThoslispatsikey
.cindex "host list" "lookup of IP address"
When a host is to be identified by a single-key lookup of its complete IP
address, the pattern takes this form:
.display
&`net-<`&&'single-key-search-type'&&`>;<`&&'search-data'&&`>`&
.endd
For example:
.code
hosts_lookup = net-cdb;/hosts-by-ip.db
.endd
The text form of the IP address of the subject host is used as the lookup key.
IPv6 addresses are converted to an unabbreviated form, using lower case
letters, with dots as separators because colon is the key terminator in
&(lsearch)& files. [Colons can in fact be used in keys in &(lsearch)& files by
quoting the keys, but this is a facility that was added later.] The data
returned by the lookup is not used.

.cindex "IP address" "masking"
.cindex "host list" "masked IP address"
Single-key lookups can also be performed using masked IP addresses, using
patterns of this form:
.display
&`net<`&&'number'&&`>-<`&&'single-key-search-type'&&`>;<`&&'search-data'&&`>`&
.endd
For example:
.code
net24-dbm;/networks.db
.endd
The IP address of the subject host is masked using <&'number'&> as the mask
length. A textual string is constructed from the masked value, followed by the
mask, and this is used as the lookup key. For example, if the host's IP address
is 192.168.34.6, the key that is looked up for the above example is
&"192.168.34.0/24"&.

When an IPv6 address is converted to a string, dots are normally used instead
of colons, so that keys in &(lsearch)& files need not contain colons (which
terminate &(lsearch)& keys). This was implemented some time before the ability
to quote keys was made available in &(lsearch)& files. However, the more
recently implemented &(iplsearch)& files do require colons in IPv6 keys
(notated using the quoting facility) so as to distinguish them from IPv4 keys.
For this reason, when the lookup type is &(iplsearch)&, IPv6 addresses are
converted using colons and not dots.
In all cases except IPv4-mapped IPv6, full, unabbreviated IPv6
addresses are always used.
The latter are converted to IPv4 addresses, in dotted-quad form.

Ideally, it would be nice to tidy up this anomalous situation by changing to
colons in all cases, given that quoting is now available for &(lsearch)&.
However, this would be an incompatible change that might break some existing
configurations.

&*Warning*&: Specifying &%net32-%& (for an IPv4 address) or &%net128-%& (for an
IPv6 address) is not the same as specifying just &%net-%& without a number. In
the former case the key strings include the mask value, whereas in the latter
case the IP address is used on its own.



.subsection "Host list patterns that match by host name" SECThoslispatnam
.cindex "host" "lookup failures"
.cindex "unknown host name"
.cindex "host list" "matching host name"
There are several types of pattern that require Exim to know the name of the
remote host. These are either wildcard patterns or lookups by name. (If a
complete hostname is given without any wildcarding, it is used to find an IP
address to match against, as described in section &<<SECThoslispatip>>&
above.)

If the remote host name is not already known when Exim encounters one of these
patterns, it has to be found from the IP address.
Although many sites on the Internet are conscientious about maintaining reverse
DNS data for their hosts, there are also many that do not do this.
Consequently, a name cannot always be found, and this may lead to unwanted
effects. Take care when configuring host lists with wildcarded name patterns.
Consider what will happen if a name cannot be found.

Because of the problems of determining host names from IP addresses, matching
against host names is not as common as matching against IP addresses.

By default, in order to find a host name, Exim first does a reverse DNS lookup;
if no name is found in the DNS, the system function (&[gethostbyaddr()]& or
&[getipnodebyaddr()]& if available) is tried. The order in which these lookups
are done can be changed by setting the &%host_lookup_order%& option. For
security, once Exim has found one or more names, it looks up the IP addresses
for these names and compares them with the IP address that it started with.
Only those names whose IP addresses match are accepted. Any other names are
discarded. If no names are left, Exim behaves as if the host name cannot be
found. In the most common case there is only one name and one IP address.

There are some options that control what happens if a host name cannot be
found. These are described in section &<<SECTbehipnot>>& below.

.cindex "host" "alias for"
.cindex "alias for host"
As a result of aliasing, hosts may have more than one name. When processing any
of the following types of pattern, all the host's names are checked:

.ilist
.cindex "asterisk" "in host list"
If a pattern starts with &"*"& the remainder of the item must match the end of
the host name. For example, &`*.b.c`& matches all hosts whose names end in
&'.b.c'&. This special simple form is provided because this is a very common
requirement. Other kinds of wildcarding require the use of a regular
expression.
.next
.cindex "regular expressions" "in host list"
.cindex "host list" "regular expression in"
If the item starts with &"^"& it is taken to be a regular expression which is
matched against the host name. Host names are case-independent, so this regular
expression match is by default case-independent, but you can make it
case-dependent by starting it with &`(?-i)`&. References to descriptions of the
syntax of regular expressions are given in chapter &<<CHAPregexp>>&. For
example,
.code
^(a|b)\.c\.d$
.endd
is a regular expression that matches either of the two hosts &'a.c.d'& or
&'b.c.d'&. When a regular expression is used in a host list, you must take care
that backslash and dollar characters are not misinterpreted as part of the
string expansion. The simplest way to do this is to use &`\N`& to mark that
part of the string as non-expandable. For example:
.code
sender_unqualified_hosts = \N^(a|b)\.c\.d$\N : ....
.endd
&*Warning*&: If you want to match a complete host name, you must include the
&`$`& terminating metacharacter in the regular expression, as in the above
example. Without it, a match at the start of the host name is all that is
required.
.endlist




.subsection "Behaviour when an IP address or name cannot be found" SECTbehipnot
.cindex "host" "lookup failures, permanent"
While processing a host list, Exim may need to look up an IP address from a
name (see section &<<SECThoslispatip>>&), or it may need to look up a host name
from an IP address (see section &<<SECThoslispatnam>>&). In either case, the
behaviour when it fails to find the information it is seeking is the same.

&*Note*&: This section applies to permanent lookup failures. It does &'not'&
apply to temporary DNS errors, whose handling is described in the next section.

.cindex "&`+include_unknown`&"
.cindex "&`+ignore_unknown`&"
Exim parses a host list from left to right. If it encounters a permanent
lookup failure in any item in the host list before it has found a match,
Exim treats it as a failure and the default behavior is as if the host
does not match the list. This may not always be what you want to happen.
To change Exim's behaviour, the special items &`+include_unknown`& or
&`+ignore_unknown`& may appear in the list (at top level &-- they are
not recognized in an indirected file).

.ilist
If any item that follows &`+include_unknown`& requires information that
cannot found, Exim behaves as if the host does match the list. For example,
.code
host_reject_connection = +include_unknown:*.enemy.ex
.endd
rejects connections from any host whose name matches &`*.enemy.ex`&, and also
any hosts whose name it cannot find.

.next
If any item that follows &`+ignore_unknown`& requires information that cannot
be found, Exim ignores that item and proceeds to the rest of the list. For
example:
.code
accept hosts = +ignore_unknown : friend.example : \
               192.168.4.5
.endd
accepts from any host whose name is &'friend.example'& and from 192.168.4.5,
whether or not its host name can be found. Without &`+ignore_unknown`&, if no
name can be found for 192.168.4.5, it is rejected.
.endlist

Both &`+include_unknown`& and &`+ignore_unknown`& may appear in the same
list. The effect of each one lasts until the next, or until the end of the
list.

.subsection "Mixing wildcarded host names and addresses in host lists" &&&
         SECTmixwilhos
.cindex "host list" "mixing names and addresses in"

This section explains the host/ip processing logic with the same concepts
as the previous section, but specifically addresses what happens when a
wildcarded hostname is one of the items in the hostlist.

.ilist
If you have name lookups or wildcarded host names and
IP addresses in the same host list, you should normally put the IP
addresses first. For example, in an ACL you could have:
.code
accept hosts = 10.9.8.7 : *.friend.example
.endd
The reason you normally would order it this way lies in the
left-to-right way that Exim processes lists.  It can test IP addresses
without doing any DNS lookups, but when it reaches an item that requires
a host name, it fails if it cannot find a host name to compare with the
pattern. If the above list is given in the opposite order, the
&%accept%& statement fails for a host whose name cannot be found, even
if its IP address is 10.9.8.7.

.next
If you really do want to do the name check first, and still recognize the IP
address, you can rewrite the ACL like this:
.code
accept hosts = *.friend.example
accept hosts = 10.9.8.7
.endd
If the first &%accept%& fails, Exim goes on to try the second one. See chapter
&<<CHAPACL>>& for details of ACLs. Alternatively, you can use
&`+ignore_unknown`&, which was discussed in depth in the first example in
this section.
.endlist


.subsection "Temporary DNS errors when looking up host information" &&&
         SECTtemdnserr
.cindex "host" "lookup failures, temporary"
.cindex "&`+include_defer`&"
.cindex "&`+ignore_defer`&"
A temporary DNS lookup failure normally causes a defer action (except when
&%dns_again_means_nonexist%& converts it into a permanent error). However,
host lists can include &`+ignore_defer`& and &`+include_defer`&, analogous to
&`+ignore_unknown`& and &`+include_unknown`&, as described in the previous
section. These options should be used with care, probably only in non-critical
host lists such as whitelists.



.subsection "Host list patterns for single-key lookups by host name" &&&
         SECThoslispatnamsk
.cindex "unknown host name"
.cindex "host list" "matching host name"
If a pattern is of the form
.display
<&'single-key-search-type'&>;<&'search-data'&>
.endd
for example
.code
dbm;/host/accept/list
.endd
a single-key lookup is performed, using the host name as its key. If the
lookup succeeds, the host matches the item. The actual data that is looked up
is not used.

&*Reminder*&: With this kind of pattern, you must have host &'names'& as
keys in the file, not IP addresses. If you want to do lookups based on IP
addresses, you must precede the search type with &"net-"&
(see section &<<SECThoslispatsikey>>&).
There is, however, no reason why you could not use
two items in the same list, one doing an address lookup and one doing a name
lookup, both using the same file.



.subsection "Host list patterns for query-style lookups" SECID81
If a pattern is of the form
.display
<&'query-style-search-type'&>;<&'query'&>
.endd
the query is obeyed, and if it succeeds, the host matches the item. The actual
data that is looked up is not used. The variables &$sender_host_address$& and
&$sender_host_name$& can be used in the query. For example:
.code
hosts_lookup = pgsql;\
  select ip from hostlist where ip='$sender_host_address'
.endd
The value of &$sender_host_address$& for an IPv6 address contains colons. You
can use the &%sg%& expansion item to change this if you need to. If you want to
use masked IP addresses in database queries, you can use the &%mask%& expansion
operator.

If the query contains a reference to &$sender_host_name$&, Exim automatically
looks up the host name if it has not already done so. (See section
&<<SECThoslispatnam>>& for comments on finding host names.)

Historical note: prior to release 4.30, Exim would always attempt to find a
host name before running the query, unless the search type was preceded by
&`net-`&. This is no longer the case. For backwards compatibility, &`net-`& is
still recognized for query-style lookups, but its presence or absence has no
effect. (Of course, for single-key lookups, &`net-`& &'is'& important.
See section &<<SECThoslispatsikey>>&.)





.section "Address lists" "SECTaddresslist"
.cindex "list" "address list"
.cindex "address list" "empty item"
.cindex "address list" "patterns"
Address lists contain patterns that are matched against mail addresses. There
is one special case to be considered: the sender address of a bounce message is
always empty. You can test for this by providing an empty item in an address
list. For example, you can set up a router to process bounce messages by
using this option setting:
.code
senders = :
.endd
The presence of the colon creates an empty item. If you do not provide any
data, the list is empty and matches nothing. The empty sender can also be
detected by a regular expression that matches an empty string,
and by a query-style lookup that succeeds when &$sender_address$& is empty.

Non-empty items in an address list can be straightforward email addresses. For
example:
.code
senders = jbc@askone.example : hs@anacreon.example
.endd
A certain amount of wildcarding is permitted. If a pattern contains an @
character, but is not a regular expression and does not begin with a
semicolon-terminated lookup type (described below), the local part of the
subject address is compared with the local part of the pattern, which may start
with an asterisk. If the local parts match, the domain is checked in exactly
the same way as for a pattern in a domain list. For example, the domain can be
wildcarded, refer to a named list, or be a lookup:
.code
deny senders = *@*.spamming.site:\
               *@+hostile_domains:\
               bozo@partial-lsearch;/list/of/dodgy/sites:\
               *@dbm;/bad/domains.db
.endd
.cindex "local part" "starting with !"
.cindex "address list" "local part starting with !"
If a local part that begins with an exclamation mark is required, it has to be
specified using a regular expression, because otherwise the exclamation mark is
treated as a sign of negation, as is standard in lists.

If a non-empty pattern that is not a regular expression or a lookup does not
contain an @ character, it is matched against the domain part of the subject
address. The only two formats that are recognized this way are a literal
domain, or a domain pattern that starts with *. In both these cases, the effect
is the same as if &`*@`& preceded the pattern. For example:
.code
deny senders = enemy.domain : *.enemy.domain
.endd

The following kinds of more complicated address list pattern can match any
address, including the empty address that is characteristic of bounce message
senders:

.ilist
.cindex "regular expressions" "in address list"
.cindex "address list" "regular expression in"
If (after expansion) a pattern starts with &"^"&, a regular expression match is
done against the complete address, with the pattern as the regular expression.
You must take care that backslash and dollar characters are not misinterpreted
as part of the string expansion. The simplest way to do this is to use &`\N`&
to mark that part of the string as non-expandable. For example:
.code
deny senders = \N^.*this.*@example\.com$\N : \
               \N^\d{8}.+@spamhaus.example$\N : ...
.endd
The &`\N`& sequences are removed by the expansion, so these items do indeed
start with &"^"& by the time they are being interpreted as address patterns.

.next
.cindex "address list" "lookup for complete address"
Complete addresses can be looked up by using a pattern that starts with a
lookup type terminated by a semicolon, followed by the data for the lookup. For
example:
.code
deny senders = cdb;/etc/blocked.senders : \
  mysql;select address from blocked where \
  address='${quote_mysql:$sender_address}'
.endd
Both query-style and single-key lookup types can be used. For a single-key
lookup type, Exim uses the complete address as the key. However, empty keys are
not supported for single-key lookups, so a match against the empty address
always fails. This restriction does not apply to query-style lookups.

Partial matching for single-key lookups (section &<<SECTpartiallookup>>&)
cannot be used, and is ignored if specified, with an entry being written to the
panic log.
.cindex "*@ with single-key lookup"
However, you can configure lookup defaults, as described in section
&<<SECTdefaultvaluelookups>>&, but this is useful only for the &"*@"& type of
default. For example, with this lookup:
.code
accept senders = lsearch*@;/some/file
.endd
the file could contains lines like this:
.code
user1@domain1.example
*@domain2.example
.endd
and for the sender address &'nimrod@jaeger.example'&, the sequence of keys
that are tried is:
.code
nimrod@jaeger.example
*@jaeger.example
*
.endd
&*Warning 1*&: Do not include a line keyed by &"*"& in the file, because that
would mean that every address matches, thus rendering the test useless.

&*Warning 2*&: Do not confuse these two kinds of item:
.code
deny recipients = dbm*@;/some/file
deny recipients = *@dbm;/some/file
.endd
The first does a whole address lookup, with defaulting, as just described,
because it starts with a lookup type. The second matches the local part and
domain independently, as described in a bullet point below.
.endlist


The following kinds of address list pattern can match only non-empty addresses.
If the subject address is empty, a match against any of these pattern types
always fails.


.ilist
.cindex "@@ with single-key lookup"
.cindex "address list" "@@ lookup type"
.cindex "address list" "split local part and domain"
If a pattern starts with &"@@"& followed by a single-key lookup item
(for example, &`@@lsearch;/some/file`&), the address that is being checked is
split into a local part and a domain. The domain is looked up in the file. If
it is not found, there is no match. If it is found, the data that is looked up
from the file is treated as a colon-separated list of local part patterns, each
of which is matched against the subject local part in turn.

.cindex "asterisk" "in address list"
The lookup may be a partial one, and/or one involving a search for a default
keyed by &"*"& (see section &<<SECTdefaultvaluelookups>>&). The local part
patterns that are looked up can be regular expressions or begin with &"*"&, or
even be further lookups. They may also be independently negated. For example,
with
.code
deny senders = @@dbm;/etc/reject-by-domain
.endd
the data from which the DBM file is built could contain lines like
.code
baddomain.com:  !postmaster : *
.endd
to reject all senders except &%postmaster%& from that domain.

.cindex "local part" "starting with !"
If a local part that actually begins with an exclamation mark is required, it
has to be specified using a regular expression. In &(lsearch)& files, an entry
may be split over several lines by indenting the second and subsequent lines,
but the separating colon must still be included at line breaks. White space
surrounding the colons is ignored. For example:
.code
aol.com:  spammer1 : spammer2 : ^[0-9]+$ :
  spammer3 : spammer4
.endd
As in all colon-separated lists in Exim, a colon can be included in an item by
doubling.

If the last item in the list starts with a right angle-bracket, the remainder
of the item is taken as a new key to look up in order to obtain a continuation
list of local parts. The new key can be any sequence of characters. Thus one
might have entries like
.code
aol.com: spammer1 : spammer 2 : >*
xyz.com: spammer3 : >*
*:       ^\d{8}$
.endd
in a file that was searched with &%@@dbm*%&, to specify a match for 8-digit
local parts for all domains, in addition to the specific local parts listed for
each domain. Of course, using this feature costs another lookup each time a
chain is followed, but the effort needed to maintain the data is reduced.

.cindex "loop" "in lookups"
It is possible to construct loops using this facility, and in order to catch
them, the chains may be no more than fifty items long.

.next
The @@<&'lookup'&> style of item can also be used with a query-style
lookup, but in this case, the chaining facility is not available. The lookup
can only return a single list of local parts.
.endlist

&*Warning*&: There is an important difference between the address list items
in these two examples:
.code
senders = +my_list
senders = *@+my_list
.endd
In the first one, &`my_list`& is a named address list, whereas in the second
example it is a named domain list.




.subsection "Case of letters in address lists" SECTcasletadd
.cindex "case of local parts"
.cindex "address list" "case forcing"
.cindex "case forcing in address lists"
Domains in email addresses are always handled caselessly, but for local parts
case may be significant on some systems (see &%caseful_local_part%& for how
Exim deals with this when routing addresses). However,
&url(https://www.rfc-editor.org/rfc/rfc2505,RFC 2505) (&'Anti-Spam
Recommendations for SMTP MTAs'&) suggests that matching of addresses to
blocking lists should be done in a case-independent manner. Since most address
lists in Exim are used for this kind of control, Exim attempts to do this by
default.

The domain portion of an address is always lowercased before matching it to an
address list. The local part is lowercased by default, and any string
comparisons that take place are done caselessly. This means that the data in
the address list itself, in files included as plain filenames, and in any file
that is looked up using the &"@@"& mechanism, can be in any case. However, the
keys in files that are looked up by a search type other than &(lsearch)& (which
works caselessly) must be in lower case, because these lookups are not
case-independent.

.cindex "&`+caseful`&"
To allow for the possibility of caseful address list matching, if an item in
an address list is the string &"+caseful"&, the original case of the local
part is restored for any comparisons that follow, and string comparisons are no
longer case-independent. This does not affect the domain, which remains in
lower case. However, although independent matches on the domain alone are still
performed caselessly, regular expressions that match against an entire address
become case-sensitive after &"+caseful"& has been seen.



.section "Local part lists" "SECTlocparlis"
.cindex "list" "local part list"
.cindex "local part" "list"
These behave in the same way as domain and host lists, with the following
changes:

Case-sensitivity in local part lists is handled in the same way as for address
lists, as just described. The &"+caseful"& item can be used if required. In a
setting of the &%local_parts%& option in a router with &%caseful_local_part%&
set false, the subject is lowercased and the matching is initially
case-insensitive. In this case, &"+caseful"& will restore case-sensitive
matching in the local part list, but not elsewhere in the router. If
&%caseful_local_part%& is set true in a router, matching in the &%local_parts%&
option is case-sensitive from the start.

If a local part list is indirected to a file (see section &<<SECTfilnamlis>>&),
comments are handled in the same way as address lists &-- they are recognized
only if the # is preceded by white space or the start of the line.
Otherwise, local part lists are matched in the same way as domain lists, except
that the special items that refer to the local host (&`@`&, &`@[]`&,
&`@mx_any`&, &`@mx_primary`&, and &`@mx_secondary`&) are not recognized.
Refer to section &<<SECTdomainlist>>& for details of the other available item
types.
.ecindex IIDdohoadli




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "String expansions" "CHAPexpand"
.scindex IIDstrexp "expansion" "of strings"
Many strings in Exim's runtime configuration are expanded before use. Some of
them are expanded every time they are used; others are expanded only once.

When a string is being expanded it is copied verbatim from left to right except
.cindex expansion "string concatenation"
when a dollar or backslash character is encountered. A dollar specifies the
start of a portion of the string that is interpreted and replaced as described
below in section &<<SECTexpansionitems>>& onwards. Backslash is used as an
escape character, as described in the following section.

.cindex "tainted data" tracking
If any porttion of the result string is tainted, the entire result is.

Whether a string is expanded depends upon the context.  Usually this is solely
dependent upon the option for which a value is sought; in this documentation,
options for which string expansion is performed are marked with &dagger; after
the data type.  ACL rules always expand strings.  A couple of expansion
conditions do not expand some of the brace-delimited branches, for security
reasons,
.cindex "tainted data" expansion
.cindex "tainted data" definition
.cindex expansion "tainted data"
and expansion of data deriving from the sender (&"tainted data"&)
is not permitted (including acessing a file using a tainted name).

Common ways of obtaining untainted equivalents of variables with
tainted values
.cindex "tainted data" "de-tainting"
come down to using the tainted value as a lookup key in a trusted database.
This database could be the filesystem structure,
or the password file,
or accessed via a DBMS.
Specific methods are indexed under &"de-tainting"&.



.section "Literal text in expanded strings" "SECTlittext"
.cindex "expansion" "including literal text"
An uninterpreted dollar can be included in an expanded string by putting a
backslash in front of it. A backslash can be used to prevent any special
character being treated specially in an expansion, including backslash itself.
If the string appears in quotes in the configuration file, two backslashes are
required because the quotes themselves cause interpretation of backslashes when
the string is read in (see section &<<SECTstrings>>&).

.cindex "expansion" "non-expandable substrings"
A portion of the string can specified as non-expandable by placing it between
two occurrences of &`\N`&. This is particularly useful for protecting regular
expressions, which often contain backslashes and dollar signs. For example:
.code
deny senders = \N^\d{8}[a-z]@some\.site\.example$\N
.endd
On encountering the first &`\N`&, the expander copies subsequent characters
without interpretation until it reaches the next &`\N`& or the end of the
string.



.section "Character escape sequences in expanded strings" "SECID82"
.cindex "expansion" "escape sequences"
A backslash followed by one of the letters &"n"&, &"r"&, or &"t"& in an
expanded string is recognized as an escape sequence for the character newline,
carriage return, or tab, respectively. A backslash followed by up to three
octal digits is recognized as an octal encoding for a single character, and a
backslash followed by &"x"& and up to two hexadecimal digits is a hexadecimal
encoding.

These escape sequences are also recognized in quoted strings when they are read
in. Their interpretation in expansions as well is useful for unquoted strings,
and for other cases such as looked-up strings that are then expanded.


.section "Testing string expansions" "SECID83"
.cindex "expansion" "testing"
.cindex "testing" "string expansion"
.oindex "&%-be%&"
Many expansions can be tested by calling Exim with the &%-be%& option. This
takes the command arguments, or lines from the standard input if there are no
arguments, runs them through the string expansion code, and writes the results
to the standard output. Variables based on configuration values are set up, but
since no message is being processed, variables such as &$local_part$& have no
value. Nevertheless the &%-be%& option can be useful for checking out file and
database lookups, and the use of expansion operators such as &%sg%&, &%substr%&
and &%nhash%&.

When reading lines from the standard input,
macros can be defined and ACL variables can be set.
For example:
.code
MY_MACRO = foo
set acl_m_myvar = bar
.endd
Such macros and variables can then be used in later input lines.

Exim gives up its root privilege when it is called with the &%-be%& option, and
instead runs under the uid and gid it was called with, to prevent users from
using &%-be%& for reading files to which they do not have access.

.oindex "&%-bem%&"
If you want to test expansions that include variables whose values are taken
from a message, there are two other options that can be used. The &%-bem%&
option is like &%-be%& except that it is followed by a filename. The file is
read as a message before doing the test expansions. For example:
.code
exim -bem /tmp/test.message '$h_subject:'
.endd
The &%-Mset%& option is used in conjunction with &%-be%& and is followed by an
Exim message identifier. For example:
.code
exim -be -Mset 1GrA8W-0004WS-LQ '$recipients'
.endd
This loads the message from Exim's spool before doing the test expansions, and
is therefore restricted to admin users.


.section "Forced expansion failure" "SECTforexpfai"
.cindex "expansion" "forced failure"
A number of expansions that are described in the following section have
alternative &"true"& and &"false"& substrings, enclosed in brace characters
(which are sometimes called &"curly brackets"&). Which of the two strings is
used depends on some condition that is evaluated as part of the expansion. If,
instead of a &"false"& substring, the word &"fail"& is used (not in braces),
the entire string expansion fails in a way that can be detected by the code
that requested the expansion. This is called &"forced expansion failure"&, and
its consequences depend on the circumstances. In some cases it is no different
from any other expansion failure, but in others a different action may be
taken. Such variations are mentioned in the documentation of the option that is
being expanded.




.section "Expansion items" "SECTexpansionitems"
The following items are recognized in expanded strings. White space may be used
between sub-items that are keywords or substrings enclosed in braces inside an
outer set of braces, to improve readability. &*Warning*&: Within braces,
white space is significant.

.vlist
.vitem &*$*&<&'variable&~name'&>&~or&~&*${*&<&'variable&~name'&>&*}*&
.cindex "expansion" "variables"
Substitute the contents of the named variable, for example:
.code
$local_part
${domain}
.endd
The second form can be used to separate the name from subsequent alphanumeric
characters. This form (using braces) is available only for variables; it does
&'not'& apply to message headers. The names of the variables are given in
section &<<SECTexpvar>>& below. If the name of a non-existent variable is
given, the expansion fails.

.vitem &*${*&<&'op'&>&*:*&<&'string'&>&*}*&
.cindex "expansion" "operators"
The string is first itself expanded, and then the operation specified by
<&'op'&> is applied to it. For example:
.code
${lc:$local_part}
.endd
The string starts with the first character after the colon, which may be
leading white space. A list of operators is given in section &<<SECTexpop>>&
below. The operator notation is used for simple expansion items that have just
one argument, because it reduces the number of braces and therefore makes the
string easier to understand.

.vitem &*$bheader_*&<&'header&~name'&>&*:*&&~or&~&*$bh_*&<&'header&~name'&>&*:*&
This item inserts &"basic"& header lines. It is described with the &%header%&
expansion item below.


.vitem "&*${acl{*&<&'name'&>&*}{*&<&'arg'&>&*}...}*&"
.cindex "expansion" "calling an acl"
.cindex "&%acl%&" "call from expansion"
The name and zero to nine argument strings are first expanded separately.  The expanded
arguments are assigned to the variables &$acl_arg1$& to &$acl_arg9$& in order.
Any unused are made empty.  The variable &$acl_narg$& is set to the number of
arguments.  The named ACL (see chapter &<<CHAPACL>>&) is called
and may use the variables; if another acl expansion is used the values
are restored after it returns.  If the ACL sets
a value using a "message =" modifier and returns accept or deny, the value becomes
the result of the expansion.
If no message is set and the ACL returns accept or deny
the expansion result is an empty string.
If the ACL returns defer the result is a forced-fail.  Otherwise the expansion fails.


.vitem "&*${authresults{*&<&'authserv-id'&>&*}}*&"
.cindex authentication "results header"
.chindex Authentication-Results:
.cindex authentication "expansion item"
This item returns a string suitable for insertion as an
&'Authentication-Results:'&
header line.
The given <&'authserv-id'&> is included in the result; typically this
will be a domain name identifying the system performing the authentications.
Methods that might be present in the result include:
.code
none
iprev
auth
spf
dkim
.endd

Example use (as an ACL modifier):
.code
      add_header = :at_start:${authresults {$primary_hostname}}
.endd
This is safe even if no authentication results are available
and would generally be placed in the DATA ACL.


.vitem "&*${certextract{*&<&'field'&>&*}{*&<&'certificate'&>&*}&&&
       {*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&"
.cindex "expansion" "extracting certificate fields"
.cindex "&%certextract%&" "certificate fields"
.cindex "certificate" "extracting fields"
The <&'certificate'&> must be a variable of type certificate.
The field name is expanded and used to retrieve the relevant field from
the certificate.  Supported fields are:
.display
&`version        `&
&`serial_number  `&
&`subject        `& RFC4514 DN
&`issuer         `& RFC4514 DN
&`notbefore      `& time
&`notafter       `& time
&`sig_algorithm  `&
&`signature      `&
&`subj_altname   `& tagged list
&`ocsp_uri       `& list
&`crl_uri        `& list
.endd
If the field is found,
<&'string2'&> is expanded, and replaces the whole item;
otherwise <&'string3'&> is used. During the expansion of <&'string2'&> the
variable &$value$& contains the value that has been extracted. Afterwards, it
is restored to any previous value it might have had.

If {<&'string3'&>} is omitted, the item is replaced by an empty string if the
key is not found. If {<&'string2'&>} is also omitted, the value that was
extracted is used.

Some field names take optional modifiers, appended and separated by commas.

The field selectors marked as "RFC4514" above
output a Distinguished Name string which is
not quite
parseable by Exim as a comma-separated tagged list
(the exceptions being elements containing commas).
RDN elements of a single type may be selected by
a modifier of the type label; if so the expansion
result is a list (newline-separated by default).
The separator may be changed by another modifier of
a right angle-bracket followed immediately by the new separator.
Recognised RDN type labels include "CN", "O", "OU" and "DC".

The field selectors marked as "time" above
take an optional modifier of "int"
for which the result is the number of seconds since epoch.
Otherwise the result is a human-readable string
in the timezone selected by the main "timezone" option.

The field selectors marked as "list" above return a list,
newline-separated by default,
(embedded separator characters in elements are doubled).
The separator may be changed by a modifier of
a right angle-bracket followed immediately by the new separator.

The field selectors marked as "tagged" above
prefix each list element with a type string and an equals sign.
Elements of only one type may be selected by a modifier
which is one of "dns", "uri" or "mail";
if so the element tags are omitted.

If not otherwise noted field values are presented in human-readable form.

.vitem "&*${dlfunc{*&<&'file'&>&*}{*&<&'function'&>&*}{*&<&'arg'&>&*}&&&
       {*&<&'arg'&>&*}...}*&"
.cindex &%dlfunc%&
This expansion dynamically loads and then calls a locally-written C function.
This functionality is available only if Exim is compiled with
.code
EXPAND_DLFUNC=yes
.endd
set in &_Local/Makefile_&. Once loaded, Exim remembers the dynamically loaded
object so that it doesn't reload the same object file in the same Exim process
(but of course Exim does start new processes frequently).

There may be from zero to eight arguments to the function.

When compiling
a local function that is to be called in this way,
first &_DLFUNC_IMPL_& should be defined,
and second &_local_scan.h_& should be included.
The Exim variables and functions that are defined by that API
are also available for dynamically loaded functions. The function itself
must have the following type:
.code
int dlfunction(uschar **yield, int argc, uschar *argv[])
.endd
Where &`uschar`& is a typedef for &`unsigned char`& in &_local_scan.h_&. The
function should return one of the following values:

&`OK`&: Success. The string that is placed in the variable &'yield'& is put
into the expanded string that is being built.

&`FAIL`&: A non-forced expansion failure occurs, with the error message taken
from &'yield'&, if it is set.

&`FAIL_FORCED`&: A forced expansion failure occurs, with the error message
taken from &'yield'& if it is set.

&`ERROR`&: Same as &`FAIL`&, except that a panic log entry is written.

When compiling a function that is to be used in this way with gcc,
you need to add &%-shared%& to the gcc command. Also, in the Exim build-time
configuration, you must add &%-export-dynamic%& to EXTRALIBS.


.vitem "&*${env{*&<&'key'&>&*}{*&<&'string1'&>&*}{*&<&'string2'&>&*}}*&"
.cindex "expansion" "extracting value from environment"
.cindex "environment" "values from"
The key is first expanded separately, and leading and trailing white space
removed.
This is then searched for as a name in the environment.
If a variable is found then its value is placed in &$value$&
and <&'string1'&> is expanded, otherwise <&'string2'&> is expanded.

Instead of {<&'string2'&>} the word &"fail"& (not in curly brackets) can
appear, for example:
.code
${env{USER}{$value} fail }
.endd
This forces an expansion failure (see section &<<SECTforexpfai>>&);
{<&'string1'&>} must be present for &"fail"& to be recognized.

If {<&'string2'&>} is omitted an empty string is substituted on
search failure.
If {<&'string1'&>} is omitted the search result is substituted on
search success.

The environment is adjusted by the &%keep_environment%& and
&%add_environment%& main section options.


.vitem "&*${extract{*&<&'key'&>&*}{*&<&'string1'&>&*}{*&<&'string2'&>&*}&&&
       {*&<&'string3'&>&*}}*&"
.cindex "expansion" "extracting substrings by key"
.cindex "&%extract%&" "substrings by key"
The key and <&'string1'&> are first expanded separately. Leading and trailing
white space is removed from the key (but not from any of the strings). The key
must not be empty and must not consist entirely of digits.
The expanded <&'string1'&> must be of the form:
.display
<&'key1'&> = <&'value1'&>  <&'key2'&> = <&'value2'&> ...
.endd
.vindex "&$value$&"
where the equals signs and spaces (but not both) are optional. If any of the
values contain white space, they must be enclosed in double quotes, and any
values that are enclosed in double quotes are subject to escape processing as
described in section &<<SECTstrings>>&. The expanded <&'string1'&> is searched
for the value that corresponds to the key. The search is case-insensitive. If
the key is found, <&'string2'&> is expanded, and replaces the whole item;
otherwise <&'string3'&> is used. During the expansion of <&'string2'&> the
variable &$value$& contains the value that has been extracted. Afterwards, it
is restored to any previous value it might have had.

If {<&'string3'&>} is omitted, the item is replaced by an empty string if the
key is not found. If {<&'string2'&>} is also omitted, the value that was
extracted is used. Thus, for example, these two expansions are identical, and
yield &"2001"&:
.code
${extract{gid}{uid=1984 gid=2001}}
${extract{gid}{uid=1984 gid=2001}{$value}}
.endd
Instead of {<&'string3'&>} the word &"fail"& (not in curly brackets) can
appear, for example:
.code
${extract{Z}{A=... B=...}{$value} fail }
.endd
This forces an expansion failure (see section &<<SECTforexpfai>>&);
{<&'string2'&>} must be present for &"fail"& to be recognized.

.vitem "&*${extract json{*&<&'key'&>&*}{*&<&'string1'&>&*}{*&<&'string2'&>&*}&&&
       {*&<&'string3'&>&*}}*&" &&&
       "&*${extract jsons{*&<&'key'&>&*}{*&<&'string1'&>&*}{*&<&'string2'&>&*}&&&
       {*&<&'string3'&>&*}}*&"
.cindex "expansion" "extracting from JSON object"
.cindex JSON expansions
The key and <&'string1'&> are first expanded separately. Leading and trailing
white space is removed from the key (but not from any of the strings). The key
must not be empty and must not consist entirely of digits.
The expanded <&'string1'&> must be of the form:
.display
{ <&'"key1"'&> : <&'value1'&> ,  <&'"key2"'&> , <&'value2'&> ... }
.endd
.vindex "&$value$&"
The braces, commas and colons, and the quoting of the member name are required;
the spaces are optional.
Matching of the key against the member names is done case-sensitively.
For the &"json"& variant,
if a returned value is a JSON string, it retains its leading and
trailing quotes.
For the &"jsons"& variant, which is intended for use with JSON strings, the
leading and trailing quotes are removed from the returned value.
. XXX should be a UTF-8 compare

The results of matching are handled as above.


.vitem "&*${extract{*&<&'number'&>&*}{*&<&'separators'&>&*}&&&
        {*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&"
.cindex "expansion" "extracting substrings by number"
.cindex "&%extract%&" "substrings by number"
The <&'number'&> argument must consist entirely of decimal digits,
apart from leading and trailing white space, which is ignored.
This is what distinguishes this form of &%extract%& from the previous kind. It
behaves in the same way, except that, instead of extracting a named field, it
extracts from <&'string1'&> the field whose number is given as the first
argument. You can use &$value$& in <&'string2'&> or &`fail`& instead of
<&'string3'&> as before.

The fields in the string are separated by any one of the characters in the
separator string. These may include space or tab characters.
The first field is numbered one. If the number is negative, the fields are
counted from the end of the string, with the rightmost one numbered -1. If the
number given is zero, the entire string is returned. If the modulus of the
number is greater than the number of fields in the string, the result is the
expansion of <&'string3'&>, or the empty string if <&'string3'&> is not
provided. For example:
.code
${extract{2}{:}{x:42:99:& Mailer::/bin/bash}}
.endd
yields &"42"&, and
.code
${extract{-4}{:}{x:42:99:& Mailer::/bin/bash}}
.endd
yields &"99"&. Two successive separators mean that the field between them is
empty (for example, the fifth field above).


.vitem "&*${extract json {*&<&'number'&>&*}}&&&
        {*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&" &&&
       "&*${extract jsons{*&<&'number'&>&*}}&&&
        {*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&"
.cindex "expansion" "extracting from JSON array"
.cindex JSON expansions
The <&'number'&> argument must consist entirely of decimal digits,
apart from leading and trailing white space, which is ignored.

Field selection and result handling is as above;
there is no choice of field separator.
For the &"json"& variant,
if a returned value is a JSON string, it retains its leading and
trailing quotes.
For the &"jsons"& variant, which is intended for use with JSON strings, the
leading and trailing quotes are removed from the returned value.


.vitem &*${filter{*&<&'string'&>&*}{*&<&'condition'&>&*}}*&
.cindex "list" "selecting by condition"
.cindex "expansion" "selecting from list by condition"
.vindex "&$item$&"
.new
<&'string1'&> first has the part after any change-of-list-separator
(see &<<SECTlistsepchange>>&) expanded,
then the whole is taken as a list.
.wen
The default separator for the list is a colon.

For each item in this list,
its value is placed in &$item$&, and then the condition is evaluated.
Any modification of &$value$& by this evaluation is discarded.
If the condition is true, &$item$& is added to the output as an
item in a new list; if the condition is false, the item is discarded. The
separator used for the output list is the same as the one used for the
input, but a separator setting is not included in the output. For example:
.code
${filter{a:b:c}{!eq{$item}{b}}}
.endd
yields &`a:c`&. At the end of the expansion, the value of &$item$& is restored
to what it was before.
See also the &%map%& and &%reduce%& expansion items.


.vitem &*${hash{*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&
.cindex "hash function" "textual"
.cindex "expansion" "textual hash"
This is a textual hashing function, and was the first to be implemented in
early versions of Exim. In current releases, there are other hashing functions
(numeric, MD5, and SHA-1), which are described below.

The first two strings, after expansion, must be numbers. Call them <&'m'&> and
<&'n'&>. If you are using fixed values for these numbers, that is, if
<&'string1'&> and <&'string2'&> do not change when they are expanded, you can
use the simpler operator notation that avoids some of the braces:
.code
${hash_<n>_<m>:<string>}
.endd
The second number is optional (in both notations). If <&'n'&> is greater than
or equal to the length of the string, the expansion item returns the string.
Otherwise it computes a new string of length <&'n'&> by applying a hashing
function to the string. The new string consists of characters taken from the
first <&'m'&> characters of the string
.code
abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQWRSTUVWXYZ0123456789
.endd
If <&'m'&> is not present the value 26 is used, so that only lower case
letters appear. For example:
.display
&`$hash{3}{monty}}           `&   yields  &`jmg`&
&`$hash{5}{monty}}           `&   yields  &`monty`&
&`$hash{4}{62}{monty python}}`&   yields  &`fbWx`&
.endd

.vitem "&*$header_*&<&'header&~name'&>&*:*&&~or&~&&&
        &*$h_*&<&'header&~name'&>&*:*&" &&&
       "&*$bheader_*&<&'header&~name'&>&*:*&&~or&~&&&
        &*$bh_*&<&'header&~name'&>&*:*&" &&&
       "&*$lheader_*&<&'header&~name'&>&*:*&&~or&~&&&
        &*$lh_*&<&'header&~name'&>&*:*&" &&&
       "&*$rheader_*&<&'header&~name'&>&*:*&&~or&~&&&
        &*$rh_*&<&'header&~name'&>&*:*&"
.cindex "expansion" "header insertion"
.vindex "&$header_$&"
.vindex "&$bheader_$&"
.vindex "&$lheader_$&"
.vindex "&$rheader_$&"
.cindex "header lines" "in expansion strings"
.cindex "header lines" "character sets"
.cindex "header lines" "decoding"
Substitute the contents of the named message header line, for example
.code
$header_reply-to:
.endd
The newline that terminates a header line is not included in the expansion, but
internal newlines (caused by splitting the header line over several physical
lines) may be present.

The difference between the four pairs of expansions is in the way
the data in the header line is interpreted.

.ilist
.cindex "white space" "in header lines"
&%rheader%& gives the original &"raw"& content of the header line, with no
processing at all, and without the removal of leading and trailing white space.

.next
.cindex "list" "of header lines"
&%lheader%& gives a colon-separated list, one element per header when there
are multiple headers with a given name.
Any embedded colon characters within an element are doubled, so normal Exim
list-processing facilities can be used.
The terminating newline of each element is removed; in other respects
the content is &"raw"&.

.next
.cindex "base64 encoding" "in header lines"
&%bheader%& removes leading and trailing white space, and then decodes base64
or quoted-printable MIME &"words"& within the header text, but does no
character set translation. If decoding of what looks superficially like a MIME
&"word"& fails, the raw string is returned. If decoding
.cindex "binary zero" "in header line"
produces a binary zero character, it is replaced by a question mark &-- this is
what Exim does for binary zeros that are actually received in header lines.

.next
&%header%& tries to translate the string as decoded by &%bheader%& to a
standard character set. This is an attempt to produce the same string as would
be displayed on a user's MUA. If translation fails, the &%bheader%& string is
returned. Translation is attempted only on operating systems that support the
&[iconv()]& function. This is indicated by the compile-time macro HAVE_ICONV in
a system Makefile or in &_Local/Makefile_&.
.endlist ilist

In a filter file, the target character set for &%header%& can be specified by a
command of the following form:
.code
headers charset "UTF-8"
.endd
This command affects all references to &$h_$& (or &$header_$&) expansions in
subsequently obeyed filter commands. In the absence of this command, the target
character set in a filter is taken from the setting of the &%headers_charset%&
option in the runtime configuration. The value of this option defaults to the
value of HEADERS_CHARSET in &_Local/Makefile_&. The ultimate default is
ISO-8859-1.

Header names follow the syntax of
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822),
which states that they may contain
any printing characters except space and colon. Consequently, curly brackets
&'do not'& terminate header names, and should not be used to enclose them as
if they were variables. Attempting to do so causes a syntax error.

Only header lines that are common to all copies of a message are visible to
this mechanism. These are the original header lines that are received with the
message, and any that are added by an ACL statement or by a system
filter. Header lines that are added to a particular copy of a message by a
router or transport are not accessible.

For incoming SMTP messages, no header lines are visible in
ACLs that are obeyed before the data phase completes,
because the header structure is not set up until the message is received.
They are visible in DKIM, PRDR and DATA ACLs.
Header lines that are added in a RCPT ACL (for example)
are saved until the message's incoming header lines are available, at which
point they are added.
When any of the above ACLs are
running, however, header lines added by earlier ACLs are visible.

Upper case and lower case letters are synonymous in header names. If the
following character is white space, the terminating colon may be omitted, but
this is not recommended, because you may then forget it when it is needed. When
white space terminates the header name, this white space is included in the
expanded string.  If the message does not contain the given header, the
expansion item is replaced by an empty string. (See the &%def%& condition in
section &<<SECTexpcond>>& for a means of testing for the existence of a
header.)

If there is more than one header with the same name, they are all concatenated
to form the substitution string, up to a maximum length of 64K. Unless
&%rheader%& is being used, leading and trailing white space is removed from
each header before concatenation, and a completely empty header is ignored. A
newline character is then inserted between non-empty headers, but there is no
newline at the very end. For the &%header%& and &%bheader%& expansion, for
those headers that contain lists of addresses, a comma is also inserted at the
junctions between headers. This does not happen for the &%rheader%& expansion.

.cindex "tainted data" "message headers"
When the headers are from an incoming message,
the result of expanding any of these variables is tainted.


.vitem &*${hmac{*&<&'hashname'&>&*}{*&<&'secret'&>&*}{*&<&'string'&>&*}}*&
.cindex "expansion" "hmac hashing"
.cindex &%hmac%&
This function uses cryptographic hashing (either MD5 or SHA-1) to convert a
shared secret and some text into a message authentication code, as specified in
&url(https://www.rfc-editor.org/rfc/rfc2104.RFC 2104).
This differs from &`${md5:secret_text...}`& or
&`${sha1:secret_text...}`& in that the hmac step adds a signature to the
cryptographic hash, allowing for authentication that is not possible with MD5
or SHA-1 alone. The hash name must expand to either &`md5`& or &`sha1`& at
present. For example:
.code
${hmac{md5}{somesecret}{$primary_hostname $tod_log}}
.endd
For the hostname &'mail.example.com'& and time 2002-10-17 11:30:59, this
produces:
.code
dd97e3ba5d1a61b5006108f8c8252953
.endd
As an example of how this might be used, you might put in the main part of
an Exim configuration:
.code
SPAMSCAN_SECRET=cohgheeLei2thahw
.endd
In a router or a transport you could then have:
.code
headers_add = \
  X-Spam-Scanned: ${primary_hostname} ${message_exim_id} \
  ${hmac{md5}{SPAMSCAN_SECRET}\
  {${primary_hostname},${message_exim_id},$h_message-id:}}
.endd
Then given a message, you can check where it was scanned by looking at the
&'X-Spam-Scanned:'& header line. If you know the secret, you can check that
this header line is authentic by recomputing the authentication code from the
host name, message ID and the &'Message-id:'& header line. This can be done
using Exim's &%-be%& option, or by other means, for example, by using the
&'hmac_md5_hex()'& function in Perl.


.vitem &*${if&~*&<&'condition'&>&*&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}}*&
.cindex "expansion" "conditional"
.cindex "&%if%&, expansion item"
If <&'condition'&> is true, <&'string1'&> is expanded and replaces the whole
item; otherwise <&'string2'&> is used. The available conditions are described
in section &<<SECTexpcond>>& below. For example:
.code
${if eq {$local_part}{postmaster} {yes}{no} }
.endd
The second string need not be present; if it is not and the condition is not
true, the item is replaced with nothing. Alternatively, the word &"fail"& may
be present instead of the second string (without any curly brackets). In this
case, the expansion is forced to fail if the condition is not true (see section
&<<SECTforexpfai>>&).

If both strings are omitted, the result is the string &`true`& if the condition
is true, and the empty string if the condition is false. This makes it less
cumbersome to write custom ACL and router conditions. For example, instead of
.code
condition = ${if >{$acl_m4}{3}{true}{false}}
.endd
you can use
.code
condition = ${if >{$acl_m4}{3}}
.endd



.vitem &*${imapfolder{*&<&'foldername'&>&*}}*&
.cindex expansion "imap folder"
.cindex "&%imapfolder%& expansion item"
This item converts a (possibly multilevel, or with non-ASCII characters)
folder specification to a Maildir name for filesystem use.
For information on internationalisation support see &<<SECTi18nMDA>>&.



.vitem &*${length{*&<&'string1'&>&*}{*&<&'string2'&>&*}}*&
.cindex "expansion" "string truncation"
.cindex "&%length%& expansion item"
The &%length%& item is used to extract the initial portion of a string. Both
strings are expanded, and the first one must yield a number, <&'n'&>, say. If
you are using a fixed value for the number, that is, if <&'string1'&> does not
change when expanded, you can use the simpler operator notation that avoids
some of the braces:
.code
${length_<n>:<string>}
.endd
The result of this item is either the first <&'n'&> bytes or the whole
of <&'string2'&>, whichever is the shorter. Do not confuse &%length%& with
&%strlen%&, which gives the length of a string.
All measurement is done in bytes and is not UTF-8 aware.


.vitem "&*${listextract{*&<&'number'&>&*}&&&
        {*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&"
.cindex "expansion" "extracting list elements by number"
.cindex "&%listextract%&" "extract list elements by number"
.cindex "list" "extracting elements by number"
The <&'number'&> argument must consist entirely of decimal digits,
apart from an optional leading minus,
and leading and trailing white space (which is ignored).

.new
The <&'string1'&> argument, after any leading change-of-separator
(see &<<SECTlistsepchange>>&),
is expanded and the whole forms the list.
.wen
By default, the list separator is a colon.

The first field of the list is numbered one.
If the number is negative, the fields are
counted from the end of the list, with the rightmost one numbered -1.
The numbered element of the list is extracted and placed in &$value$&,
then <&'string2'&> is expanded as the result.

If the modulus of the
number is zero or greater than the number of fields in the string,
the result is the expansion of <&'string3'&>.

For example:
.code
${listextract{2}{x:42:99}}
.endd
yields &"42"&, and
.code
${listextract{-3}{<, x,42,99,& Mailer,,/bin/bash}{result: $value}}
.endd
yields &"result: 42"&.

If {<&'string3'&>} is omitted, an empty string is used for string3.
If {<&'string2'&>} is also omitted, the value that was
extracted is used.
You can use &`fail`& instead of {<&'string3'&>} as in a string extract.


.vitem &*${listquote{*&<&'separator'&>&*}{*&<&'string'&>&*}}*&
.cindex quoting "for list"
.cindex list quoting
This item doubles any occurrence of the separator character
in the given string.
An empty string is replaced with a single space.
This converts the string into a safe form for use as a list element,
in a list using the given separator.


.vitem "&*${lookup&~{*&<&'key'&>&*}&~*&<&'search&~type'&>&*&~&&&
        {*&<&'file'&>&*}&~{*&<&'string1'&>&*}&~{*&<&'string2'&>&*}}*&" &&&
       "&*${lookup&~*&<&'search&~type'&>&*&~{*&<&'query'&>&*}&~&&&
        {*&<&'string1'&>&*}&~{*&<&'string2'&>&*}}*&"
.cindex "expansion" "lookup in"
.cindex "file" "lookups"
.cindex "lookup" "in expanded string"
The two forms of lookup item specify data lookups in files and databases, as
discussed in chapter &<<CHAPfdlookup>>&. The first form is used for single-key
lookups, and the second is used for query-style lookups. The <&'key'&>,
<&'file'&>, and <&'query'&> strings are expanded before use.

If there is any white space in a lookup item which is part of a filter command,
a retry or rewrite rule, a routing rule for the &(manualroute)& router, or any
other place where white space is significant, the lookup item must be enclosed
in double quotes. The use of data lookups in users' filter files may be locked
out by the system administrator.

.vindex "&$value$&"
If the lookup succeeds, <&'string1'&> is expanded and replaces the entire item.
During its expansion, the variable &$value$& contains the data returned by the
lookup. Afterwards it reverts to the value it had previously (at the outer
level it is empty). If the lookup fails, <&'string2'&> is expanded and replaces
the entire item. If {<&'string2'&>} is omitted, the replacement is the empty
string on failure. If <&'string2'&> is provided, it can itself be a nested
lookup, thus providing a mechanism for looking up a default value when the
original lookup fails.

If a nested lookup is used as part of <&'string1'&>, &$value$& contains the
data for the outer lookup while the parameters of the second lookup are
expanded, and also while <&'string2'&> of the second lookup is expanded, should
the second lookup fail. Instead of {<&'string2'&>} the word &"fail"& can
appear, and in this case, if the lookup fails, the entire expansion is forced
to fail (see section &<<SECTforexpfai>>&). If both {<&'string1'&>} and
{<&'string2'&>} are omitted, the result is the looked up value in the case of a
successful lookup, and nothing in the case of failure.

For single-key lookups, the string &"partial"& is permitted to precede the
search type in order to do partial matching, and * or *@ may follow a search
type to request default lookups if the key does not match (see sections
&<<SECTdefaultvaluelookups>>& and &<<SECTpartiallookup>>& for details).

.cindex "numerical variables (&$1$& &$2$& etc)" "in lookup expansion"
If a partial search is used, the variables &$1$& and &$2$& contain the wild
and non-wild parts of the key during the expansion of the replacement text.
They return to their previous values at the end of the lookup item.

This example looks up the postmaster alias in the conventional alias file:
.code
${lookup {postmaster} lsearch {/etc/aliases} {$value}}
.endd
This example uses NIS+ to look up the full name of the user corresponding to
the local part of an address, forcing the expansion to fail if it is not found:
.code
${lookup nisplus {[name=$local_part],passwd.org_dir:gcos} \
  {$value}fail}
.endd


.vitem &*${map{*&<&'string1'&>&*}{*&<&'string2'&>&*}}*&
.cindex "expansion" "list creation"
.vindex "&$item$&"
.new
<&'string1'&> first has the part after any change-of-list-separator
(see &<<SECTlistsepchange>>&) expanded,
then the whole is taken as a list.
.wen
The default separator for the list is a colon.

For each item in this list,
its value is place in &$item$&, and then <&'string2'&> is
expanded and added to the output as an item in a new list. The separator used
for the output list is the same as the one used for the input, but a separator
setting is not included in the output. For example:
.code
${map{a:b:c}{[$item]}} ${map{<- x-y-z}{($item)}}
.endd
expands to &`[a]:[b]:[c] (x)-(y)-(z)`&. At the end of the expansion, the
value of &$item$& is restored to what it was before. See also the &%filter%&
and &%reduce%& expansion items.

.vitem &*${nhash{*&<&'string1'&>&*}{*&<&'string2'&>&*}{*&<&'string3'&>&*}}*&
.cindex "expansion" "numeric hash"
.cindex "hash function" "numeric"
The three strings are expanded; the first two must yield numbers. Call them
<&'n'&> and <&'m'&>. If you are using fixed values for these numbers, that is,
if <&'string1'&> and <&'string2'&> do not change when they are expanded, you
can use the simpler operator notation that avoids some of the braces:
.code
${nhash_<n>_<m>:<string>}
.endd
The second number is optional (in both notations). If there is only one number,
the result is a number in the range 0&--<&'n'&>-1. Otherwise, the string is
processed by a div/mod hash function that returns two numbers, separated by a
slash, in the ranges 0 to <&'n'&>-1 and 0 to <&'m'&>-1, respectively. For
example,
.code
${nhash{8}{64}{supercalifragilisticexpialidocious}}
.endd
returns the string &"6/33"&.



.vitem &*${perl{*&<&'subroutine'&>&*}{*&<&'arg'&>&*}{*&<&'arg'&>&*}...}*&
.cindex "Perl" "use in expanded string"
.cindex "expansion" "calling Perl from"
This item is available only if Exim has been built to include an embedded Perl
interpreter. The subroutine name and the arguments are first separately
expanded, and then the Perl subroutine is called with those arguments. No
additional arguments need be given; the maximum number permitted, including the
name of the subroutine, is nine.

The return value of the subroutine is inserted into the expanded string, unless
the return value is &%undef%&. In that case, the entire expansion is
forced to fail, in the same way as an explicit &"fail"& on a lookup item
does (see section &<<SECTforexpfai>>&).  Whatever you return is evaluated
in a scalar context, thus the return value is a scalar.  For example, if you
return a Perl vector, the return value is the size of the vector,
not its contents.

If the subroutine exits by calling Perl's &%die%& function, the expansion fails
with the error message that was passed to &%die%&. More details of the embedded
Perl facility are given in chapter &<<CHAPperl>>&.

The &(redirect)& router has an option called &%forbid_filter_perl%& which locks
out the use of this expansion item in filter files.


.vitem &*${prvs{*&<&'address'&>&*}{*&<&'secret'&>&*}{*&<&'keynumber'&>&*}}*&
.cindex "&%prvs%& expansion item"
The first argument is a complete email address and the second is secret
keystring. The third argument, specifying a key number, is optional. If absent,
it defaults to 0. The result of the expansion is a prvs-signed email address,
to be typically used with the &%return_path%& option on an &(smtp)& transport
as part of a bounce address tag validation (BATV) scheme. For more discussion
and an example, see section &<<SECTverifyPRVS>>&.

.vitem "&*${prvscheck{*&<&'address'&>&*}{*&<&'secret'&>&*}&&&
        {*&<&'string'&>&*}}*&"
.cindex "&%prvscheck%& expansion item"
This expansion item is the complement of the &%prvs%& item. It is used for
checking prvs-signed addresses. If the expansion of the first argument does not
yield a syntactically valid prvs-signed address, the whole item expands to the
empty string. When the first argument does expand to a syntactically valid
prvs-signed address, the second argument is expanded, with the prvs-decoded
version of the address and the key number extracted from the address in the
variables &$prvscheck_address$& and &$prvscheck_keynum$&, respectively.

These two variables can be used in the expansion of the second argument to
retrieve the secret. The validity of the prvs-signed address is then checked
against the secret. The result is stored in the variable &$prvscheck_result$&,
which is empty for failure or &"1"& for success.

The third argument is optional; if it is missing, it defaults to an empty
string. This argument is now expanded. If the result is an empty string, the
result of the expansion is the decoded version of the address. This is the case
whether or not the signature was valid. Otherwise, the result of the expansion
is the expansion of the third argument.

All three variables can be used in the expansion of the third argument.
However, once the expansion is complete, only &$prvscheck_result$& remains set.
For more discussion and an example, see section &<<SECTverifyPRVS>>&.

.vitem &*${readfile{*&<&'file&~name'&>&*}{*&<&'eol&~string'&>&*}}*&
.cindex "expansion" "inserting an entire file"
.cindex "file" "inserting into expansion"
.cindex "&%readfile%& expansion item"
The filename and end-of-line (eol) string are first expanded separately. The file is
then read, and its contents replace the entire item. All newline characters in
the file are replaced by the end-of-line string if it is present. Otherwise,
newlines are left in the string.
String expansion is not applied to the contents of the file. If you want this,
you must wrap the item in an &%expand%& operator. If the file cannot be read,
the string expansion fails.

The &(redirect)& router has an option called &%forbid_filter_readfile%& which
locks out the use of this expansion item in filter files.



.vitem "&*${readsocket{*&<&'name'&>&*}{*&<&'request'&>&*}&&&
        {*&<&'options'&>&*}{*&<&'eol&~string'&>&*}{*&<&'fail&~string'&>&*}}*&"
.cindex "expansion" "inserting from a socket"
.cindex "socket, use of in expansion"
.cindex "&%readsocket%& expansion item"
This item inserts data from a Unix domain or TCP socket into the expanded
string. The minimal way of using it uses just two arguments, as in these
examples:
.code
${readsocket{/socket/name}{request string}}
${readsocket{inet:some.host:1234}{request string}}
.endd
For a Unix domain socket, the first substring must be the path to the socket.
For an Internet socket, the first substring must contain &`inet:`& followed by
a host name or IP address, followed by a colon and a port, which can be a
number or the name of a TCP port in &_/etc/services_&. An IP address may
optionally be enclosed in square brackets. This is best for IPv6 addresses. For
example:
.code
${readsocket{inet:[::1]:1234}{request string}}
.endd
Only a single host name may be given, but if looking it up yields more than
one IP address, they are each tried in turn until a connection is made. For
both kinds of socket, Exim makes a connection, writes the request string
(unless it is an empty string; no terminating NUL is ever sent)
and reads from the socket until an end-of-file
is read. A timeout of 5 seconds is applied. Additional, optional arguments
extend what can be done. Firstly, you can vary the timeout. For example:
.code
${readsocket{/socket/name}{request string}{3s}}
.endd

The third argument is a list of options, of which the first element is the timeout
and must be present if any options are given.
Further elements are options of form &'name=value'&.
Example:
.code
${readsocket{/socket/name}{request string}{3s:shutdown=no}}
.endd

The following option names are recognised:
.ilist
&*cache*&
Defines if the result data can be cached for use by a later identical
request in the same process.
Values are &"yes"& or &"no"& (the default).
If not, all cached results for this connection specification
will be invalidated.

.next
&*shutdown*&
Defines whether or not a write-shutdown is done on the connection after
sending the request. Values are &"yes"& (the default) or &"no"&
(preferred, eg. by some webservers).

.next
&*sni*&
Controls the use of Server Name Identification on the connection.
Any nonempty value will be the SNI sent; TLS will be forced.

.next
&*tls*&
Controls the use of TLS on the connection.
Values are &"yes"& or &"no"& (the default).
If it is enabled, a shutdown as described above is never done.
.endlist


A fourth argument allows you to change any newlines that are in the data
that is read, in the same way as for &%readfile%& (see above). This example
turns them into spaces:
.code
${readsocket{inet:127.0.0.1:3294}{request string}{3s}{ }}
.endd
As with all expansions, the substrings are expanded before the processing
happens. Errors in these sub-expansions cause the expansion to fail. In
addition, the following errors can occur:

.ilist
Failure to create a socket file descriptor;
.next
Failure to connect the socket;
.next
Failure to write the request string;
.next
Timeout on reading from the socket.
.endlist

By default, any of these errors causes the expansion to fail. However, if
you supply a fifth substring, it is expanded and used when any of the above
errors occurs. For example:
.code
${readsocket{/socket/name}{request string}{3s}{\n}\
  {socket failure}}
.endd
You can test for the existence of a Unix domain socket by wrapping this
expansion in &`${if exists`&, but there is a race condition between that test
and the actual opening of the socket, so it is safer to use the fifth argument
if you want to be absolutely sure of avoiding an expansion error for a
non-existent Unix domain socket, or a failure to connect to an Internet socket.

The &(redirect)& router has an option called &%forbid_filter_readsocket%& which
locks out the use of this expansion item in filter files.


.vitem &*${reduce{*&<&'string1'&>}{<&'string2'&>&*}{*&<&'string3'&>&*}}*&
.cindex "expansion" "reducing a list to a scalar"
.cindex "list" "reducing to a scalar"
.vindex "&$value$&"
.vindex "&$item$&"
This operation reduces a list to a single, scalar string.

.new
<&'string1'&> first has the part after any change-of-list-separator
(see &<<SECTlistsepchange>>&) expanded,
then the whole is taken as a list.
.wen
The default separator for the list is a colon.

Then <&'string2'&> is expanded and
assigned to the &$value$& variable. After this, each item in the <&'string1'&>
list is assigned to &$item$&, in turn, and <&'string3'&> is expanded for each of
them. The result of that expansion is assigned to &$value$& before the next
iteration. When the end of the list is reached, the final value of &$value$& is
added to the expansion output. The &%reduce%& expansion item can be used in a
number of ways. For example, to add up a list of numbers:
.code
${reduce {<, 1,2,3}{0}{${eval:$value+$item}}}
.endd
The result of that expansion would be &`6`&. The maximum of a list of numbers
can be found:
.code
${reduce {3:0:9:4:6}{0}{${if >{$item}{$value}{$item}{$value}}}}
.endd
At the end of a &*reduce*& expansion, the values of &$item$& and &$value$& are
restored to what they were before. See also the &%filter%& and &%map%&
expansion items.

. A bit of a special-case logic error in writing an expansion;
. probably not worth including in the mainline of documentation.
. If only we had footnotes (the html output variant is the problem).
.
. .new
. &*Note*&: if an &'expansion condition'& is used in <&'string3'&>
. and that condition modifies &$value$&,
. then the string expansions dependent on the condition cannot use
. the &$value$& of the reduce iteration.
. .wen

.vitem &*$rheader_*&<&'header&~name'&>&*:*&&~or&~&*$rh_*&<&'header&~name'&>&*:*&
This item inserts &"raw"& header lines. It is described with the &%header%&
expansion item in section &<<SECTexpansionitems>>& above.

.vitem "&*${run<&'options'&> {*&<&'command&~string'&>&*}{*&<&'string1'&>&*}&&&
        {*&<&'string2'&>&*}}*&"
.cindex "expansion" "running a command"
.cindex "&%run%& expansion item"
This item runs an external command, as a subprocess.
One option is supported after the word &'run'&, comma-separated
and without whitespace.

If the option &'preexpand'& is not used,
the command string before expansion is split into individual arguments by spaces
and then each argument is separately expanded.
Then the command is run
in a separate process, but under the same uid and gid.  As in other command
executions from Exim, a shell is not used by default. If the command requires
a shell, you must explicitly code it.
The command name may not be tainted, but the remaining arguments can be.

&*Note*&: if tainted arguments are used, they are supplied by a
potential attacker;
a careful assessment for security vulnerabilities should be done.

If the option &'preexpand'& is used,
the command string is first expanded as a whole.
The expansion result is split apart into individual arguments by spaces,
and then the command is run as above.
Since the arguments are split by spaces, when there is a variable expansion
which has an empty result, it will cause the situation that the argument will
simply be omitted when the program is actually executed by Exim. If the
script/program requires a specific number of arguments and the expanded
variable could possibly result in this empty expansion, the variable must be
quoted. This is more difficult if the expanded variable itself could result
in a string containing quotes, because it would interfere with the quotes
around the command arguments. A possible guard against this is to wrap the
variable in the &%sg%& operator to change any quote marks to some other
character.
Neither the command nor any argument may be tainted.

The standard input for the command exists, but is empty. The standard output
and standard error are set to the same file descriptor.
.cindex "return code" "from &%run%& expansion"
.vindex "&$value$&"
If the command succeeds (gives a zero return code) <&'string1'&> is expanded
and replaces the entire item; during this expansion, the standard output/error
from the command is in the variable &$value$&. If the command fails,
<&'string2'&>, if present, is expanded and used. Once again, during the
expansion, the standard output/error from the command is in the variable
&$value$&.

If <&'string2'&> is absent, the result is empty. Alternatively, <&'string2'&>
can be the word &"fail"& (not in braces) to force expansion failure if the
command does not succeed. If both strings are omitted, the result is contents
of the standard output/error on success, and nothing on failure.

.vindex "&$run_in_acl$&"
The standard output/error of the command is put in the variable &$value$&.
In this ACL example, the output of a command is logged for the admin to
troubleshoot:
.code
warn  condition    = ${run{/usr/bin/id}{yes}{no}}
      log_message  = Output of id: $value
.endd
If the command requires shell idioms, such as the > redirect operator, the
shell must be invoked directly, such as with:
.code
${run{/bin/bash -c "/usr/bin/id >/tmp/id"}{yes}{yes}}
.endd
Note that &$value$& will not persist beyond the reception of a single message.

.vindex "&$runrc$&"
The return code from the command is put in the variable &$runrc$&, and this
remains set afterwards, so in a filter file you can do things like this:
.code
if "${run{x y z}{}}$runrc" is 1 then ...
  elif $runrc is 2 then ...
  ...
endif
.endd
If execution of the command fails (for example, the command does not exist),
the return code is 127 &-- the same code that shells use for non-existent
commands.

&*Warning*&: In a router or transport, you cannot assume the order in which
option values are expanded, except for those preconditions whose order of
testing is documented. Therefore, you cannot reliably expect to set &$runrc$&
by the expansion of one option, and use it in another.

The &(redirect)& router has an option called &%forbid_filter_run%& which locks
out the use of this expansion item in filter files.


.vitem &*${sg{*&<&'subject'&>&*}{*&<&'regex'&>&*}{*&<&'replacement'&>&*}}*&
.cindex "expansion" "string substitution"
.cindex "&%sg%& expansion item"
This item works like Perl's substitution operator (s) with the global (/g)
option; hence its name. However, unlike the Perl equivalent, Exim does not
modify the subject string; instead it returns the modified string for insertion
into the overall expansion. The item takes three arguments: the subject string,
a regular expression, and a substitution string. For example:
.code
${sg{abcdefabcdef}{abc}{xyz}}
.endd
yields &"xyzdefxyzdef"&. Because all three arguments are expanded before use,
if any $, } or \ characters are required in the regular expression or in the
substitution string, they have to be escaped. For example:
.code
${sg{abcdef}{^(...)(...)\$}{\$2\$1}}
.endd
yields &"defabc"&, and
.code
${sg{1=A 4=D 3=C}{\N(\d+)=\N}{K\$1=}}
.endd
yields &"K1=A K4=D K3=C"&. Note the use of &`\N`& to protect the contents of
the regular expression from string expansion.

The regular expression is compiled in 8-bit mode, working against bytes
rather than any Unicode-aware character handling.


.vitem &*${sort{*&<&'string'&>&*}{*&<&'comparator'&>&*}{*&<&'extractor'&>&*}}*&
.cindex sorting "a list"
.cindex list sorting
.cindex expansion "list sorting"
.new
<&'string'&> first has the part after any change-of-list-separator
(see &<<SECTlistsepchange>>&) expanded,
then the whole is taken as a list.
.wen
The default separator for the list is a colon.

The <&'comparator'&> argument is interpreted as the operator
of a two-argument expansion condition.
The numeric operators plus ge, gt, le, lt (and ~i variants) are supported.
The comparison should return true when applied to two values
if the first value should sort before the second value.
The <&'extractor'&> expansion is applied repeatedly to elements of the list,
the element being placed in &$item$&,
to give values for comparison.

The item result is a sorted list,
with the original list separator,
of the list elements (in full) of the original.

Examples:
.code
${sort{3:2:1:4}{<}{$item}}
.endd
sorts a list of numbers, and
.code
${sort {${lookup dnsdb{>:,,mx=example.com}}} {<} {${listextract{1}{<,$item}}}}
.endd
will sort an MX lookup into priority order.



.vitem &*${srs_encode&~{*&<&'secret'&>&*}{*&<&'return&~path'&>&*}{*&<&'original&~domain'&>&*}}*&
SRS encoding.  See section &<<SECTSRS>>& for details.



.vitem &*${substr{*&<&'start'&>&*}{*&<&'len'&>&*}{*&<&'subject'&>&*}}*&
.cindex "&%substr%& expansion item"
.cindex "substring extraction"
.cindex "expansion" "substring extraction"
The three strings are expanded; the first two must yield numbers. Call them
<&'n'&> and <&'m'&>. If you are using fixed values for these numbers, that is,
if <&'start'&> and <&'len'&> do not change when they are expanded, you
can use the simpler operator notation that avoids some of the braces:
.code
${substr_<n>_<m>:<subject>}
.endd
The second number is optional (in both notations).
If it is absent in the simpler format, the preceding underscore must also be
omitted.

The &%substr%& item can be used to extract more general substrings than
&%length%&. The first number, <&'n'&>, is a starting offset, and <&'m'&> is the
length required. For example
.code
${substr{3}{2}{$local_part}}
.endd
If the starting offset is greater than the string length the result is the
null string; if the length plus starting offset is greater than the string
length, the result is the right-hand part of the string, starting from the
given offset. The first byte (character) in the string has offset zero.

The &%substr%& expansion item can take negative offset values to count
from the right-hand end of its operand. The last byte (character) is offset -1,
the second-last is offset -2, and so on. Thus, for example,
.code
${substr{-5}{2}{1234567}}
.endd
yields &"34"&. If the absolute value of a negative offset is greater than the
length of the string, the substring starts at the beginning of the string, and
the length is reduced by the amount of overshoot. Thus, for example,
.code
${substr{-5}{2}{12}}
.endd
yields an empty string, but
.code
${substr{-3}{2}{12}}
.endd
yields &"1"&.

When the second number is omitted from &%substr%&, the remainder of the string
is taken if the offset is positive. If it is negative, all bytes (characters) in the
string preceding the offset point are taken. For example, an offset of -1 and
no length, as in these semantically identical examples:
.code
${substr_-1:abcde}
${substr{-1}{abcde}}
.endd
yields all but the last character of the string, that is, &"abcd"&.

All measurement is done in bytes and is not UTF-8 aware.



.vitem "&*${tr{*&<&'subject'&>&*}{*&<&'characters'&>&*}&&&
        {*&<&'replacements'&>&*}}*&"
.cindex "expansion" "character translation"
.cindex "&%tr%& expansion item"
This item does single-character (in bytes) translation on its subject string. The second
argument is a list of characters to be translated in the subject string. Each
matching character is replaced by the corresponding character from the
replacement list. For example
.code
${tr{abcdea}{ac}{13}}
.endd
yields &`1b3de1`&. If there are duplicates in the second character string, the
last occurrence is used. If the third string is shorter than the second, its
last character is replicated. However, if it is empty, no translation takes
place.

All character handling is done in bytes and is not UTF-8 aware.

.endlist



.section "Expansion operators" "SECTexpop"
.cindex "expansion" "operators"
For expansion items that perform transformations on a single argument string,
the &"operator"& notation is used because it is simpler and uses fewer braces.
The substring is first expanded before the operation is applied to it. The
following operations can be performed:

.vlist
.vitem &*${address:*&<&'string'&>&*}*&
.cindex "expansion" "RFC 2822 address handling"
.cindex "&%address%& expansion item"
The string is interpreted as an
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address, as it might appear in a
header line, and the effective address is extracted from it. If the string does
not parse successfully, the result is empty.

The parsing correctly handles SMTPUTF8 Unicode in the string.


.vitem &*${addresses:*&<&'string'&>&*}*&
.cindex "expansion" "RFC 2822 address handling"
.cindex "&%addresses%& expansion item"
The string (after expansion) is interpreted as a list of addresses in
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
format, such as can be found in a &'To:'& or &'Cc:'& header line. The
operative address (&'local-part@domain'&) is extracted from each item, and the
result of the expansion is a colon-separated list, with appropriate
doubling of colons should any happen to be present in the email addresses.
Syntactically invalid RFC2822 address items are omitted from the output.

It is possible to specify a character other than colon for the output
separator by starting the string with > followed by the new separator
character. For example:
.code
${addresses:>& Chief <ceo@up.stairs>, sec@base.ment (dogsbody)}
.endd
expands to &`ceo@up.stairs&&sec@base.ment`&. The string is expanded
first, so if the expanded string starts with >, it may change the output
separator unintentionally. This can be avoided by setting the output
separator explicitly:
.code
${addresses:>:$h_from:}
.endd

Compare the &%address%& (singular)
expansion item, which extracts the working address from a single RFC2822
address. See the &%filter%&, &%map%&, and &%reduce%& items for ways of
processing lists.

To clarify "list of addresses in RFC 2822 format" mentioned above, Exim follows
a strict interpretation of header line formatting.  Exim parses the bare,
unquoted portion of an email address and if it finds a comma, treats it as an
email address separator. For the example header line:
.code
From: =?iso-8859-2?Q?Last=2C_First?= <user@example.com>
.endd
The first example below demonstrates that Q-encoded email addresses are parsed
properly if it is given the raw header (in this example, &`$rheader_from:`&).
It does not see the comma because it's still encoded as "=2C".  The second
example below is passed the contents of &`$header_from:`&, meaning it gets
de-mimed. Exim sees the decoded "," so it treats it as &*two*& email addresses.
The third example shows that the presence of a comma is skipped when it is
quoted.  The fourth example shows SMTPUTF8 handling.
.code
# exim -be '${addresses:From: \
=?iso-8859-2?Q?Last=2C_First?= <user@example.com>}'
user@example.com
# exim -be '${addresses:From: Last, First <user@example.com>}'
Last:user@example.com
# exim -be '${addresses:From: "Last, First" <user@example.com>}'
user@example.com
# exim -be '${addresses:フィル <フィリップ@example.jp>}'
フィリップ@example.jp
.endd

.vitem &*${base32:*&<&'digits'&>&*}*&
.cindex "&%base32%& expansion item"
.cindex "expansion" "conversion to base 32"
The string must consist entirely of decimal digits. The number is converted to
base 32 and output as a (empty, for zero) string of characters.
Only lowercase letters are used.

.vitem &*${base32d:*&<&'base-32&~digits'&>&*}*&
.cindex "&%base32d%& expansion item"
.cindex "expansion" "conversion to base 32"
The string must consist entirely of base-32 digits.
The number is converted to decimal and output as a string.

.vitem &*${base62:*&<&'digits'&>&*}*&
.cindex "&%base62%& expansion item"
.cindex "expansion" "conversion to base 62"
The string must consist entirely of decimal digits. The number is converted to
base 62 and output as a string of six characters, including leading zeros. In
the few operating environments where Exim uses base 36 instead of base 62 for
its message identifiers (because those systems do not have case-sensitive
filenames), base 36 is used by this operator, despite its name. &*Note*&: Just
to be absolutely clear: this is &'not'& base64 encoding.

.vitem &*${base62d:*&<&'base-62&~digits'&>&*}*&
.cindex "&%base62d%& expansion item"
.cindex "expansion" "conversion to base 62"
The string must consist entirely of base-62 digits, or, in operating
environments where Exim uses base 36 instead of base 62 for its message
identifiers, base-36 digits. The number is converted to decimal and output as a
string.

.vitem &*${base64:*&<&'string'&>&*}*&
.cindex "expansion" "base64 encoding"
.cindex "base64 encoding" "in string expansion"
.cindex "&%base64%& expansion item"
.cindex certificate "base64 of DER"
This operator converts a string into one that is base64 encoded.

If the string is a single variable of type certificate,
returns the base64 encoding of the DER form of the certificate.


.vitem &*${base64d:*&<&'string'&>&*}*&
.cindex "expansion" "base64 decoding"
.cindex "base64 decoding" "in string expansion"
.cindex "&%base64d%& expansion item"
This operator converts a base64-encoded string into the un-coded form.


.vitem &*${domain:*&<&'string'&>&*}*&
.cindex "domain" "extraction"
.cindex "expansion" "domain extraction"
The string is interpreted as an
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address and the domain is extracted from it.
If the string does not parse successfully, the result is empty.


.vitem &*${escape:*&<&'string'&>&*}*&
.cindex "expansion" "escaping non-printing characters"
.cindex "&%escape%& expansion item"
If the string contains any non-printing characters, they are converted to
escape sequences starting with a backslash. Whether characters with the most
significant bit set (so-called &"8-bit characters"&) count as printing or not
is controlled by the &%print_topbitchars%& option.

.vitem &*${escape8bit:*&<&'string'&>&*}*&
.cindex "expansion" "escaping 8-bit characters"
.cindex "&%escape8bit%& expansion item"
If the string contains any characters with the most significant bit set,
they are converted to escape sequences starting with a backslash.
Backslashes and DEL characters are also converted.


.vitem &*${eval:*&<&'string'&>&*}*&&~and&~&*${eval10:*&<&'string'&>&*}*&
.cindex "expansion" "expression evaluation"
.cindex "expansion" "arithmetic expression"
.cindex "&%eval%& expansion item"
These items supports simple arithmetic and bitwise logical operations in
expansion strings. The string (after expansion) must be a conventional
arithmetic expression, but it is limited to basic arithmetic operators, bitwise
logical operators, and parentheses. All operations are carried out using
integer arithmetic. The operator priorities are as follows (the same as in the
C programming language):
.table2 70pt 300pt
.irow &'highest:'& "not (~), negate (-)"
.irow ""   "multiply (*), divide (/), remainder (%)"
.irow ""   "plus (+), minus (-)"
.irow ""   "shift-left (<<), shift-right (>>)"
.irow ""   "and (&&)"
.irow ""   "xor (^)"
.irow &'lowest:'&  "or (|)"
.endtable
Binary operators with the same priority are evaluated from left to right. White
space is permitted before or after operators.

For &%eval%&, numbers may be decimal, octal (starting with &"0"&) or
hexadecimal (starting with &"0x"&). For &%eval10%&, all numbers are taken as
decimal, even if they start with a leading zero; hexadecimal numbers are not
permitted. This can be useful when processing numbers extracted from dates or
times, which often do have leading zeros.

A number may be followed by &"K"&, &"M"& or &"G"& to multiply it by 1024, 1024*1024
or 1024*1024*1024,
respectively. Negative numbers are supported. The result of the computation is
a decimal representation of the answer (without &"K"&, &"M"& or &"G"&). For example:

.display
&`${eval:1+1}            `&  yields 2
&`${eval:1+2*3}          `&  yields 7
&`${eval:(1+2)*3}        `&  yields 9
&`${eval:2+42%5}         `&  yields 4
&`${eval:0xc&amp;5}          `&  yields 4
&`${eval:0xc|5}          `&  yields 13
&`${eval:0xc^5}          `&  yields 9
&`${eval:0xc>>1}         `&  yields 6
&`${eval:0xc<<1}         `&  yields 24
&`${eval:~255&amp;0x1234}    `&  yields 4608
&`${eval:-(~255&amp;0x1234)} `&  yields -4608
.endd

As a more realistic example, in an ACL you might have
.code
deny   condition =                    \
         ${if and {                   \
           {>{$rcpt_count}{10}}       \
           {                          \
           <                          \
             {$recipients_count}      \
             {${eval:$rcpt_count/2}}  \
           }                          \
         }{yes}{no}}
       message = Too many bad recipients
.endd
The condition is true if there have been more than 10 RCPT commands and
fewer than half of them have resulted in a valid recipient.


.vitem &*${expand:*&<&'string'&>&*}*&
.cindex "expansion" "re-expansion of substring"
The &%expand%& operator causes a string to be expanded for a second time. For
example,
.code
${expand:${lookup{$domain}dbm{/some/file}{$value}}}
.endd
first looks up a string in a file while expanding the operand for &%expand%&,
and then re-expands what it has found.


.vitem &*${from_utf8:*&<&'string'&>&*}*&
.cindex "Unicode"
.cindex "UTF-8" "conversion from"
.cindex "expansion" "UTF-8 conversion"
.cindex "&%from_utf8%& expansion item"
The world is slowly moving towards Unicode, although there are no standards for
email yet. However, other applications (including some databases) are starting
to store data in Unicode, using UTF-8 encoding. This operator converts from a
UTF-8 string to an ISO-8859-1 string. UTF-8 code values greater than 255 are
converted to underscores. The input must be a valid UTF-8 string. If it is not,
the result is an undefined sequence of bytes.

Unicode code points with values less than 256 are compatible with ASCII and
ISO-8859-1 (also known as Latin-1).
For example, character 169 is the copyright symbol in both cases, though the
way it is encoded is different. In UTF-8, more than one byte is needed for
characters with code values greater than 127, whereas ISO-8859-1 is a
single-byte encoding (but thereby limited to 256 characters). This makes
translation from UTF-8 to ISO-8859-1 straightforward.


.vitem &*${hash_*&<&'n'&>&*_*&<&'m'&>&*:*&<&'string'&>&*}*&
.cindex "hash function" "textual"
.cindex "expansion" "textual hash"
The &%hash%& operator is a simpler interface to the hashing function that can
be used when the two parameters are fixed numbers (as opposed to strings that
change when expanded). The effect is the same as
.code
${hash{<n>}{<m>}{<string>}}
.endd
See the description of the general &%hash%& item above for details. The
abbreviation &%h%& can be used when &%hash%& is used as an operator.



.vitem &*${headerwrap_*&<&'cols'&>&*_*&<&'limit'&>&*:*&<&'string'&>&*}*&
.cindex header "wrapping operator"
.cindex expansion "header wrapping"
This operator line-wraps its argument in a way useful for headers.
The &'cols'& value gives the column number to wrap after,
the &'limit'& gives a limit number of result characters to truncate at.
Either just the &'limit'& and the preceding underbar, or both, can be omitted;
the defaults are 80 and 998.
Wrapping will be inserted at a space if possible before the
column number is reached.
Whitespace at a chosen wrap point is removed.
A line-wrap consists of a newline followed by a tab,
and the tab is counted as 8 columns.



.vitem &*${hex2b64:*&<&'hexstring'&>&*}*&
.cindex "base64 encoding" "conversion from hex"
.cindex "expansion" "hex to base64"
.cindex "&%hex2b64%& expansion item"
This operator converts a hex string into one that is base64 encoded. This can
be useful for processing the output of the various hashing functions.



.vitem &*${hexquote:*&<&'string'&>&*}*&
.cindex "quoting" "hex-encoded unprintable characters"
.cindex "&%hexquote%& expansion item"
This operator converts non-printable characters in a string into a hex
escape form. Byte values between 33 (!) and 126 (~) inclusive are left
as is, and other byte values are converted to &`\xNN`&, for example, a
byte value 127 is converted to &`\x7f`&.


.vitem &*${ipv6denorm:*&<&'string'&>&*}*&
.cindex "&%ipv6denorm%& expansion item"
.cindex "IP address" normalisation
This expands an IPv6 address to a full eight-element colon-separated set
of hex digits including leading zeroes.
A trailing ipv4-style dotted-decimal set is converted to hex.
Pure IPv4 addresses are converted to IPv4-mapped IPv6.

.vitem &*${ipv6norm:*&<&'string'&>&*}*&
.cindex "&%ipv6norm%& expansion item"
.cindex "IP address" normalisation
.cindex "IP address" "canonical form"
This converts an IPv6 address to canonical form.
Leading zeroes of groups are omitted, and the longest
set of zero-valued groups is replaced with a double colon.
A trailing ipv4-style dotted-decimal set is converted to hex.
Pure IPv4 addresses are converted to IPv4-mapped IPv6.


.vitem &*${lc:*&<&'string'&>&*}*&
.cindex "case forcing in strings"
.cindex "string" "case forcing"
.cindex "lower casing"
.cindex "expansion" "case forcing"
.cindex "&%lc%& expansion item"
This forces the letters in the string into lower-case, for example:
.code
${lc:$local_part}
.endd
Case is defined per the system C locale.

.vitem &*${length_*&<&'number'&>&*:*&<&'string'&>&*}*&
.cindex "expansion" "string truncation"
.cindex "&%length%& expansion item"
The &%length%& operator is a simpler interface to the &%length%& function that
can be used when the parameter is a fixed number (as opposed to a string that
changes when expanded). The effect is the same as
.code
${length{<number>}{<string>}}
.endd
See the description of the general &%length%& item above for details. Note that
&%length%& is not the same as &%strlen%&. The abbreviation &%l%& can be used
when &%length%& is used as an operator.
All measurement is done in bytes and is not UTF-8 aware.


.vitem &*${listcount:*&<&'string'&>&*}*&
.cindex "expansion" "list item count"
.cindex "list" "item count"
.cindex "list" "count of items"
.cindex "&%listcount%& expansion item"
The part of the string after any leading change-of-separator is expanded,
then the whole is interpreted as a list and the number of items is returned.


.vitem &*${listnamed:*&<&'name'&>&*}*&&~and&~&*${listnamed_*&<&'type'&>&*:*&<&'name'&>&*}*&
.cindex "expansion" "named list"
.cindex "&%listnamed%& expansion item"
The name is interpreted as a named list and the content of the list is returned,
expanding any referenced lists, re-quoting as needed for colon-separation.
If the optional type is given it must be one of "a", "d", "h" or "l"
and selects address-, domain-, host- or localpart- lists to search among respectively.
Otherwise all types are searched in an undefined order and the first
matching list is returned.
&*Note*&: Neither string-expansion of lists referenced by named-list syntax elements,
nor expansion of lookup elements, is done by the &%listnamed%& operator.


.vitem &*${local_part:*&<&'string'&>&*}*&
.cindex "expansion" "local part extraction"
.cindex "&%local_part%& expansion item"
The string is interpreted as an
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address and the local part is extracted from it.
If the string does not parse successfully, the result is empty.
The parsing correctly handles SMTPUTF8 Unicode in the string.


.vitem &*${mask:*&<&'IP&~address'&>&*/*&<&'bit&~count'&>&*}*& &&&
       &*${mask_n:*&<&'IP&~address'&>&*/*&<&'bit&~count'&>&*}*&
.cindex "masked IP address"
.cindex "IP address" "masking"
.cindex "CIDR notation"
.cindex "expansion" "IP address masking"
.cindex "&%mask%& expansion item"
If the form of the string to be operated on is not an IP address followed by a
slash and an integer (that is, a network address in CIDR notation), the
expansion fails. Otherwise, this operator converts the IP address to binary,
masks off the least significant bits according to the bit count, and converts
the result back to text, with mask appended. For example,
.code
${mask:10.111.131.206/28}
.endd
returns the string &"10.111.131.192/28"&.

Since this operation is expected to
be mostly used for looking up masked addresses in files, the
normal
result for an IPv6
address uses dots to separate components instead of colons, because colon
terminates a key string in lsearch files. So, for example,
.code
${mask:3ffe:ffff:836f:0a00:000a:0800:200a:c031/99}
.endd
returns the string
.code
3ffe.ffff.836f.0a00.000a.0800.2000.0000/99
.endd
If the optional form &*mask_n*& is used, IPv6 address result are instead
returned in normailsed form, using colons and with zero-compression.
Letters in IPv6 addresses are always output in lower case.


.vitem &*${md5:*&<&'string'&>&*}*&
.cindex "MD5 hash"
.cindex "expansion" "MD5 hash"
.cindex certificate fingerprint
.cindex "&%md5%& expansion item"
The &%md5%& operator computes the MD5 hash value of the string, and returns it
as a 32-digit hexadecimal number, in which any letters are in lower case.

If the string is a single variable of type certificate,
returns the MD5 hash fingerprint of the certificate.


.vitem &*${nhash_*&<&'n'&>&*_*&<&'m'&>&*:*&<&'string'&>&*}*&
.cindex "expansion" "numeric hash"
.cindex "hash function" "numeric"
The &%nhash%& operator is a simpler interface to the numeric hashing function
that can be used when the two parameters are fixed numbers (as opposed to
strings that change when expanded). The effect is the same as
.code
${nhash{<n>}{<m>}{<string>}}
.endd
See the description of the general &%nhash%& item above for details.


.vitem &*${quote:*&<&'string'&>&*}*&
.cindex "quoting" "in string expansions"
.cindex "expansion" "quoting"
.cindex "&%quote%& expansion item"
The &%quote%& operator puts its argument into double quotes if it
is an empty string or
contains anything other than letters, digits, underscores, dots, and hyphens.
Any occurrences of double quotes and backslashes are escaped with a backslash.
Newlines and carriage returns are converted to &`\n`& and &`\r`&,
respectively For example,
.code
${quote:ab"*"cd}
.endd
becomes
.code
"ab\"*\"cd"
.endd
The place where this is useful is when the argument is a substitution from a
variable or a message header.

.vitem &*${quote_local_part:*&<&'string'&>&*}*&
.cindex "&%quote_local_part%& expansion item"
This operator is like &%quote%&, except that it quotes the string only if
required to do so by the rules of
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
for quoting local parts. For example,
a plus sign would not cause quoting (but it would for &%quote%&).
If you are creating a new email address from the contents of &$local_part$&
(or any other unknown data), you should always use this operator.

This quoting determination is not SMTPUTF8-aware, thus quoting non-ASCII data
will likely use the quoting form.
Thus &'${quote_local_part:フィル}'& will always become &'"フィル"'&.


.vitem &*${quote_*&<&'lookup-type'&>&*:*&<&'string'&>&*}*&
.cindex "quoting" "lookup-specific"
This operator applies lookup-specific quoting rules to the string. Each
query-style lookup type has its own quoting rules which are described with
the lookups in chapter &<<CHAPfdlookup>>&. For example,
.code
${quote_ldap:two * two}
.endd
returns
.code
two%20%5C2A%20two
.endd
For single-key lookup types, no quoting is ever necessary and this operator
yields an unchanged string.


.vitem &*${randint:*&<&'n'&>&*}*&
.cindex "random number"
This operator returns a somewhat random number which is less than the
supplied number and is at least 0.  The quality of this randomness depends
on how Exim was built; the values are not suitable for keying material.
If Exim is linked against OpenSSL then RAND_pseudo_bytes() is used.
If Exim is linked against GnuTLS then gnutls_rnd(GNUTLS_RND_NONCE) is used,
for versions of GnuTLS with that function.
Otherwise, the implementation may be arc4random(), random() seeded by
srandomdev() or srandom(), or a custom implementation even weaker than
random().


.vitem &*${reverse_ip:*&<&'ipaddr'&>&*}*&
.cindex "expansion" "IP address"
This operator reverses an IP address; for IPv4 addresses, the result is in
dotted-quad decimal form, while for IPv6 addresses the result is in
dotted-nibble hexadecimal form.  In both cases, this is the "natural" form
for DNS.  For example,
.code
${reverse_ip:192.0.2.4}
${reverse_ip:2001:0db8:c42:9:1:abcd:192.0.2.127}
.endd
returns
.code
4.2.0.192
f.7.2.0.0.0.0.c.d.c.b.a.1.0.0.0.9.0.0.0.2.4.c.0.8.b.d.0.1.0.0.2
.endd


.vitem &*${rfc2047:*&<&'string'&>&*}*&
.cindex "expansion" "RFC 2047"
.cindex "RFC 2047" "expansion operator"
.cindex "&%rfc2047%& expansion item"
This operator encodes text according to the rules of
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047). This is an
encoding that is used in header lines to encode non-ASCII characters. It is
assumed that the input string is in the encoding specified by the
&%headers_charset%& option, which gets its default at build time. If the string
contains only characters in the range 33&--126, and no instances of the
characters
.code
? = ( ) < > @ , ; : \ " . [ ] _
.endd
it is not modified. Otherwise, the result is the RFC 2047 encoding of the
string, using as many &"encoded words"& as necessary to encode all the
characters.


.vitem &*${rfc2047d:*&<&'string'&>&*}*&
.cindex "expansion" "RFC 2047"
.cindex "RFC 2047" "decoding"
.cindex "&%rfc2047d%& expansion item"
This operator decodes strings that are encoded as per
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047).
Binary zero
bytes are replaced by question marks. Characters are converted into the
character set defined by &%headers_charset%&. Overlong RFC 2047 &"words"& are
not recognized unless &%check_rfc2047_length%& is set false.

&*Note*&: If you use &%$header%&_&'xxx'&&*:*& (or &%$h%&_&'xxx'&&*:*&) to
access a header line, RFC 2047 decoding is done automatically. You do not need
to use this operator as well.



.vitem &*${rxquote:*&<&'string'&>&*}*&
.cindex "quoting" "in regular expressions"
.cindex "regular expressions" "quoting"
.cindex "&%rxquote%& expansion item"
The &%rxquote%& operator inserts a backslash before any non-alphanumeric
characters in its argument. This is useful when substituting the values of
variables or headers inside regular expressions.


.vitem &*${sha1:*&<&'string'&>&*}*&
.cindex "SHA-1 hash"
.cindex "expansion" "SHA-1 hashing"
.cindex certificate fingerprint
.cindex "&%sha1%& expansion item"
The &%sha1%& operator computes the SHA-1 hash value of the string, and returns
it as a 40-digit hexadecimal number, in which any letters are in upper case.

If the string is a single variable of type certificate,
returns the SHA-1 hash fingerprint of the certificate.


.vitem &*${sha256:*&<&'string'&>&*}*& &&&
       &*${sha2:*&<&'string'&>&*}*& &&&
       &*${sha2_<n>:*&<&'string'&>&*}*&
.cindex "SHA-256 hash"
.cindex "SHA-2 hash"
.cindex certificate fingerprint
.cindex "expansion" "SHA-256 hashing"
.cindex "&%sha256%& expansion item"
.cindex "&%sha2%& expansion item"
The &%sha256%& operator computes the SHA-256 hash value of the string
and returns
it as a 64-digit hexadecimal number, in which any letters are in upper case.

If the string is a single variable of type certificate,
returns the SHA-256 hash fingerprint of the certificate.

The operator can also be spelled &%sha2%& and does the same as &%sha256%&
(except for certificates, which are not supported).
Finally, if an underbar
and a number is appended it specifies the output length, selecting a
member of the SHA-2 family of hash functions.
Values of 256, 384 and 512 are accepted, with 256 being the default.


.vitem &*${sha3:*&<&'string'&>&*}*& &&&
       &*${sha3_<n>:*&<&'string'&>&*}*&
.cindex "SHA3 hash"
.cindex "expansion" "SHA3 hashing"
.cindex "&%sha3%& expansion item"
The &%sha3%& operator computes the SHA3-256 hash value of the string
and returns
it as a 64-digit hexadecimal number, in which any letters are in upper case.

If a number is appended, separated by an underbar, it specifies
the output length.  Values of 224, 256, 384 and 512 are accepted;
with 256 being the default.

The &%sha3%& expansion item is only supported if Exim has been
compiled with GnuTLS 3.5.0 or later,
or OpenSSL 1.1.1 or later.
The macro "_CRYPTO_HASH_SHA3" will be defined if it is supported.


.vitem &*${stat:*&<&'string'&>&*}*&
.cindex "expansion" "statting a file"
.cindex "file" "extracting characteristics"
.cindex "&%stat%& expansion item"
The string, after expansion, must be a file path. A call to the &[stat()]&
function is made for this path. If &[stat()]& fails, an error occurs and the
expansion fails. If it succeeds, the data from the stat replaces the item, as a
series of <&'name'&>=<&'value'&> pairs, where the values are all numerical,
except for the value of &"smode"&. The names are: &"mode"& (giving the mode as
a 4-digit octal number), &"smode"& (giving the mode in symbolic format as a
10-character string, as for the &'ls'& command), &"inode"&, &"device"&,
&"links"&, &"uid"&, &"gid"&, &"size"&, &"atime"&, &"mtime"&, and &"ctime"&. You
can extract individual fields using the &%extract%& expansion item.

The use of the &%stat%& expansion in users' filter files can be locked out by
the system administrator. &*Warning*&: The file size may be incorrect on 32-bit
systems for files larger than 2GB.

.vitem &*${str2b64:*&<&'string'&>&*}*&
.cindex "&%str2b64%& expansion item"
Now deprecated, a synonym for the &%base64%& expansion operator.



.vitem &*${strlen:*&<&'string'&>&*}*&
.cindex "expansion" "string length"
.cindex "string" "length in expansion"
.cindex "&%strlen%& expansion item"
The item is replaced by the length of the expanded string, expressed as a
decimal number. &*Note*&: Do not confuse &%strlen%& with &%length%&.
All measurement is done in bytes and is not UTF-8 aware.


.vitem &*${substr_*&<&'start'&>&*_*&<&'length'&>&*:*&<&'string'&>&*}*&
.cindex "&%substr%& expansion item"
.cindex "substring extraction"
.cindex "expansion" "substring expansion"
The &%substr%& operator is a simpler interface to the &%substr%& function that
can be used when the two parameters are fixed numbers (as opposed to strings
that change when expanded). The effect is the same as
.code
${substr{<start>}{<length>}{<string>}}
.endd
See the description of the general &%substr%& item above for details. The
abbreviation &%s%& can be used when &%substr%& is used as an operator.
All measurement is done in bytes and is not UTF-8 aware.

.vitem &*${time_eval:*&<&'string'&>&*}*&
.cindex "&%time_eval%& expansion item"
.cindex "time interval" "decoding"
This item converts an Exim time interval such as &`2d4h5m`& into a number of
seconds.

.vitem &*${time_interval:*&<&'string'&>&*}*&
.cindex "&%time_interval%& expansion item"
.cindex "time interval" "formatting"
The argument (after sub-expansion) must be a sequence of decimal digits that
represents an interval of time as a number of seconds. It is converted into a
number of larger units and output in Exim's normal time format, for example,
&`1w3d4h2m6s`&.

.vitem &*${uc:*&<&'string'&>&*}*&
.cindex "case forcing in strings"
.cindex "string" "case forcing"
.cindex "upper casing"
.cindex "expansion" "case forcing"
.cindex "&%uc%& expansion item"
This forces the letters in the string into upper-case.
Case is defined per the system C locale.

.vitem &*${utf8clean:*&<&'string'&>&*}*&
.cindex "correction of invalid utf-8 sequences in strings"
.cindex "utf-8" "utf-8 sequences"
.cindex "incorrect utf-8"
.cindex "expansion" "utf-8 forcing"
.cindex "&%utf8clean%& expansion item"
This replaces any invalid utf-8 sequence in the string by the character &`?`&.
In versions of Exim before 4.92, this did not correctly do so for a truncated
final codepoint's encoding, and the character would be silently dropped.
If you must handle detection of this scenario across both sets of Exim behavior,
the complexity will depend upon the task.
For instance, to detect if the first character is multibyte and a 1-byte
extraction can be successfully used as a path component (as is common for
dividing up delivery folders), you might use:
.code
condition = ${if inlist{${utf8clean:${length_1:$local_part}}}{:?}{yes}{no}}
.endd
(which will false-positive if the first character of the local part is a
literal question mark).

.vitem "&*${utf8_domain_to_alabel:*&<&'string'&>&*}*&" &&&
       "&*${utf8_domain_from_alabel:*&<&'string'&>&*}*&" &&&
       "&*${utf8_localpart_to_alabel:*&<&'string'&>&*}*&" &&&
       "&*${utf8_localpart_from_alabel:*&<&'string'&>&*}*&"
.cindex expansion UTF-8
.cindex UTF-8 expansion
.cindex EAI
.cindex internationalisation
.cindex "&%utf8_domain_to_alabel%& expansion item"
.cindex "&%utf8_domain_from_alabel%& expansion item"
.cindex "&%utf8_localpart_to_alabel%& expansion item"
.cindex "&%utf8_localpart_from_alabel%& expansion item"
These convert EAI mail name components between UTF-8 and a-label forms.
For information on internationalisation support see &<<SECTi18nMTA>>&.


.vitem &*${xtextd:*&<&'string'&>&*}*&
.cindex "text forcing in strings"
.cindex "string" "xtext decoding"
.cindex "xtext"
.cindex "&%xtextd%& expansion item"
This performs xtext decoding of the string (per
&url(https://www.rfc-editor.org/rfc/rfc3461,RFC 3461) section 4).



.endlist






.section "Expansion conditions" "SECTexpcond"
.scindex IIDexpcond "expansion" "conditions"
The following conditions are available for testing by the &%${if%& construct
while expanding strings:

.vlist
.vitem &*!*&<&'condition'&>
.cindex "expansion" "negating a condition"
.cindex "negation" "in expansion condition"
Preceding any condition with an exclamation mark negates the result of the
condition.

.vitem <&'symbolic&~operator'&>&~&*{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "numeric comparison"
.cindex "expansion" "numeric comparison"
There are a number of symbolic operators for doing numeric comparisons. They
are:
.itable none 0 0 2 10* left 90* left
.irow "=   " "equal"
.irow "==  " "equal"
.irow ">   " "greater"
.irow ">=  " "greater or equal"
.irow "<   " "less"
.irow "<=  " "less or equal"
.endtable
For example:
.code
${if >{$message_size}{10M} ...
.endd
Note that the general negation operator provides for inequality testing. The
two strings must take the form of optionally signed decimal integers,
optionally followed by one of the letters &"K"&, &"M"& or &"G"& (in either upper or
lower case), signifying multiplication by 1024, 1024*1024 or 1024*1024*1024, respectively.
As a special case, the numerical value of an empty string is taken as
zero.

In all cases, a relative comparator OP is testing if <&'string1'&> OP
<&'string2'&>; the above example is checking if &$message_size$& is larger than
10M, not if 10M is larger than &$message_size$&.


.vitem &*acl&~{{*&<&'name'&>&*}{*&<&'arg1'&>&*}&&&
        {*&<&'arg2'&>&*}...}*&
.cindex "expansion" "calling an acl"
.cindex "&%acl%&" "expansion condition"
The name and zero to nine argument strings are first expanded separately.  The expanded
arguments are assigned to the variables &$acl_arg1$& to &$acl_arg9$& in order.
Any unused are made empty.  The variable &$acl_narg$& is set to the number of
arguments.  The named ACL (see chapter &<<CHAPACL>>&) is called
and may use the variables; if another acl expansion is used the values
are restored after it returns.  If the ACL sets
a value using a "message =" modifier the variable $value becomes
the result of the expansion, otherwise it is empty.
If the ACL returns accept the condition is true; if deny, false.
If the ACL returns defer the result is a forced-fail.

.vitem &*bool&~{*&<&'string'&>&*}*&
.cindex "expansion" "boolean parsing"
.cindex "&%bool%& expansion condition"
This condition turns a string holding a true or false representation into
a boolean state.  It parses &"true"&, &"false"&, &"yes"& and &"no"&
(case-insensitively); also integer numbers map to true if non-zero,
false if zero.
An empty string is treated as false.
Leading and trailing whitespace is ignored;
thus a string consisting only of whitespace is false.
All other string values will result in expansion failure.

When combined with ACL variables, this expansion condition will let you
make decisions in one place and act on those decisions in another place.
For example:
.code
${if bool{$acl_m_privileged_sender} ...
.endd


.vitem &*bool_lax&~{*&<&'string'&>&*}*&
.cindex "expansion" "boolean parsing"
.cindex "&%bool_lax%& expansion condition"
Like &%bool%&, this condition turns a string into a boolean state.  But
where &%bool%& accepts a strict set of strings, &%bool_lax%& uses the same
loose definition that the Router &%condition%& option uses.  The empty string
and the values &"false"&, &"no"& and &"0"& map to false, all others map to
true.  Leading and trailing whitespace is ignored.

Note that where &"bool{00}"& is false, &"bool_lax{00}"& is true.

.vitem &*crypteq&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "expansion" "encrypted comparison"
.cindex "encrypted strings, comparing"
.cindex "&%crypteq%& expansion condition"
This condition is included in the Exim binary if it is built to support any
authentication mechanisms (see chapter &<<CHAPSMTPAUTH>>&). Otherwise, it is
necessary to define SUPPORT_CRYPTEQ in &_Local/Makefile_& to get &%crypteq%&
included in the binary.

The &%crypteq%& condition has two arguments. The first is encrypted and
compared against the second, which is already encrypted. The second string may
be in the LDAP form for storing encrypted strings, which starts with the
encryption type in curly brackets, followed by the data. If the second string
does not begin with &"{"& it is assumed to be encrypted with &[crypt()]& or
&[crypt16()]& (see below), since such strings cannot begin with &"{"&.
Typically this will be a field from a password file. An example of an encrypted
string in LDAP form is:
.code
{md5}CY9rzUYh03PK3k6DJie09g==
.endd
If such a string appears directly in an expansion, the curly brackets have to
be quoted, because they are part of the expansion syntax. For example:
.code
${if crypteq {test}{\{md5\}CY9rzUYh03PK3k6DJie09g==}{yes}{no}}
.endd
The following encryption types (whose names are matched case-independently) are
supported:

.ilist
.cindex "MD5 hash"
.cindex "base64 encoding" "in encrypted password"
&%{md5}%& computes the MD5 digest of the first string, and expresses this as
printable characters to compare with the remainder of the second string. If the
length of the comparison string is 24, Exim assumes that it is base64 encoded
(as in the above example). If the length is 32, Exim assumes that it is a
hexadecimal encoding of the MD5 digest. If the length not 24 or 32, the
comparison fails.

.next
.cindex "SHA-1 hash"
&%{sha1}%& computes the SHA-1 digest of the first string, and expresses this as
printable characters to compare with the remainder of the second string. If the
length of the comparison string is 28, Exim assumes that it is base64 encoded.
If the length is 40, Exim assumes that it is a hexadecimal encoding of the
SHA-1 digest. If the length is not 28 or 40, the comparison fails.

.next
.cindex "&[crypt()]&"
&%{crypt}%& calls the &[crypt()]& function, which traditionally used to use
only the first eight characters of the password. However, in modern operating
systems this is no longer true, and in many cases the entire password is used,
whatever its length.

.next
.cindex "&[crypt16()]&"
&%{crypt16}%& calls the &[crypt16()]& function, which was originally created to
use up to 16 characters of the password in some operating systems. Again, in
modern operating systems, more characters may be used.
.endlist
Exim has its own version of &[crypt16()]&, which is just a double call to
&[crypt()]&. For operating systems that have their own version, setting
HAVE_CRYPT16 in &_Local/Makefile_& when building Exim causes it to use the
operating system version instead of its own. This option is set by default in
the OS-dependent &_Makefile_& for those operating systems that are known to
support &[crypt16()]&.

Some years after Exim's &[crypt16()]& was implemented, a user discovered that
it was not using the same algorithm as some operating systems' versions. It
turns out that as well as &[crypt16()]& there is a function called
&[bigcrypt()]& in some operating systems. This may or may not use the same
algorithm, and both of them may be different to Exim's built-in &[crypt16()]&.

However, since there is now a move away from the traditional &[crypt()]&
functions towards using SHA1 and other algorithms, tidying up this area of
Exim is seen as very low priority.

If you do not put a encryption type (in curly brackets) in a &%crypteq%&
comparison, the default is usually either &`{crypt}`& or &`{crypt16}`&, as
determined by the setting of DEFAULT_CRYPT in &_Local/Makefile_&. The default
default is &`{crypt}`&. Whatever the default, you can always use either
function by specifying it explicitly in curly brackets.

.vitem &*def:*&<&'variable&~name'&>
.cindex "expansion" "checking for empty variable"
.cindex "&%def%& expansion condition"
The &%def%& condition must be followed by the name of one of the expansion
variables defined in section &<<SECTexpvar>>&. The condition is true if the
variable does not contain the empty string. For example:
.code
${if def:sender_ident {from $sender_ident}}
.endd
Note that the variable name is given without a leading &%$%& character. If the
variable does not exist, the expansion fails.

.vitem "&*def:header_*&<&'header&~name'&>&*:*&&~&~or&~&&&
        &~&*def:h_*&<&'header&~name'&>&*:*&"
.cindex "expansion" "checking header line existence"
This condition is true if a message is being processed and the named header
exists in the message. For example,
.code
${if def:header_reply-to:{$h_reply-to:}{$h_from:}}
.endd
&*Note*&: No &%$%& appears before &%header_%& or &%h_%& in the condition, and
the header name must be terminated by a colon if white space does not follow.

.vitem &*eq&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
       &*eqi&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "string" "comparison"
.cindex "expansion" "string comparison"
.cindex "&%eq%& expansion condition"
.cindex "&%eqi%& expansion condition"
The two substrings are first expanded. The condition is true if the two
resulting strings are identical. For &%eq%& the comparison includes the case of
letters, whereas for &%eqi%& the comparison is case-independent, where
case is defined per the system C locale.

.vitem &*exists&~{*&<&'file&~name'&>&*}*&
.cindex "expansion" "file existence test"
.cindex "file" "existence test"
.cindex "&%exists%&, expansion condition"
The substring is first expanded and then interpreted as an absolute path. The
condition is true if the named file (or directory) exists. The existence test
is done by calling the &[stat()]& function. The use of the &%exists%& test in
users' filter files may be locked out by the system administrator.

&*Note:*& Testing a path using this condition is not a sufficient way of
de-tainting it.
Consider using a dsearch lookup.

.vitem &*first_delivery*&
.cindex "delivery" "first"
.cindex "first delivery"
.cindex "expansion" "first delivery test"
.cindex "&%first_delivery%& expansion condition"
.cindex retry condition
This condition, which has no data, is true during a message's first delivery
attempt. It is false during any subsequent delivery attempts.


.vitem "&*forall{*&<&'a list'&>&*}{*&<&'a condition'&>&*}*&" &&&
       "&*forany{*&<&'a list'&>&*}{*&<&'a condition'&>&*}*&"
.cindex "list" "iterative conditions"
.cindex "expansion" "&*forall*& condition"
.cindex "expansion" "&*forany*& condition"
.vindex "&$item$&"
These conditions iterate over a list.
.new
The first argument, after any leading change-of-separator
(see &<<SECTlistsepchange>>&),
is expanded and the whole forms the list.
.wen
By default, the list separator is a colon.

The second argument is interpreted as a condition that is to
be applied to each item in the list in turn. During the interpretation of the
condition, the current list item is placed in a variable called &$item$&.
.ilist
For &*forany*&, interpretation stops if the condition is true for any item, and
the result of the whole condition is true. If the condition is false for all
items in the list, the overall condition is false.
.next
For &*forall*&, interpretation stops if the condition is false for any item,
and the result of the whole condition is false. If the condition is true for
all items in the list, the overall condition is true.
.endlist
Note that negation of &*forany*& means that the condition must be false for all
items for the overall condition to succeed, and negation of &*forall*& means
that the condition must be false for at least one item.

Example:
.code
${if forany{$recipients_list}{match{$item}{^user3@}}{yes}{no}}
.endd
The value of &$item$& is saved and restored while &%forany%& or &%forall%& is
being processed, to enable these expansion items to be nested.

To scan a named list, expand it with the &*listnamed*& operator.

.vitem "&*forall_json{*&<&'a JSON array'&>&*}{*&<&'a condition'&>&*}*&" &&&
       "&*forany_json{*&<&'a JSON array'&>&*}{*&<&'a condition'&>&*}*&" &&&
       "&*forall_jsons{*&<&'a JSON array'&>&*}{*&<&'a condition'&>&*}*&" &&&
       "&*forany_jsons{*&<&'a JSON array'&>&*}{*&<&'a condition'&>&*}*&"
.cindex JSON "iterative conditions"
.cindex JSON expansions
.cindex expansion "&*forall_json*& condition"
.cindex expansion "&*forany_json*& condition"
.cindex expansion "&*forall_jsons*& condition"
.cindex expansion "&*forany_jsons*& condition"
As for the above, except that the first argument must, after expansion,
be a JSON array.
The array separator is not changeable.
For the &"jsons"& variants the elements are expected to be JSON strings
and have their quotes removed before the evaluation of the condition.



.vitem &*ge&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
       &*gei&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "string" "comparison"
.cindex "expansion" "string comparison"
.cindex "&%ge%& expansion condition"
.cindex "&%gei%& expansion condition"
The two substrings are first expanded. The condition is true if the first
string is lexically greater than or equal to the second string. For &%ge%& the
comparison includes the case of letters, whereas for &%gei%& the comparison is
case-independent.
Case and collation order are defined per the system C locale.

.vitem &*gt&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
       &*gti&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "string" "comparison"
.cindex "expansion" "string comparison"
.cindex "&%gt%& expansion condition"
.cindex "&%gti%& expansion condition"
The two substrings are first expanded. The condition is true if the first
string is lexically greater than the second string. For &%gt%& the comparison
includes the case of letters, whereas for &%gti%& the comparison is
case-independent.
Case and collation order are defined per the system C locale.


.vitem &*inbound_srs&~{*&<&'local&~part'&>&*}{*&<&'secret'&>&*}*&
SRS decode.  See SECT &<<SECTSRS>>& for details.


.vitem &*inlist&~{*&<&'subject'&>&*}{*&<&'list'&>&*}*& &&&
       &*inlisti&~{*&<&'subject'&>&*}{*&<&'list'&>&*}*&
.cindex "string" "comparison"
.cindex "list" "iterative conditions"
The <&'subject'&> string is expanded.
.new
The <&'list'&> first has any change-of-list-separator
+(see &<<SECTlistsepchange>>&) retained verbatim,
+then the remainder is expanded.
+.wen
The whole is treated as a list of simple strings;
if the subject string is a member of that list, then the condition is true.
For the case-independent &%inlisti%& condition, case is defined per the system C locale.

These are simpler to use versions of the more powerful &*forany*& condition.
Examples, and the &*forany*& equivalents:
.code
${if inlist{needle}{foo:needle:bar}}
  ${if forany{foo:needle:bar}{eq{$item}{needle}}}
${if inlisti{Needle}{fOo:NeeDLE:bAr}}
  ${if forany{fOo:NeeDLE:bAr}{eqi{$item}{Needle}}}
.endd

The variable &$value$& will be set for a successful match and can be
used in the success clause of an &%if%& expansion item using the condition.
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using an inlist expansion condition"
It will have the same taint status as the list; expansions such as
.code
${if inlist {$h_mycode:} {0 : 1 : 42} {$value}}
.endd
can be used for de-tainting.
Any previous &$value$& is restored after the if.


.vitem &*isip&~{*&<&'string'&>&*}*&  &&&
       &*isip4&~{*&<&'string'&>&*}*& &&&
       &*isip6&~{*&<&'string'&>&*}*&
.cindex "IP address" "testing string format"
.cindex "string" "testing for IP address"
.cindex "&%isip%& expansion condition"
.cindex "&%isip4%& expansion condition"
.cindex "&%isip6%& expansion condition"
The substring is first expanded, and then tested to see if it has the form of
an IP address. Both IPv4 and IPv6 addresses are valid for &%isip%&, whereas
&%isip4%& and &%isip6%& test specifically for IPv4 or IPv6 addresses.

For an IPv4 address, the test is for four dot-separated components, each of
which consists of from one to three digits. For an IPv6 address, up to eight
colon-separated components are permitted, each containing from one to four
hexadecimal digits. There may be fewer than eight components if an empty
component (adjacent colons) is present. Only one empty component is permitted.

&*Note*&: The checks used to be just on the form of the address; actual numerical
values were not considered. Thus, for example, 999.999.999.999 passed the IPv4
check.
This is no longer the case.

The main use of these tests is to distinguish between IP addresses and
host names, or between IPv4 and IPv6 addresses. For example, you could use
.code
${if isip4{$sender_host_address}...
.endd
to test which IP version an incoming SMTP connection is using.

.vitem &*ldapauth&~{*&<&'ldap&~query'&>&*}*&
.cindex "LDAP" "use for authentication"
.cindex "expansion" "LDAP authentication test"
.cindex "&%ldapauth%& expansion condition"
This condition supports user authentication using LDAP. See section
&<<SECTldap>>& for details of how to use LDAP in lookups and the syntax of
queries. For this use, the query must contain a user name and password. The
query itself is not used, and can be empty. The condition is true if the
password is not empty, and the user name and password are accepted by the LDAP
server. An empty password is rejected without calling LDAP because LDAP binds
with an empty password are considered anonymous regardless of the username, and
will succeed in most configurations. See chapter &<<CHAPSMTPAUTH>>& for details
of SMTP authentication, and chapter &<<CHAPplaintext>>& for an example of how
this can be used.


.vitem &*le&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
       &*lei&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "string" "comparison"
.cindex "expansion" "string comparison"
.cindex "&%le%& expansion condition"
.cindex "&%lei%& expansion condition"
The two substrings are first expanded. The condition is true if the first
string is lexically less than or equal to the second string. For &%le%& the
comparison includes the case of letters, whereas for &%lei%& the comparison is
case-independent.
Case and collation order are defined per the system C locale.

.vitem &*lt&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*& &&&
       &*lti&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "string" "comparison"
.cindex "expansion" "string comparison"
.cindex "&%lt%& expansion condition"
.cindex "&%lti%& expansion condition"
The two substrings are first expanded. The condition is true if the first
string is lexically less than the second string. For &%lt%& the comparison
includes the case of letters, whereas for &%lti%& the comparison is
case-independent.
Case and collation order are defined per the system C locale.


.vitem &*match&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "expansion" "regular expression comparison"
.cindex "regular expressions" "match in expanded string"
.cindex "&%match%& expansion condition"
The two substrings are first expanded. The second is then treated as a regular
expression and applied to the first. Because of the pre-expansion, if the
regular expression contains dollar, or backslash characters, they must be
escaped. Care must also be taken if the regular expression contains braces
(curly brackets). A closing brace must be escaped so that it is not taken as a
premature termination of <&'string2'&>. The easiest approach is to use the
&`\N`& feature to disable expansion of the regular expression.
For example,
.code
${if match {$local_part}{\N^\d{3}\N} ...
.endd
If the whole expansion string is in double quotes, further escaping of
backslashes is also required.

The condition is true if the regular expression match succeeds.
The regular expression is not required to begin with a circumflex
metacharacter, but if there is no circumflex, the expression is not anchored,
and it may match anywhere in the subject, not just at the start. If you want
the pattern to match at the end of the subject, you must include the &`$`&
metacharacter at an appropriate point.
All character handling is done in bytes and is not UTF-8 aware,
but we might change this in a future Exim release.

.cindex "numerical variables (&$1$& &$2$& etc)" "in &%if%& expansion"
At the start of an &%if%& expansion the values of the numeric variable
substitutions &$1$& etc. are remembered. Obeying a &%match%& condition that
succeeds causes them to be reset to the substrings of that condition and they
will have these values during the expansion of the success string. At the end
of the &%if%& expansion, the previous values are restored. After testing a
combination of conditions using &%or%&, the subsequent values of the numeric
variables are those of the condition that succeeded.

.vitem &*match_address&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "&%match_address%& expansion condition"
See &*match_local_part*&.

.vitem &*match_domain&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "&%match_domain%& expansion condition"
See &*match_local_part*&.

.vitem &*match_ip&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "&%match_ip%& expansion condition"
This condition matches an IP address to a list of IP address patterns. It must
be followed by two argument strings. The first (after expansion) must be an IP
address or an empty string. The second (not expanded) is a restricted host
list that can match only an IP address, not a host name. For example:
.code
${if match_ip{$sender_host_address}{1.2.3.4:5.6.7.8}{...}{...}}
.endd
The specific types of host list item that are permitted in the list are:

.ilist
An IP address, optionally with a CIDR mask.
.next
A single asterisk, which matches any IP address.
.next
An empty item, which matches only if the IP address is empty. This could be
useful for testing for a locally submitted message or one from specific hosts
in a single test such as
. ==== As this is a nested list, any displays it contains must be indented
. ==== as otherwise they are too far to the left. This comment applies to
. ==== the use of xmlto plus fop. There's no problem when formatting with
. ==== sdop, with or without the extra indent.
.code
  ${if match_ip{$sender_host_address}{:4.3.2.1:...}{...}{...}}
.endd
where the first item in the list is the empty string.
.next
The item @[] matches any of the local host's interface addresses.
.next
Single-key lookups are assumed to be like &"net-"& style lookups in host lists
(see section &<<SECThoslispatsikey>>&),
even if &`net-`& is not specified. There is never any attempt to turn the IP
address into a host name. The most common type of linear search for
&*match_ip*& is likely to be &*iplsearch*&, in which the file can contain CIDR
masks. For example:
.code
  ${if match_ip{$sender_host_address}{iplsearch;/some/file}...
.endd
It is of course possible to use other kinds of lookup, and in such a case, you
do need to specify the &`net-`& prefix if you want to specify a specific
address mask, for example:
.code
  ${if match_ip{$sender_host_address}{net24-dbm;/some/file}...
.endd
However, unless you are combining a &%match_ip%& condition with others, it is
just as easy to use the fact that a lookup is itself a condition, and write:
.code
  ${lookup{${mask:$sender_host_address/24}}dbm{/a/file}...
.endd
.endlist ilist

Note that <&'string2'&> is not itself subject to string expansion, unless
Exim was built with the EXPAND_LISTMATCH_RHS option.
.new
For the latter case, only the part after any leading
change-of-separator specification is expanded.
.wen

Consult section &<<SECThoslispatip>>& for further details of these patterns.

The variable &$value$& will be set for a successful match and can be
used in the success clause of an &%if%& expansion item using the condition.
Any previous &$value$& is restored after the if.

.vitem &*match_local_part&~{*&<&'string1'&>&*}{*&<&'string2'&>&*}*&
.cindex "domain list" "in expansion condition"
.cindex "address list" "in expansion condition"
.cindex "local part" "list, in expansion condition"
.cindex "&%match_local_part%& expansion condition"
This condition, together with &%match_address%& and &%match_domain%&, make it
possible to test domain, address, and local part lists within expansions. Each
condition requires two arguments: an item and a list to match. A trivial
example is:
.code
${if match_domain{a.b.c}{x.y.z:a.b.c:p.q.r}{yes}{no}}
.endd
In each case, the second argument may contain any of the allowable items for a
list of the appropriate type. Also, because the second argument
is a standard form of list, it is possible to refer to a named list.
Thus, you can use conditions like this:
.code
${if match_domain{$domain}{+local_domains}{...
.endd
.cindex "&`+caseful`&"
For address lists, the matching starts off caselessly, but the &`+caseful`&
item can be used, as in all address lists, to cause subsequent items
(including those of referenced named lists)
to have their local parts matched casefully.
Domains are always matched caselessly.

The variable &$value$& will be set for a successful match and can be
used in the success clause of an &%if%& expansion item using the condition.
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using a match_local_part expansion condition"
It will have the same taint status as the list; expansions such as
.code
${if match_local_part {$local_part} {alice : bill : charlotte : dave} {$value}}
.endd
can be used for de-tainting.
Any previous &$value$& is restored after the if.

Note that <&'string2'&> is not itself subject to string expansion, unless
Exim was built with the EXPAND_LISTMATCH_RHS option.
.new
For the latter case, only the part after any leading
change-of-separator specification is expanded.
.wen

&*Note*&: Host lists are &'not'& supported in this way. This is because
hosts have two identities: a name and an IP address, and it is not clear
how to specify cleanly how such a test would work. However, IP addresses can be
matched using &%match_ip%&.

.vitem &*pam&~{*&<&'string1'&>&*:*&<&'string2'&>&*:...}*&
.cindex "PAM authentication"
.cindex "AUTH" "with PAM"
.cindex "Solaris" "PAM support"
.cindex "expansion" "PAM authentication test"
.cindex "&%pam%& expansion condition"
&'Pluggable Authentication Modules'&
(&url(https://mirrors.edge.kernel.org/pub/linux/libs/pam/)) are a facility that is
available in Solaris
and in some GNU/Linux distributions.
The Exim support, which is intended for use in conjunction with
the SMTP AUTH command, is available only if Exim is compiled with
.code
SUPPORT_PAM=yes
.endd
in &_Local/Makefile_&. You probably need to add &%-lpam%& to EXTRALIBS, and
in some releases of GNU/Linux &%-ldl%& is also needed.

The argument string is first expanded, and the result must be a
colon-separated list of strings. Leading and trailing white space is ignored.
The PAM module is initialized with the service name &"exim"& and the user name
taken from the first item in the colon-separated data string (<&'string1'&>).
The remaining items in the data string are passed over in response to requests
from the authentication function. In the simple case there will only be one
request, for a password, so the data consists of just two strings.

There can be problems if any of the strings are permitted to contain colon
characters. In the usual way, these have to be doubled to avoid being taken as
separators.
The &%listquote%& expansion item can be used for this.
For example, the configuration
of a LOGIN authenticator might contain this setting:
.code
server_condition = ${if pam{$auth1:${listquote{:}{$auth2}}}}
.endd
In some operating systems, PAM authentication can be done only from a process
running as root. Since Exim is running as the Exim user when receiving
messages, this means that PAM cannot be used directly in those systems.
. --- 2018-09-07: the pam_exim modified variant has gone, removed claims re using Exim via that


.vitem &*pwcheck&~{*&<&'string1'&>&*:*&<&'string2'&>&*}*&
.cindex "&'pwcheck'& daemon"
.cindex "Cyrus"
.cindex "expansion" "&'pwcheck'& authentication test"
.cindex "&%pwcheck%& expansion condition"
This condition supports user authentication using the Cyrus &'pwcheck'& daemon.
This is one way of making it possible for passwords to be checked by a process
that is not running as root. &*Note*&: The use of &'pwcheck'& is now
deprecated. Its replacement is &'saslauthd'& (see below).

The pwcheck support is not included in Exim by default. You need to specify
the location of the pwcheck daemon's socket in &_Local/Makefile_& before
building Exim. For example:
.code
CYRUS_PWCHECK_SOCKET=/var/pwcheck/pwcheck
.endd
You do not need to install the full Cyrus software suite in order to use
the pwcheck daemon. You can compile and install just the daemon alone
from the Cyrus SASL library. Ensure that &'exim'& is the only user that has
access to the &_/var/pwcheck_& directory.

The &%pwcheck%& condition takes one argument, which must be the user name and
password, separated by a colon. For example, in a LOGIN authenticator
configuration, you might have this:
.code
server_condition = ${if pwcheck{$auth1:$auth2}}
.endd
Again, for a PLAIN authenticator configuration, this would be:
.code
server_condition = ${if pwcheck{$auth2:$auth3}}
.endd
.vitem &*queue_running*&
.cindex "queue runner" "detecting when delivering from"
.cindex "expansion" "queue runner test"
.cindex "&%queue_running%& expansion condition"
This condition, which has no data, is true during delivery attempts that are
initiated by queue runner processes, and false otherwise.


.vitem &*radius&~{*&<&'authentication&~string'&>&*}*&
.cindex "Radius"
.cindex "expansion" "Radius authentication"
.cindex "&%radius%& expansion condition"
Radius authentication
(&url(https://www.rfc-editor.org/rfc/rfc2865,RFC 2865))
is supported in a similar way to PAM. You must
set RADIUS_CONFIG_FILE in &_Local/Makefile_& to specify the location of
the Radius client configuration file in order to build Exim with Radius
support.

With just that one setting, Exim expects to be linked with the &%radiusclient%&
library, using the original API. If you are using release 0.4.0 or later of
this library, you need to set
.code
RADIUS_LIB_TYPE=RADIUSCLIENTNEW
.endd
in &_Local/Makefile_& when building Exim. You can also link Exim with the
&%libradius%& library that comes with FreeBSD. To do this, set
.code
RADIUS_LIB_TYPE=RADLIB
.endd
in &_Local/Makefile_&, in addition to setting RADIUS_CONFIGURE_FILE.
You may also have to supply a suitable setting in EXTRALIBS so that the
Radius library can be found when Exim is linked.

The string specified by RADIUS_CONFIG_FILE is expanded and passed to the
Radius client library, which calls the Radius server. The condition is true if
the authentication is successful. For example:
.code
server_condition = ${if radius{<arguments>}}
.endd


.vitem "&*saslauthd&~{{*&<&'user'&>&*}{*&<&'password'&>&*}&&&
        {*&<&'service'&>&*}{*&<&'realm'&>&*}}*&"
.cindex "&'saslauthd'& daemon"
.cindex "Cyrus"
.cindex "expansion" "&'saslauthd'& authentication test"
.cindex "&%saslauthd%& expansion condition"
This condition supports user authentication using the Cyrus &'saslauthd'&
daemon. This replaces the older &'pwcheck'& daemon, which is now deprecated.
Using this daemon is one way of making it possible for passwords to be checked
by a process that is not running as root.

The saslauthd support is not included in Exim by default. You need to specify
the location of the saslauthd daemon's socket in &_Local/Makefile_& before
building Exim. For example:
.code
CYRUS_SASLAUTHD_SOCKET=/var/state/saslauthd/mux
.endd
You do not need to install the full Cyrus software suite in order to use
the saslauthd daemon. You can compile and install just the daemon alone
from the Cyrus SASL library.

Up to four arguments can be supplied to the &%saslauthd%& condition, but only
two are mandatory. For example:
.code
server_condition = ${if saslauthd{{$auth1}{$auth2}}}
.endd
The service and the realm are optional (which is why the arguments are enclosed
in their own set of braces). For details of the meaning of the service and
realm, and how to run the daemon, consult the Cyrus documentation.
.endlist vlist



.section "Combining expansion conditions" "SECID84"
.cindex "expansion" "combining conditions"
Several conditions can be tested at once by combining them using the &%and%&
and &%or%& combination conditions. Note that &%and%& and &%or%& are complete
conditions on their own, and precede their lists of sub-conditions. Each
sub-condition must be enclosed in braces within the overall braces that contain
the list. No repetition of &%if%& is used.


.vlist
.vitem &*or&~{{*&<&'cond1'&>&*}{*&<&'cond2'&>&*}...}*&
.cindex "&""or""& expansion condition"
.cindex "expansion" "&""or""& of conditions"
The sub-conditions are evaluated from left to right. The condition is true if
any one of the sub-conditions is true.
For example,
.code
${if or {{eq{$local_part}{spqr}}{eq{$domain}{testing.com}}}...
.endd
When a true sub-condition is found, the following ones are parsed but not
evaluated. If there are several &"match"& sub-conditions the values of the
numeric variables afterwards are taken from the first one that succeeds.

.vitem &*and&~{{*&<&'cond1'&>&*}{*&<&'cond2'&>&*}...}*&
.cindex "&""and""& expansion condition"
.cindex "expansion" "&""and""& of conditions"
The sub-conditions are evaluated from left to right. The condition is true if
all of the sub-conditions are true. If there are several &"match"&
sub-conditions, the values of the numeric variables afterwards are taken from
the last one. When a false sub-condition is found, the following ones are
parsed but not evaluated.
.endlist
.ecindex IIDexpcond




.section "Expansion variables" "SECTexpvar"
.cindex "expansion" "variables, list of"
This section contains an alphabetical list of all the expansion variables. Some
of them are available only when Exim is compiled with specific options such as
support for TLS or the content scanning extension.
.cindex "tainted data"
Variables marked as &'tainted'& are likely to carry data supplied by
a potential attacker.
Variables without such marking may also, depending on how their
values are created.
Such variables should not be further expanded,
used as filenames
or used as command-line arguments for external commands.

.vlist
.vitem "&$0$&, &$1$&, etc"
.cindex "numerical variables (&$1$& &$2$& etc)"
When a &%match%& expansion condition succeeds, these variables contain the
captured substrings identified by the regular expression during subsequent
processing of the success string of the containing &%if%& expansion item.
In the expansion condition case
they do not retain their values afterwards; in fact, their previous
values are restored at the end of processing an &%if%& item. The numerical
variables may also be set externally by some other matching process which
precedes the expansion of the string. For example, the commands available in
Exim filter files include an &%if%& command with its own regular expression
matching condition.
If the subject string was tainted then any captured substring will also be.

.vitem "&$acl_arg1$&, &$acl_arg2$&, etc"
Within an acl condition, expansion condition or expansion item
any arguments are copied to these variables,
any unused variables being made empty.

.vitem "&$acl_c...$&"
Values can be placed in these variables by the &%set%& modifier in an ACL. They
can be given any name that starts with &$acl_c$& and is at least six characters
long, but the sixth character must be either a digit or an underscore. For
example: &$acl_c5$&, &$acl_c_mycount$&. The values of the &$acl_c...$&
variables persist throughout the lifetime of an SMTP connection. They can be
used to pass information between ACLs and between different invocations of the
same ACL. When a message is received, the values of these variables are saved
with the message, and can be accessed by filters, routers, and transports
during subsequent delivery.

.vitem "&$acl_m...$&"
These variables are like the &$acl_c...$& variables, except that their values
are reset after a message has been received. Thus, if several messages are
received in one SMTP connection, &$acl_m...$& values are not passed on from one
message to the next, as &$acl_c...$& values are. The &$acl_m...$& variables are
also reset by MAIL, RSET, EHLO, HELO, and after starting a TLS session. When a
message is received, the values of these variables are saved with the message,
and can be accessed by filters, routers, and transports during subsequent
delivery.

.vitem &$acl_narg$&
Within an acl condition, expansion condition or expansion item
this variable has the number of arguments.

.vitem &$acl_verify_message$&
.vindex "&$acl_verify_message$&"
After an address verification has failed, this variable contains the failure
message. It retains its value for use in subsequent modifiers of the verb.
The message can be preserved by coding like this:
.code
warn !verify = sender
     set acl_m0 = $acl_verify_message
.endd
You can use &$acl_verify_message$& during the expansion of the &%message%& or
&%log_message%& modifiers, to include information about the verification
failure.
&*Note*&: The variable is cleared at the end of processing the ACL verb.

.vitem &$address_data$&
.vindex "&$address_data$&"
This variable is set by means of the &%address_data%& option in routers. The
value then remains with the address while it is processed by subsequent routers
and eventually a transport. If the transport is handling multiple addresses,
the value from the first address is used. See chapter &<<CHAProutergeneric>>&
for more details. &*Note*&: The contents of &$address_data$& are visible in
user filter files.

If &$address_data$& is set when the routers are called from an ACL to verify
a recipient address, the final value is still in the variable for subsequent
conditions and modifiers of the ACL statement. If routing the address caused it
to be redirected to just one address, the child address is also routed as part
of the verification, and in this case the final value of &$address_data$& is
from the child's routing.

If &$address_data$& is set when the routers are called from an ACL to verify a
sender address, the final value is also preserved, but this time in
&$sender_address_data$&, to distinguish it from data from a recipient
address.

In both cases (recipient and sender verification), the value does not persist
after the end of the current ACL statement. If you want to preserve
these values for longer, you can save them in ACL variables.

.vitem &$address_file$&
.vindex "&$address_file$&"
When, as a result of aliasing, forwarding, or filtering, a message is directed
to a specific file, this variable holds the name of the file when the transport
is running. At other times, the variable is empty. For example, using the
default configuration, if user &%r2d2%& has a &_.forward_& file containing
.code
/home/r2d2/savemail
.endd
then when the &(address_file)& transport is running, &$address_file$&
contains the text string &`/home/r2d2/savemail`&.
.cindex "Sieve filter" "value of &$address_file$&"
For Sieve filters, the value may be &"inbox"& or a relative folder name. It is
then up to the transport configuration to generate an appropriate absolute path
to the relevant file.

.vitem &$address_pipe$&
.vindex "&$address_pipe$&"
When, as a result of aliasing or forwarding, a message is directed to a pipe,
this variable holds the pipe command when the transport is running.

.new
.vitem &$atrn_host$&
.vindex ATRN "data for routing"
When an ATRN command is accepted, this variable is filled in with the client
IP and port, for use in a manualroute router.
.wen

.vitem "&$auth1$& &-- &$auth4$&"
.vindex "&$auth1$&, &$auth2$&, etc"
These variables are used in SMTP authenticators (see chapters
&<<CHAPplaintext>>&&--&<<CHAPtlsauth>>&). Elsewhere, they are empty.

.vitem &$authenticated_id$&
.cindex "authentication" "id"
.vindex "&$authenticated_id$&"
When a server successfully authenticates a client it may be configured to
preserve some of the authentication information in the variable
&$authenticated_id$& (see chapter &<<CHAPSMTPAUTH>>&). For example, a
user/password authenticator configuration might preserve the user name for use
in the routers. Note that this is not the same information that is saved in
&$sender_host_authenticated$&.

When a message is submitted locally (that is, not over a TCP connection)
the value of &$authenticated_id$& is normally the login name of the calling
process. However, a trusted user can override this by means of the &%-oMai%&
command line option.
This second case also sets up information used by the
&$authresults$& expansion item.

.vitem &$authenticated_fail_id$&
.cindex "authentication" "fail" "id"
.vindex "&$authenticated_fail_id$&"
When an authentication attempt fails, the variable &$authenticated_fail_id$&
will contain the failed authentication id. If more than one authentication
id is attempted, it will contain only the last one. The variable is
available for processing in the ACL's, generally the quit or notquit ACL.
A message to a local recipient could still be accepted without requiring
authentication, which means this variable could also be visible in all of
the ACL's as well.


.tvar &$authenticated_sender$&
.cindex "sender" "authenticated"
.cindex "authentication" "sender"
.cindex "AUTH" "on MAIL command"
When acting as a server, Exim takes note of the AUTH= parameter on an incoming
SMTP MAIL command if it believes the sender is sufficiently trusted, as
described in section &<<SECTauthparamail>>&. Unless the data is the string
&"<>"&, it is set as the authenticated sender of the message, and the value is
available during delivery in the &$authenticated_sender$& variable. If the
sender is not trusted, Exim accepts the syntax of AUTH=, but ignores the data.

.vindex "&$qualify_domain$&"
When a message is submitted locally (that is, not over a TCP connection), the
value of &$authenticated_sender$& is an address constructed from the login
name of the calling process and &$qualify_domain$&, except that a trusted user
can override this by means of the &%-oMas%& command line option.


.vitem &$authentication_failed$&
.cindex "authentication" "failure"
.vindex "&$authentication_failed$&"
This variable is set to &"1"& in an Exim server if a client issues an AUTH
command that does not succeed. Otherwise it is set to &"0"&. This makes it
possible to distinguish between &"did not try to authenticate"&
(&$sender_host_authenticated$& is empty and &$authentication_failed$& is set to
&"0"&) and &"tried to authenticate but failed"& (&$sender_host_authenticated$&
is empty and &$authentication_failed$& is set to &"1"&).
Failure includes cancellation of a authentication attempt,
and any negative response to an AUTH command,
(including, for example, an attempt to use an undefined mechanism).

.vitem &$av_failed$&
.cindex "content scanning" "AV scanner failure"
This variable is available when Exim is compiled with the content-scanning
extension. It is set to &"0"& by default, but will be set to &"1"& if any
problem occurs with the virus scanner (specified by &%av_scanner%&) during
the ACL malware condition.

.vitem &$body_linecount$&
.cindex "message body" "line count"
.cindex "body of message" "line count"
.vindex "&$body_linecount$&"
When a message is being received or delivered, this variable contains the
number of lines in the message's body. See also &$message_linecount$&.

.vitem &$body_zerocount$&
.cindex "message body" "binary zero count"
.cindex "body of message" "binary zero count"
.cindex "binary zero" "in message body"
.vindex "&$body_zerocount$&"
When a message is being received or delivered, this variable contains the
number of binary zero bytes (ASCII NULs) in the message's body.

.vitem &$bounce_recipient$&
.vindex "&$bounce_recipient$&"
This is set to the recipient address of a bounce message while Exim is creating
it. It is useful if a customized bounce message text file is in use (see
chapter &<<CHAPemsgcust>>&).

.vitem &$bounce_return_size_limit$&
.vindex "&$bounce_return_size_limit$&"
This contains the value set in the &%bounce_return_size_limit%& option, rounded
up to a multiple of 1000. It is useful when a customized error message text
file is in use (see chapter &<<CHAPemsgcust>>&).

.vitem &$caller_gid$&
.cindex "gid (group id)" "caller"
.vindex "&$caller_gid$&"
The real group id under which the process that called Exim was running. This is
not the same as the group id of the originator of a message (see
&$originator_gid$&). If Exim re-execs itself, this variable in the new
incarnation normally contains the Exim gid.

.vitem &$caller_uid$&
.cindex "uid (user id)" "caller"
.vindex "&$caller_uid$&"
The real user id under which the process that called Exim was running. This is
not the same as the user id of the originator of a message (see
&$originator_uid$&). If Exim re-execs itself, this variable in the new
incarnation normally contains the Exim uid.

.vitem &$callout_address$&
.vindex "&$callout_address$&"
After a callout for verification, spamd or malware daemon service, the
address that was connected to.

.vitem &$compile_number$&
.vindex "&$compile_number$&"
The building process for Exim keeps a count of the number
of times it has been compiled. This serves to distinguish different
compilations of the same version of Exim.

.vitem &$config_dir$&
.vindex "&$config_dir$&"
The directory name of the main configuration file. That is, the content of
&$config_file$& with the last component stripped. The value does not
contain the trailing slash. If &$config_file$& does not contain a slash,
&$config_dir$& is ".".

.vitem &$config_file$&
.vindex "&$config_file$&"
The name of the main configuration file Exim is using.

.vitem &$connection_id$&
.vindex "&$connection_id$&"
.cindex connection "identifier logging"
An identifier for the accepted connection, for use in custom logging.

.vitem &$dkim_verify_status$&
Results of DKIM verification.
For details see section &<<SECDKIMVFY>>&.

.vitem &$dkim_cur_signer$& &&&
       &$dkim_verify_reason$& &&&
       &$dkim_domain$& &&&
       &$dkim_identity$& &&&
       &$dkim_selector$& &&&
       &$dkim_algo$& &&&
       &$dkim_canon_body$& &&&
       &$dkim_canon_headers$& &&&
       &$dkim_copiedheaders$& &&&
       &$dkim_bodylength$& &&&
       &$dkim_created$& &&&
       &$dkim_expires$& &&&
       &$dkim_headernames$& &&&
       &$dkim_key_testing$& &&&
       &$dkim_key_nosubdomains$& &&&
       &$dkim_key_srvtype$& &&&
       &$dkim_key_granularity$& &&&
       &$dkim_key_notes$& &&&
       &$dkim_key_length$&
These variables are only available within the DKIM ACL.
For details see section &<<SECDKIMVFY>>&.

.vitem &$dkim_signers$&
.vindex &$dkim_signers$&
When a message has been received this variable contains
a colon-separated list of signer domains and identities for the message.
For details see section &<<SECDKIMVFY>>&.

.vitem &$dmarc_alignment_spf$& &&&
       &$dmarc_alignment_dkim$& &&&
       &$dmarc_domain_policy$& &&&
       &$dmarc_status$& &&&
       &$dmarc_status_text$& &&&
       &$dmarc_used_domains$&
Results of DMARC verification.
For details see section &<<SECDMARC>>&.

.vitem &$dnslist_domain$& &&&
       &$dnslist_matched$& &&&
       &$dnslist_text$& &&&
       &$dnslist_value$&
.vindex "&$dnslist_domain$&"
.vindex "&$dnslist_matched$&"
.vindex "&$dnslist_text$&"
.vindex "&$dnslist_value$&"
.cindex "black list (DNS)"
When a DNS (black) list lookup succeeds, these variables are set to contain
the following data from the lookup: the list's domain name, the key that was
looked up, the contents of any associated TXT record, and the value from the
main A record. See section &<<SECID204>>& for more details.

.tvar &$domain$&
When an address is being routed, or delivered on its own, this variable
contains the domain. Uppercase letters in the domain are converted into lower
case for &$domain$&.

Global address rewriting happens when a message is received, so the value of
&$domain$& during routing and delivery is the value after rewriting. &$domain$&
is set during user filtering, but not during system filtering, because a
message may have many recipients and the system filter is called just once.

When more than one address is being delivered at once (for example, several
RCPT commands in one SMTP delivery), &$domain$& is set only if they all
have the same domain. Transports can be restricted to handling only one domain
at a time if the value of &$domain$& is required at transport time &-- this is
the default for local transports. For further details of the environment in
which local transports are run, see chapter &<<CHAPenvironment>>&.

.oindex "&%delay_warning_condition%&"
At the end of a delivery, if all deferred addresses have the same domain, it is
set in &$domain$& during the expansion of &%delay_warning_condition%&.

The &$domain$& variable is also used in some other circumstances:

.ilist
When an ACL is running for a RCPT command, &$domain$& contains the domain of
the recipient address. The domain of the &'sender'& address is in
&$sender_address_domain$& at both MAIL time and at RCPT time. &$domain$& is not
normally set during the running of the MAIL ACL. However, if the sender address
is verified with a callout during the MAIL ACL, the sender domain is placed in
&$domain$& during the expansions of &%hosts%&, &%interface%&, and &%port%& in
the &(smtp)& transport.

.next
When a rewrite item is being processed (see chapter &<<CHAPrewrite>>&),
&$domain$& contains the domain portion of the address that is being rewritten;
it can be used in the expansion of the replacement address, for example, to
rewrite domains by file lookup.

.next
With one important exception, whenever a domain list is being scanned,
&$domain$& contains the subject domain. &*Exception*&: When a domain list in
a &%sender_domains%& condition in an ACL is being processed, the subject domain
is in &$sender_address_domain$& and not in &$domain$&. It works this way so
that, in a RCPT ACL, the sender domain list can be dependent on the
recipient domain (which is what is in &$domain$& at this time).

.next
.cindex "ETRN" "value of &$domain$&"
.oindex "&%smtp_etrn_command%&"
When the &%smtp_etrn_command%& option is being expanded, &$domain$& contains
the complete argument of the ETRN command (see section &<<SECTETRN>>&).
.endlist

.cindex "tainted data"
If the origin of the data is an incoming message,
the result of expanding this variable is tainted and may not
be further expanded or used as a filename.
When an untainted version is needed, one should be obtained from
looking up the value in a local (therefore trusted) database.
Often &$domain_data$& is usable in this role.


.vitem &$domain_data$&
.vindex "&$domain_data$&"
When the &%domains%& condition on a router
or an ACL
matches a domain
against a list, the match value is copied to &$domain_data$&.
This is an enhancement over previous versions of Exim, when it only
applied to the data read by a lookup.
For details on match values see section &<<SECTlistresults>>& et. al.

If the router routes the
address to a transport, the value is available in that transport. If the
transport is handling multiple addresses, the value from the first address is
used.

&$domain_data$& set in an ACL is available during
the rest of the ACL statement.

.vitem &$exim_gid$&
.vindex "&$exim_gid$&"
This variable contains the numerical value of the Exim group id.

.vitem &$exim_path$&
.vindex "&$exim_path$&"
This variable contains the path to the Exim binary.

.vitem &$exim_uid$&
.vindex "&$exim_uid$&"
This variable contains the numerical value of the Exim user id.

.vitem &$exim_version$&
.vindex "&$exim_version$&"
This variable contains the version string of the Exim build.
The first character is a major version number, currently 4.
Then after a dot, the next group of digits is a minor version number.
There may be other characters following the minor version.
This value may be overridden by the &%exim_version%& main config option.

.vitem &$header_$&<&'name'&>
.tmark
This is not strictly an expansion variable. It is expansion syntax for
inserting the message header line with the given name. Note that the name must
be terminated by colon or white space, because it may contain a wide variety of
characters. Note also that braces must &'not'& be used.
See the full description in section &<<SECTexpansionitems>>& above.

.vitem &$headers_added$&
.vindex "&$headers_added$&"
Within an ACL this variable contains the headers added so far by
the ACL modifier add_header (section &<<SECTaddheadacl>>&).
The headers are a newline-separated list.

.vitem &$home$&
.vindex "&$home$&"
When the &%check_local_user%& option is set for a router, the user's home
directory is placed in &$home$& when the check succeeds. In particular, this
means it is set during the running of users' filter files. A router may also
explicitly set a home directory for use by a transport; this can be overridden
by a setting on the transport itself.

When running a filter test via the &%-bf%& option, &$home$& is set to the value
of the environment variable HOME, which is subject to the
&%keep_environment%& and &%add_environment%& main config options.

.vitem &$host$&
.vindex "&$host$&"
If a router assigns an address to a transport (any transport), and passes a
list of hosts with the address, the value of &$host$& when the transport starts
to run is the name of the first host on the list. Note that this applies both
to local and remote transports.

.cindex "transport" "filter"
.cindex "filter" "transport filter"
For the &(smtp)& transport, if there is more than one host, the value of
&$host$& changes as the transport works its way through the list. In
particular, when the &(smtp)& transport is expanding its options for encryption
using TLS, or for specifying a transport filter (see chapter
&<<CHAPtransportgeneric>>&), &$host$& contains the name of the host to which it
is connected.

When used in the client part of an authenticator configuration (see chapter
&<<CHAPSMTPAUTH>>&), &$host$& contains the name of the server to which the
client is connected.


.vitem &$host_address$&
.vindex "&$host_address$&"
This variable is set to the remote host's IP address whenever &$host$& is set
for a remote connection. It is also set to the IP address that is being checked
when the &%ignore_target_hosts%& option is being processed.

.vitem &$host_data$&
.vindex "&$host_data$&"
If a &%hosts%& condition in an ACL is satisfied by means of a lookup, the
result of the lookup is made available in the &$host_data$& variable. This
allows you, for example, to do things like this:
.code
deny  hosts = net-lsearch;/some/file
      message = $host_data
.endd

.vitem &$host_lookup_deferred$&
.cindex "host name" "lookup, failure of"
.vindex "&$host_lookup_deferred$&"
This variable normally contains &"0"&, as does &$host_lookup_failed$&. When a
message comes from a remote host and there is an attempt to look up the host's
name from its IP address, and the attempt is not successful, one of these
variables is set to &"1"&.

.ilist
If the lookup receives a definite negative response (for example, a DNS lookup
succeeded, but no records were found), &$host_lookup_failed$& is set to &"1"&.

.next
If there is any kind of problem during the lookup, such that Exim cannot
tell whether or not the host name is defined (for example, a timeout for a DNS
lookup), &$host_lookup_deferred$& is set to &"1"&.
.endlist ilist

Looking up a host's name from its IP address consists of more than just a
single reverse lookup. Exim checks that a forward lookup of at least one of the
names it receives from a reverse lookup yields the original IP address. If this
is not the case, Exim does not accept the looked up name(s), and
&$host_lookup_failed$& is set to &"1"&. Thus, being able to find a name from an
IP address (for example, the existence of a PTR record in the DNS) is not
sufficient on its own for the success of a host name lookup. If the reverse
lookup succeeds, but there is a lookup problem such as a timeout when checking
the result, the name is not accepted, and &$host_lookup_deferred$& is set to
&"1"&. See also &$sender_host_name$&.

.cindex authentication "expansion item"
Performing these checks sets up information used by the
&%authresults%& expansion item.


.vitem &$host_lookup_failed$&
.vindex "&$host_lookup_failed$&"
See &$host_lookup_deferred$&.

.vitem &$host_port$&
.vindex "&$host_port$&"
This variable is set to the remote host's TCP port whenever &$host$& is set
for an outbound connection.

.vitem &$initial_cwd$&
.vindex "&$initial_cwd$&
This variable contains the full path name of the initial working
directory of the current Exim process. This may differ from the current
working directory, as Exim changes this to "/" during early startup, and
to &$spool_directory$& later.

.vitem &$inode$&
.vindex "&$inode$&"
The only time this variable is set is while expanding the &%directory_file%&
option in the &(appendfile)& transport. The variable contains the inode number
of the temporary file which is about to be renamed. It can be used to construct
a unique name for the file.

.vitem &$interface_address$& &&&
       &$interface_port$&
.vindex "&$interface_address$&"
.vindex "&$interface_port$&"
These are obsolete names for &$received_ip_address$& and &$received_port$&.

.vitem &$item$&
.vindex "&$item$&"
This variable is used during the expansion of &*forall*& and &*forany*&
conditions (see section &<<SECTexpcond>>&), and &*filter*&, &*map*&, and
&*reduce*& items (see section &<<SECTexpcond>>&). In other circumstances, it is
empty.

.vitem &$ldap_dn$&
.vindex "&$ldap_dn$&"
This variable, which is available only when Exim is compiled with LDAP support,
contains the DN from the last entry in the most recently successful LDAP
lookup.

.vitem &$load_average$&
.vindex "&$load_average$&"
This variable contains the system load average, multiplied by 1000 so that it
is an integer. For example, if the load average is 0.21, the value of the
variable is 210. The value is recomputed every time the variable is referenced.

.tvar &$local_part$&
When an address is being routed, or delivered on its own, this
variable contains the local part. When a number of addresses are being
delivered together (for example, multiple RCPT commands in an SMTP
session), &$local_part$& is not set.

Global address rewriting happens when a message is received, so the value of
&$local_part$& during routing and delivery is the value after rewriting.
&$local_part$& is set during user filtering, but not during system filtering,
because a message may have many recipients and the system filter is called just
once.

.cindex "tainted data"
If the origin of the data is an incoming message,
the result of expanding this variable is tainted and
may not be further expanded or used as a filename.

&*Warning*&: the content of this variable is usually provided by a potential
attacker.
Consider carefully the implications of using it unvalidated as a name
for file access.
This presents issues for users' &_.forward_& and filter files.
For traditional full user accounts, use &%check_local_users%& and the
&$local_part_data$& variable rather than this one.
For virtual users, store a suitable pathname component in the database
which is used for account name validation, and use that retrieved value
rather than this variable.
Often &$local_part_data$& is usable in this role.
If needed, use a router &%address_data%& or &%set%& option for
the retrieved data.

When a message is being delivered to a file, pipe, or autoreply transport as a
result of aliasing or forwarding, &$local_part$& is set to the local part of
the parent address, not to the filename or command (see &$address_file$& and
&$address_pipe$&).

When an ACL is running for a RCPT command, &$local_part$& contains the
local part of the recipient address.

When a rewrite item is being processed (see chapter &<<CHAPrewrite>>&),
&$local_part$& contains the local part of the address that is being rewritten;
it can be used in the expansion of the replacement address, for example.

In all cases, all quoting is removed from the local part. For example, for both
the addresses
.code
"abc:xyz"@test.example
abc\:xyz@test.example
.endd
the value of &$local_part$& is
.code
abc:xyz
.endd
If you use &$local_part$& to create another address, you should always wrap it
inside a quoting operator. For example, in a &(redirect)& router you could
have:
.code
data = ${quote_local_part:$local_part}@new.domain.example
.endd
&*Note*&: The value of &$local_part$& is normally lower cased. If you want
to process local parts in a case-dependent manner in a router, you can set the
&%caseful_local_part%& option (see chapter &<<CHAProutergeneric>>&).

.vitem &$local_part_data$&
.vindex "&$local_part_data$&"
When the &%local_parts%& condition on a router or ACL
matches a local part list
the match value is copied to &$local_part_data$&.
This is an enhancement over previous versions of Exim, when it only
applied to the data read by a lookup.
For details on match values see section &<<SECTlistresults>>& et. al.

The &%check_local_user%& router option also sets this variable.

.vindex &$local_part_prefix$& &&&
        &$local_part_prefix_v$& &&&
        &$local_part_suffix$& &&&
        &$local_part_suffix_v$&
.cindex affix variables
If a local part prefix or suffix has been recognized, it is not included in the
value of &$local_part$& during routing and subsequent delivery. The values of
any prefix or suffix are in &$local_part_prefix$& and
&$local_part_suffix$&, respectively.
.cindex "tainted data"
If the specification did not include a wildcard then
the affix variable value is not tainted.

If the affix specification included a wildcard then the portion of
the affix matched by the wildcard is in
&$local_part_prefix_v$& or &$local_part_suffix_v$& as appropriate,
and both the whole and varying values are tainted.

.vitem &$local_scan_data$&
.vindex "&$local_scan_data$&"
This variable contains the text returned by the &[local_scan()]& function when
a message is received. See chapter &<<CHAPlocalscan>>& for more details.

.vitem &$local_user_gid$&
.vindex "&$local_user_gid$&"
See &$local_user_uid$&.

.vitem &$local_user_uid$&
.vindex "&$local_user_uid$&"
This variable and &$local_user_gid$& are set to the uid and gid after the
&%check_local_user%& router precondition succeeds. This means that their values
are available for the remaining preconditions (&%senders%&, &%require_files%&,
and &%condition%&), for the &%address_data%& expansion, and for any
router-specific expansions. At all other times, the values in these variables
are &`(uid_t)(-1)`& and &`(gid_t)(-1)`&, respectively.

.vitem &$localhost_number$&
.vindex "&$localhost_number$&"
This contains the expanded value of the
&%localhost_number%& option. The expansion happens after the main options have
been read.

.vitem &$log_inodes$&
.vindex "&$log_inodes$&"
The number of free inodes in the disk partition where Exim's
log files are being written. The value is recalculated whenever the variable is
referenced. If the relevant file system does not have the concept of inodes,
the value of is -1. See also the &%check_log_inodes%& option.

.vitem &$log_space$&
.vindex "&$log_space$&"
The amount of free space (as a number of kilobytes) in the disk
partition where Exim's log files are being written. The value is recalculated
whenever the variable is referenced. If the operating system does not have the
ability to find the amount of free space (only true for experimental systems),
the space value is -1. See also the &%check_log_space%& option.


.vitem &$lookup_dnssec_authenticated$&
.vindex "&$lookup_dnssec_authenticated$&"
This variable is set after a DNS lookup done by
a dnsdb lookup expansion, dnslookup router or smtp transport.
.cindex "DNS" "DNSSEC"
It will be empty if &(DNSSEC)& was not requested,
&"no"& if the result was not labelled as authenticated data
and &"yes"& if it was.
Results that are labelled as authoritative answer that match
the &%dns_trust_aa%& configuration variable count also
as authenticated data.

.vitem &$mailstore_basename$&
.vindex "&$mailstore_basename$&"
This variable is set only when doing deliveries in &"mailstore"& format in the
&(appendfile)& transport. During the expansion of the &%mailstore_prefix%&,
&%mailstore_suffix%&, &%message_prefix%&, and &%message_suffix%& options, it
contains the basename of the files that are being written, that is, the name
without the &".tmp"&, &".env"&, or &".msg"& suffix. At all other times, this
variable is empty.

.vitem &$malware_name$&
.vindex "&$malware_name$&"
This variable is available when Exim is compiled with the
content-scanning extension. It is set to the name of the virus that was found
when the ACL &%malware%& condition is true (see section &<<SECTscanvirus>>&).

.vitem &$max_received_linelength$&
.vindex "&$max_received_linelength$&"
.cindex "maximum" "line length"
.cindex "line length" "maximum"
This variable contains the number of bytes in the longest line that was
received as part of the message, not counting the line termination
character(s).
It is not valid if the &%spool_wireformat%& option is used.

.vitem &$message_age$&
.cindex "message" "age of"
.vindex "&$message_age$&"
This variable is set at the start of a delivery attempt to contain the number
of seconds since the message was received. It does not change during a single
delivery attempt.

.tvar &$message_body$&
.cindex "body of message" "expansion variable"
.cindex "message body" "in expansion"
.cindex "binary zero" "in message body"
.oindex "&%message_body_visible%&"
This variable contains the initial portion of a message's body while it is
being delivered, and is intended mainly for use in filter files. The maximum
number of characters of the body that are put into the variable is set by the
&%message_body_visible%& configuration option; the default is 500.

.oindex "&%message_body_newlines%&"
By default, newlines are converted into spaces in &$message_body$&, to make it
easier to search for phrases that might be split over a line break. However,
this can be disabled by setting &%message_body_newlines%& to be true. Binary
zeros are always converted into spaces.

.tvar &$message_body_end$&
.cindex "body of message" "expansion variable"
.cindex "message body" "in expansion"
This variable contains the final portion of a message's
body while it is being delivered. The format and maximum size are as for
&$message_body$&.

.vitem &$message_body_size$&
.cindex "body of message" "size"
.cindex "message body" "size"
.vindex "&$message_body_size$&"
When a message is being delivered, this variable contains the size of the body
in bytes. The count starts from the character after the blank line that
separates the body from the header. Newlines are included in the count. See
also &$message_size$&, &$body_linecount$&, and &$body_zerocount$&.

If the spool file is wireformat
(see the &%spool_wireformat%& main option)
the CRLF line-terminators are included in the count.

.vitem &$message_exim_id$&
.vindex "&$message_exim_id$&"
When a message is being received or delivered, this variable contains the
unique message id that is generated and used by Exim to identify the message.
An id is not created for a message until after its header has been successfully
received. &*Note*&: This is &'not'& the contents of the &'Message-ID:'& header
line; it is the local id that Exim assigns to the message, for example:
&`1BXTIK-0001yO-VA`&.

.tvar &$message_headers$&
This variable contains a concatenation of all the header lines when a message
is being processed, except for lines added by routers or transports. The header
lines are separated by newline characters. Their contents are decoded in the
same way as a header line that is inserted by &%bheader%&.

.tvar &$message_headers_raw$&
This variable is like &$message_headers$& except that no processing of the
contents of header lines is done.

.vitem &$message_id$&
This is an old name for &$message_exim_id$&. It is now deprecated.

.vitem &$message_linecount$&
.vindex "&$message_linecount$&"
This variable contains the total number of lines in the header and body of the
message. Compare &$body_linecount$&, which is the count for the body only.
During the DATA and content-scanning ACLs, &$message_linecount$& contains the
number of lines received. Before delivery happens (that is, before filters,
routers, and transports run) the count is increased to include the
&'Received:'& header line that Exim standardly adds, and also any other header
lines that are added by ACLs. The blank line that separates the message header
from the body is not counted.

As with the special case of &$message_size$&, during the expansion of the
appendfile transport's maildir_tag option in maildir format, the value of
&$message_linecount$& is the precise size of the number of newlines in the
file that has been written (minus one for the blank line between the
header and the body).

Here is an example of the use of this variable in a DATA ACL:
.code
deny condition = \
      ${if <{250}{${eval:$message_linecount - $body_linecount}}}
     message   = Too many lines in message header
.endd
In the MAIL and RCPT ACLs, the value is zero because at that stage the
message has not yet been received.

This variable is not valid if the &%spool_wireformat%& option is used.

.vitem &$message_size$&
.cindex "size" "of message"
.cindex "message" "size"
.vindex "&$message_size$&"
When a message is being processed, this variable contains its size in bytes. In
most cases, the size includes those headers that were received with the
message, but not those (such as &'Envelope-to:'&) that are added to individual
deliveries as they are written. However, there is one special case: during the
expansion of the &%maildir_tag%& option in the &(appendfile)& transport while
doing a delivery in maildir format, the value of &$message_size$& is the
precise size of the file that has been written. See also
&$message_body_size$&, &$body_linecount$&, and &$body_zerocount$&.

.cindex "RCPT" "value of &$message_size$&"
While running a per message ACL (mail/rcpt/predata), &$message_size$&
contains the size supplied on the MAIL command, or -1 if no size was given. The
value may not, of course, be truthful.

.vitem &$mime_anomaly_level$& &&&
       &$mime_anomaly_text$& &&&
       &$mime_boundary$& &&&
       &$mime_charset$& &&&
       &$mime_content_description$& &&&
       &$mime_content_disposition$& &&&
       &$mime_content_id$& &&&
       &$mime_content_size$& &&&
       &$mime_content_transfer_encoding$& &&&
       &$mime_content_type$& &&&
       &$mime_decoded_filename$& &&&
       &$mime_filename$& &&&
       &$mime_is_coverletter$& &&&
       &$mime_is_multipart$& &&&
       &$mime_is_rfc822$& &&&
       &$mime_part_count$&
A number of variables whose names start with &$mime$& are
available when Exim is compiled with the content-scanning extension. For
details, see section &<<SECTscanmimepart>>&.

.vitem "&$n0$& &-- &$n9$&"
These variables are counters that can be incremented by means
of the &%add%& command in filter files.

.tvar &$original_domain$&
.vindex "&$domain$&"
When a top-level address is being processed for delivery, this contains the
same value as &$domain$&. However, if a &"child"& address (for example,
generated by an alias, forward, or filter file) is being processed, this
variable contains the domain of the original address (lower cased). This
differs from &$parent_domain$& only when there is more than one level of
aliasing or forwarding. When more than one address is being delivered in a
single transport run, &$original_domain$& is not set.

If a new address is created by means of a &%deliver%& command in a system
filter, it is set up with an artificial &"parent"& address. This has the local
part &'system-filter'& and the default qualify domain.

.tvar &$original_local_part$&
.vindex "&$local_part$&"
When a top-level address is being processed for delivery, this contains the
same value as &$local_part$&, unless a prefix or suffix was removed from the
local part, because &$original_local_part$& always contains the full local
part. When a &"child"& address (for example, generated by an alias, forward, or
filter file) is being processed, this variable contains the full local part of
the original address.

If the router that did the redirection processed the local part
case-insensitively, the value in &$original_local_part$& is in lower case.
This variable differs from &$parent_local_part$& only when there is more than
one level of aliasing or forwarding. When more than one address is being
delivered in a single transport run, &$original_local_part$& is not set.

If a new address is created by means of a &%deliver%& command in a system
filter, it is set up with an artificial &"parent"& address. This has the local
part &'system-filter'& and the default qualify domain.

.vitem &$originator_gid$&
.cindex "gid (group id)" "of originating user"
.cindex "sender" "gid"
.vindex "&$caller_gid$&"
.vindex "&$originator_gid$&"
This variable contains the value of &$caller_gid$& that was set when the
message was received. For messages received via the command line, this is the
gid of the sending user. For messages received by SMTP over TCP/IP, this is
normally the gid of the Exim user.

.vitem &$originator_uid$&
.cindex "uid (user id)" "of originating user"
.cindex "sender" "uid"
.vindex "&$caller_uid$&"
.vindex "&$originator_uid$&"
The value of &$caller_uid$& that was set when the message was received. For
messages received via the command line, this is the uid of the sending user.
For messages received by SMTP over TCP/IP, this is normally the uid of the Exim
user.

.tvar &$parent_domain$&
This variable is similar to &$original_domain$& (see
above), except that it refers to the immediately preceding parent address.

.tvar &$parent_local_part$&
This variable is similar to &$original_local_part$&
(see above), except that it refers to the immediately preceding parent address.

.vitem &$pid$&
.cindex "pid (process id)" "of current process"
.vindex "&$pid$&"
This variable contains the current process id.

.vitem &$pipe_addresses$&
.cindex "filter" "transport filter"
.cindex "transport" "filter"
.vindex "&$pipe_addresses$&"
This is not an expansion variable, but is mentioned here because the string
&`$pipe_addresses`& is handled specially in the command specification for the
&(pipe)& transport (chapter &<<CHAPpipetransport>>&) and in transport filters
(described under &%transport_filter%& in chapter &<<CHAPtransportgeneric>>&).
It cannot be used in general expansion strings, and provokes an &"unknown
variable"& error if encountered.
&*Note*&: This value permits data supplied by a potential attacker to
be used in the command for a &(pipe)& transport.
Such configurations should be carefully assessed for security vulnerbilities.

.vitem &$primary_hostname$&
.vindex "&$primary_hostname$&"
This variable contains the value set by &%primary_hostname%& in the
configuration file, or read by the &[uname()]& function. If &[uname()]& returns
a single-component name, Exim calls &[gethostbyname()]& (or
&[getipnodebyname()]& where available) in an attempt to acquire a fully
qualified host name. See also &$smtp_active_hostname$&.


.vitem &$proxy_external_address$& &&&
       &$proxy_external_port$& &&&
       &$proxy_local_address$& &&&
       &$proxy_local_port$& &&&
       &$proxy_session$&
These variables are only available when built with Proxy Protocol
or SOCKS5 support.
For details see chapter &<<SECTproxyInbound>>&.

.vitem &$prdr_requested$&
.cindex "PRDR" "variable for"
This variable is set to &"yes"& if PRDR was requested by the client for the
current message, otherwise &"no"&.

.vitem &$prvscheck_address$& &&&
       &$prvscheck_keynum$& &&&
       &$prvscheck_result$&
These variables are used in conjunction with the &%prvscheck%& expansion item,
which is described in sections &<<SECTexpansionitems>>& and
&<<SECTverifyPRVS>>&.

.vitem &$qualify_domain$&
.vindex "&$qualify_domain$&"
The value set for the &%qualify_domain%& option in the configuration file.

.vitem &$qualify_recipient$&
.vindex "&$qualify_recipient$&"
The value set for the &%qualify_recipient%& option in the configuration file,
or if not set, the value of &$qualify_domain$&.

.vitem &$queue_name$&
.vindex &$queue_name$&
.cindex "named queues" variable
.cindex queues named
The name of the spool queue in use; empty for the default queue.

.vitem &$queue_size$&
.vindex "&$queue_size$&"
.cindex "queue" "size of"
.cindex "spool" "number of messages"
This variable contains the number of messages queued.
It is evaluated on demand, but no more often than once every minute.
If there is no daemon notifier socket open, the value will be
an empty string.

.vitem &$r_...$&
.vindex &$r_...$&
.cindex router variables
Values can be placed in these variables by the &%set%& option of a router.
They can be given any name that starts with &$r_$&.
The values persist for the address being handled through subsequent routers
and the eventual transport.

.vitem &$rcpt_count$&
.vindex "&$rcpt_count$&"
When a message is being received by SMTP, this variable contains the number of
RCPT commands received for the current message. If this variable is used in a
RCPT ACL, its value includes the current command.

.vitem &$rcpt_defer_count$&
.vindex "&$rcpt_defer_count$&"
.cindex "4&'xx'& responses" "count of"
When a message is being received by SMTP, this variable contains the number of
RCPT commands in the current message that have previously been rejected with a
temporary (4&'xx'&) response.

.vitem &$rcpt_fail_count$&
.vindex "&$rcpt_fail_count$&"
When a message is being received by SMTP, this variable contains the number of
RCPT commands in the current message that have previously been rejected with a
permanent (5&'xx'&) response.

.vitem &$received_count$&
.vindex "&$received_count$&"
This variable contains the number of &'Received:'& header lines in the message,
including the one added by Exim (so its value is always greater than zero). It
is available in the DATA ACL, the non-SMTP ACL, and while routing and
delivering.

.tvar &$received_for$&
If there is only a single recipient address in an incoming message, this
variable contains that address when the &'Received:'& header line is being
built. The value is copied after recipient rewriting has happened, but before
the &[local_scan()]& function is run.

.vitem &$received_ip_address$& &&&
       &$received_port$&
.vindex "&$received_ip_address$&"
.vindex "&$received_port$&"
As soon as an Exim server starts processing an incoming TCP/IP connection, these
variables are set to the address and port on the local IP interface.
(The remote IP address and port are in
&$sender_host_address$& and &$sender_host_port$&.) When testing with &%-bh%&,
the port value is -1 unless it has been set using the &%-oMi%& command line
option.

As well as being useful in ACLs (including the &"connect"& ACL), these variable
could be used, for example, to make the filename for a TLS certificate depend
on which interface and/or port is being used for the incoming connection. The
values of &$received_ip_address$& and &$received_port$& are saved with any
messages that are received, thus making these variables available at delivery
time.
For outbound connections see &$sending_ip_address$&.

.vitem &$received_protocol$&
.vindex "&$received_protocol$&"
When a message is being processed, this variable contains the name of the
protocol by which it was received. Most of the names used by Exim are defined
by RFCs 821, 2821, and 3848. They start with &"smtp"& (the client used HELO) or
&"esmtp"& (the client used EHLO). This can be followed by &"s"& for secure
(encrypted) and/or &"a"& for authenticated. Thus, for example, if the protocol
is set to &"esmtpsa"&, the message was received over an encrypted SMTP
connection and the client was successfully authenticated.

Exim uses the protocol name &"smtps"& for the case when encryption is
automatically set up on connection without the use of STARTTLS (see
&%tls_on_connect_ports%&), and the client uses HELO to initiate the
encrypted SMTP session. The name &"smtps"& is also used for the rare situation
where the client initially uses EHLO, sets up an encrypted connection using
STARTTLS, and then uses HELO afterwards.

The &%-oMr%& option provides a way of specifying a custom protocol name for
messages that are injected locally by trusted callers. This is commonly used to
identify messages that are being re-injected after some kind of scanning.

.vitem &$received_time$&
.vindex "&$received_time$&"
This variable contains the date and time when the current message was received,
as a number of seconds since the start of the Unix epoch.

.vitem &$recipient_data$&
.vindex "&$recipient_data$&"
This variable is set after an indexing lookup success in an ACL &%recipients%&
condition. It contains the data from the lookup, and the value remains set
until the next &%recipients%& test. Thus, you can do things like this:
.display
&`require recipients  = cdb*@;/some/file`&
&`deny    `&&'some further test involving'& &`$recipient_data`&
.endd
&*Warning*&: This variable is set only when a lookup is used as an indexing
method in the address list, using the semicolon syntax as in the example above.
The variable is not set for a lookup that is used as part of the string
expansion that all such lists undergo before being interpreted.

.vitem &$recipient_verify_failure$&
.vindex "&$recipient_verify_failure$&"
In an ACL, when a recipient verification fails, this variable contains
information about the failure. It is set to one of the following words:

.ilist
&"qualify"&: The address was unqualified (no domain), and the message
was neither local nor came from an exempted host.

.next
&"route"&: Routing failed.

.next
&"mail"&: Routing succeeded, and a callout was attempted; rejection occurred at
or before the MAIL command (that is, on initial connection, HELO, or
MAIL).

.next
&"recipient"&: The RCPT command in a callout was rejected.
.next

&"postmaster"&: The postmaster check in a callout was rejected.
.endlist

The main use of this variable is expected to be to distinguish between
rejections of MAIL and rejections of RCPT.

.tvar &$recipients$&
.tvar &$recipients_list$&
These variables both contain the envelope recipients for a message.

The first uses a comma and a space separate the addresses in the replacement text.
&*Note*&: an address can legitimately contain a comma;
this variable is not intended for further processing.

The second is a proper Exim list; colon-separated.

However, the variables
are not generally available, to prevent exposure of Bcc recipients in
unprivileged users' filter files. You can use either of them  only in these
cases:

.olist
In a system filter file.
.next
In the ACLs associated with the DATA command and with non-SMTP messages, that
is, the ACLs defined by &%acl_smtp_predata%&, &%acl_smtp_data%&,
&%acl_smtp_mime%&, &%acl_not_smtp_start%&, &%acl_not_smtp%&, and
&%acl_not_smtp_mime%&.
.next
From within a &[local_scan()]& function.
.endlist


.vitem &$recipients_count$&
.vindex "&$recipients_count$&"
When a message is being processed, this variable contains the number of
envelope recipients that came with the message. Duplicates are not excluded
from the count. While a message is being received over SMTP, the number
increases for each accepted recipient. It can be referenced in an ACL.


.vitem &$regex_match_string$&
.vindex "&$regex_match_string$&"
This variable is set to contain the matching regular expression after a
&%regex%& ACL condition has matched (see section &<<SECTscanregex>>&).

.vitem "&$regex1$&, &$regex2$&, etc"
.cindex "regex submatch variables (&$1regex$& &$2regex$& etc)"
When a &%regex%& or &%mime_regex%& ACL condition succeeds,
these variables contain the
captured substrings identified by the regular expression.
If the subject string was tainted then so will any captured substring.


.tvar &$reply_address$&
When a message is being processed, this variable contains the contents of the
&'Reply-To:'& header line if one exists and it is not empty, or otherwise the
contents of the &'From:'& header line. Apart from the removal of leading
white space, the value is not processed in any way. In particular, no RFC 2047
decoding or character code translation takes place.

.vitem &$return_path$&
.vindex "&$return_path$&"
When a message is being delivered, this variable contains the return path &--
the sender field that will be sent as part of the envelope. It is not enclosed
in <> characters. At the start of routing an address, &$return_path$& has the
same value as &$sender_address$&, but if, for example, an incoming message to a
mailing list has been expanded by a router which specifies a different address
for bounce messages, &$return_path$& subsequently contains the new bounce
address, whereas &$sender_address$& always contains the original sender address
that was received with the message. In other words, &$sender_address$& contains
the incoming envelope sender, and &$return_path$& contains the outgoing
envelope sender.

.vitem &$return_size_limit$&
.vindex "&$return_size_limit$&"
This is an obsolete name for &$bounce_return_size_limit$&.

.vitem &$router_name$&
.cindex "router" "name"
.cindex "name" "of router"
.vindex "&$router_name$&"
During the running of a router, or a transport called,
this variable contains the router name.

.vitem &$runrc$&
.cindex "return code" "from &%run%& expansion"
.vindex "&$runrc$&"
This variable contains the return code from a command that is run by the
&%${run...}%& expansion item. &*Warning*&: In a router or transport, you cannot
assume the order in which option values are expanded, except for those
preconditions whose order of testing is documented. Therefore, you cannot
reliably expect to set &$runrc$& by the expansion of one option, and use it in
another.

.vitem &$self_hostname$&
.oindex "&%self%&" "value of host name"
.vindex "&$self_hostname$&"
When an address is routed to a supposedly remote host that turns out to be the
local host, what happens is controlled by the &%self%& generic router option.
One of its values causes the address to be passed to another router. When this
happens, &$self_hostname$& is set to the name of the local host that the
original router encountered. In other circumstances its contents are null.

.tvar &$sender_address$&
When a message is being processed, this variable contains the sender's address
that was received in the message's envelope. The case of letters in the address
is retained, in both the local part and the domain. For bounce messages, the
value of this variable is the empty string. See also &$return_path$&.

.vitem &$sender_address_data$&
.vindex "&$address_data$&"
.vindex "&$sender_address_data$&"
If &$address_data$& is set when the routers are called from an ACL to verify a
sender address, the final value is preserved in &$sender_address_data$&, to
distinguish it from data from a recipient address. The value does not persist
after the end of the current ACL statement. If you want to preserve it for
longer, you can save it in an ACL variable.

.tvar &$sender_address_domain$&
The domain portion of &$sender_address$&.

.tvar &$sender_address_local_part$&
The local part portion of &$sender_address$&.

.vitem &$sender_data$&
.vindex "&$sender_data$&"
This variable is set after a lookup success in an ACL &%senders%& condition or
in a router &%senders%& option. It contains the data from the lookup, and the
value remains set until the next &%senders%& test. Thus, you can do things like
this:
.display
&`require senders      = cdb*@;/some/file`&
&`deny    `&&'some further test involving'& &`$sender_data`&
.endd
&*Warning*&: This variable is set only when a lookup is used as an indexing
method in the address list, using the semicolon syntax as in the example above.
The variable is not set for a lookup that is used as part of the string
expansion that all such lists undergo before being interpreted.

.vitem &$sender_fullhost$&
.vindex "&$sender_fullhost$&"
When a message is received from a remote host, this variable contains the host
name and IP address in a single string. It ends with the IP address in square
brackets, followed by a colon and a port number if the logging of ports is
enabled. The format of the rest of the string depends on whether the host
issued a HELO or EHLO SMTP command, and whether the host name was verified by
looking up its IP address. (Looking up the IP address can be forced by the
&%host_lookup%& option, independent of verification.) A plain host name at the
start of the string is a verified host name; if this is not present,
verification either failed or was not requested. A host name in parentheses is
the argument of a HELO or EHLO command. This is omitted if it is identical to
the verified host name or to the host's IP address in square brackets.

.vitem &$sender_helo_dnssec$&
.vindex "&$sender_helo_dnssec$&"
This boolean variable is true if a successful HELO verification was
.cindex "DNS" "DNSSEC"
done using DNS information the resolver library stated was authenticated data.

.tvar &$sender_helo_name$&
When a message is received from a remote host that has issued a HELO or EHLO
command, the argument of that command is placed in this variable. It is also
set if HELO or EHLO is used when a message is received using SMTP locally via
the &%-bs%& or &%-bS%& options.

.vitem &$sender_host_address$&
.vindex "&$sender_host_address$&"
When a message is received from a remote host using SMTP,
this variable contains that
host's IP address. For locally non-SMTP submitted messages, it is empty.

.vitem &$sender_host_authenticated$&
.vindex "&$sender_host_authenticated$&"
This variable contains the name (not the public name) of the authenticator
driver that successfully authenticated the client from which the message was
received. It is empty if there was no successful authentication. See also
&$authenticated_id$&.

.vitem &$sender_host_dnssec$&
.vindex "&$sender_host_dnssec$&"
If an attempt to populate &$sender_host_name$& has been made
(by reference, &%hosts_lookup%& or
otherwise) then this boolean will have been set true if, and only if, the
resolver library states that both
the reverse and forward DNS were authenticated data.  At all
other times, this variable is false.

.cindex "DNS" "DNSSEC"
It is likely that you will need to coerce DNSSEC support on in the resolver
library, by setting:
.code
dns_dnssec_ok = 1
.endd

In addition, on Linux with glibc 2.31 or newer the resolver library will
default to stripping out a successful validation status.
This will break a previously working Exim installation.
Provided that you do trust the resolver (ie, is on localhost) you can tell
glibc to pass through any successful validation with a new option in
&_/etc/resolv.conf_&:
.code
options trust-ad
.endd

Exim does not perform DNSSEC validation itself, instead leaving that to a
validating resolver (e.g. unbound, or bind with suitable configuration).

If you have changed &%host_lookup_order%& so that &`bydns`& is not the first
mechanism in the list, then this variable will be false.

This requires that your system resolver library support EDNS0 (and that
DNSSEC flags exist in the system headers).  If the resolver silently drops
all EDNS0 options, then this will have no effect.  OpenBSD's asr resolver
is known to currently ignore EDNS0, documented in CAVEATS of asr_run(3).


.tvar &$sender_host_name$&
When a message is received from a remote host, this variable contains the
host's name as obtained by looking up its IP address. For messages received by
other means, this variable is empty.

.vindex "&$host_lookup_failed$&"
If the host name has not previously been looked up, a reference to
&$sender_host_name$& triggers a lookup (for messages from remote hosts).
A looked up name is accepted only if it leads back to the original IP address
via a forward lookup. If either the reverse or the forward lookup fails to find
any data, or if the forward lookup does not yield the original IP address,
&$sender_host_name$& remains empty, and &$host_lookup_failed$& is set to &"1"&.

.vindex "&$host_lookup_deferred$&"
However, if either of the lookups cannot be completed (for example, there is a
DNS timeout), &$host_lookup_deferred$& is set to &"1"&, and
&$host_lookup_failed$& remains set to &"0"&.

Once &$host_lookup_failed$& is set to &"1"&, Exim does not try to look up the
host name again if there is a subsequent reference to &$sender_host_name$&
in the same Exim process, but it does try again if &$host_lookup_deferred$&
is set to &"1"&.

Exim does not automatically look up every calling host's name. If you want
maximum efficiency, you should arrange your configuration so that it avoids
these lookups altogether. The lookup happens only if one or more of the
following are true:

.ilist
A string containing &$sender_host_name$& is expanded.
.next
The calling host matches the list in &%host_lookup%&. In the default
configuration, this option is set to *, so it must be changed if lookups are
to be avoided. (In the code, the default for &%host_lookup%& is unset.)
.next
Exim needs the host name in order to test an item in a host list. The items
that require this are described in sections &<<SECThoslispatnam>>& and
&<<SECThoslispatnamsk>>&.
.next
The calling host matches &%helo_try_verify_hosts%& or &%helo_verify_hosts%&.
In this case, the host name is required to compare with the name quoted in any
EHLO or HELO commands that the client issues.
.next
The remote host issues a EHLO or HELO command that quotes one of the
domains in &%helo_lookup_domains%&. The default value of this option is
. ==== As this is a nested list, any displays it contains must be indented
. ==== as otherwise they are too far to the left.
.code
  helo_lookup_domains = @ : @[]
.endd
which causes a lookup if a remote host (incorrectly) gives the server's name or
IP address in an EHLO or HELO command.
.endlist


.vitem &$sender_host_port$&
.vindex "&$sender_host_port$&"
When a message is received from a remote host, this variable contains the port
number that was used on the remote host.

.vitem &$sender_ident$&
.vindex "&$sender_ident$&"
When a message is received from a remote host, this variable contains the
identification received in response to an RFC 1413 request. When a message has
been received locally, this variable contains the login name of the user that
called Exim.

.vitem &$sender_rate_$&&'xxx'&
A number of variables whose names begin &$sender_rate_$& are set as part of the
&%ratelimit%& ACL condition. Details are given in section
&<<SECTratelimiting>>&.

.vitem &$sender_rcvhost$&
.cindex "DNS" "reverse lookup"
.cindex "reverse DNS lookup"
.vindex "&$sender_rcvhost$&"
This is provided specifically for use in &'Received:'& headers. It starts with
either the verified host name (as obtained from a reverse DNS lookup) or, if
there is no verified host name, the IP address in square brackets. After that
there may be text in parentheses. When the first item is a verified host name,
the first thing in the parentheses is the IP address in square brackets,
followed by a colon and a port number if port logging is enabled. When the
first item is an IP address, the port is recorded as &"port=&'xxxx'&"& inside
the parentheses.

There may also be items of the form &"helo=&'xxxx'&"& if HELO or EHLO
was used and its argument was not identical to the real host name or IP
address, and &"ident=&'xxxx'&"& if an RFC 1413 ident string is available. If
all three items are present in the parentheses, a newline and tab are inserted
into the string, to improve the formatting of the &'Received:'& header.

.vitem &$sender_verify_failure$&
.vindex "&$sender_verify_failure$&"
In an ACL, when a sender verification fails, this variable contains information
about the failure. The details are the same as for
&$recipient_verify_failure$&.

.vitem &$sending_ip_address$&
.vindex "&$sending_ip_address$&"
This variable is set whenever an outgoing SMTP connection to another host has
been set up. It contains the IP address of the local interface that is being
used. This is useful if a host that has more than one IP address wants to take
on different personalities depending on which one is being used. For incoming
connections, see &$received_ip_address$&.

.vitem &$sending_port$&
.vindex "&$sending_port$&"
This variable is set whenever an outgoing SMTP connection to another host has
been set up. It contains the local port that is being used. For incoming
connections, see &$received_port$&.

.vitem &$smtp_active_hostname$&
.vindex "&$smtp_active_hostname$&"
During an incoming SMTP session, this variable contains the value of the active
host name, as specified by the &%smtp_active_hostname%& option. The value of
&$smtp_active_hostname$& is saved with any message that is received, so its
value can be consulted during routing and delivery.

.tvar &$smtp_command$&
During the processing of an incoming SMTP command, this variable contains the
entire command. This makes it possible to distinguish between HELO and EHLO in
the HELO ACL, and also to distinguish between commands such as these:
.code
MAIL FROM:<>
MAIL FROM: <>
.endd
For a MAIL command, extra parameters such as SIZE can be inspected. For a RCPT
command, the address in &$smtp_command$& is the original address before any
rewriting, whereas the values in &$local_part$& and &$domain$& are taken from
the address after SMTP-time rewriting.

.tvar &$smtp_command_argument$&
.cindex "SMTP" "command, argument for"
While an ACL is running to check an SMTP command, this variable contains the
argument, that is, the text that follows the command name, with leading white
space removed. Following the introduction of &$smtp_command$&, this variable is
somewhat redundant, but is retained for backwards compatibility.

.vitem &$smtp_command_history$&
.cindex SMTP "command history"
.vindex "&$smtp_command_history$&"
A comma-separated list (with no whitespace) of the most-recent SMTP commands
received, in time-order left to right.  Only a limited number of commands
are remembered.

.vitem &$smtp_count_at_connection_start$&
.vindex "&$smtp_count_at_connection_start$&"
This variable is set greater than zero only in processes spawned by the Exim
daemon for handling incoming SMTP connections. The name is deliberately long,
in order to emphasize what the contents are. When the daemon accepts a new
connection, it increments this variable. A copy of the variable is passed to
the child process that handles the connection, but its value is fixed, and
never changes. It is only an approximation of how many incoming connections
there actually are, because many other connections may come and go while a
single connection is being processed. When a child process terminates, the
daemon decrements its copy of the variable.

.vitem &$smtp_notquit_reason$&
.vindex "&$smtp_notquit_reason$&"
When the not-QUIT ACL is running, this variable is set to a string
that indicates the reason for the termination of the SMTP connection.

.vitem "&$sn0$& &-- &$sn9$&"
These variables are copies of the values of the &$n0$& &-- &$n9$& accumulators
that were current at the end of the system filter file. This allows a system
filter file to set values that can be tested in users' filter files. For
example, a system filter could set a value indicating how likely it is that a
message is junk mail.

.vitem &$spam_score$& &&&
       &$spam_score_int$& &&&
       &$spam_bar$& &&&
       &$spam_report$& &&&
       &$spam_action$&
A number of variables whose names start with &$spam$& are available when Exim
is compiled with the content-scanning extension. For details, see section
&<<SECTscanspamass>>&.

.vitem &$spf_header_comment$& &&&
       &$spf_received$& &&&
       &$spf_result$& &&&
       &$spf_result_guessed$& &&&
       &$spf_smtp_comment$&
These variables are only available if Exim is built with SPF support.
For details see section &<<SECSPF>>&.

.vitem &$spool_directory$&
.vindex "&$spool_directory$&"
The name of Exim's spool directory.

.vitem &$spool_inodes$&
.vindex "&$spool_inodes$&"
The number of free inodes in the disk partition where Exim's spool files are
being written. The value is recalculated whenever the variable is referenced.
If the relevant file system does not have the concept of inodes, the value of
is -1. See also the &%check_spool_inodes%& option.

.vitem &$spool_space$&
.vindex "&$spool_space$&"
The amount of free space (as a number of kilobytes) in the disk partition where
Exim's spool files are being written. The value is recalculated whenever the
variable is referenced. If the operating system does not have the ability to
find the amount of free space (only true for experimental systems), the space
value is -1. For example, to check in an ACL that there is at least 50
megabytes free on the spool, you could write:
.code
condition = ${if > {$spool_space}{50000}}
.endd
See also the &%check_spool_space%& option.


.vitem &$thisaddress$&
.vindex "&$thisaddress$&"
This variable is set only during the processing of the &%foranyaddress%&
command in a filter file. Its use is explained in the description of that
command, which can be found in the separate document entitled &'Exim's
interfaces to mail filtering'&.

.vitem &$tls_in_bits$&
.vindex "&$tls_in_bits$&"
Contains an approximation of the TLS cipher's bit-strength
on the inbound connection; the meaning of
this depends upon the TLS implementation used.
If TLS has not been negotiated, the value will be 0.
The value of this is automatically fed into the Cyrus SASL authenticator
when acting as a server, to specify the "external SSF" (a SASL term).

The deprecated &$tls_bits$& variable refers to the inbound side
except when used in the context of an outbound SMTP delivery, when it refers to
the outbound.

.vitem &$tls_out_bits$&
.vindex "&$tls_out_bits$&"
Contains an approximation of the TLS cipher's bit-strength
on an outbound SMTP connection; the meaning of
this depends upon the TLS implementation used.
If TLS has not been negotiated, the value will be 0.

.vitem &$tls_in_ourcert$&
.vindex "&$tls_in_ourcert$&"
.cindex certificate variables
This variable refers to the certificate presented to the peer of an
inbound connection when the message was received.
It is only useful as the argument of a
&%certextract%& expansion item, &%md5%&, &%sha1%& or &%sha256%& operator,
or a &%def%& condition.

&*Note*&: Under versions of OpenSSL preceding 1.1.1,
when a list of more than one
file is used for &%tls_certificate%&, this variable is not reliable.
The macro "_TLS_BAD_MULTICERT_IN_OURCERT" will be defined for those versions.

.vitem &$tls_in_peercert$&
.vindex "&$tls_in_peercert$&"
This variable refers to the certificate presented by the peer of an
inbound connection when the message was received.
It is only useful as the argument of a
&%certextract%& expansion item, &%md5%&, &%sha1%& or &%sha256%& operator,
or a &%def%& condition.
If certificate verification fails it may refer to a failing chain element
which is not the leaf.

.vitem &$tls_out_ourcert$&
.vindex "&$tls_out_ourcert$&"
This variable refers to the certificate presented to the peer of an
outbound connection.  It is only useful as the argument of a
&%certextract%& expansion item, &%md5%&, &%sha1%& or &%sha256%& operator,
or a &%def%& condition.

.vitem &$tls_out_peercert$&
.vindex "&$tls_out_peercert$&"
This variable refers to the certificate presented by the peer of an
outbound connection.  It is only useful as the argument of a
&%certextract%& expansion item, &%md5%&, &%sha1%& or &%sha256%& operator,
or a &%def%& condition.
If certificate verification fails it may refer to a failing chain element
which is not the leaf.

.vitem &$tls_in_certificate_verified$&
.vindex "&$tls_in_certificate_verified$&"
This variable is set to &"1"& if a TLS certificate was verified when the
message was received, and &"0"& otherwise.

The deprecated &$tls_certificate_verified$& variable refers to the inbound side
except when used in the context of an outbound SMTP delivery, when it refers to
the outbound.

.vitem &$tls_out_certificate_verified$&
.vindex "&$tls_out_certificate_verified$&"
This variable is set to &"1"& if a TLS certificate was verified when an
outbound SMTP connection was made,
and &"0"& otherwise.

.vitem &$tls_in_cipher$&
.vindex "&$tls_in_cipher$&"
.vindex "&$tls_cipher$&"
When a message is received from a remote host over an encrypted SMTP
connection, this variable is set to the cipher suite that was negotiated, for
example DES-CBC3-SHA. In other circumstances, in particular, for message
received over unencrypted connections, the variable is empty. Testing
&$tls_in_cipher$& for emptiness is one way of distinguishing between encrypted and
non-encrypted connections during ACL processing.

The deprecated &$tls_cipher$& variable is the same as &$tls_in_cipher$& during message reception,
but in the context of an outward SMTP delivery taking place via the &(smtp)& transport
becomes the same as &$tls_out_cipher$&.

.vitem &$tls_in_cipher_std$&
.vindex "&$tls_in_cipher_std$&"
As above, but returning the RFC standard name for the cipher suite.

.vitem &$tls_out_cipher$&
.vindex "&$tls_out_cipher$&"
This variable is
cleared before any outgoing SMTP connection is made,
and then set to the outgoing cipher suite if one is negotiated. See chapter
&<<CHAPTLS>>& for details of TLS support and chapter &<<CHAPsmtptrans>>& for
details of the &(smtp)& transport.

.vitem &$tls_out_cipher_std$&
.vindex "&$tls_out_cipher_std$&"
As above, but returning the RFC standard name for the cipher suite.

.vitem &$tls_out_dane$&
.vindex &$tls_out_dane$&
DANE active status.  See section &<<SECDANE>>&.

.vitem &$tls_in_ocsp$&
.vindex "&$tls_in_ocsp$&"
When a message is received from a remote client connection
the result of any OCSP request from the client is encoded in this variable:
.code
0 OCSP proof was not requested (default value)
1 No response to request
2 Response not verified
3 Verification failed
4 Verification succeeded
.endd

.vitem &$tls_out_ocsp$&
.vindex "&$tls_out_ocsp$&"
When a message is sent to a remote host connection
the result of any OCSP request made is encoded in this variable.
See &$tls_in_ocsp$& for values.

.vitem &$tls_in_peerdn$&
.vindex "&$tls_in_peerdn$&"
.vindex "&$tls_peerdn$&"
.cindex certificate "extracting fields"
When a message is received from a remote host over an encrypted SMTP
connection, and Exim is configured to request a certificate from the client,
the value of the Distinguished Name of the certificate is made available in the
&$tls_in_peerdn$& during subsequent processing.
If certificate verification fails it may refer to a failing chain element
which is not the leaf.

The deprecated &$tls_peerdn$& variable refers to the inbound side
except when used in the context of an outbound SMTP delivery, when it refers to
the outbound.

.vitem &$tls_out_peerdn$&
.vindex "&$tls_out_peerdn$&"
When a message is being delivered to a remote host over an encrypted SMTP
connection, and Exim is configured to request a certificate from the server,
the value of the Distinguished Name of the certificate is made available in the
&$tls_out_peerdn$& during subsequent processing.
If certificate verification fails it may refer to a failing chain element
which is not the leaf.


.vitem &$tls_in_resumption$& &&&
       &$tls_out_resumption$&
.vindex &$tls_in_resumption$&
.vindex &$tls_out_resumption$&
.cindex TLS resumption
Observability for TLS session resumption.  See &<<SECTresumption>>& for details.


.tvar &$tls_in_sni$&
.vindex "&$tls_sni$&"
.cindex "TLS" "Server Name Indication"
.cindex "TLS" SNI
.cindex SNI "observability on server"
When a TLS session is being established, if the client sends the Server
Name Indication extension, the value will be placed in this variable.
If the variable appears in &%tls_certificate%& then this option and
some others, described in &<<SECTtlssni>>&,
will be re-expanded early in the TLS session, to permit
a different certificate to be presented (and optionally a different key to be
used) to the client, based upon the value of the SNI extension.

The deprecated &$tls_sni$& variable refers to the inbound side
except when used in the context of an outbound SMTP delivery, when it refers to
the outbound.

.vitem &$tls_out_sni$&
.vindex "&$tls_out_sni$&"
.cindex "TLS" "Server Name Indication"
.cindex "TLS" SNI
.cindex SNI "observability in client"
During outbound
SMTP deliveries, this variable reflects the value of the &%tls_sni%& option on
the transport.

.vitem &$tls_out_tlsa_usage$&
.vindex &$tls_out_tlsa_usage$&
Bitfield of TLSA record types found.  See section &<<SECDANE>>&.

.vitem &$tls_in_ver$&
.vindex "&$tls_in_ver$&"
When a message is received from a remote host over an encrypted SMTP connection
this variable is set to the protocol version, eg &'TLS1.2'&.

.vitem &$tls_out_ver$&
.vindex "&$tls_out_ver$&"
When a message is being delivered to a remote host over an encrypted SMTP connection
this variable is set to the protocol version.


.vitem &$tod_bsdinbox$&
.vindex "&$tod_bsdinbox$&"
The time of day and the date, in the format required for BSD-style mailbox
files, for example: Thu Oct 17 17:14:09 1995.

.vitem &$tod_epoch$&
.vindex "&$tod_epoch$&"
The time and date as a number of seconds since the start of the Unix epoch.

.vitem &$tod_epoch_l$&
.vindex "&$tod_epoch_l$&"
The time and date as a number of microseconds since the start of the Unix epoch.

.vitem &$tod_full$&
.vindex "&$tod_full$&"
A full version of the time and date, for example: Wed, 16 Oct 1995 09:51:40
+0100. The timezone is always given as a numerical offset from UTC, with
positive values used for timezones that are ahead (east) of UTC, and negative
values for those that are behind (west).

.vitem &$tod_log$&
.vindex "&$tod_log$&"
The time and date in the format used for writing Exim's log files, for example:
1995-10-12 15:32:29, but without a timezone.

.vitem &$tod_logfile$&
.vindex "&$tod_logfile$&"
This variable contains the date in the format yyyymmdd. This is the format that
is used for datestamping log files when &%log_file_path%& contains the &`%D`&
flag.

.vitem &$tod_zone$&
.vindex "&$tod_zone$&"
This variable contains the numerical value of the local timezone, for example:
-0500.

.vitem &$tod_zulu$&
.vindex "&$tod_zulu$&"
This variable contains the UTC date and time in &"Zulu"& format, as specified
by ISO 8601, for example: 20030221154023Z.

.vitem &$transport_name$&
.cindex "transport" "name"
.cindex "name" "of transport"
.vindex "&$transport_name$&"
During the running of a transport, this variable contains its name.

.vitem &$value$&
.vindex "&$value$&"
This variable contains the result of an expansion lookup, extraction operation,
or external command, as described above. It is also used during a
&*reduce*& expansion.

.vitem &$verify_mode$&
.vindex "&$verify_mode$&"
While a router or transport is being run in verify mode or for cutthrough delivery,
contains "S" for sender-verification or "R" for recipient-verification.
Otherwise, empty.

.vitem &$version_number$&
.vindex "&$version_number$&"
The version number of Exim. Same as &$exim_version$&, may be overridden
by the &%exim_version%& main config option.

.vitem &$warn_message_delay$&
.vindex "&$warn_message_delay$&"
This variable is set only during the creation of a message warning about a
delivery delay. Details of its use are explained in section &<<SECTcustwarn>>&.

.vitem &$warn_message_recipients$&
.vindex "&$warn_message_recipients$&"
This variable is set only during the creation of a message warning about a
delivery delay. Details of its use are explained in section &<<SECTcustwarn>>&.
.endlist
.ecindex IIDstrexp



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Embedded Perl" "CHAPperl"
.scindex IIDperl "Perl" "calling from Exim"
Exim can be built to include an embedded Perl interpreter. When this is done,
Perl subroutines can be called as part of the string expansion process. To make
use of the Perl support, you need version 5.004 or later of Perl installed on
your system. To include the embedded interpreter in the Exim binary, include
the line
.code
EXIM_PERL = perl.o
.endd
in your &_Local/Makefile_& and then build Exim in the normal way.


.section "Setting up so Perl can be used" "SECID85"
.oindex "&%perl_startup%&"
Access to Perl subroutines is via a global configuration option called
&%perl_startup%& and an expansion string operator &%${perl ...}%&. If there is
no &%perl_startup%& option in the Exim configuration file then no Perl
interpreter is started and there is almost no overhead for Exim (since none of
the Perl library will be paged in unless used). If there is a &%perl_startup%&
option then the associated value is taken to be Perl code which is executed in
a newly created Perl interpreter.

The value of &%perl_startup%& is not expanded in the Exim sense, so you do not
need backslashes before any characters to escape special meanings. The option
should usually be something like
.code
perl_startup = do '/etc/exim.pl'
.endd
where &_/etc/exim.pl_& is Perl code which defines any subroutines you want to
use from Exim. Exim can be configured either to start up a Perl interpreter as
soon as it is entered, or to wait until the first time it is needed. Starting
the interpreter at the beginning ensures that it is done while Exim still has
its setuid privilege, but can impose an unnecessary overhead if Perl is not in
fact used in a particular run. Also, note that this does not mean that Exim is
necessarily running as root when Perl is called at a later time. By default,
the interpreter is started only when it is needed, but this can be changed in
two ways:

.ilist
.oindex "&%perl_at_start%&"
Setting &%perl_at_start%& (a boolean option) in the configuration requests
a startup when Exim is entered.
.next
The command line option &%-ps%& also requests a startup when Exim is entered,
overriding the setting of &%perl_at_start%&.
.endlist

There is also a command line option &%-pd%& (for delay) which suppresses the
initial startup, even if &%perl_at_start%& is set.

.ilist
.oindex "&%perl_taintmode%&"
.cindex "Perl" "taintmode"
To provide more security executing Perl code via the embedded Perl
interpreter, the &%perl_taintmode%& option can be set. This enables the
taint mode of the Perl interpreter. You are encouraged to set this
option to a true value. To avoid breaking existing installations, it
defaults to false.

&*Note*&: This is entirely separate from Exim's tainted-data tracking.


.section "Calling Perl subroutines" "SECID86"
When the configuration file includes a &%perl_startup%& option you can make use
of the string expansion item to call the Perl subroutines that are defined
by the &%perl_startup%& code. The operator is used in any of the following
forms:
.code
${perl{foo}}
${perl{foo}{argument}}
${perl{foo}{argument1}{argument2} ... }
.endd
which calls the subroutine &%foo%& with the given arguments. A maximum of eight
arguments may be passed. Passing more than this results in an expansion failure
with an error message of the form
.code
Too many arguments passed to Perl subroutine "foo" (max is 8)
.endd
The return value of the Perl subroutine is evaluated in a scalar context before
it is passed back to Exim to be inserted into the expanded string. If the
return value is &'undef'&, the expansion is forced to fail in the same way as
an explicit &"fail"& on an &%if%& or &%lookup%& item. If the subroutine aborts
by obeying Perl's &%die%& function, the expansion fails with the error message
that was passed to &%die%&.


.section "Calling Exim functions from Perl" "SECID87"
Within any Perl code called from Exim, the function &'Exim::expand_string()'&
is available to call back into Exim's string expansion function. For example,
the Perl code
.code
my $lp = Exim::expand_string('$local_part');
.endd
makes the current Exim &$local_part$& available in the Perl variable &$lp$&.
Note those are single quotes and not double quotes to protect against
&$local_part$& being interpolated as a Perl variable.

If the string expansion is forced to fail by a &"fail"& item, the result of
&'Exim::expand_string()'& is &%undef%&. If there is a syntax error in the
expansion string, the Perl call from the original expansion string fails with
an appropriate error message, in the same way as if &%die%& were used.

.cindex "debugging" "from embedded Perl"
.cindex "log" "writing from embedded Perl"
Two other Exim functions are available for use from within Perl code.
&'Exim::debug_write()'& writes a string to the standard error stream if Exim's
debugging is enabled. If you want a newline at the end, you must supply it.
&'Exim::log_write()'& writes a string to Exim's main log, adding a leading
timestamp. In this case, you should not supply a terminating newline.


.section "Use of standard output and error by Perl" "SECID88"
.cindex "Perl" "standard output and error"
You should not write to the standard error or output streams from within your
Perl code, as it is not defined how these are set up. In versions of Exim
before 4.50, it is possible for the standard output or error to refer to the
SMTP connection during message reception via the daemon. Writing to this stream
is certain to cause chaos. From Exim 4.50 onwards, the standard output and
error streams are connected to &_/dev/null_& in the daemon. The chaos is
avoided, but the output is lost.

.cindex "Perl" "use of &%warn%&"
The Perl &%warn%& statement writes to the standard error stream by default.
Calls to &%warn%& may be embedded in Perl modules that you use, but over which
you have no control. When Exim starts up the Perl interpreter, it arranges for
output from the &%warn%& statement to be written to the Exim main log. You can
change this by including appropriate Perl magic somewhere in your Perl code.
For example, to discard &%warn%& output completely, you need this:
.code
$SIG{__WARN__} = sub { };
.endd
Whenever a &%warn%& is obeyed, the anonymous subroutine is called. In this
example, the code for the subroutine is empty, so it does nothing, but you can
include any Perl code that you like. The text of the &%warn%& message is passed
as the first subroutine argument.
.ecindex IIDperl


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Starting the daemon and the use of network interfaces" &&&
         "CHAPinterfaces" &&&
         "Starting the daemon"
.cindex "daemon" "starting"
.cindex "interface" "listening"
.cindex "network interface"
.cindex "interface" "network"
.cindex "IP address" "for listening"
.cindex "daemon" "listening IP addresses"
.cindex "TCP/IP" "setting listening interfaces"
.cindex "TCP/IP" "setting listening ports"
A host that is connected to a TCP/IP network may have one or more physical
hardware network interfaces. Each of these interfaces may be configured as one
or more &"logical"& interfaces, which are the entities that a program actually
works with. Each of these logical interfaces is associated with an IP address.
In addition, TCP/IP software supports &"loopback"& interfaces (127.0.0.1 in
IPv4 and ::1 in IPv6), which do not use any physical hardware. Exim requires
knowledge about the host's interfaces for use in three different circumstances:

.olist
When a listening daemon is started, Exim needs to know which interfaces
and ports to listen on.
.next
When Exim is routing an address, it needs to know which IP addresses
are associated with local interfaces. This is required for the correct
processing of MX lists by removing the local host and others with the
same or higher priority values. Also, Exim needs to detect cases
when an address is routed to an IP address that in fact belongs to the
local host. Unless the &%self%& router option or the &%allow_localhost%&
option of the smtp transport is set (as appropriate), this is treated
as an error situation.
.next
When Exim connects to a remote host, it may need to know which interface to use
for the outgoing connection.
.endlist


Exim's default behaviour is likely to be appropriate in the vast majority
of cases. If your host has only one interface, and you want all its IP
addresses to be treated in the same way, and you are using only the
standard SMTP port, you should not need to take any special action. The
rest of this chapter does not apply to you.

In a more complicated situation you may want to listen only on certain
interfaces, or on different ports, and for this reason there are a number of
options that can be used to influence Exim's behaviour. The rest of this
chapter describes how they operate.

When a message is received over TCP/IP, the interface and port that were
actually used are set in &$received_ip_address$& and &$received_port$&.



.section "Starting a listening daemon" "SECID89"
When a listening daemon is started (by means of the &%-bd%& command line
option), the interfaces and ports on which it listens are controlled by the
following options:

.ilist
&%daemon_smtp_ports%& contains a list of default ports
or service names.
(For backward compatibility, this option can also be specified in the singular.)
.next
&%local_interfaces%& contains list of interface IP addresses on which to
listen. Each item may optionally also specify a port.
.endlist

The default list separator in both cases is a colon, but this can be changed as
described in section &<<SECTlistsepchange>>&. When IPv6 addresses are involved,
it is usually best to change the separator to avoid having to double all the
colons. For example:
.code
local_interfaces = <; 127.0.0.1 ; \
                      192.168.23.65 ; \
                      ::1 ; \
                      3ffe:ffff:836f::fe86:a061
.endd
There are two different formats for specifying a port along with an IP address
in &%local_interfaces%&:

.olist
The port is added onto the address with a dot separator. For example, to listen
on port 1234 on two different IP addresses:
.code
local_interfaces = <; 192.168.23.65.1234 ; \
                      3ffe:ffff:836f::fe86:a061.1234
.endd
.next
The IP address is enclosed in square brackets, and the port is added
with a colon separator, for example:
.code
local_interfaces = <; [192.168.23.65]:1234 ; \
                      [3ffe:ffff:836f::fe86:a061]:1234
.endd
.endlist

When a port is not specified, the value of &%daemon_smtp_ports%& is used. The
default setting contains just one port:
.code
daemon_smtp_ports = smtp
.endd
If more than one port is listed, each interface that does not have its own port
specified listens on all of them. Ports that are listed in
&%daemon_smtp_ports%& can be identified either by name (defined in
&_/etc/services_&) or by number. However, when ports are given with individual
IP addresses in &%local_interfaces%&, only numbers (not names) can be used.



.section "Special IP listening addresses" "SECID90"
The addresses 0.0.0.0 and ::0 are treated specially. They are interpreted
as &"all IPv4 interfaces"& and &"all IPv6 interfaces"&, respectively. In each
case, Exim tells the TCP/IP stack to &"listen on all IPv&'x'& interfaces"&
instead of setting up separate listening sockets for each interface. The
default value of &%local_interfaces%& is
.code
local_interfaces = 0.0.0.0
.endd
when Exim is built without IPv6 support; otherwise it is:
.code
local_interfaces = <; ::0 ; 0.0.0.0
.endd
Thus, by default, Exim listens on all available interfaces, on the SMTP port.



.section "Overriding local_interfaces and daemon_smtp_ports" "SECID91"
The &%-oX%& command line option can be used to override the values of
&%daemon_smtp_ports%& and/or &%local_interfaces%& for a particular daemon
instance. Another way of doing this would be to use macros and the &%-D%&
option. However, &%-oX%& can be used by any admin user, whereas modification of
the runtime configuration by &%-D%& is allowed only when the caller is root or
exim.

The value of &%-oX%& is a list of items. The default colon separator can be
changed in the usual way (&<<SECTlistsepchange>>&) if required.
If there are any items that do not
contain dots or colons (that is, are not IP addresses), the value of
&%daemon_smtp_ports%& is replaced by the list of those items. If there are any
items that do contain dots or colons, the value of &%local_interfaces%& is
replaced by those items. Thus, for example,
.code
-oX 1225
.endd
overrides &%daemon_smtp_ports%&, but leaves &%local_interfaces%& unchanged,
whereas
.code
-oX 192.168.34.5.1125
.endd
overrides &%local_interfaces%&, leaving &%daemon_smtp_ports%& unchanged.
(However, since &%local_interfaces%& now contains no items without ports, the
value of &%daemon_smtp_ports%& is no longer relevant in this example.)



.section "Support for the submissions (aka SSMTP or SMTPS) protocol" "SECTsupobssmt"
.cindex "submissions protocol"
.cindex "ssmtp protocol"
.cindex "smtps protocol"
.cindex "SMTP" "ssmtp protocol"
.cindex "SMTP" "smtps protocol"
Exim supports the use of TLS-on-connect, used by mail clients in the
&"submissions"& protocol, historically also known as SMTPS or SSMTP.
For some years, IETF Standards Track documents only blessed the
STARTTLS-based Submission service (port 587) while common practice was to support
the same feature set on port 465, but using TLS-on-connect.
If your installation needs to provide service to mail clients
(Mail User Agents, MUAs) then you should provide service on both the 587 and
the 465 TCP ports.

If the &%tls_on_connect_ports%& option is set to a list of port numbers or
service names, connections to those ports must first establish TLS, before
proceeding to the application layer use of the SMTP protocol.

The common use of this option is expected to be
.code
tls_on_connect_ports = 465
.endd
per &url(https://www.rfc-editor.org/rfc/rfc8314,RFC 8314).
There is also a command line option &%-tls-on-connect%&, which forces all ports
to behave in this way when a daemon is started.

&*Warning*&: Setting &%tls_on_connect_ports%& does not of itself cause the
daemon to listen on those ports. You must still specify them in
&%daemon_smtp_ports%&, &%local_interfaces%&, or the &%-oX%& option. (This is
because &%tls_on_connect_ports%& applies to &%inetd%& connections as well as to
connections via the daemon.)




.section "IPv6 address scopes" "SECID92"
.cindex "IPv6" "address scopes"
IPv6 addresses have &"scopes"&, and a host with multiple hardware interfaces
can, in principle, have the same link-local IPv6 address on different
interfaces. Thus, additional information is needed, over and above the IP
address, to distinguish individual interfaces. A convention of using a
percent sign followed by something (often the interface name) has been
adopted in some cases, leading to addresses like this:
.code
fe80::202:b3ff:fe03:45c1%eth0
.endd
To accommodate this usage, a percent sign followed by an arbitrary string is
allowed at the end of an IPv6 address. By default, Exim calls &[getaddrinfo()]&
to convert a textual IPv6 address for actual use. This function recognizes the
percent convention in operating systems that support it, and it processes the
address appropriately. Unfortunately, some older libraries have problems with
&[getaddrinfo()]&. If
.code
IPV6_USE_INET_PTON=yes
.endd
is set in &_Local/Makefile_& (or an OS-dependent Makefile) when Exim is built,
Exim uses &'inet_pton()'& to convert a textual IPv6 address for actual use,
instead of &[getaddrinfo()]&. (Before version 4.14, it always used this
function.) Of course, this means that the additional functionality of
&[getaddrinfo()]& &-- recognizing scoped addresses &-- is lost.

.section "Disabling IPv6" "SECID93"
.cindex "IPv6" "disabling"
Sometimes it happens that an Exim binary that was compiled with IPv6 support is
run on a host whose kernel does not support IPv6. The binary will fall back to
using IPv4, but it may waste resources looking up AAAA records, and trying to
connect to IPv6 addresses, causing delays to mail delivery. If you set the
.oindex "&%disable_ipv6%&"
&%disable_ipv6%& option true, even if the Exim binary has IPv6 support, no IPv6
activities take place. AAAA records are never looked up, and any IPv6 addresses
that are listed in &%local_interfaces%&, data for the &(manualroute)& router,
etc. are ignored. If IP literals are enabled, the &(ipliteral)& router declines
to handle IPv6 literal addresses.

On the other hand, when IPv6 is in use, there may be times when you want to
disable it for certain hosts or domains. You can use the &%dns_ipv4_lookup%&
option to globally suppress the lookup of AAAA records for specified domains,
and you can use the &%ignore_target_hosts%& generic router option to ignore
IPv6 addresses in an individual router.



.section "Examples of starting a listening daemon" "SECID94"
The default case in an IPv6 environment is
.code
daemon_smtp_ports = smtp
local_interfaces = <; ::0 ; 0.0.0.0
.endd
This specifies listening on the smtp port on all IPv6 and IPv4 interfaces.
Either one or two sockets may be used, depending on the characteristics of
the TCP/IP stack. (This is complicated and messy; for more information,
read the comments in the &_daemon.c_& source file.)

To specify listening on ports 25 and 26 on all interfaces:
.code
daemon_smtp_ports = 25 : 26
.endd
(leaving &%local_interfaces%& at the default setting) or, more explicitly:
.code
local_interfaces = <; ::0.25     ; ::0.26 \
                      0.0.0.0.25 ; 0.0.0.0.26
.endd
To listen on the default port on all IPv4 interfaces, and on port 26 on the
IPv4 loopback address only:
.code
local_interfaces = 0.0.0.0 : 127.0.0.1.26
.endd
To specify listening on the default port on specific interfaces only:
.code
local_interfaces = 10.0.0.67 : 192.168.34.67
.endd
&*Warning*&: Such a setting excludes listening on the loopback interfaces.



.section "Recognizing the local host" "SECTreclocipadd"
The &%local_interfaces%& option is also used when Exim needs to determine
whether or not an IP address refers to the local host. That is, the IP
addresses of all the interfaces on which a daemon is listening are always
treated as local.

For this usage, port numbers in &%local_interfaces%& are ignored. If either of
the items 0.0.0.0 or ::0 are encountered, Exim gets a complete list of
available interfaces from the operating system, and extracts the relevant
(that is, IPv4 or IPv6) addresses to use for checking.

Some systems set up large numbers of virtual interfaces in order to provide
many virtual web servers. In this situation, you may want to listen for
email on only a few of the available interfaces, but nevertheless treat all
interfaces as local when routing. You can do this by setting
&%extra_local_interfaces%& to a list of IP addresses, possibly including the
&"all"& wildcard values. These addresses are recognized as local, but are not
used for listening. Consider this example:
.code
local_interfaces = <; 127.0.0.1 ; ::1 ; \
                      192.168.53.235 ; \
                      3ffe:2101:12:1:a00:20ff:fe86:a061

extra_local_interfaces = <; ::0 ; 0.0.0.0
.endd
The daemon listens on the loopback interfaces and just one IPv4 and one IPv6
address, but all available interface addresses are treated as local when
Exim is routing.

In some environments the local host name may be in an MX list, but with an IP
address that is not assigned to any local interface. In other cases it may be
desirable to treat other host names as if they referred to the local host. Both
these cases can be handled by setting the &%hosts_treat_as_local%& option.
This contains host names rather than IP addresses. When a host is referenced
during routing, either via an MX record or directly, it is treated as the local
host if its name matches &%hosts_treat_as_local%&, or if any of its IP
addresses match &%local_interfaces%& or &%extra_local_interfaces%&.



.section "Delivering to a remote host" "SECID95"
Delivery to a remote host is handled by the smtp transport. By default, it
allows the system's TCP/IP functions to choose which interface to use (if
there is more than one) when connecting to a remote host. However, the
&%interface%& option can be set to specify which interface is used. See the
description of the smtp transport in chapter &<<CHAPsmtptrans>>& for more
details.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Main configuration" "CHAPmainconfig"
.scindex IIDconfima "configuration file" "main section"
.scindex IIDmaiconf "main configuration"
The first part of the runtime configuration file contains three types of item:

.ilist
Macro definitions: These lines start with an upper case letter. See section
&<<SECTmacrodefs>>& for details of macro processing.
.next
Named list definitions: These lines start with one of the words &"domainlist"&,
&"hostlist"&, &"addresslist"&, or &"localpartlist"&. Their use is described in
section &<<SECTnamedlists>>&.
.next
Main configuration settings: Each setting occupies one line of the file
(with possible continuations). If any setting is preceded by the word
&"hide"&, the &%-bP%& command line option displays its value to admin users
only. See section &<<SECTcos>>& for a description of the syntax of these option
settings.
.endlist

This chapter specifies all the main configuration options, along with their
types and default values. For ease of finding a particular option, they appear
in alphabetical order in section &<<SECTalomo>>& below. However, because there
are now so many options, they are first listed briefly in functional groups, as
an aid to finding the name of the option you are looking for. Some options are
listed in more than one group.

.section "Miscellaneous" "SECID96"
.table2
.row &%add_environment%&             "environment variables"
.row &%bi_command%&                  "to run for &%-bi%& command line option"
.row &%debug_store%&                 "do extra internal checks"
.row &%disable_ipv6%&                "do no IPv6 processing"
.row &%keep_environment%&            "environment variables"
.row &%keep_malformed%&              "for broken files &-- should not happen"
.row &%localhost_number%&            "for unique message ids in clusters"
.row &%message_body_newlines%&       "retain newlines in &$message_body$&"
.row &%message_body_visible%&        "how much to show in &$message_body$&"
.row &%mua_wrapper%&                 "run in &""MUA wrapper""& mode"
.row &%print_topbitchars%&           "top-bit characters are printing"
.row &%spool_wireformat%&            "use wire-format spool data files when possible"
.row &%timezone%&                    "force time zone"
.endtable


.section "Exim parameters" "SECID97"
.table2
.row &%exim_group%&                  "override compiled-in value"
.row &%exim_path%&                   "override compiled-in value"
.row &%exim_user%&                   "override compiled-in value"
.row &%primary_hostname%&            "default from &[uname()]&"
.row &%split_spool_directory%&       "use multiple directories"
.row &%spool_directory%&             "override compiled-in value"
.endtable



.section "Privilege controls" "SECID98"
.table2
.row &%admin_groups%&                "groups that are Exim admin users"
.row &%commandline_checks_require_admin%& "require admin for various checks"
.row &%deliver_drop_privilege%&      "drop root for delivery processes"
.row &%local_from_check%&            "insert &'Sender:'& if necessary"
.row &%local_from_prefix%&           "for testing &'From:'& for local sender"
.row &%local_from_suffix%&           "for testing &'From:'& for local sender"
.row &%local_sender_retain%&         "keep &'Sender:'& from untrusted user"
.row &%never_users%&                 "do not run deliveries as these"
.row &%prod_requires_admin%&         "forced delivery requires admin user"
.row &%queue_list_requires_admin%&   "queue listing requires admin user"
.row &%trusted_groups%&              "groups that are trusted"
.row &%trusted_users%&               "users that are trusted"
.endtable



.section "Logging" "SECID99"
.table2
.row &%event_action%&                "custom logging"
.row &%hosts_connection_nolog%&      "exemption from connect logging"
.row &%log_file_path%&               "override compiled-in value"
.row &%log_selector%&                "set/unset optional logging"
.row &%log_timezone%&                "add timezone to log lines"
.row &%message_logs%&                "create per-message logs"
.row &%preserve_message_logs%&       "after message completion"
.row &%panic_coredump%&              "request coredump on fatal errors"
.row &%process_log_path%&            "for SIGUSR1 and &'exiwhat'&"
.row &%slow_lookup_log%&             "control logging of slow DNS lookups"
.row &%syslog_duplication%&          "controls duplicate log lines on syslog"
.row &%syslog_facility%&             "set syslog &""facility""& field"
.row &%syslog_pid%&                  "pid in syslog lines"
.row &%syslog_processname%&          "set syslog &""ident""& field"
.row &%syslog_timestamp%&            "timestamp syslog lines"
.row &%write_rejectlog%&             "control use of message log"
.endtable



.section "Frozen messages" "SECID100"
.table2
.row &%auto_thaw%&                   "sets time for retrying frozen messages"
.row &%freeze_tell%&                 "send message when freezing"
.row &%move_frozen_messages%&        "to another directory"
.row &%timeout_frozen_after%&        "keep frozen messages only so long"
.endtable



.section "Data lookups" "SECID101"
.table2
.row &%ibase_servers%&               "InterBase servers"
.row &%ldap_ca_cert_dir%&            "dir of CA certs to verify LDAP server's"
.row &%ldap_ca_cert_file%&           "file of CA certs to verify LDAP server's"
.row &%ldap_cert_file%&              "client cert file for LDAP"
.row &%ldap_cert_key%&               "client key file for LDAP"
.row &%ldap_cipher_suite%&           "TLS negotiation preference control"
.row &%ldap_default_servers%&        "used if no server in query"
.row &%ldap_require_cert%&           "action to take without LDAP server cert"
.row &%ldap_start_tls%&              "require TLS within LDAP"
.row &%ldap_version%&                "set protocol version"
.row &%lookup_open_max%&             "lookup files held open"
.row &%mysql_servers%&               "default MySQL servers"
.row &%oracle_servers%&              "Oracle servers"
.row &%pgsql_servers%&               "default PostgreSQL servers"
.row &%sqlite_lock_timeout%&         "as it says"
.endtable



.section "Message ids" "SECID102"
.table2
.row &%message_id_header_domain%&    "used to build &'Message-ID:'& header"
.row &%message_id_header_text%&      "ditto"
.endtable



.section "Embedded Perl Startup" "SECID103"
.table2
.row &%perl_at_start%&               "always start the interpreter"
.row &%perl_startup%&                "code to obey when starting Perl"
.row &%perl_taintmode%&              "enable taint mode in Perl"
.endtable



.section "Daemon" "SECID104"
.table2
.row &%daemon_smtp_ports%&           "default ports"
.row &%daemon_startup_retries%&      "number of times to retry"
.row &%daemon_startup_sleep%&        "time to sleep between tries"
.row &%extra_local_interfaces%&      "not necessarily listened on"
.row &%local_interfaces%&            "on which to listen, with optional ports"
.row &%notifier_socket%&             "override compiled-in value"
.row &%pid_file_path%&               "override compiled-in value"
.row &%queue_run_max%&               "maximum simultaneous queue runners"
.row &%smtp_backlog_monitor%&        "level to log listen backlog"
.endtable



.section "Resource control" "SECID105"
.table2
.row &%check_log_inodes%&            "before accepting a message"
.row &%check_log_space%&             "before accepting a message"
.row &%check_spool_inodes%&          "before accepting a message"
.row &%check_spool_space%&           "before accepting a message"
.row &%deliver_queue_load_max%&      "no queue deliveries if load high"
.row &%queue_only_load%&             "queue incoming if load high"
.row &%queue_only_load_latch%&       "don't re-evaluate load for each message"
.row &%queue_run_max%&               "maximum simultaneous queue runners"
.row &%remote_max_parallel%&         "parallel SMTP delivery per message"
.row &%smtp_accept_max%&             "simultaneous incoming connections"
.row &%smtp_accept_max_nonmail%&     "non-mail commands"
.row &%smtp_accept_max_nonmail_hosts%& "hosts to which the limit applies"
.row &%smtp_accept_max_per_connection%& "messages per connection"
.row &%smtp_accept_max_per_host%&    "connections from one host"
.row &%smtp_accept_queue%&           "queue mail if more connections"
.row &%smtp_accept_queue_per_connection%& "queue if more messages per &&&
                                           connection"
.row &%smtp_accept_reserve%&         "only reserve hosts if more connections"
.row &%smtp_check_spool_space%&      "from SIZE on MAIL command"
.row &%smtp_connect_backlog%&        "passed to TCP/IP stack"
.row &%smtp_load_reserve%&           "SMTP from reserved hosts if load high"
.row &%smtp_reserve_hosts%&          "these are the reserve hosts"
.endtable



.section "Policy controls" "SECID106"
.table2
.row &%acl_not_smtp%&                "ACL for non-SMTP messages"
.row &%acl_not_smtp_mime%&           "ACL for non-SMTP MIME parts"
.row &%acl_not_smtp_start%&          "ACL for start of non-SMTP message"
.row &%acl_smtp_atrn%&               "ACL for ATRN"
.row &%acl_smtp_auth%&               "ACL for AUTH"
.row &%acl_smtp_connect%&            "ACL for connection"
.row &%acl_smtp_data%&               "ACL for DATA"
.row &%acl_smtp_data_prdr%&          "ACL for DATA, per-recipient"
.row &%acl_smtp_dkim%&               "ACL for DKIM verification"
.row &%acl_smtp_etrn%&               "ACL for ETRN"
.row &%acl_smtp_expn%&               "ACL for EXPN"
.row &%acl_smtp_helo%&               "ACL for EHLO or HELO"
.row &%acl_smtp_mail%&               "ACL for MAIL"
.row &%acl_smtp_mailauth%&           "ACL for AUTH on MAIL command"
.row &%acl_smtp_mime%&               "ACL for MIME parts"
.row &%acl_smtp_notquit%&            "ACL for non-QUIT terminations"
.row &%acl_smtp_predata%&            "ACL for start of data"
.row &%acl_smtp_quit%&               "ACL for QUIT"
.row &%acl_smtp_rcpt%&               "ACL for RCPT"
.row &%acl_smtp_starttls%&           "ACL for STARTTLS"
.row &%acl_smtp_vrfy%&               "ACL for VRFY"
.row &%acl_smtp_wellknown%&          "ACL for WELLKNOWN"
.row &%av_scanner%&                  "specify virus scanner"
.row &%check_rfc2047_length%&        "check length of RFC 2047 &""encoded &&&
                                      words""&"
.row &%dns_cname_loops%&             "follow CNAMEs returned by resolver"
.row &%dns_csa_search_limit%&        "control CSA parent search depth"
.row &%dns_csa_use_reverse%&         "en/disable CSA IP reverse search"
.row &%header_maxsize%&              "total size of message header"
.row &%header_line_maxsize%&         "individual header line limit"
.row &%helo_accept_junk_hosts%&      "allow syntactic junk from these hosts"
.row &%helo_allow_chars%&            "allow illegal chars in HELO names"
.row &%helo_lookup_domains%&         "lookup hostname for these HELO names"
.row &%helo_try_verify_hosts%&       "HELO soft-checked for these hosts"
.row &%helo_verify_hosts%&           "HELO hard-checked for these hosts"
.row &%host_lookup%&                 "host name looked up for these hosts"
.row &%host_lookup_order%&           "order of DNS and local name lookups"
.row &%hosts_proxy%&                 "use proxy protocol for these hosts"
.row &%host_reject_connection%&      "reject connection from these hosts"
.row &%hosts_treat_as_local%&        "useful in some cluster configurations"
.row &%local_scan_timeout%&          "timeout for &[local_scan()]&"
.row &%message_size_limit%&          "for all messages"
.row &%percent_hack_domains%&        "recognize %-hack for these domains"
.row &%proxy_protocol_timeout%&      "timeout for proxy protocol negotiation"
.row &%spamd_address%&               "set interface to SpamAssassin"
.row &%strict_acl_vars%&             "object to unset ACL variables"
.row &%spf_smtp_comment_template%&   "template for &$spf_smtp_comment$&"
.endtable



.section "Callout cache" "SECID107"
.table2
.row &%callout_domain_negative_expire%& "timeout for negative domain cache &&&
                                         item"
.row &%callout_domain_positive_expire%& "timeout for positive domain cache &&&
                                         item"
.row &%callout_negative_expire%&     "timeout for negative address cache item"
.row &%callout_positive_expire%&     "timeout for positive address cache item"
.row &%callout_random_local_part%&   "string to use for &""random""& testing"
.endtable



.section "TLS" "SECID108"
.table2
.row &%gnutls_compat_mode%&          "use GnuTLS compatibility mode"
.row &%gnutls_allow_auto_pkcs11%&    "allow GnuTLS to autoload PKCS11 modules"
.row &%hosts_require_alpn%&          "mandatory ALPN"
.row &%hosts_require_helo%&          "mandatory HELO/EHLO"
.row &%openssl_options%&             "adjust OpenSSL compatibility options"
.row &%tls_advertise_hosts%&         "advertise TLS to these hosts"
.row &%tls_alpn%&                    "acceptable protocol names"
.row &%tls_certificate%&             "location of server certificate"
.row &%tls_crl%&                     "certificate revocation list"
.row &%tls_dh_max_bits%&             "clamp D-H bit count suggestion"
.row &%tls_dhparam%&                 "DH parameters for server"
.row &%tls_eccurve%&                 "EC curve selection for server"
.row &%tls_ocsp_file%&               "location of server certificate status proof"
.row &%tls_on_connect_ports%&        "specify SSMTP (SMTPS) ports"
.row &%tls_privatekey%&              "location of server private key"
.row &%tls_remember_esmtp%&          "don't reset after starting TLS"
.row &%tls_require_ciphers%&         "specify acceptable ciphers"
.row &%tls_try_verify_hosts%&        "try to verify client certificate"
.row &%tls_verify_certificates%&     "expected client certificates"
.row &%tls_verify_hosts%&            "insist on client certificate verify"
.endtable



.section "Local user handling" "SECID109"
.table2
.row &%finduser_retries%&            "useful in NIS environments"
.row &%gecos_name%&                  "used when creating &'Sender:'&"
.row &%gecos_pattern%&               "ditto"
.row &%max_username_length%&         "for systems that truncate"
.row &%unknown_login%&               "used when no login name found"
.row &%unknown_username%&            "ditto"
.row &%uucp_from_pattern%&           "for recognizing &""From ""& lines"
.row &%uucp_from_sender%&            "ditto"
.endtable



.section "All incoming messages (SMTP and non-SMTP)" "SECID110"
.table2
.row &%header_maxsize%&              "total size of message header"
.row &%header_line_maxsize%&         "individual header line limit"
.row &%message_size_limit%&          "applies to all messages"
.row &%percent_hack_domains%&        "recognize %-hack for these domains"
.row &%received_header_text%&        "expanded to make &'Received:'&"
.row &%received_headers_max%&        "for mail loop detection"
.row &%recipients_max%&              "limit per message"
.row &%recipients_max_reject%&       "permanently reject excess recipients"
.endtable




.section "Non-SMTP incoming messages" "SECID111"
.table2
.row &%receive_timeout%&             "for non-SMTP messages"
.endtable





.section "Incoming SMTP messages" "SECID112"
See also the &'Policy controls'& section above.

.table2
.row &%dkim_verify_hashes%&          "DKIM hash methods accepted for signatures"
.row &%dkim_verify_keytypes%&        "DKIM key types accepted for signatures"
.row &%dkim_verify_min_keysizes%&    "DKIM key sizes accepted for signatures"
.row &%dkim_verify_signers%&         "DKIM domains for which DKIM ACL is run"
.row &%dmarc_forensic_sender%&       "DMARC sender for report messages"
.row &%dmarc_history_file%&          "DMARC results log"
.row &%dmarc_tld_file%&              "DMARC toplevel domains file"
.row &%host_lookup%&                 "host name looked up for these hosts"
.row &%host_lookup_order%&           "order of DNS and local name lookups"
.row &%recipient_unqualified_hosts%& "may send unqualified recipients"
.row &%rfc1413_hosts%&               "make ident calls to these hosts"
.row &%rfc1413_query_timeout%&       "zero disables ident calls"
.row &%sender_unqualified_hosts%&    "may send unqualified senders"
.row &%smtp_accept_keepalive%&       "some TCP/IP magic"
.row &%smtp_accept_max%&             "simultaneous incoming connections"
.row &%smtp_accept_max_nonmail%&     "non-mail commands"
.row &%smtp_accept_max_nonmail_hosts%& "hosts to which the limit applies"
.row &%smtp_accept_max_per_connection%& "messages per connection"
.row &%smtp_accept_max_per_host%&    "connections from one host"
.row &%smtp_accept_queue%&           "queue mail if more connections"
.row &%smtp_accept_queue_per_connection%& "queue if more messages per &&&
                                           connection"
.row &%smtp_accept_reserve%&         "only reserve hosts if more connections"
.row &%smtp_active_hostname%&        "host name to use in messages"
.row &%smtp_banner%&                 "text for welcome banner"
.row &%smtp_check_spool_space%&      "from SIZE on MAIL command"
.row &%smtp_connect_backlog%&        "passed to TCP/IP stack"
.row &%smtp_enforce_sync%&           "of SMTP command/responses"
.row &%smtp_etrn_command%&           "what to run for ETRN"
.row &%smtp_etrn_serialize%&         "only one at once"
.row &%smtp_load_reserve%&           "only reserve hosts if this load"
.row &%smtp_max_unknown_commands%&   "before dropping connection"
.row &%smtp_ratelimit_hosts%&        "apply ratelimiting to these hosts"
.row &%smtp_ratelimit_mail%&         "ratelimit for MAIL commands"
.row &%smtp_ratelimit_rcpt%&         "ratelimit for RCPT commands"
.row &%smtp_receive_timeout%&        "per command or data line"
.row &%smtp_reserve_hosts%&          "these are the reserve hosts"
.row &%smtp_return_error_details%&   "give detail on rejections"
.endtable



.section "SMTP extensions" "SECID113"
.table2
.row &%accept_8bitmime%&             "advertise 8BITMIME"
.row &%auth_advertise_hosts%&        "advertise AUTH to these hosts"
.row &%chunking_advertise_hosts%&    "advertise CHUNKING to these hosts"
.row &%dsn_advertise_hosts%&         "advertise DSN extensions to these hosts"
.row &%ignore_fromline_hosts%&       "allow &""From ""& from these hosts"
.row &%ignore_fromline_local%&       "allow &""From ""& from local SMTP"
.row &%limits_advertise_hosts%&      "advertise LIMITS to these hosts"
.row &%pipelining_advertise_hosts%&  "advertise pipelining to these hosts"
.row &%pipelining_connect_advertise_hosts%& "advertise pipelining to these hosts"
.row &%prdr_enable%&                 "advertise PRDR to all hosts"
.row &%smtputf8_advertise_hosts%&    "advertise SMTPUTF8 to these hosts"
.row &%tls_advertise_hosts%&         "advertise TLS to these hosts"
.row &%wellknown_advertise_hosts%&   "advertise WELLKNOWN to these hosts"
.endtable



.section "Processing messages" "SECID114"
.table2
.row &%allow_domain_literals%&       "recognize domain literal syntax"
.row &%allow_mx_to_ip%&              "allow MX to point to IP address"
.row &%allow_utf8_domains%&          "in addresses"
.row &%check_rfc2047_length%&        "check length of RFC 2047 &""encoded &&&
                                      words""&"
.row &%delivery_date_remove%&        "from incoming messages"
.row &%envelope_to_remove%&          "from incoming messages"
.row &%extract_addresses_remove_arguments%& "affects &%-t%& processing"
.row &%headers_charset%&             "default for translations"
.row &%qualify_domain%&              "default for senders"
.row &%qualify_recipient%&           "default for recipients"
.row &%return_path_remove%&          "from incoming messages"
.row &%strip_excess_angle_brackets%& "in addresses"
.row &%strip_trailing_dot%&          "at end of addresses"
.row &%untrusted_set_sender%&        "untrusted can set envelope sender"
.endtable



.section "System filter" "SECID115"
.table2
.row &%system_filter%&               "locate system filter"
.row &%system_filter_directory_transport%& "transport for delivery to a &&&
                                            directory"
.row &%system_filter_file_transport%& "transport for delivery to a file"
.row &%system_filter_group%&         "group for filter running"
.row &%system_filter_pipe_transport%& "transport for delivery to a pipe"
.row &%system_filter_reply_transport%& "transport for autoreply delivery"
.row &%system_filter_user%&          "user for filter running"
.endtable



.section "Routing and delivery" "SECID116"
.table2
.row &%disable_ipv6%&                "do no IPv6 processing"
.row &%dns_again_means_nonexist%&    "for broken domains"
.row &%dns_check_names_pattern%&     "pre-DNS syntax check"
.row &%dns_dnssec_ok%&               "parameter for resolver"
.row &%dns_ipv4_lookup%&             "only v4 lookup for these domains"
.row &%dns_retrans%&                 "parameter for resolver"
.row &%dns_retry%&                   "parameter for resolver"
.row &%dns_trust_aa%&                "DNS zones trusted as authentic"
.row &%dns_use_edns0%&               "parameter for resolver"
.row &%hold_domains%&                "hold delivery for these domains"
.row &%local_interfaces%&            "for routing checks"
.row &%queue_domains%&               "no immediate delivery for these"
.row &%queue_fast_ramp%&             "parallel delivery with 2-phase queue run"
.row &%queue_only%&                  "no immediate delivery at all"
.row &%queue_only_file%&             "no immediate delivery if file exists"
.row &%queue_only_load%&             "no immediate delivery if load is high"
.row &%queue_only_load_latch%&       "don't re-evaluate load for each message"
.row &%queue_only_override%&         "allow command line to override"
.row &%queue_run_in_order%&          "order of arrival"
.row &%queue_run_max%&               "of simultaneous queue runners"
.row &%queue_smtp_domains%&          "no immediate SMTP delivery for these"
.row &%remote_max_parallel%&         "parallel SMTP delivery per message"
.row &%remote_sort_domains%&         "order of remote deliveries"
.row &%retry_data_expire%&           "timeout for retry data"
.row &%retry_interval_max%&          "safety net for retry rules"
.endtable



.section "Bounce and warning messages" "SECID117"
.table2
.row &%bounce_message_file%&         "content of bounce"
.row &%bounce_message_text%&         "content of bounce"
.row &%bounce_return_body%&          "include body if returning message"
.row &%bounce_return_linesize_limit%& "limit on returned message line length"
.row &%bounce_return_message%&       "include original message in bounce"
.row &%bounce_return_size_limit%&    "limit on returned message"
.row &%bounce_sender_authentication%& "send authenticated sender with bounce"
.row &%dsn_from%&                    "set &'From:'& contents in bounces"
.row &%errors_copy%&                 "copy bounce messages"
.row &%errors_reply_to%&             "&'Reply-to:'& in bounces"
.row &%delay_warning%&               "time schedule"
.row &%delay_warning_condition%&     "condition for warning messages"
.row &%ignore_bounce_errors_after%&  "discard undeliverable bounces"
.row &%smtp_return_error_details%&   "give detail on rejections"
.row &%warn_message_file%&           "content of warning message"
.endtable



.section "Alphabetical list of main options" "SECTalomo"
Those options that undergo string expansion before use are marked with
&dagger;.

.option accept_8bitmime main boolean true
.cindex "8BITMIME"
.cindex "8-bit characters"
.cindex "log" "selectors"
.cindex "log" "8BITMIME"
.cindex "ESMTP extensions" 8BITMIME
This option causes Exim to send 8BITMIME in its response to an SMTP
EHLO command, and to accept the BODY= parameter on MAIL commands.
However, though Exim is 8-bit clean, it is not a protocol converter, and it
takes no steps to do anything special with messages received by this route.

Historically Exim kept this option off by default, but the maintainers
feel that in today's Internet, this causes more problems than it solves.
It now defaults to true.
A more detailed analysis of the issues is provided by Dan Bernstein:
.display
&url(https://cr.yp.to/smtp/8bitmime.html)
.endd

To log received 8BITMIME status use
.code
log_selector = +8bitmime
.endd

.option acl_not_smtp main string&!! unset
.cindex "&ACL;" "for non-SMTP messages"
.cindex "non-SMTP messages" "ACLs for"
This option defines the ACL that is run when a non-SMTP message has been
read and is on the point of being accepted. See section &<<SECnonSMTP>>& for
further details.

.option acl_not_smtp_mime main string&!! unset
This option defines the ACL that is run for individual MIME parts of non-SMTP
messages. It operates in exactly the same way as &%acl_smtp_mime%& operates for
SMTP messages.

.option acl_not_smtp_start main string&!! unset
.cindex "&ACL;" "at start of non-SMTP message"
.cindex "non-SMTP messages" "ACLs for"
This option defines the ACL that is run before Exim starts reading a
non-SMTP message. See section &<<SECnonSMTP>>& for further details.

.new
.option acl_smtp_atrn main string&!! unset
.cindex ATRN "ACL for"
.cindex ATRN advertisement
.cindex "ESMTP extensions" ATRN
This option defines the ACL that is run when an SMTP ATRN command is
received.
If no value is set, or the result after expansion is an empty string,
then the ATRN facility is not advertised.
See chapter &<<CHAPACL>>& for general information on ACLs,
and section &<<SECTATRN>>& for description of ATRN.
.wen

.option acl_smtp_auth main string&!! unset
.cindex "&ACL;" "setting up for SMTP commands"
.cindex "AUTH" "ACL for"
This option defines the ACL that is run when an SMTP AUTH command is
received.
See chapter &<<CHAPACL>>& for general information on ACLs, and chapter
&<<CHAPSMTPAUTH>>& for details of authentication.

.option acl_smtp_connect main string&!! unset
.cindex "&ACL;" "on SMTP connection"
This option defines the ACL that is run when an SMTP connection is received.
See section &<<SECconnectACL>>& for further details.

.option acl_smtp_data main string&!! unset
.cindex "DATA" "ACL for"
This option defines the ACL that is run after an SMTP DATA command has been
processed and the message itself has been received, but before the final
acknowledgment is sent. See section &<<SECdataACLS>>& for further details.

.option acl_smtp_data_prdr main string&!! accept
.cindex "PRDR" "ACL for"
.cindex "DATA" "PRDR ACL for"
.cindex "&ACL;" "PRDR-related"
.cindex "&ACL;" "per-user data processing"
This option defines the ACL that,
if the PRDR feature has been negotiated,
is run for each recipient after an SMTP DATA command has been
processed and the message itself has been received, but before the
acknowledgment is sent. See section &<<SECTPRDRACL>>& for further details.

.option acl_smtp_dkim main string&!! unset
.cindex DKIM "ACL for"
This option defines the ACL that is run for each DKIM signature
(by default, or as specified in the dkim_verify_signers option)
of a received message.
See section &<<SECDKIMVFY>>& for further details.

.option acl_smtp_etrn main string&!! unset
.cindex "ETRN" "ACL for"
.cindex "ETRN" advertisement
This option defines the ACL that is run when an SMTP ETRN command is
received.
If no value is set then the ETRN facility is not advertised.
See chapter &<<CHAPACL>>& for general information on ACLs,
and section &<<SECTETRN>>& for description of ETRN.

.option acl_smtp_expn main string&!! unset
.cindex "EXPN" "ACL for"
This option defines the ACL that is run when an SMTP EXPN command is
received. See chapter &<<CHAPACL>>& for further details.

.option acl_smtp_helo main string&!! unset
.cindex "EHLO" "ACL for"
.cindex "HELO" "ACL for"
This option defines the ACL that is run when an SMTP EHLO or HELO
command is received. See section &<<SECheloACL>>& for further details.


.option acl_smtp_mail main string&!! unset
.cindex "MAIL" "ACL for"
This option defines the ACL that is run when an SMTP MAIL command is
received. See chapter &<<CHAPACL>>& for further details.

.option acl_smtp_mailauth main string&!! unset
.cindex "AUTH" "on MAIL command"
This option defines the ACL that is run when there is an AUTH parameter on
a MAIL command.
See chapter &<<CHAPACL>>& for general information on ACLs, and chapter
&<<CHAPSMTPAUTH>>& for details of authentication.

.option acl_smtp_mime main string&!! unset
.cindex "MIME content scanning" "ACL for"
This option is available when Exim is built with the content-scanning
extension. It defines the ACL that is run for each MIME part in a message. See
section &<<SECTscanmimepart>>& for details.

.option acl_smtp_notquit main string&!! unset
.cindex "not-QUIT, ACL for"
This option defines the ACL that is run when an SMTP session
ends without a QUIT command being received.
See section &<<SECTNOTQUITACL>>& for further details.

.option acl_smtp_predata main string&!! unset
This option defines the ACL that is run when an SMTP DATA command is
received, before the message itself is received. See chapter &<<CHAPACL>>& for
further details.

.option acl_smtp_quit main string&!! unset
.cindex "QUIT, ACL for"
This option defines the ACL that is run when an SMTP QUIT command is
received. See chapter &<<CHAPACL>>& for further details.

.option acl_smtp_rcpt main string&!! unset
.cindex "RCPT" "ACL for"
This option defines the ACL that is run when an SMTP RCPT command is
received. See section &<<SECTQUITACL>>& for further details.

.option acl_smtp_starttls main string&!! unset
.cindex "STARTTLS, ACL for"
This option defines the ACL that is run when an SMTP STARTTLS command is
received. See chapter &<<CHAPACL>>& for further details.

.option acl_smtp_vrfy main string&!! unset
.cindex "VRFY" "ACL for"
This option defines the ACL that is run when an SMTP VRFY command is
received. See chapter &<<CHAPACL>>& for further details.

.option acl_smtp_wellknown main string&!! unset
.cindex "WELLKNOWN, ACL for"
This option defines the ACL that is run when an SMTP WELLKNOWN command is
received. See section &<<SECTWELLKNOWNACL>>& for further details.

.option add_environment main "string list" empty
.cindex "environment" "set values"
This option adds individual environment variables that the
currently linked libraries and programs in child processes may use.
Each list element should be of the form &"name=value"&.

See &<<SECTpipeenv>>& for the environment of &(pipe)& transports.

.option admin_groups main "string list&!!" unset
.cindex "admin user"
This option is expanded just once, at the start of Exim's processing. If the
current group or any of the supplementary groups of an Exim caller is in this
colon-separated list, the caller has admin privileges. If all your system
programmers are in a specific group, for example, you can give them all Exim
admin privileges by putting that group in &%admin_groups%&. However, this does
not permit them to read Exim's spool files (whose group owner is the Exim gid).
To permit this, you have to add individuals to the Exim group.

.option allow_domain_literals main boolean false
.cindex "domain literal"
If this option is set, the
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
domain literal format is permitted in
email addresses. The option is not set by default, because the domain literal
format is not normally required these days, and few people know about it. It
has, however, been exploited by mail abusers.

Unfortunately, it seems that some DNS black list maintainers are using this
format to report black listing to postmasters. If you want to accept messages
addressed to your hosts by IP address, you need to set
&%allow_domain_literals%& true, and also to add &`@[]`& to the list of local
domains (defined in the named domain list &%local_domains%& in the default
configuration). This &"magic string"& matches the domain literal form of all
the local host's IP addresses.

.option allow_mx_to_ip main boolean false
.cindex "MX record" "pointing to IP address"
It appears that more and more DNS zone administrators are breaking the rules
and putting domain names that look like IP addresses on the right hand side of
MX records. Exim follows the rules and rejects this, giving an error message
that explains the misconfiguration. However, some other MTAs support this
practice, so to avoid &"Why can't Exim do this?"& complaints,
&%allow_mx_to_ip%& exists, in order to enable this heinous activity. It is not
recommended, except when you have no other choice.

.option allow_utf8_domains main boolean false
.cindex "domain" "UTF-8 characters in"
.cindex "UTF-8" "in domain name"
Lots of discussion is going on about internationalized domain names. One
camp is strongly in favour of just using UTF-8 characters, and it seems
that at least two other MTAs permit this.
This option allows Exim users to experiment if they wish.

If it is set true, Exim's domain parsing function allows valid
UTF-8 multicharacters to appear in domain name components, in addition to
letters, digits, and hyphens.

If Exim is built with internationalization support
and the SMTPUTF8 ESMTP option is in use (see chapter &<<CHAPi18n>>&)
this option can be left as default.
Without that,
if you want to look up such domain names in the DNS, you must also
adjust the value of &%dns_check_names_pattern%& to match the extended form. A
suitable setting is:
.code
dns_check_names_pattern = (?i)^(?>(?(1)\.|())[a-z0-9\xc0-\xff]\
  (?>[-a-z0-9\x80-\xff]*[a-z0-9\x80-\xbf])?)+$
.endd
Alternatively, you can just disable this feature by setting
.code
dns_check_names_pattern =
.endd
That is, set the option to an empty string so that no check is done.


.option auth_advertise_hosts main "host list&!!" *
.cindex "authentication" "advertising"
.cindex "AUTH" "advertising"
.cindex "ESMTP extensions" AUTH
If any server authentication mechanisms are configured, Exim advertises them in
response to an EHLO command only if the calling host matches this list.
Otherwise, Exim does not advertise AUTH.
Exim does not accept AUTH commands from clients to which it has not
advertised the availability of AUTH. The advertising of individual
authentication mechanisms can be controlled by the use of the
&%server_advertise_condition%& generic authenticator option on the individual
authenticators. See chapter &<<CHAPSMTPAUTH>>& for further details.

Certain mail clients (for example, Netscape) require the user to provide a name
and password for authentication if AUTH is advertised, even though it may
not be needed (the host may accept messages from hosts on its local LAN without
authentication, for example). The &%auth_advertise_hosts%& option can be used
to make these clients more friendly by excluding them from the set of hosts to
which Exim advertises AUTH.

.cindex "AUTH" "advertising when encrypted"
If you want to advertise the availability of AUTH only when the connection
is encrypted using TLS, you can make use of the fact that the value of this
option is expanded, with a setting like this:
.code
auth_advertise_hosts = ${if eq{$tls_in_cipher}{}{}{*}}
.endd
.vindex "&$tls_in_cipher$&"
If &$tls_in_cipher$& is empty, the session is not encrypted, and the result of
the expansion is empty, thus matching no hosts. Otherwise, the result of the
expansion is *, which matches all hosts.


.option auto_thaw main time 0s
.cindex "thawing messages"
.cindex "unfreezing messages"
If this option is set to a time greater than zero, a queue runner will try a
new delivery attempt on any frozen message, other than a bounce message, if
this much time has passed since it was frozen. This may result in the message
being re-frozen if nothing has changed since the last attempt. It is a way of
saying &"keep on trying, even though there are big problems"&.

&*Note*&: This is an old option, which predates &%timeout_frozen_after%& and
&%ignore_bounce_errors_after%&. It is retained for compatibility, but it is not
thought to be very useful any more, and its use should probably be avoided.


.option av_scanner main string "see below"
This option is available if Exim is built with the content-scanning extension.
It specifies which anti-virus scanner to use. The default value is:
.code
sophie:/var/run/sophie
.endd
If the value of &%av_scanner%& starts with a dollar character, it is expanded
before use. See section &<<SECTscanvirus>>& for further details.


.option bi_command main string unset
.oindex "&%-bi%&"
This option supplies the name of a command that is run when Exim is called with
the &%-bi%& option (see chapter &<<CHAPcommandline>>&). The string value is
just the command name, it is not a complete command line. If an argument is
required, it must come from the &%-oA%& command line option.


.option bounce_message_file main string&!! unset
.cindex "bounce message" "customizing"
.cindex "customizing" "bounce message"
This option defines a template file containing paragraphs of text to be used
for constructing bounce messages.  Details of the file's contents are given in
chapter &<<CHAPemsgcust>>&.
.cindex bounce_message_file "tainted data"
The option is expanded to give the file path, which must be
absolute and untainted.
See also &%warn_message_file%&.


.option bounce_message_text main string unset
When this option is set, its contents are included in the default bounce
message immediately after &"This message was created automatically by mail
delivery software."& It is not used if &%bounce_message_file%& is set.

.option bounce_return_body main boolean true
.cindex "bounce message" "including body"
This option controls whether the body of an incoming message is included in a
bounce message when &%bounce_return_message%& is true. The default setting
causes the entire message, both header and body, to be returned (subject to the
value of &%bounce_return_size_limit%&). If this option is false, only the
message header is included. In the case of a non-SMTP message containing an
error that is detected during reception, only those header lines preceding the
point at which the error was detected are returned.
.cindex "bounce message" "including original"

.option bounce_return_linesize_limit main integer 998
.cindex "size" "of bounce lines, limit"
.cindex "bounce message" "line length limit"
.cindex "limit" "bounce message line length"
This option sets a limit in bytes on the line length of messages
that are returned to senders due to delivery problems,
when &%bounce_return_message%& is true.
The default value corresponds to RFC limits.
If the message being returned has lines longer than this value it is
treated as if the &%bounce_return_size_limit%& (below) restriction was exceeded.

The option also applies to bounces returned when an error is detected
during reception of a message.
In this case lines from the original are truncated.

The option does not apply to messages generated by an &(autoreply)& transport.


.option bounce_return_message main boolean true
If this option is set false, none of the original message is included in
bounce messages generated by Exim. See also &%bounce_return_size_limit%& and
&%bounce_return_body%&.


.option bounce_return_size_limit main integer 100K
.cindex "size" "of bounce, limit"
.cindex "bounce message" "size limit"
.cindex "limit" "bounce message size"
This option sets a limit in bytes on the size of messages that are returned to
senders as part of bounce messages when &%bounce_return_message%& is true. The
limit should be less than the value of the global &%message_size_limit%& and of
any &%message_size_limit%& settings on transports, to allow for the bounce text
that Exim generates. If this option is set to zero there is no limit.

When the body of any message that is to be included in a bounce message is
greater than the limit, it is truncated, and a comment pointing this out is
added at the top. The actual cutoff may be greater than the value given, owing
to the use of buffering for transferring the message in chunks (typically 8K in
size). The idea is to save bandwidth on those undeliverable 15-megabyte
messages.

.option bounce_sender_authentication main string unset
.cindex "bounce message" "sender authentication"
.cindex "authentication" "bounce message"
.cindex "AUTH" "on bounce message"
This option provides an authenticated sender address that is sent with any
bounce messages generated by Exim that are sent over an authenticated SMTP
connection. A typical setting might be:
.code
bounce_sender_authentication = mailer-daemon@my.domain.example
.endd
which would cause bounce messages to be sent using the SMTP command:
.code
MAIL FROM:<> AUTH=mailer-daemon@my.domain.example
.endd
The value of &%bounce_sender_authentication%& must always be a complete email
address.

.option callout_domain_negative_expire main time 3h
.cindex "caching" "callout timeouts"
.cindex "callout" "caching timeouts"
This option specifies the expiry time for negative callout cache data for a
domain. See section &<<SECTcallver>>& for details of callout verification, and
section &<<SECTcallvercache>>& for details of the caching.


.option callout_domain_positive_expire main time 7d
This option specifies the expiry time for positive callout cache data for a
domain. See section &<<SECTcallver>>& for details of callout verification, and
section &<<SECTcallvercache>>& for details of the caching.


.option callout_negative_expire main time 2h
This option specifies the expiry time for negative callout cache data for an
address. See section &<<SECTcallver>>& for details of callout verification, and
section &<<SECTcallvercache>>& for details of the caching.


.option callout_positive_expire main time 24h
This option specifies the expiry time for positive callout cache data for an
address. See section &<<SECTcallver>>& for details of callout verification, and
section &<<SECTcallvercache>>& for details of the caching.


.option callout_random_local_part main string&!! "see below"
This option defines the &"random"& local part that can be used as part of
callout verification. The default value is
.code
$primary_hostname-$tod_epoch-testing
.endd
See section &<<CALLaddparcall>>& for details of how this value is used.


.options check_log_inodes main integer 100 &&&
         check_log_space main integer 10M
See &%check_spool_space%& below.

.oindex "&%check_rfc2047_length%&"
.cindex "RFC 2047" "disabling length check"
.option check_rfc2047_length main boolean true
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047)
defines a way of encoding non-ASCII characters in headers using a
system of &"encoded words"&. The RFC specifies a maximum length for an encoded
word; strings to be encoded that exceed this length are supposed to use
multiple encoded words. By default, Exim does not recognize encoded words that
exceed the maximum length. However, it seems that some software, in violation
of the RFC, generates overlong encoded words. If &%check_rfc2047_length%& is
set false, Exim recognizes encoded words of any length.


.options check_spool_inodes main integer 100 &&&
         check_spool_space main integer 10M
.cindex "checking disk space"
.cindex "disk space, checking"
.cindex "spool directory" "checking space"
The four &%check_...%& options allow for checking of disk resources before a
message is accepted.

.vindex "&$log_inodes$&"
.vindex "&$log_space$&"
.vindex "&$spool_inodes$&"
.vindex "&$spool_space$&"
When any of these options are nonzero, they apply to all incoming messages. If you
want to apply different checks to different kinds of message, you can do so by
testing the variables &$log_inodes$&, &$log_space$&, &$spool_inodes$&, and
&$spool_space$& in an ACL with appropriate additional conditions.


&%check_spool_space%& and &%check_spool_inodes%& check the spool partition if
either value is greater than zero, for example:
.code
check_spool_space = 100M
check_spool_inodes = 100
.endd
The spool partition is the one that contains the directory defined by
SPOOL_DIRECTORY in &_Local/Makefile_&. It is used for holding messages in
transit.

&%check_log_space%& and &%check_log_inodes%&  check the partition in which log
files are written if either is greater than zero. These should be set only if
&%log_file_path%& and &%spool_directory%& refer to different partitions.

If there is less space or fewer inodes than requested, Exim refuses to accept
incoming mail. In the case of SMTP input this is done by giving a 452 temporary
error response to the MAIL command. If ESMTP is in use and there was a
SIZE parameter on the MAIL command, its value is added to the
&%check_spool_space%& value, and the check is performed even if
&%check_spool_space%& is zero, unless &%no_smtp_check_spool_space%& is set.

The values for &%check_spool_space%& and &%check_log_space%& are held as a
number of kilobytes (though specified in bytes).
If a non-multiple of 1024 is specified, it is rounded up.

For non-SMTP input and for batched SMTP input, the test is done at start-up; on
failure a message is written to stderr and Exim exits with a non-zero code, as
it obviously cannot send an error message of any kind.

There is a slight performance penalty for these checks.
Versions of Exim preceding 4.88 had these disabled by default;
high-rate installations confident they will never run out of resources
may wish to deliberately disable them.

.option chunking_advertise_hosts main "host list&!!" *
.cindex CHUNKING advertisement
.cindex "RFC 3030" "CHUNKING"
.cindex "ESMTP extensions" CHUNKING
The CHUNKING extension
(&url(https://www.rfc-editor.org/rfc/rfc3030.html,RFC 3030))
will be advertised in the EHLO message to these hosts.
Hosts may use the BDAT command as an alternate to DATA.

.option commandline_checks_require_admin main boolean &`false`&
.cindex "restricting access to features"
This option restricts various basic checking features to require an
administrative user.
This affects most of the &%-b*%& options, such as &%-be%&.

.option debug_store main boolean &`false`&
.cindex debugging "memory corruption"
.cindex memory debugging
This option, when true, enables extra checking in Exim's internal memory
management.  For use when a memory corruption issue is being investigated,
it should normally be left as default.

.option daemon_smtp_ports main string &`smtp`&
.cindex "port" "for daemon"
.cindex "TCP/IP" "setting listening ports"
This option specifies one or more default SMTP ports on which the Exim daemon
listens. See chapter &<<CHAPinterfaces>>& for details of how it is used. For
backward compatibility, &%daemon_smtp_port%& (singular) is a synonym.

.options daemon_startup_retries main integer 9 &&&
         daemon_startup_sleep main time 30s
.cindex "daemon startup, retrying"
These options control the retrying done by
the daemon at startup when it cannot immediately bind a listening socket
(typically because the socket is already in use): &%daemon_startup_retries%&
defines the number of retries after the first failure, and
&%daemon_startup_sleep%& defines the length of time to wait between retries.

.option delay_warning main "time list" 24h
.cindex "warning of delay"
.cindex "delay warning, specifying"
.cindex "queue" "delay warning"
When a message is delayed, Exim sends a warning message to the sender at
intervals specified by this option. The data is a colon-separated list of times
after which to send warning messages. If the value of the option is an empty
string or a zero time, no warnings are sent. Up to 10 times may be given. If a
message has been in the queue for longer than the last time, the last interval
between the times is used to compute subsequent warning times. For example,
with
.code
delay_warning = 4h:8h:24h
.endd
the first message is sent after 4 hours, the second after 8 hours, and
the third one after 24 hours. After that, messages are sent every 16 hours,
because that is the interval between the last two times on the list. If you set
just one time, it specifies the repeat interval. For example, with:
.code
delay_warning = 6h
.endd
messages are repeated every six hours. To stop warnings after a given time, set
a very large time at the end of the list. For example:
.code
delay_warning = 2h:12h:99d
.endd
Note that the option is only evaluated at the time a delivery attempt fails,
which depends on retry and queue-runner configuration.
Typically retries will be configured more frequently than warning messages.

.option delay_warning_condition main string&!! "see below"
.vindex "&$domain$&"
The string is expanded at the time a warning message might be sent. If all the
deferred addresses have the same domain, it is set in &$domain$& during the
expansion. Otherwise &$domain$& is empty. If the result of the expansion is a
forced failure, an empty string, or a string matching any of &"0"&, &"no"& or
&"false"& (the comparison being done caselessly) then the warning message is
not sent. The default is:
.code
delay_warning_condition = ${if or {\
  { !eq{$h_list-id:$h_list-post:$h_list-subscribe:}{} }\
  { match{$h_precedence:}{(?i)bulk|list|junk} }\
  { match{$h_auto-submitted:}{(?i)auto-generated|auto-replied} }\
  } {no}{yes}}
.endd
This suppresses the sending of warnings for messages that contain &'List-ID:'&,
&'List-Post:'&, or &'List-Subscribe:'& headers, or have &"bulk"&, &"list"& or
&"junk"& in a &'Precedence:'& header, or have &"auto-generated"& or
&"auto-replied"& in an &'Auto-Submitted:'& header.

.option deliver_drop_privilege main boolean false
.cindex "unprivileged delivery"
.cindex "delivery" "unprivileged"
If this option is set true, Exim drops its root privilege at the start of a
delivery process, and runs as the Exim user throughout. This severely restricts
the kinds of local delivery that are possible, but is viable in certain types
of configuration. There is a discussion about the use of root privilege in
chapter &<<CHAPsecurity>>&.

.option deliver_queue_load_max main fixed-point unset
.cindex "load average"
.cindex "queue runner" "abandoning"
When this option is set, a queue run is abandoned if the system load average
becomes greater than the value of the option. The option has no effect on
ancient operating systems on which Exim cannot determine the load average.
See also &%queue_only_load%& and &%smtp_load_reserve%&.


.option delivery_date_remove main boolean true
.cindex "&'Delivery-date:'& header line"
Exim's transports have an option for adding a &'Delivery-date:'& header to a
message when it is delivered, in exactly the same way as &'Return-path:'& is
handled. &'Delivery-date:'& records the actual time of delivery. Such headers
should not be present in incoming messages, and this option causes them to be
removed at the time the message is received, to avoid any problems that might
occur when a delivered message is subsequently sent on to some other recipient.

.option disable_fsync main boolean false
.cindex "&[fsync()]&, disabling"
This option is available only if Exim was built with the compile-time option
ENABLE_DISABLE_FSYNC. When this is not set, a reference to &%disable_fsync%& in
a runtime configuration generates an &"unknown option"& error. You should not
build Exim with ENABLE_DISABLE_FSYNC or set &%disable_fsync%& unless you
really, really, really understand what you are doing. &'No pre-compiled
distributions of Exim should ever make this option available.'&

When &%disable_fsync%& is set true, Exim no longer calls &[fsync()]& to force
updated files' data to be written to disc before continuing. Unexpected events
such as crashes and power outages may cause data to be lost or scrambled.
Here be Dragons. &*Beware.*&


.option disable_ipv6 main boolean false
.cindex "IPv6" "disabling"
If this option is set true, even if the Exim binary has IPv6 support, no IPv6
activities take place. AAAA records are never looked up, and any IPv6 addresses
that are listed in &%local_interfaces%&, data for the &%manualroute%& router,
etc. are ignored. If IP literals are enabled, the &(ipliteral)& router declines
to handle IPv6 literal addresses.


.option dkim_verify_hashes main "string list" "sha256 : sha512"
.cindex DKIM "selecting signature algorithms"
This option gives a list of hash types which are acceptable in signatures,
and an order of processing.
Signatures with algorithms not in the list will be ignored.

Acceptable values include:
.code
sha1
sha256
sha512
.endd

Note that the acceptance of sha1 violates
&url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301).

.option dkim_verify_keytypes main "string list" "ed25519 : rsa"
This option gives a list of key types which are acceptable in signatures,
and an order of processing.
Signatures with algorithms not in the list will be ignored.


.option dkim_verify_min_keysizes main "string list" "rsa=1024 ed25519=250"
This option gives a list of key sizes which are acceptable in signatures.
The list is keyed by the algorithm type for the key; the values are in bits.
Signatures with keys smaller than given by this option will fail verification.

The default enforces the
&url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301)
minimum key size for RSA signatures.

.option dkim_verify_minimal main boolean false
If set to true, verification of signatures will terminate after the
first success.

.option dkim_verify_signers main "domain list&!!" $dkim_signers
.cindex DKIM "controlling calls to the ACL"
This option gives a list of DKIM domains for which the DKIM ACL is run.
It is expanded after the message is received; by default it runs
the ACL once for each signature in the message.
See section &<<SECDKIMVFY>>&.


.options dmarc_forensic_sender main string&!! unset &&&
         dmarc_history_file main string unset &&&
         dmarc_tld_file main string unset
.cindex DMARC "main section options"
These options control DMARC processing.
See section &<<SECDMARC>>& for details.


.option dns_again_means_nonexist main "domain list&!!" unset
.cindex "DNS" "&""try again""& response; overriding"
DNS lookups give a &"try again"& response for the DNS errors
&"non-authoritative host not found"& and &"SERVERFAIL"&. This can cause Exim to
keep trying to deliver a message, or to give repeated temporary errors to
incoming mail. Sometimes the effect is caused by a badly set up name server and
may persist for a long time. If a domain which exhibits this problem matches
anything in &%dns_again_means_nonexist%&, it is treated as if it did not exist.
This option should be used with care. You can make it apply to reverse lookups
by a setting such as this:
.code
dns_again_means_nonexist = *.in-addr.arpa
.endd
This option applies to all DNS lookups that Exim does,
except for TLSA lookups (where knowing about such failures
is security-relevant).
It also applies when the
&[gethostbyname()]& or &[getipnodebyname()]& functions give temporary errors,
since these are most likely to be caused by DNS lookup problems. The
&(dnslookup)& router has some options of its own for controlling what happens
when lookups for MX or SRV records give temporary errors. These more specific
options are applied after this global option.

.option dns_check_names_pattern main string "see below"
.cindex "DNS" "pre-check of name syntax"
When this option is set to a non-empty string, it causes Exim to check domain
names for characters that are not allowed in host names before handing them to
the DNS resolver, because some resolvers give temporary errors for names that
contain unusual characters. If a domain name contains any unwanted characters,
a &"not found"& result is forced, and the resolver is not called. The check is
done by matching the domain name against a regular expression, which is the
value of this option. The default pattern is
.code
dns_check_names_pattern = \
  (?i)^(?>(?(1)\.|())[^\W_](?>[a-z0-9/-]*[^\W_])?)+$
.endd
which permits only letters, digits, slashes, and hyphens in components, but
they must start and end with a letter or digit. Slashes are not, in fact,
permitted in host names, but they are found in certain NS records (which can be
accessed in Exim by using a &%dnsdb%& lookup). If you set
&%allow_utf8_domains%&, you must modify this pattern, or set the option to an
empty string.

.option dns_csa_search_limit main integer 5
This option controls the depth of parental searching for CSA SRV records in the
DNS, as described in more detail in section &<<SECTverifyCSA>>&.

.option dns_csa_use_reverse main boolean true
This option controls whether or not an IP address, given as a CSA domain, is
reversed and looked up in the reverse DNS, as described in more detail in
section &<<SECTverifyCSA>>&.

.option dns_cname_loops main integer 1
.cindex DNS "CNAME following"
This option controls the following of CNAME chains, needed if the resolver does
not do it internally.
As of 2018 most should, and the default can be left.
If you have an ancient one, a value of 10 is likely needed.

The default value of one CNAME-follow is needed
thanks to the observed return for an MX request,
given no MX presence but a CNAME to an A, of the CNAME.


.option dns_dnssec_ok main integer -1
.cindex "DNS" "resolver options"
.cindex "DNS" "DNSSEC"
If this option is set to a non-negative number then Exim will initialise the
DNS resolver library to either use or not use DNSSEC, overriding the system
default. A value of 0 coerces DNSSEC off, a value of 1 coerces DNSSEC on.

If the resolver library does not support DNSSEC then this option has no effect.

On Linux with glibc 2.31 or newer this is insufficient, the resolver library
will default to stripping out a successful validation status.
This will break a previously working Exim installation.
Provided that you do trust the resolver (ie, is on localhost) you can tell
glibc to pass through any successful validation with a new option in
&_/etc/resolv.conf_&:
.code
options trust-ad
.endd


.option dns_ipv4_lookup main "domain list&!!" unset
.cindex "IPv6" "DNS lookup for AAAA records"
.cindex "DNS" "IPv6 lookup for AAAA records"
.cindex DNS "IPv6 disabling"
When Exim is compiled with IPv6 support and &%disable_ipv6%& is not set, it
looks for IPv6 address records (AAAA records) as well as IPv4 address records
(A records) when trying to find IP addresses for hosts, unless the host's
domain matches this list.

This is a fudge to help with name servers that give big delays or otherwise do
not work for the AAAA record type. In due course, when the world's name
servers have all been upgraded, there should be no need for this option.
Note that all lookups, including those done for verification, are affected;
this will result in verify failure for IPv6 connections or ones using names
only valid for IPv6 addresses.


.option dns_retrans main time 0s
.cindex "DNS" "resolver options"
.cindex timeout "dns lookup"
.cindex "DNS" timeout
The options &%dns_retrans%& and &%dns_retry%& can be used to set the
retransmission and retry parameters for DNS lookups. Values of zero (the
defaults) leave the system default settings unchanged. The first value is the
time between retries, and the second is the number of retries. It isn't
totally clear exactly how these settings affect the total time a DNS lookup may
take. I haven't found any documentation about timeouts on DNS lookups; these
parameter values are available in the external resolver interface structure,
but nowhere does it seem to describe how they are used or what you might want
to set in them.
See also the &%slow_lookup_log%& option.


.option dns_retry main integer 0
See &%dns_retrans%& above.


.option dns_trust_aa main "domain list&!!" unset
.cindex "DNS" "resolver options"
.cindex "DNS" "DNSSEC"
If this option is set then lookup results marked with the AA bit
(Authoritative Answer) are trusted the same way as if they were
DNSSEC-verified. The authority section's name of the answer must
match with this expanded domain list.

Use this option only if you talk directly to a resolver that is
authoritative for some zones and does not set the AD (Authentic Data)
bit in the answer. Some DNS servers may have an configuration option to
mark the answers from their own zones as verified (they set the AD bit).
Others do not have this option. It is considered as poor practice using
a resolver that is an authoritative server for some zones.

Use this option only if you really have to (e.g. if you want
to use DANE for remote delivery to a server that is listed in the DNS
zones that your resolver is authoritative for).

If the DNS answer packet has the AA bit set and contains resource record
in the answer section, the name of the first NS record appearing in the
authority section is compared against the list. If the answer packet is
authoritative but the answer section is empty, the name of the first SOA
record in the authoritative section is used instead.

.cindex "DNS" "resolver options"
.option dns_use_edns0 main integer -1
.cindex "DNS" "resolver options"
.cindex "DNS" "EDNS0"
.cindex "DNS" "OpenBSD
If this option is set to a non-negative number then Exim will initialise the
DNS resolver library to either use or not use EDNS0 extensions, overriding
the system default. A value of 0 coerces EDNS0 off, a value of 1 coerces EDNS0
on.

If the resolver library does not support EDNS0 then this option has no effect.

OpenBSD's asr resolver routines are known to ignore the EDNS0 option; this
means that DNSSEC will not work with Exim on that platform either, unless Exim
is linked against an alternative DNS client library.


.option drop_cr main boolean false
This is an obsolete option that is now a no-op. It used to affect the way Exim
handled CR and LF characters in incoming messages. What happens now is
described in section &<<SECTlineendings>>&.

.option dsn_advertise_hosts main "host list&!!" unset
.cindex "bounce messages" "success"
.cindex "DSN" "success"
.cindex "Delivery Status Notification" "success"
.cindex "ESMTP extensions" DSN
DSN extensions (&url(https://www.rfc-editor.org/rfc/rfc3461,RFC 3461))
will be advertised in the EHLO message to,
and accepted from, these hosts.
Hosts may use the NOTIFY and ORCPT options on RCPT TO commands,
and RET and ENVID options on MAIL FROM commands.
A NOTIFY=SUCCESS option requests success-DSN messages.
A NOTIFY= option with no argument requests that no delay or failure DSNs
are sent.
&*Note*&: Supplying success-DSN messages has been criticised
on privacy grounds; it can leak details of internal forwarding.

.option dsn_from main "string&!!" "see below"
.cindex "&'From:'& header line" "in bounces"
.cindex "bounce messages" "&'From:'& line, specifying"
This option can be used to vary the contents of &'From:'& header lines in
bounces and other automatically generated messages (&"Delivery Status
Notifications"& &-- hence the name of the option). The default setting is:
.code
dsn_from = Mail Delivery System <Mailer-Daemon@$qualify_domain>
.endd
The value is expanded every time it is needed. If the expansion fails, a
panic is logged, and the default value is used.

.option envelope_to_remove main boolean true
.cindex "&'Envelope-to:'& header line"
Exim's transports have an option for adding an &'Envelope-to:'& header to a
message when it is delivered, in exactly the same way as &'Return-path:'& is
handled. &'Envelope-to:'& records the original recipient address from the
message's envelope that caused the delivery to happen. Such headers should not
be present in incoming messages, and this option causes them to be removed at
the time the message is received, to avoid any problems that might occur when a
delivered message is subsequently sent on to some other recipient.


.option errors_copy main "string list&!!" unset
.cindex "bounce message" "copy to other address"
.cindex "copy of bounce message"
Setting this option causes Exim to send bcc copies of bounce messages that it
generates to other addresses. &*Note*&: This does not apply to bounce messages
coming from elsewhere. The value of the option is a colon-separated list of
items. Each item consists of a pattern, terminated by white space, followed by
a comma-separated list of email addresses. If a pattern contains spaces, it
must be enclosed in double quotes.

Each pattern is processed in the same way as a single item in an address list
(see section &<<SECTaddresslist>>&). When a pattern matches the recipient of
the bounce message, the message is copied to the addresses on the list. The
items are scanned in order, and once a matching one is found, no further items
are examined. For example:
.code
errors_copy = spqr@mydomain   postmaster@mydomain.example :\
              rqps@mydomain   hostmaster@mydomain.example,\
                              postmaster@mydomain.example
.endd
.vindex "&$domain$&"
.vindex "&$local_part$&"
The address list is expanded before use. The expansion variables &$local_part$&
and &$domain$& are set from the original recipient of the error message, and if
there was any wildcard matching in the pattern, the expansion
.cindex "numerical variables (&$1$& &$2$& etc)" "in &%errors_copy%&"
variables &$0$&, &$1$&, etc. are set in the normal way.


.option errors_reply_to main string unset
.cindex "bounce message" "&'Reply-to:'& in"
By default, Exim's bounce and delivery warning messages contain the header line
.display
&`From: Mail Delivery System <Mailer-Daemon@`&&'qualify-domain'&&`>`&
.endd
.oindex &%quota_warn_message%&
where &'qualify-domain'& is the value of the &%qualify_domain%& option.
A warning message that is generated by the &%quota_warn_message%& option in an
&(appendfile)& transport may contain its own &'From:'& header line that
overrides the default.

Experience shows that people reply to bounce messages. If the
&%errors_reply_to%& option is set, a &'Reply-To:'& header is added to bounce
and warning messages. For example:
.code
errors_reply_to = postmaster@my.domain.example
.endd
The value of the option is not expanded. It must specify a valid
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address. However, if a warning message that is generated by the
&%quota_warn_message%& option in an &(appendfile)& transport contain its
own &'Reply-To:'& header line, the value of the &%errors_reply_to%& option is
not used.


.option event_action main string&!! unset
.cindex events
This option declares a string to be expanded for Exim's events mechanism.
For details see chapter &<<CHAPevents>>&.


.option exim_group main string "compile-time configured"
.cindex "gid (group id)" "Exim's own"
.cindex "Exim group"
This option changes the gid under which Exim runs when it gives up root
privilege. The default value is compiled into the binary. The value of this
option is used only when &%exim_user%& is also set. Unless it consists entirely
of digits, the string is looked up using &[getgrnam()]&, and failure causes a
configuration error. See chapter &<<CHAPsecurity>>& for a discussion of
security issues.


.option exim_path main string "see below"
.cindex "Exim binary, path name"
This option specifies the path name of the Exim binary, which is used when Exim
needs to re-exec itself. The default is set up to point to the file &'exim'& in
the directory configured at compile time by the BIN_DIRECTORY setting. It
is necessary to change &%exim_path%& if, exceptionally, Exim is run from some
other place.
&*Warning*&: Do not use a macro to define the value of this option, because
you will break those Exim utilities that scan the configuration file to find
where the binary is. (They then use the &%-bP%& option to extract option
settings such as the value of &%spool_directory%&.)


.option exim_user main string "compile-time configured"
.cindex "uid (user id)" "Exim's own"
.cindex "Exim user"
This option changes the uid under which Exim runs when it gives up root
privilege. The default value is compiled into the binary. Ownership of the run
time configuration file and the use of the &%-C%& and &%-D%& command line
options is checked against the values in the binary, not what is set here.

Unless it consists entirely of digits, the string is looked up using
&[getpwnam()]&, and failure causes a configuration error. If &%exim_group%& is
not also supplied, the gid is taken from the result of &[getpwnam()]& if it is
used. See chapter &<<CHAPsecurity>>& for a discussion of security issues.


.option exim_version main string "current version"
.cindex "Exim version"
.cindex customizing "version number"
.cindex "version number of Exim" override
This option overrides the &$version_number$&/&$exim_version$& that Exim reports in
various places.  Use with care; this may fool stupid security scanners.


.option extra_local_interfaces main "string list" unset
This option defines network interfaces that are to be considered local when
routing, but which are not used for listening by the daemon. See section
&<<SECTreclocipadd>>& for details.


. Allow this long option name to split; give it unsplit as a fifth argument
. for the automatic .oindex that is generated by .option.
. WAS:
. .option "extract_addresses_remove_ &~&~arguments"
. but apparently this results in searchability problems; bug 1197

.option extract_addresses_remove_arguments main boolean true
.oindex "&%-t%&"
.cindex "command line" "addresses with &%-t%&"
.cindex "Sendmail compatibility" "&%-t%& option"
According to some Sendmail documentation (Sun, IRIX, HP-UX), if any addresses
are present on the command line when the &%-t%& option is used to build an
envelope from a message's &'To:'&, &'Cc:'& and &'Bcc:'& headers, the command
line addresses are removed from the recipients list. This is also how Smail
behaves. However, other Sendmail documentation (the O'Reilly book) states that
command line addresses are added to those obtained from the header lines. When
&%extract_addresses_remove_arguments%& is true (the default), Exim subtracts
argument headers. If it is set false, Exim adds rather than removes argument
addresses.


.option finduser_retries main integer 0
.cindex "NIS, retrying user lookups"
On systems running NIS or other schemes in which user and group information is
distributed from a remote system, there can be times when &[getpwnam()]& and
related functions fail, even when given valid data, because things time out.
Unfortunately these failures cannot be distinguished from genuine &"not found"&
errors. If &%finduser_retries%& is set greater than zero, Exim will try that
many extra times to find a user or a group, waiting for one second between
retries.

.cindex "&_/etc/passwd_&" "multiple reading of"
You should not set this option greater than zero if your user information is in
a traditional &_/etc/passwd_& file, because it will cause Exim needlessly to
search the file multiple times for non-existent users, and also cause delay.



.option freeze_tell main "string list, comma separated" unset
.cindex "freezing messages" "sending a message when freezing"
.cindex "frozen messages" "sending a message when freezing"
On encountering certain errors, or when configured to do so in a system filter,
ACL, or special router, Exim freezes a message. This means that no further
delivery attempts take place until an administrator thaws the message, or the
&%auto_thaw%&, &%ignore_bounce_errors_after%&, or &%timeout_frozen_after%&
feature cause it to be processed. If &%freeze_tell%& is set, Exim generates a
warning message whenever it freezes something, unless the message it is
freezing is a locally-generated bounce message. (Without this exception there
is the possibility of looping.) The warning message is sent to the addresses
supplied as the comma-separated value of this option. If several of the
message's addresses cause freezing, only a single message is sent. If the
freezing was automatic, the reason(s) for freezing can be found in the message
log. If you configure freezing in a filter or ACL, you must arrange for any
logging that you require.


.options gecos_name main string&!! unset &&&
         gecos_pattern main string unset
.cindex "HP-UX"
.cindex "&""gecos""& field, parsing"
Some operating systems, notably HP-UX, use the &"gecos"& field in the system
password file to hold other information in addition to users' real names. Exim
looks up this field for use when it is creating &'Sender:'& or &'From:'&
headers. If either &%gecos_pattern%& or &%gecos_name%& are unset, the contents
of the field are used unchanged, except that, if an ampersand is encountered,
it is replaced by the user's login name with the first character forced to
upper case, since this is a convention that is observed on many systems.

When these options are set, &%gecos_pattern%& is treated as a regular
expression that is to be applied to the field (again with && replaced by the
login name), and if it matches, &%gecos_name%& is expanded and used as the
user's name.

.cindex "numerical variables (&$1$& &$2$& etc)" "in &%gecos_name%&"
Numeric variables such as &$1$&, &$2$&, etc. can be used in the expansion to
pick up sub-fields that were matched by the pattern. In HP-UX, where the user's
name terminates at the first comma, the following can be used:
.code
gecos_pattern = ([^,]*)
gecos_name = $1
.endd


.option gnutls_compat_mode main boolean unset
This option controls whether GnuTLS is used in compatibility mode in an Exim
server. This reduces security slightly, but improves interworking with older
implementations of TLS.


.option gnutls_allow_auto_pkcs11 main boolean unset
This option will let GnuTLS (2.12.0 or later) autoload PKCS11 modules with
the p11-kit configuration files in &_/etc/pkcs11/modules/_&.

See
&url(https://www.gnutls.org/manual/gnutls.html#Smart-cards-and-HSMs)
for documentation.



.option headers_charset main string "see below"
This option sets a default character set for translating from encoded MIME
&"words"& in header lines, when referenced by an &$h_xxx$& expansion item. The
default is the value of HEADERS_CHARSET in &_Local/Makefile_&. The
ultimate default is ISO-8859-1. For more details see the description of header
insertions in section &<<SECTexpansionitems>>&.



.option header_maxsize main integer "see below"
.cindex "header section" "maximum size of"
.cindex "limit" "size of message header section"
This option controls the overall maximum size of a message's header
section. The default is the value of HEADER_MAXSIZE in
&_Local/Makefile_&; the default for that is 1M. Messages with larger header
sections are rejected.


.option header_line_maxsize main integer 0
.cindex "header lines" "maximum size of"
.cindex "limit" "size of one header line"
This option limits the length of any individual header line in a message, after
all the continuations have been joined together. Messages with individual
header lines that are longer than the limit are rejected. The default value of
zero means &"no limit"&.




.option helo_accept_junk_hosts main "host list&!!" unset
.cindex "HELO" "accepting junk data"
.cindex "EHLO" "accepting junk data"
Exim checks the syntax of HELO and EHLO commands for incoming SMTP
mail, and gives an error response for invalid data. Unfortunately, there are
some SMTP clients that send syntactic junk. They can be accommodated by setting
this option. Note that this is a syntax check only. See &%helo_verify_hosts%&
if you want to do semantic checking.
See also &%helo_allow_chars%& for a way of extending the permitted character
set.


.option helo_allow_chars main string unset
.cindex "HELO" "underscores in"
.cindex "EHLO" "underscores in"
.cindex "underscore in EHLO/HELO"
This option can be set to a string of rogue characters that are permitted in
non-ip-literal EHLO and HELO names in addition to the standard letters, digits,
hyphens, and dots. For example if you really must allow underscores,
you can set
.code
helo_allow_chars = _
.endd
This option does not apply to names that look like ip-literals.
Note that the value is one string, not a list.


.option helo_lookup_domains main "domain list&!!" &`@:@[]`&
.cindex "HELO" "forcing reverse lookup"
.cindex "EHLO" "forcing reverse lookup"
If the domain given by a client in a HELO or EHLO command matches this
list, a reverse lookup is done in order to establish the host's true name. The
default forces a lookup if the client host gives the server's name or any of
its IP addresses (in brackets), something that broken clients have been seen to
do.


.option helo_try_verify_hosts main "host list&!!" unset
.cindex "HELO verifying" "optional"
.cindex "EHLO" "verifying, optional"
By default, Exim just checks the syntax of HELO and EHLO commands (see
&%helo_accept_junk_hosts%& and &%helo_allow_chars%&). However, some sites like
to do more extensive checking of the data supplied by these commands. The ACL
condition &`verify = helo`& is provided to make this possible.
Formerly, it was necessary also to set this option (&%helo_try_verify_hosts%&)
to force the check to occur. From release 4.53 onwards, this is no longer
necessary. If the check has not been done before &`verify = helo`& is
encountered, it is done at that time. Consequently, this option is obsolete.
Its specification is retained here for backwards compatibility.

When an EHLO or HELO command is received, if the calling host matches
&%helo_try_verify_hosts%&, Exim checks that the host name given in the HELO or
EHLO command either:

.ilist
is an IP literal matching the calling address of the host, or
.next
.cindex "DNS" "reverse lookup"
.cindex "reverse DNS lookup"
matches the host name that Exim obtains by doing a reverse lookup of the
calling host address, or
.next
when looked up in DNS yields the calling host address.
.endlist

However, the EHLO or HELO command is not rejected if any of the checks
fail. Processing continues, but the result of the check is remembered, and can
be detected later in an ACL by the &`verify = helo`& condition.

If DNS was used for successful verification, the variable
.cindex "DNS" "DNSSEC"
&$helo_verify_dnssec$& records the DNSSEC status of the lookups.

.option helo_verify_hosts main "host list&!!" unset
.cindex "HELO verifying" "mandatory"
.cindex "EHLO" "verifying, mandatory"
Like &%helo_try_verify_hosts%&, this option is obsolete, and retained only for
backwards compatibility. For hosts that match this option, Exim checks the host
name given in the HELO or EHLO in the same way as for
&%helo_try_verify_hosts%&. If the check fails, the HELO or EHLO command is
rejected with a 550 error, and entries are written to the main and reject logs.
If a MAIL command is received before EHLO or HELO, it is rejected with a 503
error.

.option hold_domains main "domain list&!!" unset
.cindex "domain" "delaying delivery"
.cindex "delivery" "delaying certain domains"
This option allows mail for particular domains to be held in the queue
manually. The option is overridden if a message delivery is forced with the
&%-M%&, &%-qf%&, &%-Rf%& or &%-Sf%& options, and also while testing or
verifying addresses using &%-bt%& or &%-bv%&. Otherwise, if a domain matches an
item in &%hold_domains%&, no routing or delivery for that address is done, and
it is deferred every time the message is looked at.

This option is intended as a temporary operational measure for delaying the
delivery of mail while some problem is being sorted out, or some new
configuration tested. If you just want to delay the processing of some
domains until a queue run occurs, you should use &%queue_domains%& or
&%queue_smtp_domains%&, not &%hold_domains%&.

A setting of &%hold_domains%& does not override Exim's code for removing
messages from the queue if they have been there longer than the longest retry
time in any retry rule. If you want to hold messages for longer than the normal
retry times, insert a dummy retry rule with a long retry time.


.option host_lookup main "host list&!!" unset
.cindex "host name" "lookup, forcing"
Exim does not look up the name of a calling host from its IP address unless it
is required to compare against some host list, or the host matches
&%helo_try_verify_hosts%& or &%helo_verify_hosts%&, or the host matches this
option (which normally contains IP addresses rather than host names). The
default configuration file contains
.code
host_lookup = *
.endd
which causes a lookup to happen for all hosts. If the expense of these lookups
is felt to be too great, the setting can be changed or removed.

After a successful reverse lookup, Exim does a forward lookup on the name it
has obtained, to verify that it yields the IP address that it started with. If
this check fails, Exim behaves as if the name lookup failed.

.vindex "&$host_lookup_failed$&"
.vindex "&$sender_host_name$&"
After any kind of failure, the host name (in &$sender_host_name$&) remains
unset, and &$host_lookup_failed$& is set to the string &"1"&. See also
&%dns_again_means_nonexist%&, &%helo_lookup_domains%&, and
&`verify = reverse_host_lookup`& in ACLs.


.option host_lookup_order main "string list" &`bydns:byaddr`&
This option specifies the order of different lookup methods when Exim is trying
to find a host name from an IP address. The default is to do a DNS lookup
first, and then to try a local lookup (using &[gethostbyaddr()]& or equivalent)
if that fails. You can change the order of these lookups, or omit one entirely,
if you want.

&*Warning*&: The &"byaddr"& method does not always yield aliases when there are
multiple PTR records in the DNS and the IP address is not listed in
&_/etc/hosts_&. Different operating systems give different results in this
case. That is why the default tries a DNS lookup first.



.option host_reject_connection main "host list&!!" unset
.cindex "host" "rejecting connections from"
If this option is set, incoming SMTP calls from the hosts listed are rejected
as soon as the connection is made.
This option is obsolete, and retained only for backward compatibility, because
nowadays the ACL specified by &%acl_smtp_connect%& can also reject incoming
connections immediately.

If the connection is on a TLS-on-connect port then the TCP connection is
just dropped.  Otherwise, an SMTP error is sent first.

The ability to give an immediate rejection (either by this option or using an
ACL) is provided for use in unusual cases. Many hosts will just try again,
sometimes without much delay. Normally, it is better to use an ACL to reject
incoming messages at a later stage, such as after RCPT commands. See
chapter &<<CHAPACL>>&.


.option hosts_connection_nolog main "host list&!!" unset
.cindex "host" "not logging connections from"
This option defines a list of hosts for which connection logging does not
happen, even though the &%smtp_connection%& log selector is set. For example,
you might want not to log SMTP connections from local processes, or from
127.0.0.1, or from your local LAN. This option is consulted in the main loop of
the daemon; you should therefore strive to restrict its value to a short inline
list of IP addresses and networks. To disable logging SMTP connections from
local processes, you must create a host list with an empty item. For example:
.code
hosts_connection_nolog = :
.endd
The hosts affected by this option also do not log "no MAIL in SMTP connection"
lines, as may commonly be produced by a monitoring system.


.option hosts_require_alpn main "host list&!!" unset
.cindex ALPN "require negotiation in server"
.cindex TLS ALPN
.cindex TLS "Application Layer Protocol Names"
If the TLS library supports ALPN
then a successful negotiation of ALPN will be required for any client
matching the list, for TLS to be used.
See also the &%tls_alpn%& option.

&*Note*&: prevention of fallback to in-clear connection is not
managed by this option, and should be done separately.


.option hosts_require_helo main "host list&!!" *
.cindex "HELO/EHLO" requiring
Exim will require an accepted HELO or EHLO command from a host matching
this list, before accepting a MAIL command.


.option hosts_proxy main "host list&!!" unset
.cindex proxy "proxy protocol"
This option enables use of Proxy Protocol proxies for incoming
connections.  For details see section &<<SECTproxyInbound>>&.


.option hosts_treat_as_local main "domain list&!!" unset
.cindex "local host" "domains treated as"
.cindex "host" "treated as local"
If this option is set, any host names that match the domain list are treated as
if they were the local host when Exim is scanning host lists obtained from MX
records
or other sources. Note that the value of this option is a domain list, not a
host list, because it is always used to check host names, not IP addresses.

This option also applies when Exim is matching the special items
&`@mx_any`&, &`@mx_primary`&, and &`@mx_secondary`& in a domain list (see
section &<<SECTdomainlist>>&), and when checking the &%hosts%& option in the
&(smtp)& transport for the local host (see the &%allow_localhost%& option in
that transport). See also &%local_interfaces%&, &%extra_local_interfaces%&, and
chapter &<<CHAPinterfaces>>&, which contains a discussion about local network
interfaces and recognizing the local host.


.option ibase_servers main "string list" unset
.cindex "InterBase" "server list"
This option provides a list of InterBase servers and associated connection data,
to be used in conjunction with &(ibase)& lookups (see section &<<SECID72>>&).
The option is available only if Exim has been built with InterBase support.



.option ignore_bounce_errors_after main time 10w
.cindex "bounce message" "discarding"
.cindex "discarding bounce message"
This option affects the processing of bounce messages that cannot be delivered,
that is, those that suffer a permanent delivery failure. (Bounce messages that
suffer temporary delivery failures are of course retried in the usual way.)

After a permanent delivery failure, bounce messages are frozen,
because there is no sender to whom they can be returned. When a frozen bounce
message has been in the queue for more than the given time, it is unfrozen at
the next queue run, and a further delivery is attempted. If delivery fails
again, the bounce message is discarded. This makes it possible to keep failed
bounce messages around for a shorter time than the normal maximum retry time
for frozen messages. For example,
.code
ignore_bounce_errors_after = 12h
.endd
retries failed bounce message deliveries after 12 hours, discarding any further
failures. If the value of this option is set to a zero time period, bounce
failures are discarded immediately. Setting a very long time (as in the default
value) has the effect of disabling this option. For ways of automatically
dealing with other kinds of frozen message, see &%auto_thaw%& and
&%timeout_frozen_after%&.


.options ignore_fromline_hosts main "host list&!!" unset &&&
         ignore_fromline_local main boolean false
.cindex "&""From""& line"
.cindex "UUCP" "&""From""& line"
Some broken SMTP clients insist on sending a UUCP-like &"From&~"& line before
the headers of a message. By default this is treated as the start of the
message's body, which means that any following headers are not recognized as
such. Exim can be made to ignore it by setting &%ignore_fromline_hosts%& to
match those hosts that insist on sending it. If the sender is actually a local
process rather than a remote host, and is using &%-bs%& to inject the messages,
&%ignore_fromline_local%& must be set to achieve this effect.



.option keep_environment main "string list" unset
.cindex "environment" "values from"
This option contains a string list of environment variables to keep.
You have to trust these variables or you have to be sure that
these variables do not impose any security risk. Keep in mind that
during the startup phase Exim is running with an effective UID 0 in most
installations. As the default value is an empty list, the default
environment for using libraries, running embedded Perl code, or running
external binaries is empty, and does not not even contain PATH or HOME.

Actually the list is interpreted as a list of patterns
(&<<SECTlistexpand>>&), except that it is not expanded first.

WARNING: Macro substitution is still done first, so having a macro
FOO and having FOO_HOME in your &%keep_environment%& option may have
unexpected results. You may work around this using a regular expression
that does not match the macro name: ^[F]OO_HOME$.

Current versions of Exim issue a warning during startup if you do not mention
&%keep_environment%& in your runtime configuration file and if your
current environment is not empty. Future versions may not issue that warning
anymore.

See the &%add_environment%& main config option for a way to set
environment variables to a fixed value.  The environment for &(pipe)&
transports is handled separately, see section &<<SECTpipeenv>>& for
details.


.option keep_malformed main time 4d
This option specifies the length of time to keep messages whose spool files
have been corrupted in some way. This should, of course, never happen. At the
next attempt to deliver such a message, it gets removed. The incident is
logged.


.option ldap_ca_cert_dir main string unset
.cindex "LDAP", "TLS CA certificate directory"
.cindex certificate "directory for LDAP"
This option indicates which directory contains CA certificates for verifying
a TLS certificate presented by an LDAP server.
While Exim does not provide a default value, your SSL library may.
Analogous to &%tls_verify_certificates%& but as a client-side option for LDAP
and constrained to be a directory.


.option ldap_ca_cert_file main string unset
.cindex "LDAP", "TLS CA certificate file"
.cindex certificate "file for LDAP"
This option indicates which file contains CA certificates for verifying
a TLS certificate presented by an LDAP server.
While Exim does not provide a default value, your SSL library may.
Analogous to &%tls_verify_certificates%& but as a client-side option for LDAP
and constrained to be a file.


.option ldap_cert_file main string unset
.cindex "LDAP" "TLS client certificate file"
.cindex certificate "file for LDAP"
This option indicates which file contains an TLS client certificate which
Exim should present to the LDAP server during TLS negotiation.
Should be used together with &%ldap_cert_key%&.


.option ldap_cert_key main string unset
.cindex "LDAP" "TLS client key file"
.cindex certificate "key for LDAP"
This option indicates which file contains the secret/private key to use
to prove identity to the LDAP server during TLS negotiation.
Should be used together with &%ldap_cert_file%&, which contains the
identity to be proven.


.option ldap_cipher_suite main string unset
.cindex "LDAP" "TLS cipher suite"
This controls the TLS cipher-suite negotiation during TLS negotiation with
the LDAP server.  See &<<SECTreqciphssl>>& for more details of the format of
cipher-suite options with OpenSSL (as used by LDAP client libraries).


.option ldap_default_servers main "string list" unset
.cindex "LDAP" "default servers"
This option provides a list of LDAP servers which are tried in turn when an
LDAP query does not contain a server. See section &<<SECTforldaque>>& for
details of LDAP queries. This option is available only when Exim has been built
with LDAP support.


.option ldap_require_cert main string unset.
.cindex "LDAP" "policy for LDAP server TLS cert presentation"
This should be one of the values "hard", "demand", "allow", "try" or "never".
A value other than one of these is interpreted as "never".
See the entry "TLS_REQCERT" in your system man page for ldap.conf(5).
Although Exim does not set a default, the LDAP library probably defaults
to hard/demand.


.option ldap_start_tls main boolean false
.cindex "LDAP" "whether or not to negotiate TLS"
If set, Exim will attempt to negotiate TLS with the LDAP server when
connecting on a regular LDAP port.  This is the LDAP equivalent of SMTP's
"STARTTLS".  This is distinct from using "ldaps", which is the LDAP form
of SSL-on-connect.
In the event of failure to negotiate TLS, the action taken is controlled
by &%ldap_require_cert%&.
This option is ignored for &`ldapi`& connections.


.option ldap_version main integer unset
.cindex "LDAP" "protocol version, forcing"
This option can be used to force Exim to set a specific protocol version for
LDAP. If it option is unset, it is shown by the &%-bP%& command line option as
-1. When this is the case, the default is 3 if LDAP_VERSION3 is defined in
the LDAP headers; otherwise it is 2. This option is available only when Exim
has been built with LDAP support.



.option limits_advertise_hosts main "host list&!!" *
.cindex LIMITS "suppressing advertising"
.cindex "ESMTP extensions" LIMITS
This option can be used to suppress the advertisement of the SMTP
LIMITS extension (&url(https://www.ietf.org/rfc/rfc9422.html,RFC 9422))
to specific hosts.
If permitted, Exim as a server will advertise in the EHLO response
the limit for RCPT commands set by the &%recipients_max%& option (if it is set)
and the limit for MAIL commands set by the &%smtp_accept_max_per_connection%&
option.

.option local_from_check main boolean true
.cindex "&'Sender:'& header line" "disabling addition of"
.cindex "&'From:'& header line" "disabling checking of"
When a message is submitted locally (that is, not over a TCP/IP connection) by
an untrusted user, Exim removes any existing &'Sender:'& header line, and
checks that the &'From:'& header line matches the login of the calling user and
the domain specified by &%qualify_domain%&.

&*Note*&: An unqualified address (no domain) in the &'From:'& header in a
locally submitted message is automatically qualified by Exim, unless the
&%-bnq%& command line option is used.

You can use &%local_from_prefix%& and &%local_from_suffix%& to permit affixes
on the local part. If the &'From:'& header line does not match, Exim adds a
&'Sender:'& header with an address constructed from the calling user's login
and the default qualify domain.

If &%local_from_check%& is set false, the &'From:'& header check is disabled,
and no &'Sender:'& header is ever added. If, in addition, you want to retain
&'Sender:'& header lines supplied by untrusted users, you must also set
&%local_sender_retain%& to be true.

.cindex "envelope from"
.cindex "envelope sender"
These options affect only the header lines in the message. The envelope sender
is still forced to be the login id at the qualify domain unless
&%untrusted_set_sender%& permits the user to supply an envelope sender.

For messages received over TCP/IP, an ACL can specify &"submission mode"& to
request similar header line checking. See section &<<SECTthesenhea>>&, which
has more details about &'Sender:'& processing.




.options local_from_prefix main string unset &&&
         local_from_suffix main string unset
When Exim checks the &'From:'& header line of locally submitted messages for
matching the login id (see &%local_from_check%& above), it can be configured to
ignore certain prefixes and suffixes in the local part of the address. This is
done by setting &%local_from_prefix%& and/or &%local_from_suffix%& to
appropriate lists, in the same form as the &%local_part_prefix%& and
&%local_part_suffix%& router options (see chapter &<<CHAProutergeneric>>&). For
example, if
.code
local_from_prefix = *-
.endd
is set, a &'From:'& line containing
.code
From: anything-user@your.domain.example
.endd
will not cause a &'Sender:'& header to be added if &'user@your.domain.example'&
matches the actual sender address that is constructed from the login name and
qualify domain.


.option local_interfaces main "string list" "see below"
This option controls which network interfaces are used by the daemon for
listening; they are also used to identify the local host when routing. Chapter
&<<CHAPinterfaces>>& contains a full description of this option and the related
options &%daemon_smtp_ports%&, &%extra_local_interfaces%&,
&%hosts_treat_as_local%&, and &%tls_on_connect_ports%&. The default value for
&%local_interfaces%& is
.code
local_interfaces = 0.0.0.0
.endd
when Exim is built without IPv6 support; otherwise it is
.code
local_interfaces = <; ::0 ; 0.0.0.0
.endd

.option local_scan_timeout main time 5m
.cindex "timeout" "for &[local_scan()]& function"
.cindex "&[local_scan()]& function" "timeout"
This timeout applies to the &[local_scan()]& function (see chapter
&<<CHAPlocalscan>>&). Zero means &"no timeout"&. If the timeout is exceeded,
the incoming message is rejected with a temporary error if it is an SMTP
message. For a non-SMTP message, the message is dropped and Exim ends with a
non-zero code. The incident is logged on the main and reject logs.



.option local_sender_retain main boolean false
.cindex "&'Sender:'& header line" "retaining from local submission"
When a message is submitted locally (that is, not over a TCP/IP connection) by
an untrusted user, Exim removes any existing &'Sender:'& header line. If you
do not want this to happen, you must set &%local_sender_retain%&, and you must
also set &%local_from_check%& to be false (Exim will complain if you do not).
See also the ACL modifier &`control = suppress_local_fixups`&. Section
&<<SECTthesenhea>>& has more details about &'Sender:'& processing.




.option localhost_number main string&!! unset
.cindex "host" "locally unique number for"
.cindex "message ids" "with multiple hosts"
.cindex multiple "systems sharing a spool"
.cindex "multiple hosts" "sharing a spool"
.cindex "shared spool directory"
.cindex "spool directory" sharing
.vindex "&$localhost_number$&"
Exim's message ids are normally unique only within the local host. If
uniqueness among a set of hosts is required
(eg. because they share a spool directory),
each host must set a different
value for the &%localhost_number%& option. The string is expanded immediately
after reading the configuration file (so that a number can be computed from the
host name, for example) and the result of the expansion must be a number in the
range 0&--16 (or 0&--10 on operating systems with case-insensitive file
systems). This is available in subsequent string expansions via the variable
&$localhost_number$&. When &%localhost_number%& is set, the final four
characters of the message id, instead of just being a fractional part of the
time, are computed from the time and the local host number as described in
section &<<SECTmessiden>>&.



.option log_file_path main "string list&!!" "set at compile time"
.cindex "log" "file path for"
This option sets the path which is used to determine the names of Exim's log
files, or indicates that logging is to be to syslog, or both. It is expanded
when Exim is entered, so it can, for example, contain a reference to the host
name. If no specific path is set for the log files at compile or runtime,
or if the option is unset at runtime (i.e.  &`log_file_path = `&)
they are written in a sub-directory called &_log_& in Exim's spool directory.
A path must start with a slash.
To send to syslog, use the word &"syslog"&.
Chapter &<<CHAPlog>>& contains further details about Exim's logging, and
section &<<SECTwhelogwri>>& describes how the contents of &%log_file_path%& are
used. If this string is fixed at your installation (contains no expansion
variables) it is recommended that you do not set this option in the
configuration file, but instead supply the path using LOG_FILE_PATH in
&_Local/Makefile_& so that it is available to Exim for logging errors detected
early on &-- in particular, failure to read the configuration file.


.option log_selector main string unset
.cindex "log" "selectors"
This option can be used to reduce or increase the number of things that Exim
writes to its log files. Its argument is made up of names preceded by plus or
minus characters. For example:
.code
log_selector = +arguments -retry_defer
.endd
A list of possible names and what they control is given in the chapter on
logging, in section &<<SECTlogselector>>&.


.option log_timezone main boolean false
.cindex "log" "timezone for entries"
.vindex "&$tod_log$&"
.vindex "&$tod_zone$&"
By default, the timestamps on log lines are in local time without the
timezone. This means that if your timezone changes twice a year, the timestamps
in log lines are ambiguous for an hour when the clocks go back. One way of
avoiding this problem is to set the timezone to UTC. An alternative is to set
&%log_timezone%& true. This turns on the addition of the timezone offset to
timestamps in log lines. Turning on this option can add quite a lot to the size
of log files because each line is extended by 6 characters. Note that the
&$tod_log$& variable contains the log timestamp without the zone, but there is
another variable called &$tod_zone$& that contains just the timezone offset.


.option lookup_open_max main integer 25
.cindex "too many open files"
.cindex "open files, too many"
.cindex "file" "too many open"
.cindex "lookup" "maximum open files"
.cindex "limit" "open files for lookups"
This option limits the number of simultaneously open files for single-key
lookups that use regular files (that is, &(lsearch)&, &(dbm)&, and &(cdb)&).
Exim normally keeps these files open during routing, because often the same
file is required several times. If the limit is reached, Exim closes the least
recently used file. Note that if you are using the &'ndbm'& library, it
actually opens two files for each logical DBM database, though it still counts
as one for the purposes of &%lookup_open_max%&. If you are getting &"too many
open files"& errors with NDBM, you need to reduce the value of
&%lookup_open_max%&.


.option max_username_length main integer 0
.cindex "length of login name"
.cindex "user name" "maximum length"
.cindex "limit" "user name length"
Some operating systems are broken in that they truncate long arguments to
&[getpwnam()]& to eight characters, instead of returning &"no such user"&. If
this option is set greater than zero, any attempt to call &[getpwnam()]& with
an argument that is longer behaves as if &[getpwnam()]& failed.


.option message_body_newlines main bool false
.cindex "message body" "newlines in variables"
.cindex "newline" "in message body variables"
.vindex "&$message_body$&"
.vindex "&$message_body_end$&"
By default, newlines in the message body are replaced by spaces when setting
the &$message_body$& and &$message_body_end$& expansion variables. If this
option is set true, this no longer happens.


.option message_body_visible main integer 500
.cindex "body of message" "visible size"
.cindex "message body" "visible size"
.vindex "&$message_body$&"
.vindex "&$message_body_end$&"
This option specifies how much of a message's body is to be included in the
&$message_body$& and &$message_body_end$& expansion variables.


.option message_id_header_domain main string&!! unset
.cindex "&'Message-ID:'& header line"
If this option is set, the string is expanded and used as the right hand side
(domain) of the &'Message-ID:'& header that Exim creates if a
locally-originated incoming message does not have one. &"Locally-originated"&
means &"not received over TCP/IP."&
Otherwise, the primary host name is used.
Only letters, digits, dot and hyphen are accepted; any other characters are
replaced by hyphens. If the expansion is forced to fail, or if the result is an
empty string, the option is ignored.


.option message_id_header_text main string&!! unset
If this variable is set, the string is expanded and used to augment the text of
the &'Message-id:'& header that Exim creates if a locally-originated incoming
message does not have one. The text of this header is required by
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
to take the form of an address. By default, Exim uses its internal message id as
the local part, and the primary host name as the domain. If this option is set,
it is expanded, and provided the expansion is not forced to fail, and does not
yield an empty string, the result is inserted into the header immediately
before the @, separated from the internal message id by a dot. Any characters
that are illegal in an address are automatically converted into hyphens. This
means that variables such as &$tod_log$& can be used, because the spaces and
colons will become hyphens.


.option message_logs main boolean true
.cindex "message logs" "disabling"
.cindex "log" "message log; disabling"
If this option is turned off, per-message log files are not created in the
&_msglog_& spool sub-directory. This reduces the amount of disk I/O required by
Exim, by reducing the number of files involved in handling a message from a
minimum of four (header spool file, body spool file, delivery journal, and
per-message log) to three. The other major I/O activity is Exim's main log,
which is not affected by this option.


.option message_size_limit main string&!! 50M
.cindex "message" "size limit"
.cindex "limit" "message size"
.cindex "size" "of message, limit"
This option limits the maximum size of message that Exim will process. The
value is expanded for each incoming connection so, for example, it can be made
to depend on the IP address of the remote host for messages arriving via
TCP/IP. After expansion, the value must be a sequence of decimal digits,
optionally followed by K or M.

.cindex "SIZE" "ESMTP extension, advertising"
.cindex "ESMTP extensions" SIZE
If nonzero the value will be advertised as a parameter to the ESMTP SIZE
service extension keyword.

&*Note*&: This limit cannot be made to depend on a message's sender or any
other properties of an individual message, because it has to be advertised in
the server's response to EHLO. String expansion failure causes a temporary
error. A value of zero means no limit, but its use is not recommended. See also
&%bounce_return_size_limit%&.

Incoming SMTP messages are failed with a 552 error if the limit is
exceeded; locally-generated messages either get a stderr message or a delivery
failure message to the sender, depending on the &%-oe%& setting. Rejection of
an oversized message is logged in both the main and the reject logs. See also
the generic transport option &%message_size_limit%&, which limits the size of
message that an individual transport can process.

If you use a virus-scanner and set this option to to a value larger than the
maximum size that your virus-scanner is configured to support, you may get
failures triggered by large mails.  The right size to configure for the
virus-scanner depends upon what data is passed and the options in use but it's
probably safest to just set it to a little larger than this value.  E.g., with a
default Exim message size of 50M and a default ClamAV StreamMaxLength of 10M,
some problems may result.

A value of 0 will disable size limit checking; Exim will still advertise the
SIZE extension in an EHLO response, but without a limit, so as to permit
SMTP clients to still indicate the message size along with the MAIL verb.


.option move_frozen_messages main boolean false
.cindex "frozen messages" "moving"
This option, which is available only if Exim has been built with the setting
.code
SUPPORT_MOVE_FROZEN_MESSAGES=yes
.endd
in &_Local/Makefile_&, causes frozen messages and their message logs to be
moved from the &_input_& and &_msglog_& directories on the spool to &_Finput_&
and &_Fmsglog_&, respectively. There is currently no support in Exim or the
standard utilities for handling such moved messages, and they do not show up in
lists generated by &%-bp%& or by the Exim monitor.


.option mua_wrapper main boolean false
Setting this option true causes Exim to run in a very restrictive mode in which
it passes messages synchronously to a smart host. Chapter &<<CHAPnonqueueing>>&
contains a full description of this facility.



.option mysql_servers main "string list" unset
.cindex "MySQL" "server list"
This option provides a list of MySQL servers and associated connection data, to
be used in conjunction with &(mysql)& lookups (see section &<<SECID72>>&). The
option is available only if Exim has been built with MySQL support.


.option never_users main "string list&!!" unset
This option is expanded just once, at the start of Exim's processing. Local
message deliveries are normally run in processes that are setuid to the
recipient, and remote deliveries are normally run under Exim's own uid and gid.
It is usually desirable to prevent any deliveries from running as root, as a
safety precaution.

When Exim is built, an option called FIXED_NEVER_USERS can be set to a
list of users that must not be used for local deliveries. This list is fixed in
the binary and cannot be overridden by the configuration file. By default, it
contains just the single user name &"root"&. The &%never_users%& runtime option
can be used to add more users to the fixed list.

If a message is to be delivered as one of the users on the fixed list or the
&%never_users%& list, an error occurs, and delivery is deferred. A common
example is
.code
never_users = root:daemon:bin
.endd
Including root is redundant if it is also on the fixed list, but it does no
harm. This option overrides the &%pipe_as_creator%& option of the &(pipe)&
transport driver.


.option notifier_socket main string "$spool_directory/exim_daemon_notify"
This option gives the name for a unix-domain socket on which the daemon
listens for work and information-requests.
Only installations running multiple daemons sharing a spool directory
should need to modify the default.

The option is expanded before use.
If the platform supports Linux-style abstract socket names, the result
is used with a nul byte prefixed.
Otherwise,
it should be a full path name and use a directory accessible
to Exim.

If this option is set as empty,
or the command line &%-oY%& option is used, or
the command line uses a &%-oX%& option and does not use &%-oP%&,
then a notifier socket is not created.


.option openssl_options main "string list" "+no_sslv2 +no_sslv3 +single_dh_use +no_ticket +no_renegotiation"
.cindex "OpenSSL "compatibility options"
This option allows an administrator to adjust the SSL options applied
by OpenSSL to connections.  It is given as a space-separated list of items,
each one to be +added or -subtracted from the current value.

This option is only available if Exim is built against OpenSSL.  The values
available for this option vary according to the age of your OpenSSL install.
The &"all"& value controls a subset of flags which are available, typically
the bug workaround options.  The &'SSL_CTX_set_options'& man page will
list the values known on your system and Exim should support all the
&"bug workaround"& options and many of the &"modifying"& options.  The Exim
names lose the leading &"SSL_OP_"& and are lower-cased.

Note that adjusting the options can have severe impact upon the security of
SSL as used by Exim.  It is possible to disable safety checks and shoot
yourself in the foot in various unpleasant ways.  This option should not be
adjusted lightly.  An unrecognised item will be detected at startup, by
invoking Exim with the &%-bV%& flag.

The option affects Exim operating both as a server and as a client.

Historical note: prior to release 4.80, Exim defaulted this value to
"+dont_insert_empty_fragments", which may still be needed for compatibility
with some clients, but which lowers security by increasing exposure to
some now infamous attacks.

Examples:
.code
# Make both old MS and old Eudora happy:
openssl_options = -all +microsoft_big_sslv3_buffer \
                       +dont_insert_empty_fragments

# Disable older protocol versions:
openssl_options = +no_sslv2 +no_sslv3
.endd

Possible options may include:
.ilist
&`all`&
.next
&`allow_unsafe_legacy_renegotiation`&
.next
&`cipher_server_preference`&
.next
&`dont_insert_empty_fragments`&
.next
&`ephemeral_rsa`&
.next
&`legacy_server_connect`&
.next
&`microsoft_big_sslv3_buffer`&
.next
&`microsoft_sess_id_bug`&
.next
&`msie_sslv2_rsa_padding`&
.next
&`netscape_challenge_bug`&
.next
&`netscape_reuse_cipher_change_bug`&
.next
&`no_compression`&
.next
&`no_session_resumption_on_renegotiation`&
.next
&`no_sslv2`&
.next
&`no_sslv3`&
.next
&`no_ticket`&
.next
&`no_tlsv1`&
.next
&`no_tlsv1_1`&
.next
&`no_tlsv1_2`&
.next
&`safari_ecdhe_ecdsa_bug`&
.next
&`single_dh_use`&
.next
&`single_ecdh_use`&
.next
&`ssleay_080_client_dh_bug`&
.next
&`sslref2_reuse_cert_type_bug`&
.next
&`tls_block_padding_bug`&
.next
&`tls_d5_bug`&
.next
&`tls_rollback_bug`&
.endlist

As an aside, the &`safari_ecdhe_ecdsa_bug`& item is a misnomer and affects
all clients connecting using the MacOS SecureTransport TLS facility prior
to MacOS 10.8.4, including email clients.  If you see old MacOS clients failing
to negotiate TLS then this option value might help, provided that your OpenSSL
release is new enough to contain this work-around.  This may be a situation
where you have to upgrade OpenSSL to get buggy clients working.


.option oracle_servers main "string list" unset
.cindex "Oracle" "server list"
This option provides a list of Oracle servers and associated connection data,
to be used in conjunction with &(oracle)& lookups (see section &<<SECID72>>&).
The option is available only if Exim has been built with Oracle support.


.option panic_coredump main boolean false
This option is rarely needed but can help for some debugging investigations.
If set, when an internal error is detected by Exim which is sufficient
to terminate the process
(all such are logged in the paniclog)
then a coredump is requested.

Note that most systems require additional administrative configuration
to permit write a core file for a setuid program, which is Exim's
common installed configuration.

.option percent_hack_domains main "domain list&!!" unset
.cindex "&""percent hack""&"
.cindex "source routing" "in email address"
.cindex "address" "source-routed"
The &"percent hack"& is the convention whereby a local part containing a
percent sign is re-interpreted as a new email address, with the percent
replaced by @. This is sometimes called &"source routing"&, though that term is
also applied to
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
addresses that begin with an @ character. If this
option is set, Exim implements the percent facility for those domains listed,
but no others. This happens before an incoming SMTP address is tested against
an ACL.

&*Warning*&: The &"percent hack"& has often been abused by people who are
trying to get round relaying restrictions. For this reason, it is best avoided
if at all possible. Unfortunately, a number of less security-conscious MTAs
implement it unconditionally. If you are running Exim on a gateway host, and
routing mail through to internal MTAs without processing the local parts, it is
a good idea to reject recipient addresses with percent characters in their
local parts. Exim's default configuration does this.


.options perl_at_start main boolean false &&&
         perl_startup main string unset
.cindex "Perl"
These options are available only when Exim is built with an embedded Perl
interpreter. See chapter &<<CHAPperl>>& for details of their use.

.option perl_taintmode main boolean false
.cindex "Perl"
This option enables the taint mode of the embedded Perl interpreter.


.option pgsql_servers main "string list" unset
.cindex "PostgreSQL lookup type" "server list"
This option provides a list of PostgreSQL servers and associated connection
data, to be used in conjunction with &(pgsql)& lookups (see section
&<<SECID72>>&). The option is available only if Exim has been built with
PostgreSQL support.


.option pid_file_path main string&!! "set at compile time"
.cindex "daemon" "pid file path"
.cindex "pid file, path for"
This option sets the name of the file to which the Exim daemon writes its
process id. The string is expanded, so it can contain, for example, references
to the host name:
.code
pid_file_path = /var/log/$primary_hostname/exim.pid
.endd
If no path is set, the pid is written to the file &_exim-daemon.pid_& in Exim's
spool directory.
The value set by the option can be overridden by the &%-oP%& command line
option. A pid file is not written if a &"non-standard"& daemon is run by means
of the &%-oX%& option, unless a path is explicitly supplied by &%-oP%&.


.option pipelining_advertise_hosts main "host list&!!" *
.cindex "PIPELINING" "suppressing advertising"
.cindex "ESMTP extensions" PIPELINING
This option can be used to suppress the advertisement of the SMTP
PIPELINING extension to specific hosts. See also the &*no_pipelining*&
control in section &<<SECTcontrols>>&. When PIPELINING is not advertised and
&%smtp_enforce_sync%& is true, an Exim server enforces strict synchronization
for each SMTP command and response. When PIPELINING is advertised, Exim assumes
that clients will use it; &"out of order"& commands that are &"expected"& do
not count as protocol errors (see &%smtp_max_synprot_errors%&).

.option pipelining_connect_advertise_hosts main "host list&!!" *
.cindex "pipelining" "early connection"
.cindex "pipelining" PIPECONNECT
.cindex "ESMTP extensions" PIPECONNECT
If Exim is built without the DISABLE_PIPE_CONNECT build option
this option controls which hosts the facility is advertised to
and from which pipeline early-connection (before MAIL) SMTP
commands are acceptable.
When used, the pipelining saves on roundtrip times.

See also the &%hosts_pipe_connect%& smtp transport option.

The SMTP service extension keyword advertised is &"PIPECONNECT"&;
it permits the client to pipeline
TCP connection and hello command (cleatext phase),
or TLS-establishment and hello command (encrypted phase),
on later connections to the same host.


.option prdr_enable main boolean false
.cindex "PRDR" "enabling on server"
.cindex "ESMTP extensions" PRDR
This option can be used to enable the Per-Recipient Data Response extension
to SMTP, defined by Eric Hall.
If the option is set, PRDR is advertised by Exim when operating as a server.
If the client requests PRDR, and more than one recipient, for a message
an additional ACL is called for each recipient after the message content
is received.  See section &<<SECTPRDRACL>>&.

.option preserve_message_logs main boolean false
.cindex "message logs" "preserving"
If this option is set, message log files are not deleted when messages are
completed. Instead, they are moved to a sub-directory of the spool directory
called &_msglog.OLD_&, where they remain available for statistical or debugging
purposes. This is a dangerous option to set on systems with any appreciable
volume of mail. Use with care!


.option primary_hostname main string "see below"
.cindex "name" "of local host"
.cindex "host" "name of local"
.cindex "local host" "name of"
.vindex "&$primary_hostname$&"
This specifies the name of the current host. It is used in the default EHLO or
HELO command for outgoing SMTP messages (changeable via the &%helo_data%&
option in the &(smtp)& transport), and as the default for &%qualify_domain%&.
The value is also used by default in some SMTP response messages from an Exim
server. This can be changed dynamically by setting &%smtp_active_hostname%&.

If &%primary_hostname%& is not set, Exim calls &[uname()]& to find the host
name. If this fails, Exim panics and dies. If the name returned by &[uname()]&
contains only one component, Exim passes it to &[gethostbyname()]& (or
&[getipnodebyname()]& when available) in order to obtain the fully qualified
version. The variable &$primary_hostname$& contains the host name, whether set
explicitly by this option, or defaulted.


.option print_topbitchars main boolean false
.cindex "printing characters"
.cindex "8-bit characters"
By default, Exim considers only those characters whose codes lie in the range
32&--126 to be printing characters. In a number of circumstances (for example,
when writing log entries) non-printing characters are converted into escape
sequences, primarily to avoid messing up the layout. If &%print_topbitchars%&
is set, code values of 128 and above are also considered to be printing
characters.

This option also affects the header syntax checks performed by the
&(autoreply)& transport, and whether Exim uses
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047) encoding of
the user's full name when constructing From: and Sender: addresses (as
described in section &<<SECTconstr>>&). Setting this option can cause
Exim to generate eight bit message headers that do not conform to the
standards.


.option process_log_path main string unset
.cindex "process log path"
.cindex "log" "process log"
.cindex "&'exiwhat'&"
This option sets the name of the file to which an Exim process writes its
&"process log"& when sent a USR1 signal. This is used by the &'exiwhat'&
utility script. If this option is unset, the file called &_exim-process.info_&
in Exim's spool directory is used. The ability to specify the name explicitly
can be useful in environments where two different Exims are running, using
different spool directories.


.option prod_requires_admin main boolean true
.cindex "restricting access to features"
.oindex "&%-M%&"
.oindex "&%-R%&"
.oindex "&%-q%&"
The &%-M%&, &%-R%&, and &%-q%& command-line options require the caller to be an
admin user unless &%prod_requires_admin%& is set false. See also
&%queue_list_requires_admin%& and &%commandline_checks_require_admin%&.


.option proxy_protocol_timeout main time 3s
.cindex proxy "proxy protocol"
This option sets the timeout for proxy protocol negotiation.
For details see section &<<SECTproxyInbound>>&.


.option qualify_domain main string "see below"
.cindex "domain" "for qualifying addresses"
.cindex "address" "qualification"
This option specifies the domain name that is added to any envelope sender
addresses that do not have a domain qualification. It also applies to
recipient addresses if &%qualify_recipient%& is not set. Unqualified addresses
are accepted by default only for locally-generated messages. Qualification is
also applied to addresses in header lines such as &'From:'& and &'To:'& for
locally-generated messages, unless the &%-bnq%& command line option is used.

Messages from external sources must always contain fully qualified addresses,
unless the sending host matches &%sender_unqualified_hosts%& or
&%recipient_unqualified_hosts%& (as appropriate), in which case incoming
addresses are qualified with &%qualify_domain%& or &%qualify_recipient%& as
necessary. Internally, Exim always works with fully qualified envelope
addresses. If &%qualify_domain%& is not set, it defaults to the
&%primary_hostname%& value.


.option qualify_recipient main string "see below"
This option allows you to specify a different domain for qualifying recipient
addresses to the one that is used for senders. See &%qualify_domain%& above.



.option queue_domains main "domain list&!!" unset
.cindex "domain" "specifying non-immediate delivery"
.cindex "queueing incoming messages"
.cindex "message" "queueing certain domains"
This option lists domains for which immediate delivery is not required.
A delivery process is started whenever a message is received, but only those
domains that do not match are processed. All other deliveries wait until the
next queue run. See also &%hold_domains%& and &%queue_smtp_domains%&.


.option queue_fast_ramp main boolean true
.cindex "queue runner" "two phase"
.cindex "queue" "double scanning"
If set to true, two-phase queue runs, initiated using &%-qq%& on the
command line, may start parallel delivery processes during their first
phase.  This will be done when a threshold number of messages have been
routed for a single host.


.option queue_list_requires_admin main boolean true
.cindex "restricting access to features"
.oindex "&%-bp%&"
The &%-bp%& command-line option, which lists the messages that are on the
queue, requires the caller to be an admin user unless
&%queue_list_requires_admin%& is set false.
See also &%prod_requires_admin%& and &%commandline_checks_require_admin%&.


.option queue_only main boolean false
.cindex "queueing incoming messages"
.cindex "message" "queueing unconditionally"
If &%queue_only%& is set, a delivery process is not automatically started
whenever a message is received. Instead, the message waits in the queue for the
next queue run. Even if &%queue_only%& is false, incoming messages may not get
delivered immediately when certain conditions (such as heavy load) occur.

The &%-odq%& command line has the same effect as &%queue_only%&. The &%-odb%&
and &%-odi%& command line options override &%queue_only%& unless
&%queue_only_override%& is set false. See also &%queue_only_file%&,
&%queue_only_load%&, and &%smtp_accept_queue%&.


.option queue_only_file main "string list" unset
.cindex "queueing incoming messages"
.cindex "message" "queueing by file existence"
This option can be set to a colon-separated list of absolute path names, each
one optionally preceded by &"smtp"&. When Exim is receiving a message,
it tests for the existence of each listed path using a call to &[stat()]&. For
each path that exists, the corresponding queueing option is set.
For paths with no prefix, &%queue_only%& is set; for paths prefixed by
&"smtp"&, &%queue_smtp_domains%& is set to match all domains. So, for example,
.code
queue_only_file = smtp/some/file
.endd
causes Exim to behave as if &%queue_smtp_domains%& were set to &"*"& whenever
&_/some/file_& exists.


.option queue_only_load main fixed-point unset
.cindex "load average"
.cindex "queueing incoming messages"
.cindex "message" "queueing by load"
If the system load average is higher than this value, incoming messages from
all sources are queued, and no automatic deliveries are started. If this
happens during local or remote SMTP input, all subsequent messages received on
the same SMTP connection are queued by default, whatever happens to the load in
the meantime, but this can be changed by setting &%queue_only_load_latch%&
false.

Deliveries will subsequently be performed by queue runner processes. This
option has no effect on ancient operating systems on which Exim cannot
determine the load average. See also &%deliver_queue_load_max%& and
&%smtp_load_reserve%&.


.option queue_only_load_latch main boolean true
.cindex "load average" "re-evaluating per message"
When this option is true (the default), once one message has been queued
because the load average is higher than the value set by &%queue_only_load%&,
all subsequent messages received on the same SMTP connection are also queued.
This is a deliberate choice; even though the load average may fall below the
threshold, it doesn't seem right to deliver later messages on the same
connection when not delivering earlier ones. However, there are special
circumstances such as very long-lived connections from scanning appliances
where this is not the best strategy. In such cases, &%queue_only_load_latch%&
should be set false. This causes the value of the load average to be
re-evaluated for each message.


.option queue_only_override main boolean true
.cindex "queueing incoming messages"
When this option is true, the &%-od%&&'x'& command line options override the
setting of &%queue_only%& or &%queue_only_file%& in the configuration file. If
&%queue_only_override%& is set false, the &%-od%&&'x'& options cannot be used
to override; they are accepted, but ignored.


.option queue_run_in_order main boolean false
.cindex "queue runner" "processing messages in order"
If this option is set, queue runs happen in order of message arrival instead of
in an arbitrary order. For this to happen, a complete list of the entire queue
must be set up before the deliveries start. When the queue is all held in a
single directory (the default), a single list is created for both the ordered
and the non-ordered cases. However, if &%split_spool_directory%& is set, a
single list is not created when &%queue_run_in_order%& is false. In this case,
the sub-directories are processed one at a time (in a random order), and this
avoids setting up one huge list for the whole queue. Thus, setting
&%queue_run_in_order%& with &%split_spool_directory%& may degrade performance
when the queue is large, because of the extra work in setting up the single,
large list. In most situations, &%queue_run_in_order%& should not be set.



.option queue_run_max main integer&!! 5
.cindex "queue runner" "maximum number of"
This controls the maximum number of queue runner processes that an Exim daemon
can run simultaneously. This does not mean that it starts them all at once,
but rather that if the maximum number are still running when the time comes to
start another one, it refrains from starting another one. This can happen with
very large queues and/or very sluggish deliveries. This option does not,
however, interlock with other processes, so additional queue runners can be
started by other means, or by killing and restarting the daemon.

Setting this option to zero does not suppress queue runs; rather, it disables
the limit, allowing any number of simultaneous queue runner processes to be
run. If you do not want queue runs to occur, omit the &%-q%&&'xx'& setting on
the daemon's command line.

.cindex queues named
.cindex "named queues" "resource limit"
To set limits for different named queues use
an expansion depending on the &$queue_name$& variable.

.option queue_smtp_domains main "domain list&!!" unset
.cindex "queueing incoming messages"
.cindex "message" "queueing remote deliveries"
.cindex "first pass routing"
When this option is set, a delivery process is started whenever a message is
received, routing is performed, and local deliveries take place.
However, if any SMTP deliveries are required for domains that match
&%queue_smtp_domains%&, they are not immediately delivered, but instead the
message waits in the queue for the next queue run. Since routing of the message
has taken place, Exim knows to which remote hosts it must be delivered, and so
when the queue run happens, multiple messages for the same host are delivered
over a single SMTP connection. The &%-odqs%& command line option causes all
SMTP deliveries to be queued in this way, and is equivalent to setting
&%queue_smtp_domains%& to &"*"&. See also &%hold_domains%& and
&%queue_domains%&.


.option receive_timeout main time 0s
.cindex "timeout" "for non-SMTP input"
This option sets the timeout for accepting a non-SMTP message, that is, the
maximum time that Exim waits when reading a message on the standard input. If
the value is zero, it will wait forever. This setting is overridden by the
&%-or%& command line option. The timeout for incoming SMTP messages is
controlled by &%smtp_receive_timeout%&.

.option received_header_text main string&!! "see below"
.cindex "customizing" "&'Received:'& header"
.cindex "&'Received:'& header line" "customizing"
This string defines the contents of the &'Received:'& message header that is
added to each message, except for the timestamp, which is automatically added
on at the end (preceded by a semicolon). The string is expanded each time it is
used. If the expansion yields an empty string, no &'Received:'& header line is
added to the message. Otherwise, the string should start with the text
&"Received:"& and conform to the
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
specification for &'Received:'& header lines.
The default setting is:

.code
received_header_text = Received: \
  ${if def:sender_rcvhost {from $sender_rcvhost\n\t}\
    {${if def:sender_ident \
      {from ${quote_local_part:$sender_ident} }}\
        ${if def:sender_helo_name {(helo=$sender_helo_name)\n\t}}}}\
  by $primary_hostname \
  ${if def:received_protocol {with $received_protocol }}\
  ${if def:tls_in_ver        { ($tls_in_ver)}}\
  ${if def:tls_in_cipher_std { tls $tls_in_cipher_std\n\t}}\
  (Exim $version_number)\n\t\
  ${if def:sender_address \
  {(envelope-from <$sender_address>)\n\t}}\
  id $message_exim_id\
  ${if def:received_for {\n\tfor $received_for}}
.endd

The references to the TLS version and cipher are
omitted when Exim is built without TLS
support. The use of conditional expansions ensures that this works for both
locally generated messages and messages received from remote hosts, giving
header lines such as the following:
.code
Received: from scrooge.carol.example ([192.168.12.25] ident=root)
by marley.carol.example with esmtp (Exim 4.00)
(envelope-from <bob@carol.example>)
id 16IOWa-00019l-00
for chas@dickens.example; Tue, 25 Dec 2001 14:43:44 +0000
Received: by scrooge.carol.example with local (Exim 4.00)
id 16IOWW-000083-00; Tue, 25 Dec 2001 14:43:41 +0000
.endd
Until the body of the message has been received, the timestamp is the time when
the message started to be received. Once the body has arrived, and all policy
checks have taken place, the timestamp is updated to the time at which the
message was accepted.


.option received_headers_max main integer 30
.cindex "loop" "prevention"
.cindex "mail loop prevention"
.cindex "&'Received:'& header line" "counting"
When a message is to be delivered, the number of &'Received:'& headers is
counted, and if it is greater than this parameter, a mail loop is assumed to
have occurred, the delivery is abandoned, and an error message is generated.
This applies to both local and remote deliveries.


.option recipient_unqualified_hosts main "host list&!!" unset
.cindex "unqualified addresses"
.cindex "host" "unqualified addresses from"
This option lists those hosts from which Exim is prepared to accept unqualified
recipient addresses in message envelopes. The addresses are made fully
qualified by the addition of the &%qualify_recipient%& value. This option also
affects message header lines. Exim does not reject unqualified recipient
addresses in headers, but it qualifies them only if the message came from a
host that matches &%recipient_unqualified_hosts%&,
or if the message was submitted locally (not using TCP/IP), and the &%-bnq%&
option was not set.


.option recipients_max main integer&!! 50000
.cindex "limit" "number of recipients"
.cindex "recipient" "maximum number"
If the value resulting from expanding this option
is set greater than zero, it specifies the maximum number of
original recipients for any message. Additional recipients that are generated
by aliasing or forwarding do not count. SMTP messages get a 452 response for
all recipients over the limit; earlier recipients are delivered as normal.
Non-SMTP messages with too many recipients are failed, and no deliveries are
done.

For SMTP message the expansion is done after the connection is
accepted (but before any SMTP conversation) and may depend on
the IP addresses and port numbers of the connection.
&*Note*&: If an expansion is used for the option,
care should be taken that a resonable value results for
non-SMTP messages.

.cindex "RCPT" "maximum number of incoming"
&*Note*&: The RFCs specify that an SMTP server should accept at least 100
RCPT commands in a single message.


.option recipients_max_reject main boolean false
If this option is set true, Exim rejects SMTP messages containing too many
recipients by giving 552 errors to the surplus RCPT commands, and a 554
error to the eventual DATA command. Otherwise (the default) it gives a 452
error to the surplus RCPT commands and accepts the message on behalf of the
initial set of recipients. The remote server should then re-send the message
for the remaining recipients at a later time.


.option remote_max_parallel main integer 4
.cindex "delivery" "parallelism for remote"
This option controls parallel delivery of one message to a number of remote
hosts. If the value is less than 2, parallel delivery is disabled, and Exim
does all the remote deliveries for a message one by one. Otherwise, if a single
message has to be delivered to more than one remote host, or if several copies
have to be sent to the same remote host, up to &%remote_max_parallel%&
deliveries are done simultaneously. If more than &%remote_max_parallel%&
deliveries are required, the maximum number of processes are started, and as
each one finishes, another is begun. The order of starting processes is the
same as if sequential delivery were being done, and can be controlled by the
&%remote_sort_domains%& option. If parallel delivery takes place while running
with debugging turned on, the debugging output from each delivery process is
tagged with its process id.

This option controls only the maximum number of parallel deliveries for one
message in one Exim delivery process. Because Exim has no central queue
manager, there is no way of controlling the total number of simultaneous
deliveries if the configuration allows a delivery attempt as soon as a message
is received.

See also the &%max_parallel%& generic transport option,
and the &%serialize_hosts%& smtp transport option.

.cindex "number of deliveries"
.cindex "delivery" "maximum number of"
If you want to control the total number of deliveries on the system, you
need to set the &%queue_only%& option. This ensures that all incoming messages
are added to the queue without starting a delivery process. Then set up an Exim
daemon to start queue runner processes at appropriate intervals (probably
fairly often, for example, every minute), and limit the total number of queue
runners by setting the &%queue_run_max%& parameter. Because each queue runner
delivers only one message at a time, the maximum number of deliveries that can
then take place at once is &%queue_run_max%& multiplied by
&%remote_max_parallel%&.

If it is purely remote deliveries you want to control, use
&%queue_smtp_domains%& instead of &%queue_only%&. This has the added benefit of
doing the SMTP routing before queueing, so that several messages for the same
host will eventually get delivered down the same connection.


.option remote_sort_domains main "domain list&!!" unset
.cindex "sorting remote deliveries"
.cindex "delivery" "sorting remote"
When there are a number of remote deliveries for a message, they are sorted by
domain into the order given by this list. For example,
.code
remote_sort_domains = *.cam.ac.uk:*.uk
.endd
would attempt to deliver to all addresses in the &'cam.ac.uk'& domain first,
then to those in the &%uk%& domain, then to any others.


.option retry_data_expire main time 7d
.cindex "hints database" "data expiry"
This option sets a &"use before"& time on retry information in Exim's hints
database. Any older retry data is ignored. This means that, for example, once a
host has not been tried for 7 days, Exim behaves as if it has no knowledge of
past failures.


.option retry_interval_max main time 24h
.cindex "retry" "limit on interval"
.cindex "limit" "on retry interval"
Chapter &<<CHAPretry>>& describes Exim's mechanisms for controlling the
intervals between delivery attempts for messages that cannot be delivered
straight away. This option sets an overall limit to the length of time between
retries. It cannot be set greater than 24 hours; any attempt to do so forces
the default value.


.option return_path_remove main boolean true
.cindex "&'Return-path:'& header line" "removing"
&url(https://www.rfc-editor.org/rfc/rfc2821,RFC 2821), section 4.4,
states that an SMTP server must insert a
&'Return-path:'& header line into a message when it makes a &"final delivery"&.
The &'Return-path:'& header preserves the sender address as received in the
MAIL command. This description implies that this header should not be present
in an incoming message. If &%return_path_remove%& is true, any existing
&'Return-path:'& headers are removed from messages at the time they are
received. Exim's transports have options for adding &'Return-path:'& headers at
the time of delivery. They are normally used only for final local deliveries.


.option return_size_limit main integer 100K
This option is an obsolete synonym for &%bounce_return_size_limit%&.


.option rfc1413_hosts main "host list&!!" @[]
.cindex "RFC 1413"
.cindex "host" "for RFC 1413 calls"
&url(https://www.rfc-editor.org/rfc/rfc1413,RFC 1413)
identification calls are made to any client host which matches
an item in the list.
The default value specifies just this host, being any local interface
for the system.

.option rfc1413_query_timeout main time 0s
.cindex "RFC 1413" "query timeout"
.cindex "timeout" "for RFC 1413 call"
This sets the timeout on
&url(https://www.rfc-editor.org/rfc/rfc1413,RFC 1413)
identification calls. If it is set to zero,
no RFC 1413 calls are ever made.


.option sender_unqualified_hosts main "host list&!!" unset
.cindex "unqualified addresses"
.cindex "host" "unqualified addresses from"
This option lists those hosts from which Exim is prepared to accept unqualified
sender addresses. The addresses are made fully qualified by the addition of
&%qualify_domain%&. This option also affects message header lines. Exim does
not reject unqualified addresses in headers that contain sender addresses, but
it qualifies them only if the message came from a host that matches
&%sender_unqualified_hosts%&, or if the message was submitted locally (not
using TCP/IP), and the &%-bnq%& option was not set.


.option slow_lookup_log main integer 0
.cindex "logging" "slow lookups"
.cindex "dns" "logging slow lookups"
This option controls logging of slow lookups.
If the value is nonzero it is taken as a number of milliseconds
and lookups taking longer than this are logged.
Currently this applies only to DNS lookups.



.option smtp_accept_keepalive main boolean true
.cindex "keepalive" "on incoming connection"
This option controls the setting of the SO_KEEPALIVE option on incoming
TCP/IP socket connections. When set, it causes the kernel to probe idle
connections periodically, by sending packets with &"old"& sequence numbers. The
other end of the connection should send an acknowledgment if the connection is
still okay or a reset if the connection has been aborted. The reason for doing
this is that it has the beneficial effect of freeing up certain types of
connection that can get stuck when the remote host is disconnected without
tidying up the TCP/IP call properly. The keepalive mechanism takes several
hours to detect unreachable hosts.



.option smtp_accept_max main integer 20
.cindex "limit" "incoming SMTP connections"
.cindex "SMTP" "incoming connection count"
.cindex "inetd"
This option specifies the maximum number of simultaneous incoming SMTP calls
that Exim will accept. It applies only to the listening daemon; there is no
control (in Exim) when incoming SMTP is being handled by &'inetd'&. If the
value is set to zero, no limit is applied. However, it is required to be
non-zero if either &%smtp_accept_max_per_host%& or &%smtp_accept_queue%& is
set. See also &%smtp_accept_reserve%& and &%smtp_load_reserve%&.

A new SMTP connection is immediately rejected if the &%smtp_accept_max%& limit
has been reached. If not, Exim first checks &%smtp_accept_max_per_host%&. If
that limit has not been reached for the client host, &%smtp_accept_reserve%&
and &%smtp_load_reserve%& are then checked before accepting the connection.


.option smtp_accept_max_nonmail main integer 10
.cindex "limit" "non-mail SMTP commands"
.cindex "SMTP" "limiting non-mail commands"
Exim counts the number of &"non-mail"& commands in an SMTP session, and drops
the connection if there are too many. This option defines &"too many"&. The
check catches some denial-of-service attacks, repeated failing AUTHs, or a mad
client looping sending EHLO, for example. The check is applied only if the
client host matches &%smtp_accept_max_nonmail_hosts%&.

When a new message is expected, one occurrence of RSET is not counted. This
allows a client to send one RSET between messages (this is not necessary,
but some clients do it). Exim also allows one uncounted occurrence of HELO
or EHLO, and one occurrence of STARTTLS between messages. After
starting up a TLS session, another EHLO is expected, and so it too is not
counted. The first occurrence of AUTH in a connection, or immediately
following STARTTLS is not counted. Otherwise, all commands other than
MAIL, RCPT, DATA, and QUIT are counted.


.option smtp_accept_max_nonmail_hosts main "host list&!!" *
You can control which hosts are subject to the &%smtp_accept_max_nonmail%&
check by setting this option. The default value makes it apply to all hosts. By
changing the value, you can exclude any badly-behaved hosts that you have to
live with.


. Allow this long option name to split; give it unsplit as a fifth argument
. for the automatic .oindex that is generated by .option.
. We insert " &~&~" which is both pretty nasty visually and results in
. non-searchable text.  HowItWorks.txt mentions an option for inserting
. zero-width-space, which would be nicer visually and results in (at least)
. html that Firefox will split on when it's forced to reflow (rather than
. inserting a horizontal scrollbar).  However, the text is still not
. searchable.  NM changed this occurrence for bug 1197 to no longer allow
. the option name to split.

.option smtp_accept_max_per_connection main integer&!! 1000
.cindex "SMTP" "limiting incoming message count"
.cindex "limit" "messages per SMTP connection"
The value of this option limits the number of MAIL commands that Exim is
prepared to accept over a single SMTP connection, whether or not each command
results in the transfer of a message. After the limit is reached, a 421
response is given to subsequent MAIL commands. This limit is a safety
precaution against a client that goes mad (incidents of this type have been
seen).
The option is expanded after the HELO or EHLO is received
and may depend on values available at that time.
An empty or zero value after expansion removes the limit.


.option smtp_accept_max_per_host main string&!! unset
.cindex "limit" "SMTP connections from one host"
.cindex "host" "limiting SMTP connections from"
This option restricts the number of simultaneous IP connections from a single
host (strictly, from a single IP address) to the Exim daemon. The option is
expanded, to enable different limits to be applied to different hosts by
reference to &$sender_host_address$&. Once the limit is reached, additional
connection attempts from the same host are rejected with error code 421. This
is entirely independent of &%smtp_accept_reserve%&. The option's default value
of zero imposes no limit. If this option is set greater than zero, it is
required that &%smtp_accept_max%& be non-zero.

&*Warning*&: When setting this option you should not use any expansion
constructions that take an appreciable amount of time. The expansion and test
happen in the main daemon loop, in order to reject additional connections
without forking additional processes (otherwise a denial-of-service attack
could cause a vast number or processes to be created). While the daemon is
doing this processing, it cannot accept any other incoming connections.



.option smtp_accept_queue main integer 0
.cindex "SMTP" "incoming connection count"
.cindex "queueing incoming messages"
.cindex "message" "queueing by SMTP connection count"
If the number of simultaneous incoming SMTP connections being handled via the
listening daemon exceeds this value, messages received by SMTP are just placed
in the queue; no delivery processes are started automatically. The count is
fixed at the start of an SMTP connection. It cannot be updated in the
subprocess that receives messages, and so the queueing or not queueing applies
to all messages received in the same connection.

A value of zero implies no limit, and clearly any non-zero value is useful only
if it is less than the &%smtp_accept_max%& value (unless that is zero). See
also &%queue_only%&, &%queue_only_load%&, &%queue_smtp_domains%&, and the
various &%-od%&&'x'& command line options.


. See the comment on smtp_accept_max_per_connection

.option smtp_accept_queue_per_connection main integer 10
.cindex "queueing incoming messages"
.cindex "message" "queueing by message count"
This option limits the number of delivery processes that Exim starts
automatically when receiving messages via SMTP, whether via the daemon or by
the use of &%-bs%& or &%-bS%&. If the value of the option is greater than zero,
and the number of messages received in a single SMTP session exceeds this
number, subsequent messages are placed in the queue, but no delivery processes
are started. This helps to limit the number of Exim processes when a server
restarts after downtime and there is a lot of mail waiting for it on other
systems. On large systems, the default should probably be increased, and on
dial-in client systems it should probably be set to zero (that is, disabled).


.option smtp_accept_reserve main integer 0
.cindex "SMTP" "incoming call count"
.cindex "host" "reserved"
When &%smtp_accept_max%& is set greater than zero, this option specifies a
number of SMTP connections that are reserved for connections from the hosts
that are specified in &%smtp_reserve_hosts%&. The value set in
&%smtp_accept_max%& includes this reserve pool. The specified hosts are not
restricted to this number of connections; the option specifies a minimum number
of connection slots for them, not a maximum. It is a guarantee that this group
of hosts can always get at least &%smtp_accept_reserve%& connections. However,
the limit specified by &%smtp_accept_max_per_host%& is still applied to each
individual host.

For example, if &%smtp_accept_max%& is set to 50 and &%smtp_accept_reserve%& is
set to 5, once there are 45 active connections (from any hosts), new
connections are accepted only from hosts listed in &%smtp_reserve_hosts%&,
provided the other criteria for acceptance are met.


.option smtp_active_hostname main string&!! unset
.cindex "host" "name in SMTP responses"
.cindex "SMTP" "host name in responses"
.vindex "&$primary_hostname$&"
This option is provided for multi-homed servers that want to masquerade as
several different hosts. At the start of an incoming SMTP connection, its value
is expanded and used instead of the value of &$primary_hostname$& in SMTP
responses. For example, it is used as domain name in the response to an
incoming HELO or EHLO command.

.vindex "&$smtp_active_hostname$&"
The active hostname is placed in the &$smtp_active_hostname$& variable, which
is saved with any messages that are received. It is therefore available for use
in routers and transports when the message is later delivered.

If this option is unset, or if its expansion is forced to fail, or if the
expansion results in an empty string, the value of &$primary_hostname$& is
used. Other expansion failures cause a message to be written to the main and
panic logs, and the SMTP command receives a temporary error. Typically, the
value of &%smtp_active_hostname%& depends on the incoming interface address.
For example:
.code
smtp_active_hostname = ${if eq{$received_ip_address}{10.0.0.1}\
  {cox.mydomain}{box.mydomain}}
.endd

Although &$smtp_active_hostname$& is primarily concerned with incoming
messages, it is also used as the default for HELO commands in callout
verification if there is no remote transport from which to obtain a
&%helo_data%& value.

.option smtp_backlog_monitor main integer 0
.cindex "connection backlog" monitoring
If this option is set to greater than zero, and the backlog of available
TCP connections on a socket listening for SMTP is larger than it, a line
is logged giving the value and the socket address and port.
The value is retrived jsut before an accept call.
This facility is only available on Linux.

.option smtp_banner main string&!! "see below"
.cindex "SMTP" "welcome banner"
.cindex "banner for SMTP"
.cindex "welcome banner for SMTP"
.cindex "customizing" "SMTP banner"
If a connect ACL does not supply a message,
this string (which is expanded every time it is used) is output as the initial
positive response to an SMTP connection. The default setting is:
.code
smtp_banner = $smtp_active_hostname ESMTP Exim \
  $version_number $tod_full
.endd
Failure to expand the string causes a panic error;
a forced fail just closes the connection.
If you want to create a
multiline response to the initial SMTP connection, use &"\n"& in the string at
appropriate points, but not at the end. Note that the 220 code is not included
in this string. Exim adds it automatically (several times in the case of a
multiline response).


.option smtp_check_spool_space main boolean true
.cindex "checking disk space"
.cindex "disk space, checking"
.cindex "spool directory" "checking space"
When this option is set, if an incoming SMTP session encounters the SIZE
option on a MAIL command, it checks that there is enough space in the
spool directory's partition to accept a message of that size, while still
leaving free the amount specified by &%check_spool_space%& (even if that value
is zero). If there isn't enough space, a temporary error code is returned.


.option smtp_connect_backlog main integer 20
.cindex "connection backlog" "set maximum"
.cindex "SMTP" "connection backlog"
.cindex "backlog of connections"
This option specifies a maximum number of waiting SMTP connections. Exim passes
this value to the TCP/IP system when it sets up its listener. Once this number
of connections are waiting for the daemon's attention, subsequent connection
attempts are refused at the TCP/IP level. At least, that is what the manuals
say; in some circumstances such connection attempts have been observed to time
out instead. For large systems it is probably a good idea to increase the
value (to 50, say). It also gives some protection against denial-of-service
attacks by SYN flooding.


.option smtp_enforce_sync main boolean true
.cindex "SMTP" "synchronization checking"
.cindex "synchronization checking in SMTP"
The SMTP protocol specification requires the client to wait for a response from
the server at certain points in the dialogue. Without PIPELINING these
synchronization points are after every command; with PIPELINING they are
fewer, but they still exist.

Some spamming sites send out a complete set of SMTP commands without waiting
for any response. Exim protects against this by rejecting a message if the
client has sent further input when it should not have. The error response &"554
SMTP synchronization error"& is sent, and the connection is dropped. Testing
for this error cannot be perfect because of transmission delays (unexpected
input may be on its way but not yet received when Exim checks). However, it
does detect many instances.

The check can be globally disabled by setting &%smtp_enforce_sync%& false.
If you want to disable the check selectively (for example, only for certain
hosts), you can do so by an appropriate use of a &%control%& modifier in an ACL
(see section &<<SECTcontrols>>&). See also &%pipelining_advertise_hosts%&.



.option smtp_etrn_command main string&!! unset
.cindex ETRN "command to be run"
.cindex "ESMTP extensions" ETRN
.vindex "&$domain$&"
If this option is set, the given command is run whenever an SMTP ETRN
command is received from a host that is permitted to issue such commands (see
chapter &<<CHAPACL>>&). The string is split up into separate arguments which
are independently expanded. The expansion variable &$domain$& is set to the
argument of the ETRN command, and no syntax checking is done on it. For
example:
.code
smtp_etrn_command = /etc/etrn_command $domain \
                    $sender_host_address
.endd
If the option is not set, the argument for the ETRN command must
be a &'#'& followed by an address string.
In this case an &'exim -R <string>'& command is used;
if the ETRN ACL has set up a named-queue then &'-MCG <queue>'& is appended.

A new process is created to run the command, but Exim does not wait for it to
complete. Consequently, its status cannot be checked. If the command cannot be
run, a line is written to the panic log, but the ETRN caller still receives
a 250 success response. Exim is normally running under its own uid when
receiving SMTP, so it is not possible for it to change the uid before running
the command.


.option smtp_etrn_serialize main boolean true
.cindex ETRN serializing
When this option is set, it prevents the simultaneous execution of more than
one identical command as a result of ETRN in an SMTP connection. See
section &<<SECTETRN>>& for details.


.option smtp_load_reserve main fixed-point unset
.cindex "load average"
If the system load average ever gets higher than this, incoming SMTP calls are
accepted only from those hosts that match an entry in &%smtp_reserve_hosts%&.
If &%smtp_reserve_hosts%& is not set, no incoming SMTP calls are accepted when
the load is over the limit. The option has no effect on ancient operating
systems on which Exim cannot determine the load average. See also
&%deliver_queue_load_max%& and &%queue_only_load%&.



.option smtp_max_synprot_errors main integer 3
.cindex "SMTP" "limiting syntax and protocol errors"
.cindex "limit" "SMTP syntax and protocol errors"
Exim rejects SMTP commands that contain syntax or protocol errors. In
particular, a syntactically invalid email address, as in this command:
.code
RCPT TO:<abc xyz@a.b.c>
.endd
causes immediate rejection of the command, before any other tests are done.
(The ACL cannot be run if there is no valid address to set up for it.) An
example of a protocol error is receiving RCPT before MAIL. If there are
too many syntax or protocol errors in one SMTP session, the connection is
dropped. The limit is set by this option.

.cindex "PIPELINING" "expected errors"
When the PIPELINING extension to SMTP is in use, some protocol errors are
&"expected"&, for instance, a RCPT command after a rejected MAIL command.
Exim assumes that PIPELINING will be used if it advertises it (see
&%pipelining_advertise_hosts%&), and in this situation, &"expected"& errors do
not count towards the limit.



.option smtp_max_unknown_commands main integer 3
.cindex "SMTP" "limiting unknown commands"
.cindex "limit" "unknown SMTP commands"
If there are too many unrecognized commands in an incoming SMTP session, an
Exim server drops the connection. This is a defence against some kinds of abuse
that subvert web
clients
into making connections to SMTP ports; in these circumstances, a number of
non-SMTP command lines are sent first.



.options smtp_ratelimit_hosts main "host list&!!" unset &&&
         smtp_ratelimit_mail main string unset &&&
         smtp_ratelimit_rcpt main string unset
.cindex "SMTP" "rate limiting"
.cindex "limit" "rate of message arrival"
.cindex "RCPT" "rate limiting"
Some sites find it helpful to be able to limit the rate at which certain hosts
can send them messages, and the rate at which an individual message can specify
recipients.

Exim has two rate-limiting facilities. This section describes the older
facility, which can limit rates within a single connection. The newer
&%ratelimit%& ACL condition can limit rates across all connections. See section
&<<SECTratelimiting>>& for details of the newer facility.

When a host matches &%smtp_ratelimit_hosts%&, the values of
&%smtp_ratelimit_mail%& and &%smtp_ratelimit_rcpt%& are used to control the
rate of acceptance of MAIL and RCPT commands in a single SMTP session,
respectively. Each option, if set, must contain a set of four comma-separated
values:

.ilist
A threshold, before which there is no rate limiting.
.next
An initial time delay. Unlike other times in Exim, numbers with decimal
fractional parts are allowed here.
.next
A factor by which to increase the delay each time.
.next
A maximum value for the delay. This should normally be less than 5 minutes,
because after that time, the client is liable to timeout the SMTP command.
.endlist

For example, these settings have been used successfully at the site which
first suggested this feature, for controlling mail from their customers:
.code
smtp_ratelimit_mail = 2,0.5s,1.05,4m
smtp_ratelimit_rcpt = 4,0.25s,1.015,4m
.endd
The first setting specifies delays that are applied to MAIL commands after
two have been received over a single connection. The initial delay is 0.5
seconds, increasing by a factor of 1.05 each time. The second setting applies
delays to RCPT commands when more than four occur in a single message.



.option smtp_receive_timeout main time&!! 5m
.cindex "timeout" "for SMTP input"
.cindex "SMTP" "input timeout"
This sets a timeout value for SMTP reception. It applies to all forms of SMTP
input, including batch SMTP. If a line of input (either an SMTP command or a
data line) is not received within this time, the SMTP connection is dropped and
the message is abandoned.
A line is written to the log containing one of the following messages:
.code
SMTP command timeout on connection from...
SMTP data timeout on connection from...
.endd
The former means that Exim was expecting to read an SMTP command; the latter
means that it was in the DATA phase, reading the contents of a message.

If the first character of the option is a &"$"& the option is
expanded before use and may depend on
&$sender_host_name$&, &$sender_host_address$& and &$sender_host_port$&.


.oindex "&%-os%&"
The value set by this option can be overridden by the
&%-os%& command-line option. A setting of zero time disables the timeout, but
this should never be used for SMTP over TCP/IP. (It can be useful in some cases
of local input using &%-bs%& or &%-bS%&.) For non-SMTP input, the reception
timeout is controlled by &%receive_timeout%& and &%-or%&.


.option smtp_reserve_hosts main "host list&!!" unset
This option defines hosts for which SMTP connections are reserved; see
&%smtp_accept_reserve%& and &%smtp_load_reserve%& above.


.option smtp_return_error_details main boolean false
.cindex "SMTP" "details policy failures"
.cindex "policy control" "rejection, returning details"
In the default state, Exim uses bland messages such as
&"Administrative prohibition"& when it rejects SMTP commands for policy
reasons. Many sysadmins like this because it gives away little information
to spammers. However, some other sysadmins who are applying strict checking
policies want to give out much fuller information about failures. Setting
&%smtp_return_error_details%& true causes Exim to be more forthcoming. For
example, instead of &"Administrative prohibition"&, it might give:
.code
550-Rejected after DATA: '>' missing at end of address:
550 failing address in "From" header is: <user@dom.ain
.endd


.option smtputf8_advertise_hosts main "host list&!!" *
.cindex "SMTPUTF8" "ESMTP extension, advertising"
.cindex "ESMTP extensions" SMTPUTF8
When Exim is built with support for internationalised mail names,
the availability thereof is advertised in
response to EHLO only to those client hosts that match this option. See
chapter &<<CHAPi18n>>& for details of Exim's support for internationalisation.


.option spamd_address main string "127.0.0.1 783"
This option is available when Exim is compiled with the content-scanning
extension. It specifies how Exim connects to SpamAssassin's &%spamd%& daemon.
See section &<<SECTscanspamass>>& for more details.



.option spf_guess main string "v=spf1 a/24 mx/24 ptr ?all"
This option is available when Exim is compiled with SPF support.
See section &<<SECSPF>>& for more details.

.option spf_smtp_comment_template main string&!! "Please%_see%_http://www.open-spf.org/Why"
This option is available when Exim is compiled with SPF support.  It
allows the customisation of the SMTP comment that the SPF library
generates.  You are strongly encouraged to link to your own explanative
site. The template must not contain spaces. If you need spaces in the
output, use the proper placeholder. If libspf2 can not parse the
template, it uses a built-in default broken link. The following placeholders
(along with Exim variables (but see below)) are allowed in the template:
.ilist
&*%_*&: A space.
.next
&*%{L}*&: Envelope sender's local part.
.next
&*%{S}*&: Envelope sender.
.next
&*%{O}*&: Envelope sender's domain.
.next
&*%{D}*&: Current(?) domain.
.next
&*%{I}*&: SMTP client Ip.
.next
&*%{C}*&: SMTP client pretty IP.
.next
&*%{T}*&: Epoch time (UTC).
.next
&*%{P}*&: SMTP client domain name.
.next
&*%{V}*&: IP version.
.next
&*%{H}*&: EHLO/HELO domain.
.next
&*%{R}*&: Receiving domain.
.endlist
The capitalized placeholders do proper URL encoding, if you use them
lowercased, no encoding takes place.  This list was compiled from the
libspf2 sources.

A note on using Exim variables: As
currently the SPF library is initialized before the SMTP EHLO phase,
the variables useful for expansion are quite limited.


.option split_spool_directory main boolean false
.cindex "multiple spool directories"
.cindex "spool directory" "split"
.cindex "directories, multiple"
If this option is set, it causes Exim to split its input directory into 62
subdirectories, each with a single alphanumeric character as its name. The
sixth character of the message id is used to allocate messages to
subdirectories; this is the least significant base-62 digit of the time of
arrival of the message.

Splitting up the spool in this way may provide better performance on systems
where there are long mail queues, by reducing the number of files in any one
directory. The msglog directory is also split up in a similar way to the input
directory; however, if &%preserve_message_logs%& is set, all old msglog files
are still placed in the single directory &_msglog.OLD_&.

It is not necessary to take any special action for existing messages when
changing &%split_spool_directory%&. Exim notices messages that are in the
&"wrong"& place, and continues to process them. If the option is turned off
after a period of being on, the subdirectories will eventually empty and be
automatically deleted.

When &%split_spool_directory%& is set, the behaviour of queue runner processes
changes. Instead of creating a list of all messages in the queue, and then
trying to deliver each one, in turn, it constructs a list of those in one
sub-directory and tries to deliver them, before moving on to the next
sub-directory. The sub-directories are processed in a random order. This
spreads out the scanning of the input directories, and uses less memory. It is
particularly beneficial when there are lots of messages in the queue. However,
if &%queue_run_in_order%& is set, none of this new processing happens. The
entire queue has to be scanned and sorted before any deliveries can start.


.option spool_directory main string&!! "set at compile time"
.cindex "spool directory" "path to"
This defines the directory in which Exim keeps its spool, that is, the messages
it is waiting to deliver. The default value is taken from the compile-time
configuration setting, if there is one. If not, this option must be set. The
string is expanded, so it can contain, for example, a reference to
&$primary_hostname$&.

If the spool directory name is fixed on your installation, it is recommended
that you set it at build time rather than from this option, particularly if the
log files are being written to the spool directory (see &%log_file_path%&).
Otherwise log files cannot be used for errors that are detected early on, such
as failures in the configuration file.

By using this option to override the compiled-in path, it is possible to run
tests of Exim without using the standard spool.

.option spool_wireformat main boolean false
.cindex "spool directory" "file formats"
If this option is set, Exim may for some messages use an alternative format
for data-files in the spool which matches the wire format.
Doing this permits more efficient message reception and transmission.
Currently it is only done for messages received using the ESMTP CHUNKING
option.

The following variables will not have useful values:
.code
$max_received_linelength
$body_linecount
$body_zerocount
.endd

Users of the local_scan() API (see &<<CHAPlocalscan>>&),
and any external programs which are passed a reference to a message data file
(except via the &"regex"&, &"malware"& or &"spam"&) ACL conditions)
will need to be aware of the different formats potentially available.

Using any of the ACL conditions noted will negate the reception benefit
(as a Unix-mbox-format file is constructed for them).
The transmission benefit is maintained.

.option sqlite_lock_timeout main time 5s
.cindex "sqlite lookup type" "lock timeout"
This option controls the timeout that the &(sqlite)& lookup uses when trying to
access an SQLite database. See section &<<SECTsqlite>>& for more details.

.option strict_acl_vars main boolean false
.cindex "&ACL;" "variables, handling unset"
This option controls what happens if a syntactically valid but undefined ACL
variable is referenced. If it is false (the default), an empty string
is substituted; if it is true, an error is generated. See section
&<<SECTaclvariables>>& for details of ACL variables.

.option strip_excess_angle_brackets main boolean false
.cindex "angle brackets, excess"
If this option is set, redundant pairs of angle brackets round &"route-addr"&
items in addresses are stripped. For example, &'<<xxx@a.b.c.d>>'& is
treated as &'<xxx@a.b.c.d>'&. If this is in the envelope and the message is
passed on to another MTA, the excess angle brackets are not passed on. If this
option is not set, multiple pairs of angle brackets cause a syntax error.


.option strip_trailing_dot main boolean false
.cindex "trailing dot on domain"
.cindex "dot" "trailing on domain"
If this option is set, a trailing dot at the end of a domain in an address is
ignored. If this is in the envelope and the message is passed on to another
MTA, the dot is not passed on. If this option is not set, a dot at the end of a
domain causes a syntax error.
However, addresses in header lines are checked only when an ACL requests header
syntax checking.


.option syslog_duplication main boolean true
.cindex "syslog" "duplicate log lines; suppressing"
When Exim is logging to syslog, it writes the log lines for its three
separate logs at different syslog priorities so that they can in principle
be separated on the logging hosts. Some installations do not require this
separation, and in those cases, the duplication of certain log lines is a
nuisance. If &%syslog_duplication%& is set false, only one copy of any
particular log line is written to syslog. For lines that normally go to
both the main log and the reject log, the reject log version (possibly
containing message header lines) is written, at LOG_NOTICE priority.
Lines that normally go to both the main and the panic log are written at
the LOG_ALERT priority.


.option syslog_facility main string unset
.cindex "syslog" "facility; setting"
This option sets the syslog &"facility"& name, used when Exim is logging to
syslog. The value must be one of the strings &"mail"&, &"user"&, &"news"&,
&"uucp"&, &"daemon"&, or &"local&'x'&"& where &'x'& is a digit between 0 and 7.
If this option is unset, &"mail"& is used. See chapter &<<CHAPlog>>& for
details of Exim's logging.


.option syslog_pid main boolean true
.cindex "syslog" "pid"
If &%syslog_pid%& is set false, the PID on Exim's log lines are
omitted when these lines are sent to syslog. (Syslog normally prefixes
the log lines with the PID of the logging process automatically.) You need
to enable the &`+pid`& log selector item, if you want Exim to write it's PID
into the logs.) See chapter &<<CHAPlog>>& for details of Exim's logging.



.option syslog_processname main string &`exim`&
.cindex "syslog" "process name; setting"
This option sets the syslog &"ident"& name, used when Exim is logging to
syslog. The value must be no longer than 32 characters. See chapter
&<<CHAPlog>>& for details of Exim's logging.



.option syslog_timestamp main boolean true
.cindex "syslog" "timestamps"
.cindex timestamps syslog
If &%syslog_timestamp%& is set false, the timestamps on Exim's log lines are
omitted when these lines are sent to syslog. See chapter &<<CHAPlog>>& for
details of Exim's logging.


.option system_filter main string&!! unset
.cindex "filter" "system filter"
.cindex "system filter" "specifying"
.cindex "Sieve filter" "not available for system filter"
This option specifies an Exim filter file that is applied to all messages at
the start of each delivery attempt, before any routing is done. System filters
must be Exim filters; they cannot be Sieve filters. If the system filter
generates any deliveries to files or pipes, or any new mail messages, the
appropriate &%system_filter_..._transport%& option(s) must be set, to define
which transports are to be used. Details of this facility are given in chapter
&<<CHAPsystemfilter>>&.
A forced expansion failure results in no filter operation.


.option system_filter_directory_transport main string&!! unset
.vindex "&$address_file$&"
This sets the name of the transport driver that is to be used when the
&%save%& command in a system message filter specifies a path ending in &"/"&,
implying delivery of each message into a separate file in some directory.
During the delivery, the variable &$address_file$& contains the path name.


.option system_filter_file_transport main string&!! unset
.cindex "file" "transport for system filter"
This sets the name of the transport driver that is to be used when the &%save%&
command in a system message filter specifies a path not ending in &"/"&. During
the delivery, the variable &$address_file$& contains the path name.

.option system_filter_group main string unset
.cindex "gid (group id)" "system filter"
This option is used only when &%system_filter_user%& is also set. It sets the
gid under which the system filter is run, overriding any gid that is associated
with the user. The value may be numerical or symbolic.

.option system_filter_pipe_transport main string&!! unset
.cindex "&(pipe)& transport" "for system filter"
.vindex "&$address_pipe$&"
This specifies the transport driver that is to be used when a &%pipe%& command
is used in a system filter. During the delivery, the variable &$address_pipe$&
contains the pipe command.


.option system_filter_reply_transport main string&!! unset
.cindex "&(autoreply)& transport" "for system filter"
This specifies the transport driver that is to be used when a &%mail%& command
is used in a system filter.


.option system_filter_user main string unset
.cindex "uid (user id)" "system filter"
If this option is set to root, the system filter is run in the main Exim
delivery process, as root.  Otherwise, the system filter runs in a separate
process, as the given user, defaulting to the Exim run-time user.
Unless the string consists entirely of digits, it
is looked up in the password data. Failure to find the named user causes a
configuration error. The gid is either taken from the password data, or
specified by &%system_filter_group%&. When the uid is specified numerically,
&%system_filter_group%& is required to be set.

If the system filter generates any pipe, file, or reply deliveries, the uid
under which the filter is run is used when transporting them, unless a
transport option overrides.


.option tcp_nodelay main boolean true
.cindex "daemon" "TCP_NODELAY on sockets"
.cindex "Nagle algorithm"
.cindex "TCP_NODELAY on listening sockets"
If this option is set false, it stops the Exim daemon setting the
TCP_NODELAY option on its listening sockets. Setting TCP_NODELAY
turns off the &"Nagle algorithm"&, which is a way of improving network
performance in interactive (character-by-character) situations. Turning it off
should improve Exim's performance a bit, so that is what happens by default.
However, it appears that some broken clients cannot cope, and time out. Hence
this option. It affects only those sockets that are set up for listening by the
daemon. Sockets created by the smtp transport for delivering mail always set
TCP_NODELAY.


.option timeout_frozen_after main time 0s
.cindex "frozen messages" "timing out"
.cindex "timeout" "frozen messages"
If &%timeout_frozen_after%& is set to a time greater than zero, a frozen
message of any kind that has been in the queue for longer than the given time
is automatically cancelled at the next queue run. If the frozen message is a
bounce message, it is just discarded; otherwise, a bounce is sent to the
sender, in a similar manner to cancellation by the &%-Mg%& command line option.
If you want to timeout frozen bounce messages earlier than other kinds of
frozen message, see &%ignore_bounce_errors_after%&.

&*Note:*& the default value of zero means no timeouts; with this setting,
frozen messages remain in the queue forever (except for any frozen bounce
messages that are released by &%ignore_bounce_errors_after%&).


.option timezone main string unset
.cindex "timezone, setting"
.cindex "environment" "values from"
The value of &%timezone%& is used to set the environment variable TZ while
running Exim (if it is different on entry). This ensures that all timestamps
created by Exim are in the required timezone. If you want all your timestamps
to be in UTC (aka GMT) you should set
.code
timezone = UTC
.endd
The default value is taken from TIMEZONE_DEFAULT in &_Local/Makefile_&,
or, if that is not set, from the value of the TZ environment variable when Exim
is built. If &%timezone%& is set to the empty string, either at build or run
time, any existing TZ variable is removed from the environment when Exim
runs. This is appropriate behaviour for obtaining wall-clock time on some, but
unfortunately not all, operating systems.


.option tls_advertise_hosts main "host list&!!" *
.cindex "TLS" "advertising"
.cindex "encryption" "on SMTP connection"
.cindex "SMTP" "encrypted connection"
.cindex "ESMTP extensions" STARTTLS
When Exim is built with support for TLS encrypted connections, the availability
of the STARTTLS command to set up an encrypted session is advertised in
response to EHLO only to those client hosts that match this option. See
chapter &<<CHAPTLS>>& for details of Exim's support for TLS.
Note that the default value requires that a certificate be supplied
using the &%tls_certificate%& option.  If TLS support for incoming connections
is not required the &%tls_advertise_hosts%& option should be set empty.


.option tls_alpn main "string list&!!" "smtp : esmtp"
.cindex TLS "Application Layer Protocol Names"
.cindex TLS ALPN
.cindex ALPN "set acceptable names for server"
If this option is set,
the TLS library supports ALPN,
and the client offers either more than one
ALPN name or a name which does not match the list,
the TLS connection is declined.


.option tls_certificate main "string list&!!" unset
.cindex "TLS" "server certificate; location of"
.cindex "certificate" "server, location of"
The value of this option is expanded, and must then be a list of absolute paths to
files which contain the server's certificates (in PEM format).
Commonly only one file is needed.
The server's private key is also
assumed to be in this file if &%tls_privatekey%& is unset. See chapter
&<<CHAPTLS>>& for further details.

&*Note*&: The certificates defined by this option are used only when Exim is
receiving incoming messages as a server. If you want to supply certificates for
use when sending messages as a client, you must set the &%tls_certificate%&
option in the relevant &(smtp)& transport.

&*Note*&: If you use filenames based on IP addresses, change the list
separator in the usual way (&<<SECTlistsepchange>>&) to avoid confusion under IPv6.

&*Note*&: Under versions of OpenSSL preceding 1.1.1,
when a list of more than one
file is used, the &$tls_in_ourcert$& variable is unreliable.
The macro "_TLS_BAD_MULTICERT_IN_OURCERT" will be defined for those versions.

.cindex SNI "selecting server certificate based on"
If the option contains &$tls_out_sni$& and Exim is built against OpenSSL, then
if the OpenSSL build supports TLS extensions and the TLS client sends the
Server Name Indication extension, then this option and others documented in
&<<SECTtlssni>>& will be re-expanded.

If this option is unset or empty a self-signed certificate will be
used.
Under Linux this is generated at daemon startup; on other platforms it will be
generated fresh for every connection.

.option tls_crl main string&!! unset
.cindex "TLS" "server certificate revocation list"
.cindex "certificate" "revocation list for server"
This option specifies a certificate revocation list. The expanded value must
be the name of a file that contains CRLs in PEM format.

Under OpenSSL the option can specify a directory with CRL files.

&*Note:*& Under OpenSSL the option must, if given, supply a CRL
for each signing element of the certificate chain (i.e. all but the leaf).
For the file variant this can be multiple PEM blocks in the one file.

See &<<SECTtlssni>>& for discussion of when this option might be re-expanded.


.option tls_dh_max_bits main integer 2236
.cindex "TLS" "D-H bit count"
The number of bits used for Diffie-Hellman key-exchange may be suggested by
the chosen TLS library.  That value might prove to be too high for
interoperability.  This option provides a maximum clamp on the value
suggested, trading off security for interoperability.

The value must be at least 1024.

The value 2236 was chosen because, at time of adding the option, it was the
hard-coded maximum value supported by the NSS cryptographic library, as used
by Thunderbird, while GnuTLS was suggesting 2432 bits as normal.

If you prefer more security and are willing to break some clients, raise this
number.

Note that the value passed to GnuTLS for *generating* a new prime may be a
little less than this figure, because GnuTLS is inexact and may produce a
larger prime than requested.


.option tls_dhparam main string&!! unset
.cindex "TLS" "D-H parameters for server"
The value of this option is expanded and indicates the source of DH parameters
to be used by Exim.

&*Note: The Exim Maintainers strongly recommend using a filename with site-generated
local DH parameters*&, which has been supported across all versions of Exim.  The
other specific constants available are a fallback so that even when
"unconfigured", Exim can offer Perfect Forward Secrecy in older ciphersuites in TLS.

If &%tls_dhparam%& is a filename starting with a &`/`&,
then it names a file from which DH
parameters should be loaded.  If the file exists, it should hold a PEM-encoded
PKCS#3 representation of the DH prime.  If the file does not exist, for
OpenSSL it is an error.  For GnuTLS, Exim will attempt to create the file and
fill it with a generated DH prime.  For OpenSSL, if the DH bit-count from
loading the file is greater than &%tls_dh_max_bits%& then it will be ignored,
and treated as though the &%tls_dhparam%& were set to "none".

If this option expands to the string "none", then no DH parameters will be
loaded by Exim.

If this option expands to the string "historic" and Exim is using GnuTLS, then
Exim will attempt to load a file from inside the spool directory.  If the file
does not exist, Exim will attempt to create it.
See section &<<SECTgnutlsparam>>& for further details.

If Exim is using OpenSSL and this option is empty or unset, then Exim will load
a default DH prime; the default is Exim-specific but lacks verifiable provenance.

In older versions of Exim the default was the 2048 bit prime described in section
2.2 of &url(https://www.rfc-editor.org/rfc/rfc5114,RFC 5114),
"2048-bit MODP Group with 224-bit Prime Order Subgroup", which
in IKE is assigned number 23.

Otherwise, the option must expand to the name used by Exim for any of a number
of DH primes specified in
&url(https://www.rfc-editor.org/rfc/rfc2409,RFC 2409),
&url(https://www.rfc-editor.org/rfc/rfc3526,RFC 3526),
&url(https://www.rfc-editor.org/rfc/rfc5114,RFC 5114),
&url(https://www.rfc-editor.org/rfc/rfc7919,RFC 7919), or from other
sources.  As names, Exim uses a standard specified name, else "ike" followed by
the number used by IKE, or "default" which corresponds to
&`exim.dev.20160529.3`&.

The available standard primes are:
&`ffdhe2048`&, &`ffdhe3072`&, &`ffdhe4096`&, &`ffdhe6144`&, &`ffdhe8192`&,
&`ike1`&, &`ike2`&, &`ike5`&,
&`ike14`&, &`ike15`&, &`ike16`&, &`ike17`&, &`ike18`&,
&`ike22`&, &`ike23`& and &`ike24`&.

The available additional primes are:
&`exim.dev.20160529.1`&, &`exim.dev.20160529.2`& and &`exim.dev.20160529.3`&.

Some of these will be too small to be accepted by clients.
Some may be too large to be accepted by clients.
The open cryptographic community has suspicions about the integrity of some
of the later IKE values, which led into
&url(https://www.rfc-editor.org/rfc/rfc7919,RFC 7919)
providing new fixed constants (the "ffdhe" identifiers).

At this point, all of the "ike" values should be considered obsolete;
they are still in Exim to avoid breaking unusual configurations, but are
candidates for removal the next time we have backwards-incompatible changes.
Two of them in particular (&`ike1`& and &`ike22`&) are called out by
&url(https://www.rfc-editor.org/rfc/rfc8247,RFC 8247)
as MUST NOT use for IPSEC, and two more (&`ike23`& and &`ike24`&) as
SHOULD NOT.
Because of this, Exim regards them as deprecated; if either of the first pair
are used, warnings will be logged in the paniclog, and if any are used then
warnings will be logged in the mainlog.
All four will be removed in a future Exim release.

The TLS protocol does not negotiate an acceptable size for this; clients tend
to hard-drop connections if what is offered by the server is unacceptable,
whether too large or too small, and there's no provision for the client to
tell the server what these constraints are.  Thus, as a server operator, you
need to make an educated guess as to what is most likely to work for your
userbase.

Some known size constraints suggest that a bit-size in the range 2048 to 2236
is most likely to maximise interoperability.  The upper bound comes from
applications using the Mozilla Network Security Services (NSS) library, which
used to set its &`DH_MAX_P_BITS`& upper-bound to 2236.  This affects many
mail user agents (MUAs). The lower bound comes from Debian installs of Exim4
prior to the 4.80 release, as Debian used to patch Exim to raise the minimum
acceptable bound from 1024 to 2048.


.option tls_eccurve main string list&!! &`auto`&
.cindex TLS "EC cryptography"
This option selects EC curves for use by Exim when used with OpenSSL.
It has no effect when Exim is used with GnuTLS
(the equivalent can be done using a priority string for the
&%tls_require_ciphers%& option).

After expansion it must contain
one or (only for OpenSSL versiona 1.1.1 onwards) more
EC curve names, such as &`prime256v1`&, &`secp384r1`&, or &`P-521`&.
Consult your OpenSSL manual for valid curve names.

For OpenSSL versions before (and not including) 1.0.2, the string
&`auto`& selects &`prime256v1`&. For more recent OpenSSL versions
&`auto`& tells the library to choose.

If the option expands to an empty string, the effect is undefined.


.option tls_ocsp_file main string&!! unset
.cindex TLS "certificate status"
.cindex TLS "OCSP proof file"
This option
must if set expand to the absolute path to a file which contains a current
status proof for the server's certificate, as obtained from the
Certificate Authority.

Usable for GnuTLS 3.4.4 or 3.3.17 or OpenSSL 1.1.0 (or later).
The macro "_HAVE_TLS_OCSP" will be defined for those versions.

For OpenSSL 1.1.0 or later, and
for GnuTLS 3.5.6 or later the expanded value of this option can be a list
of files, to match a list given for the &%tls_certificate%& option.
The ordering of the two lists must match.
The macro "_HAVE_TLS_OCSP_LIST" will be defined for those versions.

The file(s) should be in DER format,
except for GnuTLS 3.6.3 or later
or for OpenSSL,
when an optional filetype prefix can be used.
The prefix must be one of "DER" or "PEM", followed by
a single space.  If one is used it sets the format for subsequent
files in the list; the initial format is DER.
If multiple proofs are wanted, for multiple chain elements
(this only works under TLS1.3)
they must be coded as a combined OCSP response.

Although GnuTLS will accept PEM files with multiple separate
PEM blobs (ie. separate OCSP responses), it sends them in the
TLS Certificate record interleaved with the certificates of the chain;
although a GnuTLS client is happy with that, an OpenSSL client is not.

.option tls_on_connect_ports main "string list" unset
.cindex SSMTP
.cindex SMTPS
This option specifies a list of incoming SSMTP (aka SMTPS) ports that should
operate the SSMTP (SMTPS) protocol, where a TLS session is immediately
set up without waiting for the client to issue a STARTTLS command. For
further details, see section &<<SECTsupobssmt>>&.



.option tls_privatekey main "string list&!!" unset
.cindex "TLS" "server private key; location of"
The value of this option is expanded, and must then be a list of absolute paths to
files which contains the server's private keys.
If this option is unset, or if
the expansion is forced to fail, or the result is an empty string, the private
key is assumed to be in the same file as the server's certificates. See chapter
&<<CHAPTLS>>& for further details.

See &<<SECTtlssni>>& for discussion of when this option might be re-expanded.


.option tls_remember_esmtp main boolean false
.cindex "TLS" "esmtp state; remembering"
.cindex "TLS" "broken clients"
If this option is set true, Exim violates the RFCs by remembering that it is in
&"esmtp"& state after successfully negotiating a TLS session. This provides
support for broken clients that fail to send a new EHLO after starting a
TLS session.


.option tls_require_ciphers main string&!! unset
.cindex "TLS" "requiring specific ciphers"
.cindex "cipher" "requiring specific"
This option controls which ciphers can be used for incoming TLS connections.
The &(smtp)& transport has an option of the same name for controlling outgoing
connections. This option is expanded for each connection, so can be varied for
different clients if required. The value of this option must be a list of
permitted cipher suites. The OpenSSL and GnuTLS libraries handle cipher control
in somewhat different ways. If GnuTLS is being used, the client controls the
preference order of the available ciphers. Details are given in sections
&<<SECTreqciphssl>>& and &<<SECTreqciphgnu>>&.


.option tls_resumption_hosts main "host list&!!" unset
.cindex TLS resumption
This option controls which connections to offer the TLS resumption feature.
See &<<SECTresumption>>& for details.


.option tls_try_verify_hosts main "host list&!!" unset
.cindex "TLS" "client certificate verification"
.cindex "certificate" "verification of client"
See &%tls_verify_hosts%& below.


.option tls_verify_certificates main string&!! system
.cindex "TLS" "client certificate verification"
.cindex "certificate" "verification of client"
The value of this option is expanded, and must then be either the
word "system"
or the absolute path to
a file or directory containing permitted certificates for clients that
match &%tls_verify_hosts%& or &%tls_try_verify_hosts%&.

The "system" value for the option will use a
system default location compiled into the SSL library.
This is not available for GnuTLS versions preceding 3.0.20,
and will be taken as empty; an explicit location
must be specified.

The use of a directory for the option value is not available for GnuTLS versions
preceding 3.3.6 and a single file must be used.

With OpenSSL the certificates specified
explicitly
either by file or directory
are added to those given by the system default location.

These certificates should be for the certificate authorities trusted, rather
than the public cert of individual clients.  With both OpenSSL and GnuTLS, if
the value is a file then the certificates are sent by Exim as a server to
connecting clients, defining the list of accepted certificate authorities.
Thus the values defined should be considered public data.  To avoid this,
use the explicit directory version. (If your peer is Exim up to 4.85,
using GnuTLS, you may need to send the CAs (thus using the file
variant). Otherwise the peer doesn't send its certificate.)

See &<<SECTtlssni>>& for discussion of when this option might be re-expanded.

A forced expansion failure or setting to an empty string is equivalent to
being unset.


.option tls_verify_hosts main "host list&!!" unset
.cindex "TLS" "client certificate verification"
.cindex "certificate" "verification of client"
This option, along with &%tls_try_verify_hosts%&, controls the checking of
certificates from clients. The expected certificates are defined by
&%tls_verify_certificates%&, which must be set. A configuration error occurs if
either &%tls_verify_hosts%& or &%tls_try_verify_hosts%& is set and
&%tls_verify_certificates%& is not set.

Any client that matches &%tls_verify_hosts%& is constrained by
&%tls_verify_certificates%&. When the client initiates a TLS session, it must
present one of the listed certificates. If it does not, the connection is
aborted.
&*Warning*&: Including a host in &%tls_verify_hosts%& does not require
the host to use TLS. It can still send SMTP commands through unencrypted
connections. Forcing a client to use TLS has to be done separately using an
ACL to reject inappropriate commands when the connection is not encrypted.

A weaker form of checking is provided by &%tls_try_verify_hosts%&. If a client
matches this option (but not &%tls_verify_hosts%&), Exim requests a
certificate and checks it against &%tls_verify_certificates%&, but does not
abort the connection if there is no certificate or if it does not match. This
state can be detected in an ACL, which makes it possible to implement policies
such as &"accept for relay only if a verified certificate has been received,
but accept for local delivery if encrypted, even without a verified
certificate"&.

Client hosts that match neither of these lists are not asked to present
certificates.


.option trusted_groups main "string list&!!" unset
.cindex "trusted groups"
.cindex "groups" "trusted"
This option is expanded just once, at the start of Exim's processing. If this
option is set, any process that is running in one of the listed groups, or
which has one of them as a supplementary group, is trusted. The groups can be
specified numerically or by name. See section &<<SECTtrustedadmin>>& for
details of what trusted callers are permitted to do. If neither
&%trusted_groups%& nor &%trusted_users%& is set, only root and the Exim user
are trusted.

.option trusted_users main "string list&!!" unset
.cindex "trusted users"
.cindex "user" "trusted"
This option is expanded just once, at the start of Exim's processing. If this
option is set, any process that is running as one of the listed users is
trusted. The users can be specified numerically or by name. See section
&<<SECTtrustedadmin>>& for details of what trusted callers are permitted to do.
If neither &%trusted_groups%& nor &%trusted_users%& is set, only root and the
Exim user are trusted.

.option unknown_login main string&!! unset
.cindex "uid (user id)" "unknown caller"
.vindex "&$caller_uid$&"
This is a specialized feature for use in unusual configurations. By default, if
the uid of the caller of Exim cannot be looked up using &[getpwuid()]&, Exim
gives up. The &%unknown_login%& option can be used to set a login name to be
used in this circumstance. It is expanded, so values like &%user$caller_uid%&
can be set. When &%unknown_login%& is used, the value of &%unknown_username%&
is used for the user's real name (gecos field), unless this has been set by the
&%-F%& option.

.option unknown_username main string unset
See &%unknown_login%&.

.option untrusted_set_sender main "address list&!!" unset
.cindex "trusted users"
.cindex "sender" "setting by untrusted user"
.cindex "untrusted user setting sender"
.cindex "user" "untrusted setting sender"
.cindex "envelope from"
.cindex "envelope sender"
When an untrusted user submits a message to Exim using the standard input, Exim
normally creates an envelope sender address from the user's login and the
default qualification domain. Data from the &%-f%& option (for setting envelope
senders on non-SMTP messages) or the SMTP MAIL command (if &%-bs%& or &%-bS%&
is used) is ignored.

However, untrusted users are permitted to set an empty envelope sender address,
to declare that a message should never generate any bounces. For example:
.code
exim -f '<>' user@domain.example
.endd
.vindex "&$sender_ident$&"
The &%untrusted_set_sender%& option allows you to permit untrusted users to set
other envelope sender addresses in a controlled way. When it is set, untrusted
users are allowed to set envelope sender addresses that match any of the
patterns in the list. Like all address lists, the string is expanded. The
identity of the user is in &$sender_ident$&, so you can, for example, restrict
users to setting senders that start with their login ids
followed by a hyphen
by a setting like this:
.code
untrusted_set_sender = ^$sender_ident-
.endd
If you want to allow untrusted users to set envelope sender addresses without
restriction, you can use
.code
untrusted_set_sender = *
.endd
The &%untrusted_set_sender%& option applies to all forms of local input, but
only to the setting of the envelope sender. It does not permit untrusted users
to use the other options which trusted user can use to override message
parameters. Furthermore, it does not stop Exim from removing an existing
&'Sender:'& header in the message, or from adding a &'Sender:'& header if
necessary. See &%local_sender_retain%& and &%local_from_check%& for ways of
overriding these actions. The handling of the &'Sender:'& header is also
described in section &<<SECTthesenhea>>&.

The log line for a message's arrival shows the envelope sender following
&"<="&. For local messages, the user's login always follows, after &"U="&. In
&%-bp%& displays, and in the Exim monitor, if an untrusted user sets an
envelope sender address, the user's login is shown in parentheses after the
sender address.


.option uucp_from_pattern main string "see below"
.cindex "&""From""& line"
.cindex "UUCP" "&""From""& line"
Some applications that pass messages to an MTA via a command line interface use
an initial line starting with &"From&~"& to pass the envelope sender. In
particular, this is used by UUCP software. Exim recognizes such a line by means
of a regular expression that is set in &%uucp_from_pattern%&. When the pattern
matches, the sender address is constructed by expanding the contents of
&%uucp_from_sender%&, provided that the caller of Exim is a trusted user. The
default pattern recognizes lines in the following two forms:
.code
From ph10 Fri Jan  5 12:35 GMT 1996
From ph10 Fri, 7 Jan 97 14:00:00 GMT
.endd
The pattern can be seen by running
.code
exim -bP uucp_from_pattern
.endd
It checks only up to the hours and minutes, and allows for a 2-digit or 4-digit
year in the second case. The first word after &"From&~"& is matched in the
regular expression by a parenthesized subpattern. The default value for
&%uucp_from_sender%& is &"$1"&, which therefore just uses this first word
(&"ph10"& in the example above) as the message's sender. See also
&%ignore_fromline_hosts%&.


.option uucp_from_sender main string&!! &`$1`&
See &%uucp_from_pattern%& above.


.option warn_message_file main string&!! unset
.cindex "warning of delay" "customizing the message"
.cindex "customizing" "warning message"
This option defines a template file containing paragraphs of text to be used
for constructing the warning message which is sent by Exim when a message has
been in the queue for a specified amount of time, as specified by
&%delay_warning%&. Details of the file's contents are given in chapter
&<<CHAPemsgcust>>&.
.cindex warn_message_file "tainted data"
The option is expanded to give the file path, which must be
absolute and untainted.
See also &%bounce_message_file%&.


.option wellknown_advertise_hosts main boolean unset
.cindex WELLKNOWN advertisement
.cindex "ESMTP extensions" WELLKNOWN
This option enables the advertising of the SMTP WELLKNOWN extension.
See also the &%acl_smtp_wellknown%& ACL (&<<SECTWELLKNOWNACL>>&).

.option write_rejectlog main boolean true
.cindex "reject log" "disabling"
If this option is set false, Exim no longer writes anything to the reject log.
See chapter &<<CHAPlog>>& for details of what Exim writes to its logs.
.ecindex IIDconfima
.ecindex IIDmaiconf




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Generic options for routers" "CHAProutergeneric"
.scindex IIDgenoprou1 "options" "generic; for routers"
.scindex IIDgenoprou2 "generic options" "router"
This chapter describes the generic options that apply to all routers.
Those that are preconditions are marked with &Dagger; in the &"use"& field.

For a general description of how a router operates, see sections
&<<SECTrunindrou>>& and &<<SECTrouprecon>>&. The latter specifies the order in
which the preconditions are tested. The order of expansion of the options that
provide data for a transport is: &%errors_to%&, &%headers_add%&,
&%headers_remove%&, &%transport%&.

The name of a router is limited to be &drivernamemax; ASCII characters long;
prior to Exim 4.95 names would be silently truncated at this length, but now
it is enforced.


.option address_data routers string&!! unset
.cindex "router" "data attached to address"
The string is expanded just before the router is run, that is, after all the
precondition tests have succeeded. If the expansion is forced to fail, the
router declines, the value of &%address_data%& remains unchanged, and the
&%more%& option controls what happens next. Other expansion failures cause
delivery of the address to be deferred.

.vindex "&$address_data$&"
When the expansion succeeds, the value is retained with the address, and can be
accessed using the variable &$address_data$& in the current router, subsequent
routers, and the eventual transport.

&*Warning*&: If the current or any subsequent router is a &(redirect)& router
that runs a user's filter file, the contents of &$address_data$& are accessible
in the filter. This is not normally a problem, because such data is usually
either not confidential or it &"belongs"& to the current user, but if you do
put confidential data into &$address_data$& you need to remember this point.

Even if the router declines or passes, the value of &$address_data$& remains
with the address, though it can be changed by another &%address_data%& setting
on a subsequent router. If a router generates child addresses, the value of
&$address_data$& propagates to them. This also applies to the special kind of
&"child"& that is generated by a router with the &%unseen%& option.

The idea of &%address_data%& is that you can use it to look up a lot of data
for the address once, and then pick out parts of the data later. For example,
you could use a single LDAP lookup to return a string of the form
.code
uid=1234 gid=5678 mailbox=/mail/xyz forward=/home/xyz/.forward
.endd
In the transport you could pick out the mailbox by a setting such as
.code
file = ${extract{mailbox}{$address_data}}
.endd
This makes the configuration file less messy, and also reduces the number of
lookups (though Exim does cache lookups).

See also the &%set%& option below.

.vindex "&$sender_address_data$&"
.vindex "&$address_data$&"
The &%address_data%& facility is also useful as a means of passing information
from one router to another, and from a router to a transport. In addition, if
&$address_data$& is set by a router when verifying a recipient address from an
ACL, it remains available for use in the rest of the ACL statement. After
verifying a sender, the value is transferred to &$sender_address_data$&.



.option address_test routers&!? boolean true
.oindex "&%-bt%&"
.cindex "router" "skipping when address testing"
If this option is set false, the router is skipped when routing is being tested
by means of the &%-bt%& command line option. This can be a convenience when
your first router sends messages to an external scanner, because it saves you
having to set the &"already scanned"& indicator when testing real address
routing.



.option cannot_route_message routers string&!! unset
.cindex "router" "customizing &""cannot route""& message"
.cindex "customizing" "&""cannot route""& message"
This option specifies a text message that is used when an address cannot be
routed because Exim has run out of routers. The default message is
&"Unrouteable address"&. This option is useful only on routers that have
&%more%& set false, or on the very last router in a configuration, because the
value that is used is taken from the last router that is considered. This
includes a router that is skipped because its preconditions are not met, as
well as a router that declines. For example, using the default configuration,
you could put:
.code
cannot_route_message = Remote domain not found in DNS
.endd
on the first router, which is a &(dnslookup)& router with &%more%& set false,
and
.code
cannot_route_message = Unknown local user
.endd
on the final router that checks for local users. If string expansion fails for
this option, the default message is used. Unless the expansion failure was
explicitly forced, a message about the failure is written to the main and panic
logs, in addition to the normal message about the routing failure.


.option caseful_local_part routers boolean false
.cindex "case of local parts"
.cindex "router" "case of local parts"
By default, routers handle the local parts of addresses in a case-insensitive
manner, though the actual case is preserved for transmission with the message.
If you want the case of letters to be significant in a router, you must set
this option true. For individual router options that contain address or local
part lists (for example, &%local_parts%&), case-sensitive matching can be
turned on by &"+caseful"& as a list item. See section &<<SECTcasletadd>>& for
more details.

.vindex "&$local_part$&"
.vindex "&$original_local_part$&"
.vindex "&$parent_local_part$&"
The value of the &$local_part$& variable is forced to lower case while a
router is running unless &%caseful_local_part%& is set. When a router assigns
an address to a transport, the value of &$local_part$& when the transport runs
is the same as it was in the router. Similarly, when a router generates child
addresses by aliasing or forwarding, the values of &$original_local_part$&
and &$parent_local_part$& are those that were used by the redirecting router.

This option applies to the processing of an address by a router. When a
recipient address is being processed in an ACL, there is a separate &%control%&
modifier that can be used to specify case-sensitive processing within the ACL
(see section &<<SECTcontrols>>&).



.option check_local_user routers&!? boolean false
.cindex "local user, checking in router"
.cindex "router" "checking for local user"
.cindex "&_/etc/passwd_&"
.vindex "&$home$&"
When this option is true, Exim checks that the local part of the recipient
address (with affixes removed if relevant) is the name of an account on the
local system. The check is done by calling the &[getpwnam()]& function rather
than trying to read &_/etc/passwd_& directly. This means that other methods of
holding password data (such as NIS) are supported. If the local part is a local
user,
.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using router check_local_user option"
&$local_part_data$& is set to an untainted version of the local part and
&$home$& is set from the password data. The latter can be tested in other
preconditions that are evaluated after this one (the order of evaluation is
given in section &<<SECTrouprecon>>&). However, the value of &$home$& can be
overridden by &%router_home_directory%&. If the local part is not a local user,
the router is skipped.

If you want to check that the local part is either the name of a local user
or matches something else, you cannot combine &%check_local_user%& with a
setting of &%local_parts%&, because that specifies the logical &'and'& of the
two conditions. However, you can use a &(passwd)& lookup in a &%local_parts%&
setting to achieve this. For example:
.code
local_parts = passwd;$local_part : lsearch;/etc/other/users
.endd
Note, however, that the side effects of &%check_local_user%& (such as setting
up a home directory) do not occur when a &(passwd)& lookup is used in a
&%local_parts%& (or any other) precondition.



.option condition routers&!? string&!! unset
.cindex "router" "customized precondition"
This option specifies a general precondition test that has to succeed for the
router to be called. The &%condition%& option is the last precondition to be
evaluated (see section &<<SECTrouprecon>>&). The string is expanded, and if the
result is a forced failure, or an empty string, or one of the strings &"0"& or
&"no"& or &"false"& (checked without regard to the case of the letters), the
router is skipped, and the address is offered to the next one.

If the result is any other value, the router is run (as this is the last
precondition to be evaluated, all the other preconditions must be true).

This option is unusual in that multiple &%condition%& options may be present.
All &%condition%& options must succeed.

The &%condition%& option provides a means of applying custom conditions to the
running of routers. Note that in the case of a simple conditional expansion,
the default expansion values are exactly what is wanted. For example:
.code
condition = ${if >{$message_age}{600}}
.endd
Because of the default behaviour of the string expansion, this is equivalent to
.code
condition = ${if >{$message_age}{600}{true}{}}
.endd

A multiple condition example, which succeeds:
.code
condition = ${if >{$message_age}{600}}
condition = ${if !eq{${lc:$local_part}}{postmaster}}
condition = foobar
.endd

If the expansion fails (other than forced failure) delivery is deferred. Some
of the other precondition options are common special cases that could in fact
be specified using &%condition%&.

Historical note: We have &%condition%& on ACLs and on Routers.  Routers
are far older, and use one set of semantics.  ACLs are newer and when
they were created, the ACL &%condition%& process was given far stricter
parse semantics.  The &%bool{}%& expansion condition uses the same rules as
ACLs.  The &%bool_lax{}%& expansion condition uses the same rules as
Routers.  More pointedly, the &%bool_lax{}%& was written to match the existing
Router rules processing behavior.

This is best illustrated in an example:
.code
# If used in an ACL condition will fail with a syntax error, but
# in a router condition any extra characters are treated as a string

$ exim -be '${if eq {${lc:GOOGLE.com}} {google.com}} {yes} {no}}'
true {yes} {no}}

$ exim -be '${if eq {${lc:WHOIS.com}} {google.com}} {yes} {no}}'
 {yes} {no}}
.endd
In each example above, the &%if%& statement actually ends after
&"{google.com}}"&.  Since no true or false braces were defined, the
default &%if%& behavior is to return a boolean true or a null answer
(which evaluates to false). The rest of the line is then treated as a
string. So the first example resulted in the boolean answer &"true"&
with the string &" {yes} {no}}"& appended to it. The second example
resulted in the null output (indicating false) with the string
&" {yes} {no}}"& appended to it.

In fact you can put excess forward braces in too.  In the router
&%condition%&, Exim's parser only looks for &"{"& symbols when they
mean something, like after a &"$"& or when required as part of a
conditional.  But otherwise &"{"& and &"}"& are treated as ordinary
string characters.

Thus, in a Router, the above expansion strings will both always evaluate
true, as the result of expansion is a non-empty string which doesn't
match an explicit false value.  This can be tricky to debug.  By
contrast, in an ACL either of those strings will always result in an
expansion error because the result doesn't look sufficiently boolean.


.option debug_print routers string&!! unset
.cindex "testing" "variables in drivers"
If this option is set and debugging is enabled (see the &%-d%& command line
option) or in address-testing mode (see the &%-bt%& command line option),
the string is expanded and included in the debugging output.
If expansion of the string fails, the error message is written to the debugging
output, and Exim carries on processing.
This option is provided to help with checking out the values of variables and
so on when debugging router configurations. For example, if a &%condition%&
option appears not to be working, &%debug_print%& can be used to output the
variables it references. The output happens after checks for &%domains%&,
&%local_parts%&, and &%check_local_user%& but before any other preconditions
are tested. A newline is added to the text if it does not end with one.
The variable &$router_name$& contains the name of the router.



.option disable_logging routers boolean false
If this option is set true, nothing is logged for any routing errors
or for any deliveries caused by this router. You should not set this option
unless you really, really know what you are doing. See also the generic
transport option of the same name.

.option dnssec_request_domains routers "domain list&!!" *
.cindex "MX record" "security"
.cindex "DNSSEC" "MX lookup"
.cindex "security" "MX lookup"
.cindex "DNS" "DNSSEC"
DNS lookups for domains matching &%dnssec_request_domains%& will be done with
the DNSSEC request bit set.
This applies to all of the SRV, MX, AAAA, A lookup sequence.

.option dnssec_require_domains routers "domain list&!!" unset
.cindex "MX record" "security"
.cindex "DNSSEC" "MX lookup"
.cindex "security" "MX lookup"
.cindex "DNS" "DNSSEC"
DNS lookups for domains matching &%dnssec_require_domains%& will be done with
the DNSSEC request bit set.  Any returns not having the Authenticated Data bit
(AD bit) set will be ignored and logged as a host-lookup failure.
This applies to all of the SRV, MX, AAAA, A lookup sequence.


.option domains routers&!? "domain list&!!" unset
.cindex "router" "restricting to specific domains"
.vindex "&$domain_data$&"
If this option is set, the router is skipped unless the current domain matches
the list.
The data returned by the list check
is placed in &$domain_data$& for use in string
expansions of the driver's private options and in the transport.
See section &<<SECTrouprecon>>& for
a list of the order in which preconditions are evaluated.

A complex example, using a file like:
.code
alice@dom1
bill@dom1
maggie@dom1
.endd
and checking both domain and local_part
.code
domains =         ${domain:${lookup {$local_part@$domain} lseach,ret=key {/path/to/accountsfile}}}
local_parts = ${local_part:${lookup {$local_part@$domain} lseach,ret=key {/path/to/accountsfile}}}
.endd



.option driver routers string unset
This option must always be set. It specifies which of the available routers is
to be used.


.option dsn_lasthop routers boolean false
.cindex "DSN" "success"
.cindex "Delivery Status Notification" "success"
If this option is set true, and extended DSN
(&url(https://www.rfc-editor.org/rfc/rfc3461,RFC 3461)) processing is in effect,
Exim will not pass on DSN requests to downstream DSN-aware hosts but will
instead send a success DSN as if the next hop does not support DSN.
Not effective on redirect routers.



.option errors_to routers string&!! unset
.cindex "envelope from"
.cindex "envelope sender"
.cindex "router" "changing address for errors"
If a router successfully handles an address, it may assign the address to a
transport for delivery or it may generate child addresses. In both cases, if
there is a delivery problem during later processing, the resulting bounce
message is sent to the address that results from expanding this string,
provided that the address verifies successfully. The &%errors_to%& option is
expanded before &%headers_add%&, &%headers_remove%&, and &%transport%&.

The &%errors_to%& setting associated with an address can be overridden if it
subsequently passes through other routers that have their own &%errors_to%&
settings, or if the message is delivered by a transport with a &%return_path%&
setting.

If &%errors_to%& is unset, or the expansion is forced to fail, or the result of
the expansion fails to verify, the errors address associated with the incoming
address is used. At top level, this is the envelope sender. A non-forced
expansion failure causes delivery to be deferred.

If an address for which &%errors_to%& has been set ends up being delivered over
SMTP, the envelope sender for that delivery is the &%errors_to%& value, so that
any bounces that are generated by other MTAs on the delivery route are also
sent there. You can set &%errors_to%& to the empty string by either of these
settings:
.code
errors_to =
errors_to = ""
.endd
An expansion item that yields an empty string has the same effect. If you do
this, a locally detected delivery error for addresses processed by this router
no longer gives rise to a bounce message; the error is discarded. If the
address is delivered to a remote host, the return path is set to &`<>`&, unless
overridden by the &%return_path%& option on the transport.

.vindex "&$address_data$&"
If for some reason you want to discard local errors, but use a non-empty
MAIL command for remote delivery, you can preserve the original return
path in &$address_data$& in the router, and reinstate it in the transport by
setting &%return_path%&.

The most common use of &%errors_to%& is to direct mailing list bounces to the
manager of the list, as described in section &<<SECTmailinglists>>&, or to
implement VERP (Variable Envelope Return Paths) (see section &<<SECTverp>>&).



.option expn routers&!? boolean true
.cindex "address" "testing"
.cindex "testing" "addresses"
.cindex "EXPN" "router skipping"
.cindex "router" "skipping for EXPN"
If this option is turned off, the router is skipped when testing an address
as a result of processing an SMTP EXPN command. You might, for example,
want to turn it off on a router for users' &_.forward_& files, while leaving it
on for the system alias file.
See section &<<SECTrouprecon>>& for a list of the order in which preconditions
are evaluated.

The use of the SMTP EXPN command is controlled by an ACL (see chapter
&<<CHAPACL>>&). When Exim is running an EXPN command, it is similar to testing
an address with &%-bt%&. Compare VRFY, whose counterpart is &%-bv%&.



.option fail_verify routers boolean false
.cindex "router" "forcing verification failure"
Setting this option has the effect of setting both &%fail_verify_sender%& and
&%fail_verify_recipient%& to the same value.



.option fail_verify_recipient routers boolean false
If this option is true and an address is accepted by this router when
verifying a recipient, verification fails.



.option fail_verify_sender routers boolean false
If this option is true and an address is accepted by this router when
verifying a sender, verification fails.



.option fallback_hosts routers "string list" unset
.cindex "router" "fallback hosts"
.cindex "fallback" "hosts specified on router"
String expansion is not applied to this option. The argument must be a
colon-separated list of host names or IP addresses. The list separator can be
changed (see section &<<SECTlistsepchange>>&), and a port can be specified with
each name or address. In fact, the format of each item is exactly the same as
defined for the list of hosts in a &(manualroute)& router (see section
&<<SECTformatonehostitem>>&).

If a router queues an address for a remote transport, this host list is
associated with the address, and used instead of the transport's fallback host
list. If &%hosts_randomize%& is set on the transport, the order of the list is
randomized for each use. See the &%fallback_hosts%& option of the &(smtp)&
transport for further details.


.option group routers string&!! "see below"
.cindex "gid (group id)" "local delivery"
.cindex "local transports" "uid and gid"
.cindex "transport" "local"
.cindex "router" "setting group"
When a router queues an address for a transport, and the transport does not
specify a group, the group given here is used when running the delivery
process.
The group may be specified numerically or by name. If expansion fails, the
error is logged and delivery is deferred.
The default is unset, unless &%check_local_user%& is set, when the default
is taken from the password information. See also &%initgroups%& and &%user%&
and the discussion in chapter &<<CHAPenvironment>>&.



.option headers_add routers list&!! unset
.cindex "header lines" "adding"
.cindex "router" "adding header lines"
This option specifies a list of text headers,
newline-separated (by default, changeable in the usual way &<<SECTlistsepchange>>&),
that is associated with any addresses that are accepted by the router.
Each item is separately expanded, at routing time.  However, this
option has no effect when an address is just being verified. The way in which
the text is used to add header lines at transport time is described in section
&<<SECTheadersaddrem>>&. New header lines are not actually added until the
message is in the process of being transported. This means that references to
header lines in string expansions in the transport's configuration do not
&"see"& the added header lines.

The &%headers_add%& option is expanded after &%errors_to%&, but before
&%headers_remove%& and &%transport%&. If an item is empty, or if
an item expansion is forced to fail, the item has no effect. Other expansion
failures are treated as configuration errors.

Unlike most options, &%headers_add%& can be specified multiple times
for a router; all listed headers are added.

&*Warning 1*&: The &%headers_add%& option cannot be used for a &(redirect)&
router that has the &%one_time%& option set.

.cindex "duplicate addresses"
.oindex "&%unseen%&"
&*Warning 2*&: If the &%unseen%& option is set on the router, all header
additions are deleted when the address is passed on to subsequent routers.
For a &%redirect%& router, if a generated address is the same as the incoming
address, this can lead to duplicate addresses with different header
modifications. Exim does not do duplicate deliveries (except, in certain
circumstances, to pipes -- see section &<<SECTdupaddr>>&), but it is undefined
which of the duplicates is discarded, so this ambiguous situation should be
avoided. The &%repeat_use%& option of the &%redirect%& router may be of help.



.option headers_remove routers list&!! unset
.cindex "header lines" "removing"
.cindex "router" "removing header lines"
This option specifies a list of text headers,
colon-separated (by default, changeable in the usual way &<<SECTlistsepchange>>&),
that is associated with any addresses that are accepted by the router.
However, the option has no effect when an address is just being verified.
Each list item is separately expanded, at transport time.
If an item ends in *, it will match any header with the given prefix.
The way in which
the text is used to remove header lines at transport time is described in
section &<<SECTheadersaddrem>>&. Header lines are not actually removed until
the message is in the process of being transported. This means that references
to header lines in string expansions in the transport's configuration still
&"see"& the original header lines.

The &%headers_remove%& option is handled after &%errors_to%& and
&%headers_add%&, but before &%transport%&. If an item expansion is forced to fail,
the item has no effect. Other expansion failures are treated as configuration
errors.

Unlike most options, &%headers_remove%& can be specified multiple times
for a router; all listed headers are removed.

&*Warning 1*&: The &%headers_remove%& option cannot be used for a &(redirect)&
router that has the &%one_time%& option set.

&*Warning 2*&: If the &%unseen%& option is set on the router, all header
removal requests are deleted when the address is passed on to subsequent
routers, and this can lead to problems with duplicates -- see the similar
warning for &%headers_add%& above.

&*Warning 3*&: Because of the separate expansion of the list items,
items that contain a list separator must have it doubled.
To avoid this, change the list separator (&<<SECTlistsepchange>>&).



.option ignore_target_hosts routers "host list&!!" unset
.cindex "IP address" "discarding"
.cindex "router" "discarding IP addresses"
Although this option is a host list, it should normally contain IP address
entries rather than names. If any host that is looked up by the router has an
IP address that matches an item in this list, Exim behaves as if that IP
address did not exist. This option allows you to cope with rogue DNS entries
like
.code
remote.domain.example.  A  127.0.0.1
.endd
by setting
.code
ignore_target_hosts = 127.0.0.1
.endd
on the relevant router. If all the hosts found by a &(dnslookup)& router are
discarded in this way, the router declines. In a conventional configuration, an
attempt to mail to such a domain would normally provoke the &"unrouteable
domain"& error, and an attempt to verify an address in the domain would fail.
Similarly, if &%ignore_target_hosts%& is set on an &(ipliteral)& router, the
router declines if presented with one of the listed addresses.

You can use this option to disable the use of IPv4 or IPv6 for mail delivery by
means of the first or the second of the following settings, respectively:
.code
ignore_target_hosts = 0.0.0.0/0
ignore_target_hosts = <; 0::0/0
.endd
The pattern in the first line matches all IPv4 addresses, whereas the pattern
in the second line matches all IPv6 addresses.

This option may also be useful for ignoring link-local and site-local IPv6
addresses. Because, like all host lists, the value of &%ignore_target_hosts%&
is expanded before use as a list, it is possible to make it dependent on the
domain that is being routed.

.vindex "&$host_address$&"
During its expansion, &$host_address$& is set to the IP address that is being
checked.

.option initgroups routers boolean false
.cindex "additional groups"
.cindex "groups" "additional"
.cindex "local transports" "uid and gid"
.cindex "transport" "local"
If the router queues an address for a transport, and this option is true, and
the uid supplied by the router is not overridden by the transport, the
&[initgroups()]& function is called when running the transport to ensure that
any additional groups associated with the uid are set up. See also &%group%&
and &%user%& and the discussion in chapter &<<CHAPenvironment>>&.



.option local_part_prefix routers&!? "string list" unset
.cindex affix "router precondition"
.cindex "router" "prefix for local part"
.cindex "prefix" "for local part, used in router"
If this option is set, the router is skipped unless the local part starts with
one of the given strings, or &%local_part_prefix_optional%& is true. See
section &<<SECTrouprecon>>& for a list of the order in which preconditions are
evaluated.

The list is scanned from left to right, and the first prefix that matches is
used. A limited form of wildcard is available; if the prefix begins with an
asterisk, it matches the longest possible sequence of arbitrary characters at
the start of the local part. An asterisk should therefore always be followed by
some character that does not occur in normal local parts.
.cindex "multiple mailboxes"
.cindex "mailbox" "multiple"
Wildcarding can be used to set up multiple user mailboxes, as described in
section &<<SECTmulbox>>&.

.vindex "&$local_part$&"
.vindex "&$local_part_prefix$&"
During the testing of the &%local_parts%& option, and while the router is
running, the prefix is removed from the local part, and is available in the
expansion variable &$local_part_prefix$&. When a message is being delivered, if
the router accepts the address, this remains true during subsequent delivery by
a transport. In particular, the local part that is transmitted in the RCPT
command for LMTP, SMTP, and BSMTP deliveries has the prefix removed by default.
This behaviour can be overridden by setting &%rcpt_include_affixes%& true on
the relevant transport.

.vindex &$local_part_prefix_v$&
If wildcarding (above) was used then the part of the prefix matching the
wildcard is available in &$local_part_prefix_v$&.

When an address is being verified, &%local_part_prefix%& affects only the
behaviour of the router. If the callout feature of verification is in use, this
means that the full address, including the prefix, will be used during the
callout.

The prefix facility is commonly used to handle local parts of the form
&%owner-something%&. Another common use is to support local parts of the form
&%real-username%& to bypass a user's &_.forward_& file &-- helpful when trying
to tell a user their forwarding is broken &-- by placing a router like this one
immediately before the router that handles &_.forward_& files:
.code
real_localuser:
  driver = accept
  local_part_prefix = real-
  check_local_user
  transport = local_delivery
.endd
For security, it would probably be a good idea to restrict the use of this
router to locally-generated messages, using a condition such as this:
.code
  condition = ${if match_ip {$sender_host_address} \
                        {<; ; 127.0.0.1 ; ::1}}
.endd

If both &%local_part_prefix%& and &%local_part_suffix%& are set for a router,
both conditions must be met if not optional. Care must be taken if wildcards
are used in both a prefix and a suffix on the same router. Different
separator characters must be used to avoid ambiguity.


.option local_part_prefix_optional routers boolean false
See &%local_part_prefix%& above.



.option local_part_suffix routers&!? "string list" unset
.cindex "router" "suffix for local part"
.cindex "suffix for local part" "used in router"
This option operates in the same way as &%local_part_prefix%&, except that the
local part must end (rather than start) with the given string, the
&%local_part_suffix_optional%& option determines whether the suffix is
mandatory, and the wildcard * character, if present, must be the last
character of the suffix. This option facility is commonly used to handle local
parts of the form &%something-request%& and multiple user mailboxes of the form
&%username-foo%&.


.option local_part_suffix_optional routers boolean false
See &%local_part_suffix%& above.



.option local_parts routers&!? "local part list&!!" unset
.cindex "router" "restricting to specific local parts"
.cindex "local part" "checking in router"
The router is run only if the local part of the address matches the list.
See section &<<SECTrouprecon>>& for a list of the order in which preconditions
are evaluated, and
section &<<SECTlocparlis>>& for a discussion of local part lists. Because the
string is expanded, it is possible to make it depend on the domain, for
example:
.code
local_parts = dbm;/usr/local/specials/$domain_data
.endd
.vindex "&$local_part_data$&"
the data returned by the list check
for the local part is placed in the variable &$local_part_data$& for use in
expansions of the router's private options or in the transport.
You might use this option, for
example, if you have a large number of local virtual domains, and you want to
send all postmaster mail to the same place without having to set up an alias in
each virtual domain:
.code
postmaster:
  driver = redirect
  local_parts = postmaster
  data = postmaster@real.domain.example
.endd


.option log_as_local routers boolean "see below"
.cindex "log" "delivery line"
.cindex "delivery" "log line format"
Exim has two logging styles for delivery, the idea being to make local
deliveries stand out more visibly from remote ones. In the &"local"& style, the
recipient address is given just as the local part, without a domain. The use of
this style is controlled by this option. It defaults to true for the &(accept)&
router, and false for all the others. This option applies only when a
router assigns an address to a transport. It has no effect on routers that
redirect addresses.



.option more routers boolean&!! true
The result of string expansion for this option must be a valid boolean value,
that is, one of the strings &"yes"&, &"no"&, &"true"&, or &"false"&. Any other
result causes an error, and delivery is deferred. If the expansion is forced to
fail, the default value for the option (true) is used. Other failures cause
delivery to be deferred.

If this option is set false, and the router declines to handle the address, no
further routers are tried, routing fails, and the address is bounced.
.oindex "&%self%&"
However, if the router explicitly passes an address to the following router by
means of the setting
.code
self = pass
.endd
or otherwise, the setting of &%more%& is ignored. Also, the setting of &%more%&
does not affect the behaviour if one of the precondition tests fails. In that
case, the address is always passed to the next router.

Note that &%address_data%& is not considered to be a precondition. If its
expansion is forced to fail, the router declines, and the value of &%more%&
controls what happens next.


.option pass_on_timeout routers boolean false
.cindex "timeout" "of router"
.cindex "router" "timeout"
If a router times out during a host lookup, it normally causes deferral of the
address. If &%pass_on_timeout%& is set, the address is passed on to the next
router, overriding &%no_more%&. This may be helpful for systems that are
intermittently connected to the Internet, or those that want to pass to a smart
host any messages that cannot immediately be delivered.

There are occasional other temporary errors that can occur while doing DNS
lookups. They are treated in the same way as a timeout, and this option
applies to all of them.



.option pass_router routers string unset
.cindex "router" "go to after &""pass""&"
Routers that recognize the generic &%self%& option (&(dnslookup)&,
&(ipliteral)&, and &(manualroute)&) are able to return &"pass"&, forcing
routing to continue, and overriding a false setting of &%more%&. When one of
these routers returns &"pass"&, the address is normally handed on to the next
router in sequence. This can be changed by setting &%pass_router%& to the name
of another router. However (unlike &%redirect_router%&) the named router must
be below the current router, to avoid loops. Note that this option applies only
to the special case of &"pass"&. It does not apply when a router returns
&"decline"& because it cannot handle an address.



.option redirect_router routers string unset
.cindex "router" "start at after redirection"
Sometimes an administrator knows that it is pointless to reprocess addresses
generated from alias or forward files with the same router again. For
example, if an alias file translates real names into login ids there is no
point searching the alias file a second time, especially if it is a large file.

The &%redirect_router%& option can be set to the name of any router instance.
It causes the routing of any generated addresses to start at the named router
instead of at the first router. This option has no effect if the router in
which it is set does not generate new addresses.



.option require_files routers&!? "string list&!!" unset
.cindex "file" "requiring for router"
.cindex "router" "requiring file existence"
This option provides a general mechanism for predicating the running of a
router on the existence or non-existence of certain files or directories.
Before running a router, as one of its precondition tests, Exim works its way
through the &%require_files%& list, expanding each item separately.

Because the list is split before expansion, any colons in expansion items must
be doubled, or the facility for using a different list separator must be used
(&<<SECTlistsepchange>>&).
If any expansion is forced to fail, the item is ignored. Other expansion
failures cause routing of the address to be deferred.

If any expanded string is empty, it is ignored. Otherwise, except as described
below, each string must be a fully qualified file path, optionally preceded by
&"!"&. The paths are passed to the &[stat()]& function to test for the
existence of the files or directories. The router is skipped if any paths not
preceded by &"!"& do not exist, or if any paths preceded by &"!"& do exist.

.cindex "NFS"
If &[stat()]& cannot determine whether a file exists or not, delivery of
the message is deferred. This can happen when NFS-mounted filesystems are
unavailable.

This option is checked after the &%domains%&, &%local_parts%&, and &%senders%&
options, so you cannot use it to check for the existence of a file in which to
look up a domain, local part, or sender. (See section &<<SECTrouprecon>>& for a
full list of the order in which preconditions are evaluated.) However, as
these options are all expanded, you can use the &%exists%& expansion condition
to make such tests. The &%require_files%& option is intended for checking files
that the router may be going to use internally, or which are needed by a
transport (e.g., &_.procmailrc_&).

During delivery, the &[stat()]& function is run as root, but there is a
facility for some checking of the accessibility of a file by another user.
This is not a proper permissions check, but just a &"rough"& check that
operates as follows:

If an item in a &%require_files%& list does not contain any forward slash
characters, it is taken to be the user (and optional group, separated by a
comma) to be checked for subsequent files in the list. If no group is specified
but the user is specified symbolically, the gid associated with the uid is
used. For example:
.code
require_files = mail:/some/file
require_files = $local_part_data:$home/.procmailrc
.endd
If a user or group name in a &%require_files%& list does not exist, the
&%require_files%& condition fails.

Exim performs the check by scanning along the components of the file path, and
checking the access for the given uid and gid. It checks for &"x"& access on
directories, and &"r"& access on the final file. Note that this means that file
access control lists, if the operating system has them, are ignored.

&*Warning 1*&: When the router is being run to verify addresses for an
incoming SMTP message, Exim is not running as root, but under its own uid. This
may affect the result of a &%require_files%& check. In particular, &[stat()]&
may yield the error EACCES (&"Permission denied"&). This means that the Exim
user is not permitted to read one of the directories on the file's path.

&*Warning 2*&: Even when Exim is running as root while delivering a message,
&[stat()]& can yield EACCES for a file in an NFS directory that is mounted
without root access. In this case, if a check for access by a particular user
is requested, Exim creates a subprocess that runs as that user, and tries the
check again in that process.

The default action for handling an unresolved EACCES is to consider it to
be caused by a configuration error, and routing is deferred because the
existence or non-existence of the file cannot be determined. However, in some
circumstances it may be desirable to treat this condition as if the file did
not exist. If the filename (or the exclamation mark that precedes the filename
for non-existence) is preceded by a plus sign, the EACCES error is treated
as if the file did not exist. For example:
.code
require_files = +/some/file
.endd
If the router is not an essential part of verification (for example, it
handles users' &_.forward_& files), another solution is to set the &%verify%&
option false so that the router is skipped when verifying.



.option retry_use_local_part routers boolean "see below"
.cindex "hints database" "retry keys"
.cindex "local part" "in retry keys"
When a delivery suffers a temporary routing failure, a retry record is created
in Exim's hints database. For addresses whose routing depends only on the
domain, the key for the retry record should not involve the local part, but for
other addresses, both the domain and the local part should be included.
Usually, remote routing is of the former kind, and local routing is of the
latter kind.

This option controls whether the local part is used to form the key for retry
hints for addresses that suffer temporary errors while being handled by this
router. The default value is true for any router that has any of
&%check_local_user%&,
&%local_parts%&,
&%condition%&,
&%local_part_prefix%&,
&%local_part_suffix%&,
&%senders%& or
&%require_files%&
set, and false otherwise. Note that this option does not apply to hints keys
for transport delays; they are controlled by a generic transport option of the
same name.

Failing to set this option when it is needed
(because a remote router handles only some of the local-parts for a domain)
can result in incorrect error messages being generated.

The setting of &%retry_use_local_part%& applies only to the router on which it
appears. If the router generates child addresses, they are routed
independently; this setting does not become attached to them.



.option router_home_directory routers string&!! unset
.cindex "router" "home directory for"
.cindex "home directory" "for router"
.vindex "&$home$&"
This option sets a home directory for use while the router is running. (Compare
&%transport_home_directory%&, which sets a home directory for later
transporting.) In particular, if used on a &(redirect)& router, this option
sets a value for &$home$& while a filter is running. The value is expanded;
forced expansion failure causes the option to be ignored &-- other failures
cause the router to defer.

Expansion of &%router_home_directory%& happens immediately after the
&%check_local_user%& test (if configured), before any further expansions take
place.
(See section &<<SECTrouprecon>>& for a list of the order in which preconditions
are evaluated.)
While the router is running, &%router_home_directory%& overrides the value of
&$home$& that came from &%check_local_user%&.

When a router accepts an address and assigns it to a local transport (including
the cases when a &(redirect)& router generates a pipe, file, or autoreply
delivery), the home directory setting for the transport is taken from the first
of these values that is set:

.ilist
The &%home_directory%& option on the transport;
.next
The &%transport_home_directory%& option on the router;
.next
The password data if &%check_local_user%& is set on the router;
.next
The &%router_home_directory%& option on the router.
.endlist

In other words, &%router_home_directory%& overrides the password data for the
router, but not for the transport.



.option self routers string freeze
.cindex "MX record" "pointing to local host"
.cindex "local host" "MX pointing to"
This option applies to those routers that use a recipient address to find a
list of remote hosts. Currently, these are the &(dnslookup)&, &(ipliteral)&,
and &(manualroute)& routers.
Certain configurations of the &(queryprogram)& router can also specify a list
of remote hosts.
Usually such routers are configured to send the message to a remote host via an
&(smtp)& transport. The &%self%& option specifies what happens when the first
host on the list turns out to be the local host.
The way in which Exim checks for the local host is described in section
&<<SECTreclocipadd>>&.

Normally this situation indicates either an error in Exim's configuration (for
example, the router should be configured not to process this domain), or an
error in the DNS (for example, the MX should not point to this host). For this
reason, the default action is to log the incident, defer the address, and
freeze the message. The following alternatives are provided for use in special
cases:

.vlist
.vitem &%defer%&
Delivery of the message is tried again later, but the message is not frozen.

.vitem "&%reroute%&: <&'domain'&>"
The domain is changed to the given domain, and the address is passed back to
be reprocessed by the routers. No rewriting of headers takes place. This
behaviour is essentially a redirection.

.vitem "&%reroute: rewrite:%& <&'domain'&>"
The domain is changed to the given domain, and the address is passed back to be
reprocessed by the routers. Any headers that contain the original domain are
rewritten.

.vitem &%pass%&
.oindex "&%more%&"
.vindex "&$self_hostname$&"
The router passes the address to the next router, or to the router named in the
&%pass_router%& option if it is set. This overrides &%no_more%&. During
subsequent routing and delivery, the variable &$self_hostname$& contains the
name of the local host that the router encountered. This can be used to
distinguish between different cases for hosts with multiple names. The
combination
.code
self = pass
no_more
.endd
ensures that only those addresses that routed to the local host are passed on.
Without &%no_more%&, addresses that were declined for other reasons would also
be passed to the next router.

.vitem &%fail%&
Delivery fails and an error report is generated.

.vitem &%send%&
.cindex "local host" "sending to"
The anomaly is ignored and the address is queued for the transport. This
setting should be used with extreme caution. For an &(smtp)& transport, it
makes sense only in cases where the program that is listening on the SMTP port
is not this version of Exim. That is, it must be some other MTA, or Exim with a
different configuration file that handles the domain in another way.
.endlist



.option senders routers&!? "address list&!!" unset
.cindex "router" "checking senders"
If this option is set, the router is skipped unless the message's sender
address matches something on the list.
See section &<<SECTrouprecon>>& for a list of the order in which preconditions
are evaluated.

There are issues concerning verification when the running of routers is
dependent on the sender. When Exim is verifying the address in an &%errors_to%&
setting, it sets the sender to the null string. When using the &%-bt%& option
to check a configuration file, it is necessary also to use the &%-f%& option to
set an appropriate sender. For incoming mail, the sender is unset when
verifying the sender, but is available when verifying any recipients. If the
SMTP VRFY command is enabled, it must be used after MAIL if the sender address
matters.


.option set routers "string list" unset
.cindex router variables
This option may be used multiple times on a router;
because of this the list aspect is mostly irrelevant.
The list separator is a semicolon but can be changed in the
usual way.

Each list-element given must be of the form &"name = value"&
and the names used must start with the string &"r_"&.
Values containing a list-separator should have them doubled.
When a router runs, the strings are evaluated in order,
to create variables which are added to the set associated with
the address.
This is done immediately after all the preconditions, before the
evaluation of the &%address_data%& option.
The variable is set with the expansion of the value.
The variables can be used by the router options
(not including any preconditions)
and by the transport.
Later definitions of a given named variable will override former ones.
Variable use is via the usual &$r_...$& syntax.

This is similar to the &%address_data%& option, except that
many independent variables can be used, with choice of naming.


.option translate_ip_address routers string&!! unset
.cindex "IP address" "translating"
.cindex "packet radio"
.cindex "router" "IP address translation"
There exist some rare networking situations (for example, packet radio) where
it is helpful to be able to translate IP addresses generated by normal routing
mechanisms into other IP addresses, thus performing a kind of manual IP
routing. This should be done only if the normal IP routing of the TCP/IP stack
is inadequate or broken. Because this is an extremely uncommon requirement, the
code to support this option is not included in the Exim binary unless
SUPPORT_TRANSLATE_IP_ADDRESS=yes is set in &_Local/Makefile_&.

.vindex "&$host_address$&"
The &%translate_ip_address%& string is expanded for every IP address generated
by the router, with the generated address set in &$host_address$&. If the
expansion is forced to fail, no action is taken.
For any other expansion error, delivery of the message is deferred.
If the result of the expansion is an IP address, that replaces the original
address; otherwise the result is assumed to be a host name &-- this is looked
up using &[gethostbyname()]& (or &[getipnodebyname()]& when available) to
produce one or more replacement IP addresses. For example, to subvert all IP
addresses in some specific networks, this could be added to a router:
.code
translate_ip_address = \
  ${lookup{${mask:$host_address/26}}lsearch{/some/file}\
    {$value}fail}}
.endd
The file would contain lines like
.code
10.2.3.128/26    some.host
10.8.4.34/26     10.44.8.15
.endd
You should not make use of this facility unless you really understand what you
are doing.



.option transport routers string&!! unset
This option specifies the transport to be used when a router accepts an address
and sets it up for delivery. A transport is never needed if a router is used
only for verification. The value of the option is expanded at routing time,
after the expansion of &%errors_to%&, &%headers_add%&, and &%headers_remove%&,
and result must be the name of one of the configured transports. If it is not,
delivery is deferred.

The &%transport%& option is not used by the &(redirect)& router, but it does
have some private options that set up transports for pipe and file deliveries
(see chapter &<<CHAPredirect>>&).



.option transport_current_directory routers string&!! unset
.cindex "current directory for local transport"
This option associates a current directory with any address that is routed
to a local transport. This can happen either because a transport is
explicitly configured for the router, or because it generates a delivery to a
file or a pipe. During the delivery process (that is, at transport time), this
option string is expanded and is set as the current directory, unless
overridden by a setting on the transport.
If the expansion fails for any reason, including forced failure, an error is
logged, and delivery is deferred.
See chapter &<<CHAPenvironment>>& for details of the local delivery
environment.




.option transport_home_directory routers string&!! "see below"
.cindex "home directory" "for local transport"
This option associates a home directory with any address that is routed to a
local transport. This can happen either because a transport is explicitly
configured for the router, or because it generates a delivery to a file or a
pipe. During the delivery process (that is, at transport time), the option
string is expanded and is set as the home directory, unless overridden by a
setting of &%home_directory%& on the transport.
If the expansion fails for any reason, including forced failure, an error is
logged, and delivery is deferred.

If the transport does not specify a home directory, and
&%transport_home_directory%& is not set for the router, the home directory for
the transport is taken from the password data if &%check_local_user%& is set for
the router. Otherwise it is taken from &%router_home_directory%& if that option
is set; if not, no home directory is set for the transport.

See chapter &<<CHAPenvironment>>& for further details of the local delivery
environment.




.option unseen routers boolean&!! false
.cindex "router" "carrying on after success"
The result of string expansion for this option must be a valid boolean value,
that is, one of the strings &"yes"&, &"no"&, &"true"&, or &"false"&. Any other
result causes an error, and delivery is deferred. If the expansion is forced to
fail, the default value for the option (false) is used. Other failures cause
delivery to be deferred.

When this option is set true, routing does not cease if the router accepts the
address. Instead, a copy of the incoming address is passed to the next router,
overriding a false setting of &%more%&. There is little point in setting
&%more%& false if &%unseen%& is always true, but it may be useful in cases when
the value of &%unseen%& contains expansion items (and therefore, presumably, is
sometimes true and sometimes false).

.cindex "copy of message (&%unseen%& option)"
Setting the &%unseen%& option has a similar effect to the &%unseen%& command
qualifier in filter files. It can be used to cause copies of messages to be
delivered to some other destination, while also carrying out a normal delivery.
In effect, the current address is made into a &"parent"& that has two children
&-- one that is delivered as specified by this router, and a clone that goes on
to be routed further. For this reason, &%unseen%& may not be combined with the
&%one_time%& option in a &(redirect)& router.

&*Warning*&: Header lines added to the address (or specified for removal) by
this router or by previous routers affect the &"unseen"& copy of the message
only. The clone that continues to be processed by further routers starts with
no added headers and none specified for removal. For a &%redirect%& router, if
a generated address is the same as the incoming address, this can lead to
duplicate addresses with different header modifications. Exim does not do
duplicate deliveries (except, in certain circumstances, to pipes -- see section
&<<SECTdupaddr>>&), but it is undefined which of the duplicates is discarded,
so this ambiguous situation should be avoided. The &%repeat_use%& option of the
&%redirect%& router may be of help.

Unlike the handling of header modifications, any data that was set by the
&%address_data%& option in the current or previous routers &'is'& passed on to
subsequent routers.


.option user routers string&!! "see below"
.cindex "uid (user id)" "local delivery"
.cindex "local transports" "uid and gid"
.cindex "transport" "local"
.cindex "router" "user for filter processing"
.cindex "filter" "user for processing"
When a router queues an address for a transport, and the transport does not
specify a user, the user given here is used when running the delivery process.
The user may be specified numerically or by name. If expansion fails, the
error is logged and delivery is deferred.
This user is also used by the &(redirect)& router when running a filter file.
The default is unset, except when &%check_local_user%& is set. In this case,
the default is taken from the password information. If the user is specified as
a name, and &%group%& is not set, the group associated with the user is used.
See also &%initgroups%& and &%group%& and the discussion in chapter
&<<CHAPenvironment>>&.



.option verify routers&!? boolean true
Setting this option has the effect of setting &%verify_sender%& and
&%verify_recipient%& to the same value.


.option verify_only routers&!? boolean false
.cindex "EXPN" "with &%verify_only%&"
.oindex "&%-bv%&"
.cindex "router" "used only when verifying"
If this option is set, the router is used only when verifying an address,
delivering in cutthrough mode or
testing with the &%-bv%& option, not when actually doing a delivery, testing
with the &%-bt%& option, or running the SMTP EXPN command. It can be further
restricted to verifying only senders or recipients by means of
&%verify_sender%& and &%verify_recipient%&.

&*Warning*&: When the router is being run to verify addresses for an incoming
SMTP message, Exim is not running as root, but under its own uid. If the router
accesses any files, you need to make sure that they are accessible to the Exim
user or group.


.option verify_recipient routers&!? boolean true
If this option is false, the router is skipped when verifying recipient
addresses,
delivering in cutthrough mode
or testing recipient verification using &%-bv%&.
See section &<<SECTrouprecon>>& for a list of the order in which preconditions
are evaluated.
See also the &$verify_mode$& variable.


.option verify_sender routers&!? boolean true
If this option is false, the router is skipped when verifying sender addresses
or testing sender verification using &%-bvs%&.
See section &<<SECTrouprecon>>& for a list of the order in which preconditions
are evaluated.
See also the &$verify_mode$& variable.
.ecindex IIDgenoprou1
.ecindex IIDgenoprou2






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The accept router" "CHID4"
.cindex "&(accept)& router"
.cindex "routers" "&(accept)&"
The &(accept)& router has no private options of its own. Unless it is being
used purely for verification (see &%verify_only%&) a transport is required to
be defined by the generic &%transport%& option. If the preconditions that are
specified by generic options are met, the router accepts the address and queues
it for the given transport. The most common use of this router is for setting
up deliveries to local mailboxes. For example:
.code
localusers:
  driver = accept
  domains = mydomain.example
  check_local_user
  transport = local_delivery
.endd
The &%domains%& condition in this example checks the domain of the address, and
&%check_local_user%& checks that the local part is the login of a local user.
When both preconditions are met, the &(accept)& router runs, and queues the
address for the &(local_delivery)& transport.






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The dnslookup router" "CHAPdnslookup"
.scindex IIDdnsrou1 "&(dnslookup)& router"
.scindex IIDdnsrou2 "routers" "&(dnslookup)&"
The &(dnslookup)& router looks up the hosts that handle mail for the
recipient's domain in the DNS. A transport must always be set for this router,
unless &%verify_only%& is set.

If SRV support is configured (see &%check_srv%& below), Exim first searches for
SRV records. If none are found, or if SRV support is not configured,
MX records are looked up. If no MX records exist, address records are sought.
However, &%mx_domains%& can be set to disable the direct use of address
records.

MX records of equal priority are sorted by Exim into a random order. Exim then
looks for address records for the host names obtained from MX or SRV records.
When a host has more than one IP address, they are sorted into a random order,
except that IPv6 addresses are sorted before IPv4 addresses. If all the
IP addresses found are discarded by a setting of the &%ignore_target_hosts%&
generic option, the router declines.

Unless they have the highest priority (lowest MX value), MX records that point
to the local host, or to any host name that matches &%hosts_treat_as_local%&,
are discarded, together with any other MX records of equal or lower priority.

.cindex "MX record" "pointing to local host"
.cindex "local host" "MX pointing to"
.oindex "&%self%&" "in &(dnslookup)& router"
If the host pointed to by the highest priority MX record, or looked up as an
address record, is the local host, or matches &%hosts_treat_as_local%&, what
happens is controlled by the generic &%self%& option.


.section "Problems with DNS lookups" "SECTprowitdnsloo"
There have been problems with DNS servers when SRV records are looked up.
Some misbehaving servers return a DNS error or timeout when a non-existent
SRV record is sought. Similar problems have in the past been reported for
MX records. The global &%dns_again_means_nonexist%& option can help with this
problem, but it is heavy-handed because it is a global option.

For this reason, there are two options, &%srv_fail_domains%& and
&%mx_fail_domains%&, that control what happens when a DNS lookup in a
&(dnslookup)& router results in a DNS failure or a &"try again"& response. If
an attempt to look up an SRV or MX record causes one of these results, and the
domain matches the relevant list, Exim behaves as if the DNS had responded &"no
such record"&. In the case of an SRV lookup, this means that the router
proceeds to look for MX records; in the case of an MX lookup, it proceeds to
look for A or AAAA records, unless the domain matches &%mx_domains%&, in which
case routing fails.


.section "Declining addresses by dnslookup" "SECTdnslookupdecline"
.cindex "&(dnslookup)& router" "declines"
There are a few cases where a &(dnslookup)& router will decline to accept
an address; if such a router is expected to handle "all remaining non-local
domains", then it is important to set &%no_more%&.

The router will defer rather than decline if the domain
is found in the &%fail_defer_domains%& router option.

Reasons for a &(dnslookup)& router to decline currently include:
.ilist
The domain does not exist in DNS
.next
The domain exists but the MX record's host part is just "."; this is a common
convention (borrowed from SRV) used to indicate that there is no such service
for this domain and to not fall back to trying A/AAAA records.
.next
Ditto, but for SRV records, when &%check_srv%& is set on this router.
.next
MX record points to a non-existent host.
.next
MX record points to an IP address and the main section option
&%allow_mx_to_ip%& is not set.
.next
MX records exist and point to valid hosts, but all hosts resolve only to
addresses blocked by the &%ignore_target_hosts%& generic option on this router.
.next
The domain is not syntactically valid (see also &%allow_utf8_domains%& and
&%dns_check_names_pattern%& for handling one variant of this)
.next
&%check_secondary_mx%& is set on this router but the local host can
not be found in the MX records (see below)
.endlist




.section "Private options for dnslookup" "SECID118"
.cindex "options" "&(dnslookup)& router"
The private options for the &(dnslookup)& router are as follows:

.option check_secondary_mx dnslookup boolean false
.cindex "MX record" "checking for secondary"
If this option is set, the router declines unless the local host is found in
(and removed from) the list of hosts obtained by MX lookup. This can be used to
process domains for which the local host is a secondary mail exchanger
differently to other domains. The way in which Exim decides whether a host is
the local host is described in section &<<SECTreclocipadd>>&.


.option check_srv dnslookup string&!! unset
.cindex "SRV record" "enabling use of"
The &(dnslookup)& router supports the use of SRV records
(see &url(https://www.rfc-editor.org/rfc/rfc2782,RFC 2782))
in addition to MX and address records. The support is disabled by default. To
enable SRV support, set the &%check_srv%& option to the name of the service
required. For example,
.code
check_srv = smtp
.endd
looks for SRV records that refer to the normal smtp service. The option is
expanded, so the service name can vary from message to message or address
to address. This might be helpful if SRV records are being used for a
submission service. If the expansion is forced to fail, the &%check_srv%&
option is ignored, and the router proceeds to look for MX records in the
normal way.

When the expansion succeeds, the router searches first for SRV records for
the given service (it assumes TCP protocol). A single SRV record with a
host name that consists of just a single dot indicates &"no such service for
this domain"&; if this is encountered, the router declines. If other kinds of
SRV record are found, they are used to construct a host list for delivery
according to the rules of RFC 2782. MX records are not sought in this case.

When no SRV records are found, MX records (and address records) are sought in
the traditional way. In other words, SRV records take precedence over MX
records, just as MX records take precedence over address records. Note that
this behaviour is not sanctioned by RFC 2782, though a previous draft RFC
defined it. It is apparently believed that MX records are sufficient for email
and that SRV records should not be used for this purpose. However, SRV records
have an additional &"weight"& feature which some people might find useful when
trying to split an SMTP load between hosts of different power.

See section &<<SECTprowitdnsloo>>& above for a discussion of Exim's behaviour
when there is a DNS lookup error.




.option fail_defer_domains dnslookup "domain list&!!" unset
.cindex "MX record" "not found"
DNS lookups for domains matching &%fail_defer_domains%&
which find no matching record will cause the router to defer
rather than the default behaviour of decline.
This maybe be useful for queueing messages for a newly created
domain while the DNS configuration is not ready.
However, it will result in any message with mistyped domains
also being queued.


.option ipv4_only "string&!!" unset
.cindex IPv6 disabling
.cindex DNS "IPv6 disabling"
The string is expanded, and if the result is anything but a forced failure,
or an empty string, or one of the strings “0” or “no” or “false”
(checked without regard to the case of the letters),
only A records are used.

.option ipv4_prefer "string&!!" unset
.cindex IPv4 preference
.cindex DNS "IPv4 preference"
The string is expanded, and if the result is anything but a forced failure,
or an empty string, or one of the strings “0” or “no” or “false”
(checked without regard to the case of the letters),
A records are sorted before AAAA records (inverting the default).

.option mx_domains dnslookup "domain list&!!" unset
.cindex "MX record" "required to exist"
.cindex "SRV record" "required to exist"
A domain that matches &%mx_domains%& is required to have either an MX or an SRV
record in order to be recognized. (The name of this option could be improved.)
For example, if all the mail hosts in &'fict.example'& are known to have MX
records, except for those in &'discworld.fict.example'&, you could use this
setting:
.code
mx_domains = ! *.discworld.fict.example : *.fict.example
.endd
This specifies that messages addressed to a domain that matches the list but
has no MX record should be bounced immediately instead of being routed using
the address record.


.option mx_fail_domains dnslookup "domain list&!!" unset
If the DNS lookup for MX records for one of the domains in this list causes a
DNS lookup error, Exim behaves as if no MX records were found. See section
&<<SECTprowitdnsloo>>& for more discussion.




.option qualify_single dnslookup boolean true
.cindex "DNS" "resolver options"
.cindex "DNS" "qualifying single-component names"
When this option is true, the resolver option RES_DEFNAMES is set for DNS
lookups. Typically, but not standardly, this causes the resolver to qualify
single-component names with the default domain. For example, on a machine
called &'dictionary.ref.example'&, the domain &'thesaurus'& would be changed to
&'thesaurus.ref.example'& inside the resolver. For details of what your
resolver actually does, consult your man pages for &'resolver'& and
&'resolv.conf'&.



.option rewrite_headers dnslookup boolean true
.cindex "rewriting" "header lines"
.cindex "header lines" "rewriting"
If the domain name in the address that is being processed is not fully
qualified, it may be expanded to its full form by a DNS lookup. For example, if
an address is specified as &'dormouse@teaparty'&, the domain might be
expanded to &'teaparty.wonderland.fict.example'&. Domain expansion can also
occur as a result of setting the &%widen_domains%& option. If
&%rewrite_headers%& is true, all occurrences of the abbreviated domain name in
any &'Bcc:'&, &'Cc:'&, &'From:'&, &'Reply-to:'&, &'Sender:'&, and &'To:'&
header lines of the message are rewritten with the full domain name.

This option should be turned off only when it is known that no message is
ever going to be sent outside an environment where the abbreviation makes
sense.

When an MX record is looked up in the DNS and matches a wildcard record, name
servers normally return a record containing the name that has been looked up,
making it impossible to detect whether a wildcard was present or not. However,
some name servers have recently been seen to return the wildcard entry. If the
name returned by a DNS lookup begins with an asterisk, it is not used for
header rewriting.


.option same_domain_copy_routing dnslookup boolean false
.cindex "address" "copying routing"
Addresses with the same domain are normally routed by the &(dnslookup)& router
to the same list of hosts. However, this cannot be presumed, because the router
options and preconditions may refer to the local part of the address. By
default, therefore, Exim routes each address in a message independently. DNS
servers run caches, so repeated DNS lookups are not normally expensive, and in
any case, personal messages rarely have more than a few recipients.

If you are running mailing lists with large numbers of subscribers at the same
domain, and you are using a &(dnslookup)& router which is independent of the
local part, you can set &%same_domain_copy_routing%& to bypass repeated DNS
lookups for identical domains in one message. In this case, when &(dnslookup)&
routes an address to a remote transport, any other unrouted addresses in the
message that have the same domain are automatically given the same routing
without processing them independently,
provided the following conditions are met:

.ilist
No router that processed the address specified &%headers_add%& or
&%headers_remove%&.
.next
The router did not change the address in any way, for example, by &"widening"&
the domain.
.endlist




.option search_parents dnslookup boolean false
.cindex "DNS" "resolver options"
When this option is true, the resolver option RES_DNSRCH is set for DNS
lookups. This is different from the &%qualify_single%& option in that it
applies to domains containing dots. Typically, but not standardly, it causes
the resolver to search for the name in the current domain and in parent
domains. For example, on a machine in the &'fict.example'& domain, if looking
up &'teaparty.wonderland'& failed, the resolver would try
&'teaparty.wonderland.fict.example'&. For details of what your resolver
actually does, consult your man pages for &'resolver'& and &'resolv.conf'&.

Setting this option true can cause problems in domains that have a wildcard MX
record, because any domain that does not have its own MX record matches the
local wildcard.



.option srv_fail_domains dnslookup "domain list&!!" unset
If the DNS lookup for SRV records for one of the domains in this list causes a
DNS lookup error, Exim behaves as if no SRV records were found. See section
&<<SECTprowitdnsloo>>& for more discussion.




.option widen_domains dnslookup "string list" unset
.cindex "domain" "partial; widening"
If a DNS lookup fails and this option is set, each of its strings in turn is
added onto the end of the domain, and the lookup is tried again. For example,
if
.code
widen_domains = fict.example:ref.example
.endd
is set and a lookup of &'klingon.dictionary'& fails,
&'klingon.dictionary.fict.example'& is looked up, and if this fails,
&'klingon.dictionary.ref.example'& is tried. Note that the &%qualify_single%&
and &%search_parents%& options can cause some widening to be undertaken inside
the DNS resolver. &%widen_domains%& is not applied to sender addresses
when verifying, unless &%rewrite_headers%& is false (not the default).


.section "Effect of qualify_single and search_parents" "SECID119"
When a domain from an envelope recipient is changed by the resolver as a result
of the &%qualify_single%& or &%search_parents%& options, Exim rewrites the
corresponding address in the message's header lines unless &%rewrite_headers%&
is set false. Exim then re-routes the address, using the full domain.

These two options affect only the DNS lookup that takes place inside the router
for the domain of the address that is being routed. They do not affect lookups
such as that implied by
.code
domains = @mx_any
.endd
that may happen while processing a router precondition before the router is
entered. No widening ever takes place for these lookups.
.ecindex IIDdnsrou1
.ecindex IIDdnsrou2









. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The ipliteral router" "CHID5"
.cindex "&(ipliteral)& router"
.cindex "domain literal" "routing"
.cindex "routers" "&(ipliteral)&"
This router has no private options. Unless it is being used purely for
verification (see &%verify_only%&) a transport is required to be defined by the
generic &%transport%& option. The router accepts the address if its domain part
takes the form of an
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
domain literal. For example, the &(ipliteral)&
router handles the address
.code
root@[192.168.1.1]
.endd
by setting up delivery to the host with that IP address. IPv4 domain literals
consist of an IPv4 address enclosed in square brackets. IPv6 domain literals
are similar, but the address is preceded by &`ipv6:`&. For example:
.code
postmaster@[ipv6:fe80::a00:20ff:fe86:a061.5678]
.endd
Exim allows &`ipv4:`& before IPv4 addresses, for consistency, and on the
grounds that sooner or later somebody will try it.

.oindex "&%self%&" "in &(ipliteral)& router"
If the IP address matches something in &%ignore_target_hosts%&, the router
declines. If an IP literal turns out to refer to the local host, the generic
&%self%& option determines what happens.

The RFCs require support for domain literals; however, their use is
controversial in today's Internet. If you want to use this router, you must
also set the main configuration option &%allow_domain_literals%&. Otherwise,
Exim will not recognize the domain literal syntax in addresses.



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The iplookup router" "CHID6"
.cindex "&(iplookup)& router"
.cindex "routers" "&(iplookup)&"
The &(iplookup)& router was written to fulfil a specific requirement in
Cambridge University (which in fact no longer exists). For this reason, it is
not included in the binary of Exim by default. If you want to include it, you
must set
.code
ROUTER_IPLOOKUP=yes
.endd
in your &_Local/Makefile_& configuration file.

The &(iplookup)& router routes an address by sending it over a TCP or UDP
connection to one or more specific hosts. The host can then return the same or
a different address &-- in effect rewriting the recipient address in the
message's envelope. The new address is then passed on to subsequent routers. If
this process fails, the address can be passed on to other routers, or delivery
can be deferred. Since &(iplookup)& is just a rewriting router, a transport
must not be specified for it.

.cindex "options" "&(iplookup)& router"
.option hosts iplookup string unset
This option must be supplied. Its value is a colon-separated list of host
names. The hosts are looked up using &[gethostbyname()]&
(or &[getipnodebyname()]& when available)
and are tried in order until one responds to the query. If none respond, what
happens is controlled by &%optional%&.


.option optional iplookup boolean false
If &%optional%& is true, if no response is obtained from any host, the address
is passed to the next router, overriding &%no_more%&. If &%optional%& is false,
delivery to the address is deferred.


.option port iplookup integer 0
.cindex "port" "&(iplookup)& router"
This option must be supplied. It specifies the port number for the TCP or UDP
call.


.option protocol iplookup string udp
This option can be set to &"udp"& or &"tcp"& to specify which of the two
protocols is to be used.


.option query iplookup string&!! "see below"
This defines the content of the query that is sent to the remote hosts. The
default value is:
.code
$local_part@$domain $local_part@$domain
.endd
The repetition serves as a way of checking that a response is to the correct
query in the default case (see &%response_pattern%& below).


.option reroute iplookup string&!! unset
If this option is not set, the rerouted address is precisely the byte string
returned by the remote host, up to the first white space, if any. If set, the
string is expanded to form the rerouted address. It can include parts matched
in the response by &%response_pattern%& by means of numeric variables such as
&$1$&, &$2$&, etc. The variable &$0$& refers to the entire input string,
whether or not a pattern is in use. In all cases, the rerouted address must end
up in the form &'local_part@domain'&.


.option response_pattern iplookup string unset
This option can be set to a regular expression that is applied to the string
returned from the remote host. If the pattern does not match the response, the
router declines. If &%response_pattern%& is not set, no checking of the
response is done, unless the query was defaulted, in which case there is a
check that the text returned after the first white space is the original
address. This checks that the answer that has been received is in response to
the correct question. For example, if the response is just a new domain, the
following could be used:
.code
response_pattern = ^([^@]+)$
reroute = $local_part@$1
.endd

.option timeout iplookup time 5s
This specifies the amount of time to wait for a response from the remote
machine. The same timeout is used for the &[connect()]& function for a TCP
call. It does not apply to UDP.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The manualroute router" "CHID7"
.scindex IIDmanrou1 "&(manualroute)& router"
.scindex IIDmanrou2 "routers" "&(manualroute)&"
.cindex "domain" "manually routing"
The &(manualroute)& router is so-called because it provides a way of manually
routing an address according to its domain. It is mainly used when you want to
route addresses to remote hosts according to your own rules, bypassing the
normal DNS routing that looks up MX records. However, &(manualroute)& can also
route to local transports, a facility that may be useful if you want to save
messages for dial-in hosts in local files.

The &(manualroute)& router compares a list of domain patterns with the domain
it is trying to route. If there is no match, the router declines. Each pattern
has associated with it a list of hosts and some other optional data, which may
include a transport. The combination of a pattern and its data is called a
&"routing rule"&. For patterns that do not have an associated transport, the
generic &%transport%& option must specify a transport, unless the router is
being used purely for verification (see &%verify_only%&).

.vindex "&$host$&"
In the case of verification, matching the domain pattern is sufficient for the
router to accept the address. When actually routing an address for delivery,
an address that matches a domain pattern is queued for the associated
transport. If the transport is not a local one, a host list must be associated
with the pattern; IP addresses are looked up for the hosts, and these are
passed to the transport along with the mail address. For local transports, a
host list is optional. If it is present, it is passed in &$host$& as a single
text string.

The list of routing rules can be provided as an inline string in
&%route_list%&, or the data can be obtained by looking up the domain in a file
or database by setting &%route_data%&. Only one of these settings may appear in
any one instance of &(manualroute)&. The format of routing rules is described
below, following the list of private options.


.section "Private options for manualroute" "SECTprioptman"

.cindex "options" "&(manualroute)& router"
The private options for the &(manualroute)& router are as follows:

.option host_all_ignored manualroute string defer
See &%host_find_failed%&.

.option host_find_failed manualroute string freeze
This option controls what happens when &(manualroute)& tries to find an IP
address for a host, and the host does not exist. The option can be set to one
of the following values:
.code
decline
defer
fail
freeze
ignore
pass
.endd
The default (&"freeze"&) assumes that this state is a serious configuration
error. The difference between &"pass"& and &"decline"& is that the former
forces the address to be passed to the next router (or the router defined by
&%pass_router%&),
.oindex "&%more%&"
overriding &%no_more%&, whereas the latter passes the address to the next
router only if &%more%& is true.

The value &"ignore"& causes Exim to completely ignore a host whose IP address
cannot be found. If all the hosts in the list are ignored, the behaviour is
controlled by the &%host_all_ignored%& option. This takes the same values
as &%host_find_failed%&, except that it cannot be set to &"ignore"&.

The &%host_find_failed%& option applies only to a definite &"does not exist"&
state; if a host lookup gets a temporary error, delivery is deferred unless the
generic &%pass_on_timeout%& option is set.


.option hosts_randomize manualroute boolean false
.cindex "randomized host list"
.cindex "host" "list of; randomized"
If this option is set, the order of the items in a host list in a routing rule
is randomized each time the list is used, unless an option in the routing rule
overrides (see below). Randomizing the order of a host list can be used to do
crude load sharing. However, if more than one mail address is routed by the
same router to the same host list, the host lists are considered to be the same
(even though they may be randomized into different orders) for the purpose of
deciding whether to batch the deliveries into a single SMTP transaction.

When &%hosts_randomize%& is true, a host list may be split
into groups whose order is separately randomized. This makes it possible to
set up MX-like behaviour. The boundaries between groups are indicated by an
item that is just &`+`& in the host list. For example:
.code
route_list = * host1:host2:host3:+:host4:host5
.endd
The order of the first three hosts and the order of the last two hosts is
randomized for each use, but the first three always end up before the last two.
If &%hosts_randomize%& is not set, a &`+`& item in the list is ignored. If a
randomized host list is passed to an &(smtp)& transport that also has
&%hosts_randomize set%&, the list is not re-randomized.


.option route_data manualroute string&!! unset
If this option is set, it must expand to yield the data part of a routing rule.
Typically, the expansion string includes a lookup based on the domain. For
example:
.code
route_data = ${lookup{$domain}dbm{/etc/routes}}
.endd
If the expansion is forced to fail, or the result is an empty string, the
router declines. Other kinds of expansion failure cause delivery to be
deferred.


.option route_list manualroute "string list" unset
This string is a list of routing rules, in the form defined below. Note that,
unlike most string lists, the items are separated by semicolons. This is so
that they may contain colon-separated host lists.


.option same_domain_copy_routing manualroute boolean false
.cindex "address" "copying routing"
Addresses with the same domain are normally routed by the &(manualroute)&
router to the same list of hosts. However, this cannot be presumed, because the
router options and preconditions may refer to the local part of the address. By
default, therefore, Exim routes each address in a message independently. DNS
servers run caches, so repeated DNS lookups are not normally expensive, and in
any case, personal messages rarely have more than a few recipients.

If you are running mailing lists with large numbers of subscribers at the same
domain, and you are using a &(manualroute)& router which is independent of the
local part, you can set &%same_domain_copy_routing%& to bypass repeated DNS
lookups for identical domains in one message. In this case, when
&(manualroute)& routes an address to a remote transport, any other unrouted
addresses in the message that have the same domain are automatically given the
same routing without processing them independently. However, this is only done
if &%headers_add%& and &%headers_remove%& are unset.




.section "Routing rules in route_list" "SECID120"
The value of &%route_list%& is a string consisting of a sequence of routing
rules, separated by semicolons. If a semicolon is needed in a rule, it can be
entered as two semicolons. Alternatively, the list separator can be changed as
described (for colon-separated lists) in section &<<SECTlistconstruct>>&.
Empty rules are ignored. The format of each rule is
.display
<&'domain pattern'&>  <&'list of hosts'&>  <&'options'&>
.endd
The following example contains two rules, each with a simple domain pattern and
no options:
.code
route_list = \
  dict.ref.example  mail-1.ref.example:mail-2.ref.example ; \
  thes.ref.example  mail-3.ref.example:mail-4.ref.example
.endd
The three parts of a rule are separated by white space. The pattern and the
list of hosts can be enclosed in quotes if necessary, and if they are, the
usual quoting rules apply. Each rule in a &%route_list%& must start with a
single domain pattern, which is the only mandatory item in the rule. The
pattern is in the same format as one item in a domain list (see section
&<<SECTdomainlist>>&),
except that it may not be the name of an interpolated file.
That is, it may be wildcarded, or a regular expression, or a file or database
lookup (with semicolons doubled, because of the use of semicolon as a separator
in a &%route_list%&).

The rules in &%route_list%& are searched in order until one of the patterns
matches the domain that is being routed. The list of hosts and then options are
then used as described below. If there is no match, the router declines. When
&%route_list%& is set, &%route_data%& must not be set.



.section "Routing rules in route_data" "SECID121"
The use of &%route_list%& is convenient when there are only a small number of
routing rules. For larger numbers, it is easier to use a file or database to
hold the routing information, and use the &%route_data%& option instead.
The value of &%route_data%& is a list of hosts, followed by (optional) options.
Most commonly, &%route_data%& is set as a string that contains an
expansion lookup. For example, suppose we place two routing rules in a file
like this:
.code
dict.ref.example:  mail-1.ref.example:mail-2.ref.example
thes.ref.example:  mail-3.ref.example:mail-4.ref.example
.endd
This data can be accessed by setting
.code
route_data = ${lookup{$domain}lsearch{/the/file/name}}
.endd
Failure of the lookup results in an empty string, causing the router to
decline. However, you do not have to use a lookup in &%route_data%&. The only
requirement is that the result of expanding the string is a list of hosts,
possibly followed by options, separated by white space. The list of hosts must
be enclosed in quotes if it contains white space.




.section "Format of the list of hosts" "SECID122"
A list of hosts, whether obtained via &%route_data%& or &%route_list%&, is
always separately expanded before use. If the expansion fails, the router
declines. The result of the expansion must be a colon-separated list of names
and/or IP addresses, optionally also including ports.
If the list is written with spaces, it must be protected with quotes.
The format of each item
in the list is described in the next section. The list separator can be changed
as described in section &<<SECTlistsepchange>>&.

If the list of hosts was obtained from a &%route_list%& item, the following
variables are set during its expansion:

.ilist
.cindex "numerical variables (&$1$& &$2$&  etc)" "in &(manualroute)& router"
If the domain was matched against a regular expression, the numeric variables
&$1$&, &$2$&, etc. may be set. For example:
.code
route_list = ^domain(\d+)   host-$1.text.example
.endd
.next
&$0$& is always set to the entire domain.
.next
&$1$& is also set when partial matching is done in a file lookup.

.next
.vindex "&$value$&"
If the pattern that matched the domain was a lookup item, the data that was
looked up is available in the expansion variable &$value$&. For example:
.code
route_list = lsearch;;/some/file.routes  $value
.endd
.endlist

Note the doubling of the semicolon in the pattern that is necessary because
semicolon is the default route list separator.



.section "Format of one host item" "SECTformatonehostitem"
Each item in the list of hosts can be either a host name or an IP address,
optionally with an attached port number, or it can be a single "+"
(see &%hosts_randomize%&).
When no port is given, an IP address
is not enclosed in brackets. When a port is specified, it overrides the port
specification on the transport. The port is separated from the name or address
by a colon. This leads to some complications:

.ilist
Because colon is the default separator for the list of hosts, either
the colon that specifies a port must be doubled, or the list separator must
be changed. The following two examples have the same effect:
.code
route_list = * "host1.tld::1225 : host2.tld::1226"
route_list = * "<+ host1.tld:1225 + host2.tld:1226"
.endd
.next
When IPv6 addresses are involved, it gets worse, because they contain
colons of their own. To make this case easier, it is permitted to
enclose an IP address (either v4 or v6) in square brackets if a port
number follows. For example:
.code
route_list = * "</ [10.1.1.1]:1225 / [::1]:1226"
.endd
.endlist

.section "How the list of hosts is used" "SECThostshowused"
When an address is routed to an &(smtp)& transport by &(manualroute)&, each of
the hosts is tried, in the order specified, when carrying out the SMTP
delivery. However, the order can be changed by setting the &%hosts_randomize%&
option, either on the router (see section &<<SECTprioptman>>& above), or on the
transport.

Hosts may be listed by name or by IP address. An unadorned name in the list of
hosts is interpreted as a host name. A name that is followed by &`/MX`& is
interpreted as an indirection to a sublist of hosts obtained by looking up MX
records in the DNS. For example:
.code
route_list = *  x.y.z:p.q.r/MX:e.f.g
.endd
If this feature is used with a port specifier, the port must come last. For
example:
.code
route_list = *  dom1.tld/mx::1225
.endd
If the &%hosts_randomize%& option is set, the order of the items in the list is
randomized before any lookups are done. Exim then scans the list; for any name
that is not followed by &`/MX`& it looks up an IP address. If this turns out to
be an interface on the local host and the item is not the first in the list,
Exim discards it and any subsequent items. If it is the first item, what
happens is controlled by the
.oindex "&%self%&" "in &(manualroute)& router"
&%self%& option of the router.

A name on the list that is followed by &`/MX`& is replaced with the list of
hosts obtained by looking up MX records for the name. This is always a DNS
lookup; the &%bydns%& and &%byname%& options (see section &<<SECThowoptused>>&
below) are not relevant here. The order of these hosts is determined by the
preference values in the MX records, according to the usual rules. Because
randomizing happens before the MX lookup, it does not affect the order that is
defined by MX preferences.

If the local host is present in the sublist obtained from MX records, but is
not the most preferred host in that list, it and any equally or less
preferred hosts are removed before the sublist is inserted into the main list.

If the local host is the most preferred host in the MX list, what happens
depends on where in the original list of hosts the &`/MX`& item appears. If it
is not the first item (that is, there are previous hosts in the main list),
Exim discards this name and any subsequent items in the main list.

If the MX item is first in the list of hosts, and the local host is the
most preferred host, what happens is controlled by the &%self%& option of the
router.

DNS failures when lookup up the MX records are treated in the same way as DNS
failures when looking up IP addresses: &%pass_on_timeout%& and
&%host_find_failed%& are used when relevant.

The generic &%ignore_target_hosts%& option applies to all hosts in the list,
whether obtained from an MX lookup or not.



.section "How the options are used" "SECThowoptused"
The options are a sequence of words, space-separated.
One of the words can be the name of a transport; this overrides the
&%transport%& option on the router for this particular routing rule only. The
other words (if present) control randomization of the list of hosts on a
per-rule basis, and how the IP addresses of the hosts are to be found when
routing to a remote transport. These options are as follows:

.ilist
&%randomize%&: randomize the order of the hosts in this list, overriding the
setting of &%hosts_randomize%& for this routing rule only.
.next
&%no_randomize%&: do not randomize the order of the hosts in this list,
overriding the setting of &%hosts_randomize%& for this routing rule only.
.next
&%byname%&: use &[getipnodebyname()]& (&[gethostbyname()]& on older systems) to
find IP addresses. This function may ultimately cause a DNS lookup, but it may
also look in &_/etc/hosts_& or other sources of information.
.next
&%bydns%&: look up address records for the hosts directly in the DNS; fail if
no address records are found. If there is a temporary DNS error (such as a
timeout), delivery is deferred.
.next
&%ipv4_only%&: in direct DNS lookups, look up only A records.
.next
&%ipv4_prefer%&: in direct DNS lookups, sort A records before AAAA records.
.endlist

For example:
.code
route_list = domain1  host1:host2:host3  randomize bydns;\
             domain2  host4:host5
.endd
If neither &%byname%& nor &%bydns%& is given, Exim behaves as follows: First, a
DNS lookup is done. If this yields anything other than HOST_NOT_FOUND, that
result is used. Otherwise, Exim goes on to try a call to &[getipnodebyname()]&
or &[gethostbyname()]&, and the result of the lookup is the result of that
call.

&*Warning*&: It has been discovered that on some systems, if a DNS lookup
called via &[getipnodebyname()]& times out, HOST_NOT_FOUND is returned
instead of TRY_AGAIN. That is why the default action is to try a DNS
lookup first. Only if that gives a definite &"no such host"& is the local
function called.

&*Compatibility*&: From Exim 4.85 until fixed for 4.90, there was an
inadvertent constraint that a transport name as an option had to be the last
option specified.



If no IP address for a host can be found, what happens is controlled by the
&%host_find_failed%& option.

.vindex "&$host$&"
When an address is routed to a local transport, IP addresses are not looked up.
The host list is passed to the transport in the &$host$& variable.



.section "Manualroute examples" "SECID123"
In some of the examples that follow, the presence of the &%remote_smtp%&
transport, as defined in the default configuration file, is assumed:

.ilist
.cindex "smart host" "example router"
The &(manualroute)& router can be used to forward all external mail to a
&'smart host'&. If you have set up, in the main part of the configuration, a
named domain list that contains your local domains, for example:
.code
domainlist local_domains = my.domain.example
.endd
You can arrange for all other domains to be routed to a smart host by making
your first router something like this:
.code
smart_route:
  driver = manualroute
  domains = !+local_domains
  transport = remote_smtp
  route_list = * smarthost.ref.example
.endd
This causes all non-local addresses to be sent to the single host
&'smarthost.ref.example'&. If a colon-separated list of smart hosts is given,
they are tried in order
(but you can use &%hosts_randomize%& to vary the order each time).
Another way of configuring the same thing is this:
.code
smart_route:
  driver = manualroute
  transport = remote_smtp
  route_list = !+local_domains  smarthost.ref.example
.endd
There is no difference in behaviour between these two routers as they stand.
However, they behave differently if &%no_more%& is added to them. In the first
example, the router is skipped if the domain does not match the &%domains%&
precondition; the following router is always tried. If the router runs, it
always matches the domain and so can never decline. Therefore, &%no_more%&
would have no effect. In the second case, the router is never skipped; it
always runs. However, if it doesn't match the domain, it declines. In this case
&%no_more%& would prevent subsequent routers from running.

.next
.cindex "mail hub example"
A &'mail hub'& is a host which receives mail for a number of domains via MX
records in the DNS and delivers it via its own private routing mechanism. Often
the final destinations are behind a firewall, with the mail hub being the one
machine that can connect to machines both inside and outside the firewall. The
&(manualroute)& router is usually used on a mail hub to route incoming messages
to the correct hosts. For a small number of domains, the routing can be inline,
using the &%route_list%& option, but for a larger number a file or database
lookup is easier to manage.

If the domain names are in fact the names of the machines to which the mail is
to be sent by the mail hub, the configuration can be quite simple. For
example:
.code
hub_route:
  driver = manualroute
  transport = remote_smtp
  route_list = *.rhodes.tvs.example  $domain
.endd
This configuration routes domains that match &`*.rhodes.tvs.example`& to hosts
whose names are the same as the mail domains. A similar approach can be taken
if the host name can be obtained from the domain name by a string manipulation
that the expansion facilities can handle. Otherwise, a lookup based on the
domain can be used to find the host:
.code
through_firewall:
  driver = manualroute
  transport = remote_smtp
  route_data = ${lookup {$domain} cdb {/internal/host/routes}}
.endd
The result of the lookup must be the name or IP address of the host (or
hosts) to which the address is to be routed. If the lookup fails, the route
data is empty, causing the router to decline. The address then passes to the
next router.

.next
.cindex "batched SMTP output example"
.cindex "SMTP" "batched outgoing; example"
You can use &(manualroute)& to deliver messages to pipes or files in batched
SMTP format for onward transportation by some other means. This is one way of
storing mail for a dial-up host when it is not connected. The route list entry
can be as simple as a single domain name in a configuration like this:
.code
save_in_file:
  driver = manualroute
  transport = batchsmtp_appendfile
  route_list = saved.domain.example
.endd
though often a pattern is used to pick up more than one domain. If there are
several domains or groups of domains with different transport requirements,
different transports can be listed in the routing information:
.code
save_in_file:
  driver = manualroute
  route_list = \
    *.saved.domain1.example  $domain  batch_appendfile; \
    *.saved.domain2.example  \
      ${lookup{$domain}dbm{/domain2/hosts}{$value}fail} \
      batch_pipe
.endd
.vindex "&$domain$&"
.vindex "&$host$&"
The first of these just passes the domain in the &$host$& variable, which
doesn't achieve much (since it is also in &$domain$&), but the second does a
file lookup to find a value to pass, causing the router to decline to handle
the address if the lookup fails.

.next
.cindex "UUCP" "example of router for"
Routing mail directly to UUCP software is a specific case of the use of
&(manualroute)& in a gateway to another mail environment. This is an example of
one way it can be done:
.code
# Transport
uucp:
  driver = pipe
  user = nobody
  command = /usr/local/bin/uux -r - \
    ${substr_-5:$host}!rmail ${local_part}
  return_fail_output = true

# Router
uucphost:
  transport = uucp
  driver = manualroute
  route_data = \
    ${lookup{$domain}lsearch{/usr/local/exim/uucphosts}}
.endd
The file &_/usr/local/exim/uucphosts_& contains entries like
.code
darksite.ethereal.example:           darksite.UUCP
.endd
It can be set up more simply without adding and removing &".UUCP"& but this way
makes clear the distinction between the domain name
&'darksite.ethereal.example'& and the UUCP host name &'darksite'&.
.endlist
.ecindex IIDmanrou1
.ecindex IIDmanrou2








. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The queryprogram router" "CHAPdriverlast"
.scindex IIDquerou1 "&(queryprogram)& router"
.scindex IIDquerou2 "routers" "&(queryprogram)&"
.cindex "routing" "by external program"
The &(queryprogram)& router routes an address by running an external command
and acting on its output. This is an expensive way to route, and is intended
mainly for use in lightly-loaded systems, or for performing experiments.
However, if it is possible to use the precondition options (&%domains%&,
&%local_parts%&, etc) to skip this router for most addresses, it could sensibly
be used in special cases, even on a busy host. There are the following private
options:
.cindex "options" "&(queryprogram)& router"

.option command queryprogram string&!! unset
This option must be set. It specifies the command that is to be run. The
command is split up into a command name and arguments, and then each is
expanded separately (exactly as for a &(pipe)& transport, described in chapter
&<<CHAPpipetransport>>&).


.option command_group queryprogram string unset
.cindex "gid (group id)" "in &(queryprogram)& router"
This option specifies a gid to be set when running the command while routing an
address for deliver. It must be set if &%command_user%& specifies a numerical
uid. If it begins with a digit, it is interpreted as the numerical value of the
gid. Otherwise it is looked up using &[getgrnam()]&.


.option command_user queryprogram string unset
.cindex "uid (user id)" "for &(queryprogram)&"
This option must be set. It specifies the uid which is set when running the
command while routing an address for delivery. If the value begins with a digit,
it is interpreted as the numerical value of the uid. Otherwise, it is looked up
using &[getpwnam()]& to obtain a value for the uid and, if &%command_group%& is
not set, a value for the gid also.

&*Warning:*& Changing uid and gid is possible only when Exim is running as
root, which it does during a normal delivery in a conventional configuration.
However, when an address is being verified during message reception, Exim is
usually running as the Exim user, not as root. If the &(queryprogram)& router
is called from a non-root process, Exim cannot change uid or gid before running
the command. In this circumstance the command runs under the current uid and
gid.


.option current_directory queryprogram string /
This option specifies an absolute path which is made the current directory
before running the command.


.option timeout queryprogram time 1h
If the command does not complete within the timeout period, its process group
is killed and the message is frozen. A value of zero time specifies no
timeout.


The standard output of the command is connected to a pipe, which is read when
the command terminates. It should consist of a single line of output,
containing up to five fields, separated by white space. The maximum length of
the line is 1023 characters. Longer lines are silently truncated. The first
field is one of the following words (case-insensitive):

.ilist
&'Accept'&: routing succeeded; the remaining fields specify what to do (see
below).
.next
&'Decline'&: the router declines; pass the address to the next router, unless
&%no_more%& is set.
.next
&'Fail'&: routing failed; do not pass the address to any more routers. Any
subsequent text on the line is an error message. If the router is run as part
of address verification during an incoming SMTP message, the message is
included in the SMTP response.
.next
&'Defer'&: routing could not be completed at this time; try again later. Any
subsequent text on the line is an error message which is logged. It is not
included in any SMTP response.
.next
&'Freeze'&: the same as &'defer'&, except that the message is frozen.
.next
&'Pass'&: pass the address to the next router (or the router specified by
&%pass_router%&), overriding &%no_more%&.
.next
&'Redirect'&: the message is redirected. The remainder of the line is a list of
new addresses, which are routed independently, starting with the first router,
or the router specified by &%redirect_router%&, if set.
.endlist

When the first word is &'accept'&, the remainder of the line consists of a
number of keyed data values, as follows (split into two lines here, to fit on
the page):
.code
ACCEPT TRANSPORT=<transport> HOSTS=<list of hosts>
LOOKUP=byname|bydns DATA=<text>
.endd
The data items can be given in any order, and all are optional. If no transport
is included, the transport specified by the generic &%transport%& option is
used. The list of hosts and the lookup type are needed only if the transport is
an &(smtp)& transport that does not itself supply a list of hosts.

The format of the list of hosts is the same as for the &(manualroute)& router.
As well as host names and IP addresses with optional port numbers, as described
in section &<<SECTformatonehostitem>>&, it may contain names followed by
&`/MX`& to specify sublists of hosts that are obtained by looking up MX records
(see section &<<SECThostshowused>>&).

If the lookup type is not specified, Exim behaves as follows when trying to
find an IP address for each host: First, a DNS lookup is done. If this yields
anything other than HOST_NOT_FOUND, that result is used. Otherwise, Exim
goes on to try a call to &[getipnodebyname()]& or &[gethostbyname()]&, and the
result of the lookup is the result of that call.

.vindex "&$address_data$&"
If the DATA field is set, its value is placed in the &$address_data$&
variable. For example, this return line
.code
accept hosts=x1.y.example:x2.y.example data="rule1"
.endd
routes the address to the default transport, passing a list of two hosts. When
the transport runs, the string &"rule1"& is in &$address_data$&.
.ecindex IIDquerou1
.ecindex IIDquerou2




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The redirect router" "CHAPredirect"
.scindex IIDredrou1 "&(redirect)& router"
.scindex IIDredrou2 "routers" "&(redirect)&"
.cindex "alias file" "in a &(redirect)& router"
.cindex "address redirection" "&(redirect)& router"
The &(redirect)& router handles several kinds of address redirection. Its most
common uses are for resolving local part aliases from a central alias file
(usually called &_/etc/aliases_&) and for handling users' personal &_.forward_&
files, but it has many other potential uses. The incoming address can be
redirected in several different ways:

.ilist
It can be replaced by one or more new addresses which are themselves routed
independently.
.next
It can be routed to be delivered to a given file or directory.
.next
It can be routed to be delivered to a specified pipe command.
.next
It can cause an automatic reply to be generated.
.next
It can be forced to fail, optionally with a custom error message.
.next
It can be temporarily deferred, optionally with a custom message.
.next
It can be discarded.
.endlist

The generic &%transport%& option must not be set for &(redirect)& routers.
However, there are some private options which define transports for delivery to
files and pipes, and for generating autoreplies. See the &%file_transport%&,
&%pipe_transport%& and &%reply_transport%& descriptions below.

If success DSNs have been requested
.cindex "DSN" "success"
.cindex "Delivery Status Notification" "success"
redirection triggers one and the DSN options are not passed any further.



.section "Redirection data" "SECID124"
The router operates by interpreting a text string which it obtains either by
expanding the contents of the &%data%& option, or by reading the entire
contents of a file whose name is given in the &%file%& option. These two
options are mutually exclusive. The first is commonly used for handling system
aliases, in a configuration like this:
.code
system_aliases:
  driver = redirect
  data = ${lookup{$local_part}lsearch{/etc/aliases}}
.endd
If the lookup fails, the expanded string in this example is empty. When the
expansion of &%data%& results in an empty string, the router declines. A forced
expansion failure also causes the router to decline; other expansion failures
cause delivery to be deferred.

A configuration using &%file%& is commonly used for handling users'
&_.forward_& files, like this:
.code
userforward:
  driver = redirect
  check_local_user
  file = $home/.forward
  no_verify
.endd
If the file does not exist, or causes no action to be taken (for example, it is
empty or consists only of comments), the router declines. &*Warning*&: This
is not the case when the file contains syntactically valid items that happen to
yield empty addresses, for example, items containing only
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address comments.

.cindex "tainted data" "in filenames"
.cindex redirect "tainted data"
Tainted data may not be used for a filename.

&*Warning*&: It is unwise to use &$local_part$& or &$domain$&
directly for redirection,
as they are provided by a potential attacker.
In the examples above, &$local_part$& is used for looking up data held locally
on the system, and not used directly (the second example derives &$home$& via
the passsword file or database, using &$local_part$&).



.section "Forward files and address verification" "SECID125"
.cindex "address redirection" "while verifying"
It is usual to set &%no_verify%& on &(redirect)& routers which handle users'
&_.forward_& files, as in the example above. There are two reasons for this:

.ilist
When Exim is receiving an incoming SMTP message from a remote host, it is
running under the Exim uid, not as root. Exim is unable to change uid to read
the file as the user, and it may not be able to read it as the Exim user. So in
practice the router may not be able to operate.
.next
However, even when the router can operate, the existence of a &_.forward_& file
is unimportant when verifying an address. What should be checked is whether the
local part is a valid user name or not. Cutting out the redirection processing
saves some resources.
.endlist






.section "Interpreting redirection data" "SECID126"
.cindex "Sieve filter" "specifying in redirection data"
.cindex "filter" "specifying in redirection data"
The contents of the data string, whether obtained from &%data%& or &%file%&,
can be interpreted in two different ways:

.ilist
If the &%allow_filter%& option is set true, and the data begins with the text
&"#Exim filter"& or &"#Sieve filter"&, it is interpreted as a list of
&'filtering'& instructions in the form of an Exim or Sieve filter file,
respectively. Details of the syntax and semantics of filter files are described
in a separate document entitled &'Exim's interfaces to mail filtering'&; this
document is intended for use by end users.
.next
Otherwise, the data must be a comma-separated list of redirection items, as
described in the next section.
.endlist

When a message is redirected to a file (a &"mail folder"&), the filename given
in a non-filter redirection list must always be an absolute path. A filter may
generate a relative path &-- how this is handled depends on the transport's
configuration. See section &<<SECTfildiropt>>& for a discussion of this issue
for the &(appendfile)& transport.



.section "Items in a non-filter redirection list" "SECTitenonfilred"
.cindex "address redirection" "non-filter list items"
When the redirection data is not an Exim or Sieve filter, for example, if it
comes from a conventional alias or forward file, it consists of a list of
addresses, filenames, pipe commands, or certain special items (see section
&<<SECTspecitredli>>& below). The special items can be individually enabled or
disabled by means of options whose names begin with &%allow_%& or &%forbid_%&,
depending on their default values. The items in the list are separated by
commas or newlines.
If a comma is required in an item, the entire item must be enclosed in double
quotes.

Lines starting with a # character are comments, and are ignored, and # may
also appear following a comma, in which case everything between the # and the
next newline character is ignored.

If an item is entirely enclosed in double quotes, these are removed. Otherwise
double quotes are retained because some forms of mail address require their use
(but never to enclose the entire address). In the following description,
&"item"& refers to what remains after any surrounding double quotes have been
removed.

.vindex "&$local_part$&"
&*Warning*&: If you use an Exim expansion to construct a redirection address,
and the expansion contains a reference to &$local_part$&, you should make use
of the &%quote_local_part%& expansion operator, in case the local part contains
special characters. For example, to redirect all mail for the domain
&'obsolete.example'&, retaining the existing local part, you could use this
setting:
.code
data = ${quote_local_part:$local_part}@newdomain.example
.endd


.section "Redirecting to a local mailbox" "SECTredlocmai"
.cindex "routing" "loops in"
.cindex "loop" "while routing, avoidance of"
.cindex "address redirection" "to local mailbox"
A redirection item may safely be the same as the address currently under
consideration. This does not cause a routing loop, because a router is
automatically skipped if any ancestor of the address that is being processed
is the same as the current address and was processed by the current router.
Such an address is therefore passed to the following routers, so it is handled
as if there were no redirection. When making this loop-avoidance test, the
complete local part, including any prefix or suffix, is used.

.cindex "address redirection" "local part without domain"
Specifying the same local part without a domain is a common usage in personal
filter files when the user wants to have messages delivered to the local
mailbox and also forwarded elsewhere. For example, the user whose login is
&'cleo'& might have a &_.forward_& file containing this:
.code
cleo, cleopatra@egypt.example
.endd
.cindex "backslash in alias file"
.cindex "alias file" "backslash in"
For compatibility with other MTAs, such unqualified local parts may be
preceded by &"\"&, but this is not a requirement for loop prevention. However,
it does make a difference if more than one domain is being handled
synonymously.

If an item begins with &"\"& and the rest of the item parses as a valid
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address that does not include a domain, the item is qualified using the
domain of the incoming address. In the absence of a leading &"\"&, unqualified
addresses are qualified using the value in &%qualify_recipient%&, but you can
force the incoming domain to be used by setting &%qualify_preserve_domain%&.

Care must be taken if there are alias names for local users.
Consider an MTA handling a single local domain where the system alias file
contains:
.code
Sam.Reman: spqr
.endd
Now suppose that Sam (whose login id is &'spqr'&) wants to save copies of
messages in the local mailbox, and also forward copies elsewhere. He creates
this forward file:
.code
Sam.Reman, spqr@reme.elsewhere.example
.endd
With these settings, an incoming message addressed to &'Sam.Reman'& fails. The
&(redirect)& router for system aliases does not process &'Sam.Reman'& the
second time round, because it has previously routed it,
and the following routers presumably cannot handle the alias. The forward file
should really contain
.code
spqr, spqr@reme.elsewhere.example
.endd
but because this is such a common error, the &%check_ancestor%& option (see
below) exists to provide a way to get round it. This is normally set on a
&(redirect)& router that is handling users' &_.forward_& files.



.section "Special items in redirection lists" "SECTspecitredli"
In addition to addresses, the following types of item may appear in redirection
lists (that is, in non-filter redirection data):

.ilist
.cindex "pipe" "in redirection list"
.cindex "address redirection" "to pipe"
An item is treated as a pipe command if it begins with &"|"& and does not parse
as a valid
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address that includes a domain. A transport for running the
command must be specified by the &%pipe_transport%& option.
Normally, either the router or the transport specifies a user and a group under
which to run the delivery. The default is to use the Exim user and group.

Single or double quotes can be used for enclosing the individual arguments of
the pipe command; no interpretation of escapes is done for single quotes. If
the command contains a comma character, it is necessary to put the whole item
in double quotes, for example:
.code
"|/some/command ready,steady,go"
.endd
since items in redirection lists are terminated by commas. Do not, however,
quote just the command. An item such as
.code
|"/some/command ready,steady,go"
.endd
is interpreted as a pipe with a rather strange command name, and no arguments.

Note that the above example assumes that the text comes from a lookup source
of some sort, so that the quotes are part of the data.  If composing a
redirect router with a &%data%& option directly specifying this command, the
quotes will be used by the configuration parser to define the extent of one
string, but will not be passed down into the redirect router itself.  There
are two main approaches to get around this: escape quotes to be part of the
data itself, or avoid using this mechanism and instead create a custom
transport with the &%command%& option set and reference that transport from
an &%accept%& router.

.next
.cindex "file" "in redirection list"
.cindex "address redirection" "to file"
An item is interpreted as a path name if it begins with &"/"& and does not
parse as a valid
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address that includes a domain. For example,
.code
/home/world/minbari
.endd
is treated as a filename, but
.code
/s=molari/o=babylon/@x400gate.way
.endd
is treated as an address. For a filename, a transport must be specified using
the &%file_transport%& option. However, if the generated path name ends with a
forward slash character, it is interpreted as a directory name rather than a
filename, and &%directory_transport%& is used instead.

Normally, either the router or the transport specifies a user and a group under
which to run the delivery. The default is to use the Exim user and group.

.cindex "&_/dev/null_&"
However, if a redirection item is the path &_/dev/null_&, delivery to it is
bypassed at a high level, and the log entry shows &"**bypassed**"&
instead of a transport name. In this case the user and group are not used.

.next
.cindex "included address list"
.cindex "address redirection" "included external list"
If an item is of the form
.code
:include:<path name>
.endd
a list of further items is taken from the given file and included at that
point. &*Note*&: Such a file can not be a filter file; it is just an
out-of-line addition to the list. The items in the included list are separated
by commas or newlines and are not subject to expansion. If this is the first
item in an alias list in an &(lsearch)& file, a colon must be used to terminate
the alias name. This example is incorrect:
.code
list1    :include:/opt/lists/list1
.endd
It must be given as
.code
list1:   :include:/opt/lists/list1
.endd
.cindex "tainted data" "in filenames"
.cindex redirect "tainted data"
Tainted data may not be used for a filename.
.next
.cindex "address redirection" "to black hole"
.cindex "delivery" "discard"
.cindex "delivery" "blackhole"
.cindex "black hole"
.cindex "abandoning mail"
Sometimes you want to throw away mail to a particular local part.  Making the
&%data%& option expand to an empty string does not work, because that causes
the router to decline. Instead, the alias item
.code
:blackhole:
.endd
can be used. It does what its name implies. No delivery is
done, and no error message is generated. This has the same effect as specifying
&_/dev/null_& as a destination, but it can be independently disabled.

&*Warning*&: If &':blackhole:'& appears anywhere in a redirection list, no
delivery is done for the original local part, even if other redirection items
are present. If you are generating a multi-item list (for example, by reading a
database) and need the ability to provide a no-op item, you must use
&_/dev/null_&.

.next
.cindex "delivery" "forcing failure"
.cindex "delivery" "forcing deferral"
.cindex "failing delivery" "forcing"
.cindex "deferred delivery, forcing"
.cindex "customizing" "failure message"
An attempt to deliver a particular address can be deferred or forced to fail by
redirection items of the form
.code
:defer:
:fail:
.endd
respectively. When a redirection list contains such an item, it applies
to the entire redirection; any other items in the list are ignored. Any
text following &':fail:'& or &':defer:'& is placed in the error text
associated with the failure. For example, an alias file might contain:
.code
X.Employee:  :fail: Gone away, no forwarding address
.endd
In the case of an address that is being verified from an ACL or as the subject
of a
.cindex "VRFY" "error text, display of"
VRFY command, the text is included in the SMTP error response by
default.
.cindex "EXPN" "error text, display of"
The text is not included in the response to an EXPN command. In non-SMTP cases
the text is included in the error message that Exim generates.

.cindex "SMTP" "error codes"
By default for verify, Exim sends a 451 SMTP code for a &':defer:'&, and 550 for
&':fail:'&. However, if the message starts with three digits followed by a
space, optionally followed by an extended code of the form &'n.n.n'&, also
followed by a space, and the very first digit is the same as the default error
code, the code from the message is used instead. If the very first digit is
incorrect, a panic error is logged, and the default code is used. You can
suppress the use of the supplied code in a redirect router by setting the
&%forbid_smtp_code%& option true. In this case, any SMTP code is quietly
ignored.

.vindex "&$acl_verify_message$&"
In an ACL, an explicitly provided message overrides the default, but the
default message is available in the variable &$acl_verify_message$& and can
therefore be included in a custom message if this is desired.

Normally the error text is the rest of the redirection list &-- a comma does
not terminate it &-- but a newline does act as a terminator. Newlines are not
normally present in alias expansions. In &(lsearch)& lookups they are removed
as part of the continuation process, but they may exist in other kinds of
lookup and in &':include:'& files.

During routing for message delivery (as opposed to verification), a redirection
containing &':fail:'& causes an immediate failure of the incoming address,
whereas &':defer:'& causes the message to remain in the queue so that a
subsequent delivery attempt can happen at a later time. If an address is
deferred for too long, it will ultimately fail, because the normal retry
rules still apply.

.next
.cindex "alias file" "exception to default"
Sometimes it is useful to use a single-key search type with a default (see
chapter &<<CHAPfdlookup>>&) to look up aliases. However, there may be a need
for exceptions to the default. These can be handled by aliasing them to
&':unknown:'&. This differs from &':fail:'& in that it causes the &(redirect)&
router to decline, whereas &':fail:'& forces routing to fail. A lookup which
results in an empty redirection list has the same effect.
.endlist


.section "Duplicate addresses" "SECTdupaddr"
.cindex "duplicate addresses"
.cindex "address duplicate, discarding"
.cindex "pipe" "duplicated"
Exim removes duplicate addresses from the list to which it is delivering, so as
to deliver just one copy to each address. This does not apply to deliveries
routed to pipes by different immediate parent addresses, but an indirect
aliasing scheme of the type
.code
pipe:       |/some/command $local_part
localpart1: pipe
localpart2: pipe
.endd
does not work with a message that is addressed to both local parts, because
when the second is aliased to the intermediate local part &"pipe"& it gets
discarded as being the same as a previously handled address. However, a scheme
such as
.code
localpart1: |/some/command $local_part
localpart2: |/some/command $local_part
.endd
does result in two different pipe deliveries, because the immediate parents of
the pipes are distinct.



.section "Repeated redirection expansion" "SECID128"
.cindex "repeated redirection expansion"
.cindex "address redirection" "repeated for each delivery attempt"
When a message cannot be delivered to all of its recipients immediately,
leading to two or more delivery attempts, redirection expansion is carried out
afresh each time for those addresses whose children were not all previously
delivered. If redirection is being used as a mailing list, this can lead to new
members of the list receiving copies of old messages. The &%one_time%& option
can be used to avoid this.


.section "Errors in redirection lists" "SECID129"
.cindex "address redirection" "errors"
If &%skip_syntax_errors%& is set, a malformed address that causes a parsing
error is skipped, and an entry is written to the main log. This may be useful
for mailing lists that are automatically managed. Otherwise, if an error is
detected while generating the list of new addresses, the original address is
deferred. See also &%syntax_errors_to%&.



.section "Private options for the redirect router" "SECID130"

.cindex "options" "&(redirect)& router"
The private options for the &(redirect)& router are as follows:


.option allow_defer redirect boolean false
Setting this option allows the use of &':defer:'& in non-filter redirection
data, or the &%defer%& command in an Exim filter file.


.option allow_fail redirect boolean false
.cindex "failing delivery" "from filter"
If this option is true, the &':fail:'& item can be used in a redirection list,
and the &%fail%& command may be used in an Exim filter file.


.option allow_filter redirect boolean false
.cindex "filter" "enabling use of"
.cindex "Sieve filter" "enabling use of"
Setting this option allows Exim to interpret redirection data that starts with
&"#Exim filter"& or &"#Sieve filter"& as a set of filtering instructions. There
are some features of Exim filter files that some administrators may wish to
lock out; see the &%forbid_filter_%&&'xxx'& options below.

It is also possible to lock out Exim filters or Sieve filters while allowing
the other type; see &%forbid_exim_filter%& and &%forbid_sieve_filter%&.


The filter is run using the uid and gid set by the generic &%user%& and
&%group%& options. These take their defaults from the password data if
&%check_local_user%& is set, so in the normal case of users' personal filter
files, the filter is run as the relevant user. When &%allow_filter%& is set
true, Exim insists that either &%check_local_user%& or &%user%& is set.



.option allow_freeze redirect boolean false
.cindex "freezing messages" "allowing in filter"
Setting this option allows the use of the &%freeze%& command in an Exim filter.
This command is more normally encountered in system filters, and is disabled by
default for redirection filters because it isn't something you usually want to
let ordinary users do.



.option check_ancestor redirect boolean false
This option is concerned with handling generated addresses that are the same
as some address in the list of redirection ancestors of the current address.
Although it is turned off by default in the code, it is set in the default
configuration file for handling users' &_.forward_& files. It is recommended
for this use of the &(redirect)& router.

When &%check_ancestor%& is set, if a generated address (including the domain)
is the same as any ancestor of the current address, it is replaced by a copy of
the current address. This helps in the case where local part A is aliased to B,
and B has a &_.forward_& file pointing back to A. For example, within a single
domain, the local part &"Joe.Bloggs"& is aliased to &"jb"& and
&_&~jb/.forward_& contains:
.code
\Joe.Bloggs, <other item(s)>
.endd
Without the &%check_ancestor%& setting, either local part (&"jb"& or
&"joe.bloggs"&) gets processed once by each router and so ends up as it was
originally. If &"jb"& is the real mailbox name, mail to &"jb"& gets delivered
(having been turned into &"joe.bloggs"& by the &_.forward_& file and back to
&"jb"& by the alias), but mail to &"joe.bloggs"& fails. Setting
&%check_ancestor%& on the &(redirect)& router that handles the &_.forward_&
file prevents it from turning &"jb"& back into &"joe.bloggs"& when that was the
original address. See also the &%repeat_use%& option below.


.option check_group redirect boolean "see below"
When the &%file%& option is used, the group owner of the file is checked only
when this option is set. The permitted groups are those listed in the
&%owngroups%& option, together with the user's default group if
&%check_local_user%& is set. If the file has the wrong group, routing is
deferred. The default setting for this option is true if &%check_local_user%&
is set and the &%modemask%& option permits the group write bit, or if the
&%owngroups%& option is set. Otherwise it is false, and no group check occurs.



.option check_owner redirect boolean "see below"
When the &%file%& option is used, the owner of the file is checked only when
this option is set. If &%check_local_user%& is set, the local user is
permitted; otherwise the owner must be one of those listed in the &%owners%&
option. The default value for this option is true if &%check_local_user%& or
&%owners%& is set. Otherwise the default is false, and no owner check occurs.


.option data redirect string&!! unset
This option is mutually exclusive with &%file%&. One or other of them must be
set, but not both. The contents of &%data%& are expanded, and then used as the
list of forwarding items, or as a set of filtering instructions. If the
expansion is forced to fail, or the result is an empty string or a string that
has no effect (consists entirely of comments), the router declines.

When filtering instructions are used, the string must begin with &"#Exim
filter"&, and all comments in the string, including this initial one, must be
terminated with newline characters. For example:
.code
data = #Exim filter\n\
       if $h_to: contains Exim then save $home/mail/exim endif
.endd
If you are reading the data from a database where newlines cannot be included,
you can use the &${sg}$& expansion item to turn the escape string of your
choice into a newline.


.option directory_transport redirect string&!! unset
A &(redirect)& router sets up a direct delivery to a directory when a path name
ending with a slash is specified as a new &"address"&. The transport used is
specified by this option, which, after expansion, must be the name of a
configured transport. This should normally be an &(appendfile)& transport.


.option file redirect string&!! unset
This option specifies the name of a file that contains the redirection data. It
is mutually exclusive with the &%data%& option. The string is expanded before
use; if the expansion is forced to fail, the router declines. Other expansion
failures cause delivery to be deferred. The result of a successful expansion
must be an absolute path. The entire file is read and used as the redirection
data. If the data is an empty string or a string that has no effect (consists
entirely of comments), the router declines.

.cindex "NFS" "checking for file existence"
If the attempt to open the file fails with a &"does not exist"& error, Exim
runs a check on the containing directory,
unless &%ignore_enotdir%& is true (see below).
If the directory does not appear to exist, delivery is deferred. This can
happen when users' &_.forward_& files are in NFS-mounted directories, and there
is a mount problem. If the containing directory does exist, but the file does
not, the router declines.


.option file_transport redirect string&!! unset
.vindex "&$address_file$&"
A &(redirect)& router sets up a direct delivery to a file when a path name not
ending in a slash is specified as a new &"address"&. The transport used is
specified by this option, which, after expansion, must be the name of a
configured transport. This should normally be an &(appendfile)& transport. When
it is running, the filename is in &$address_file$&.


.option filter_prepend_home redirect boolean true
When this option is true, if a &(save)& command in an Exim filter specifies a
relative path, and &$home$& is defined, it is automatically prepended to the
relative path. If this option is set false, this action does not happen. The
relative path is then passed to the transport unmodified.


.option forbid_blackhole redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, the &':blackhole:'& item may not appear in a
redirection list.


.option forbid_exim_filter redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is set true, only Sieve filters are permitted when
&%allow_filter%& is true.




.option forbid_file redirect boolean false
.cindex "restricting access to features"
.cindex "delivery" "to file; forbidding"
.cindex "filter" "locking out certain features"
.cindex "Sieve filter" "forbidding delivery to a file"
.cindex "Sieve filter" "&""keep""& facility; disabling"
If this option is true, this router may not generate a new address that
specifies delivery to a local file or directory, either from a filter or from a
conventional forward file. This option is forced to be true if &%one_time%& is
set. It applies to Sieve filters as well as to Exim filters, but if true, it
locks out the Sieve's &"keep"& facility.


.option forbid_filter_dlfunc redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, string expansions in Exim filters are not allowed to
make use of the &%dlfunc%& expansion facility to run dynamically loaded
functions.

.option forbid_filter_existstest redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
.cindex "expansion" "statting a file"
If this option is true, string expansions in Exim filters are not allowed to
make use of the &%exists%& condition or the &%stat%& expansion item.

.option forbid_filter_logwrite redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, use of the logging facility in Exim filters is not
permitted. Logging is in any case available only if the filter is being run
under some unprivileged uid (which is normally the case for ordinary users'
&_.forward_& files).


.option forbid_filter_lookup redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, string expansions in Exim filter files are not allowed
to make use of &%lookup%& items.


.option forbid_filter_perl redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
This option has an effect only if Exim is built with embedded Perl support. If
it is true, string expansions in Exim filter files are not allowed to make use
of the embedded Perl support.


.option forbid_filter_readfile redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, string expansions in Exim filter files are not allowed
to make use of &%readfile%& items.


.option forbid_filter_readsocket redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, string expansions in Exim filter files are not allowed
to make use of &%readsocket%& items.


.option forbid_filter_reply redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, this router may not generate an automatic reply
message. Automatic replies can be generated only from Exim or Sieve filter
files, not from traditional forward files. This option is forced to be true if
&%one_time%& is set.


.option forbid_filter_run redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, string expansions in Exim filter files are not allowed
to make use of &%run%& items.


.option forbid_include redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is true, items of the form
.code
:include:<path name>
.endd
are not permitted in non-filter redirection lists.


.option forbid_pipe redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
.cindex "delivery" "to pipe; forbidding"
If this option is true, this router may not generate a new address which
specifies delivery to a pipe, either from an Exim filter or from a conventional
forward file. This option is forced to be true if &%one_time%& is set.


.option forbid_sieve_filter redirect boolean false
.cindex "restricting access to features"
.cindex "filter" "locking out certain features"
If this option is set true, only Exim filters are permitted when
&%allow_filter%& is true.


.cindex "SMTP" "error codes"
.option forbid_smtp_code redirect boolean false
If this option is set true, any SMTP error codes that are present at the start
of messages specified for &`:defer:`& or &`:fail:`& are quietly ignored, and
the default codes (451 and 550, respectively) are always used.




.option hide_child_in_errmsg redirect boolean false
.cindex "bounce message" "redirection details; suppressing"
If this option is true, it prevents Exim from quoting a child address if it
generates a bounce or delay message for it. Instead it says &"an address
generated from <&'the top level address'&>"&. Of course, this applies only to
bounces generated locally. If a message is forwarded to another host, &'its'&
bounce may well quote the generated address.


.option ignore_eacces redirect boolean false
.cindex "EACCES"
If this option is set and an attempt to open a redirection file yields the
EACCES error (permission denied), the &(redirect)& router behaves as if the
file did not exist.


.option ignore_enotdir redirect boolean false
.cindex "ENOTDIR"
If this option is set and an attempt to open a redirection file yields the
ENOTDIR error (something on the path is not a directory), the &(redirect)&
router behaves as if the file did not exist.

Setting &%ignore_enotdir%& has another effect as well: When a &(redirect)&
router that has the &%file%& option set discovers that the file does not exist
(the ENOENT error), it tries to &[stat()]& the parent directory, as a check
against unmounted NFS directories. If the parent can not be statted, delivery
is deferred. However, it seems wrong to do this check when &%ignore_enotdir%&
is set, because that option tells Exim to ignore &"something on the path is not
a directory"& (the ENOTDIR error). This is a confusing area, because it seems
that some operating systems give ENOENT where others give ENOTDIR.



.option include_directory redirect string unset
If this option is set, the path names of any &':include:'& items in a
redirection list must start with this directory.


.option modemask redirect "octal integer" 022
This specifies mode bits which must not be set for a file specified by the
&%file%& option. If any of the forbidden bits are set, delivery is deferred.


.option one_time redirect boolean false
.cindex "one-time aliasing/forwarding expansion"
.cindex "alias file" "one-time expansion"
.cindex "forward file" "one-time expansion"
.cindex "mailing lists" "one-time expansion"
.cindex "address redirection" "one-time expansion"
Sometimes the fact that Exim re-evaluates aliases and reprocesses redirection
files each time it tries to deliver a message causes a problem when one or more
of the generated addresses fails be delivered at the first attempt. The problem
is not one of duplicate delivery &-- Exim is clever enough to handle that &--
but of what happens when the redirection list changes during the time that the
message is on Exim's queue. This is particularly true in the case of mailing
lists, where new subscribers might receive copies of messages that were posted
before they subscribed.

If &%one_time%& is set and any addresses generated by the router fail to
deliver at the first attempt, the failing addresses are added to the message as
&"top level"& addresses, and the parent address that generated them is marked
&"delivered"&. Thus, redirection does not happen again at the next delivery
attempt.

&*Warning 1*&: Any header line addition or removal that is specified by this
router would be lost if delivery did not succeed at the first attempt. For this
reason, the &%headers_add%& and &%headers_remove%& generic options are not
permitted when &%one_time%& is set.

&*Warning 2*&: To ensure that the router generates only addresses (as opposed
to pipe or file deliveries or auto-replies) &%forbid_file%&, &%forbid_pipe%&,
and &%forbid_filter_reply%& are forced to be true when &%one_time%& is set.

&*Warning 3*&: The &%unseen%& generic router option may not be set with
&%one_time%&.

The original top-level address is remembered with each of the generated
addresses, and is output in any log messages. However, any intermediate parent
addresses are not recorded. This makes a difference to the log only if
&%all_parents%& log selector is set. It is expected that &%one_time%& will
typically be used for mailing lists, where there is normally just one level of
expansion.


.option owners redirect "string list" unset
.cindex "ownership" "alias file"
.cindex "ownership" "forward file"
.cindex "alias file" "ownership"
.cindex "forward file" "ownership"
This specifies a list of permitted owners for the file specified by &%file%&.
This list is in addition to the local user when &%check_local_user%& is set.
See &%check_owner%& above.


.option owngroups redirect "string list" unset
This specifies a list of permitted groups for the file specified by &%file%&.
The list is in addition to the local user's primary group when
&%check_local_user%& is set. See &%check_group%& above.


.option pipe_transport redirect string&!! unset
.vindex "&$address_pipe$&"
A &(redirect)& router sets up a direct delivery to a pipe when a string
starting with a vertical bar character is specified as a new &"address"&. The
transport used is specified by this option, which, after expansion, must be the
name of a configured transport. This should normally be a &(pipe)& transport.
When the transport is run, the pipe command is in &$address_pipe$&.


.option qualify_domain redirect string&!! unset
.vindex "&$qualify_recipient$&"
If this option is set, and an unqualified address (one without a domain) is
generated, and that address would normally be qualified by the global setting
in &%qualify_recipient%&, it is instead qualified with the domain specified by
expanding this string. If the expansion fails, the router declines. If you want
to revert to the default, you can have the expansion generate
&$qualify_recipient$&.

This option applies to all unqualified addresses generated by Exim filters,
but for traditional &_.forward_& files, it applies only to addresses that are
not preceded by a backslash. Sieve filters cannot generate unqualified
addresses.

.option qualify_preserve_domain redirect boolean false
.cindex "domain" "in redirection; preserving"
.cindex "preserving domain in redirection"
.cindex "address redirection" "domain; preserving"
If this option is set, the router's local &%qualify_domain%& option must not be
set (a configuration error occurs if it is). If an unqualified address (one
without a domain) is generated, it is qualified with the domain of the parent
address (the immediately preceding ancestor) instead of the global
&%qualify_recipient%& value. In the case of a traditional &_.forward_& file,
this applies whether or not the address is preceded by a backslash.


.option repeat_use redirect boolean true
If this option is set false, the router is skipped for a child address that has
any ancestor that was routed by this router. This test happens before any of
the other preconditions are tested. Exim's default anti-looping rules skip
only when the ancestor is the same as the current address. See also
&%check_ancestor%& above and the generic &%redirect_router%& option.


.option reply_transport redirect string&!! unset
A &(redirect)& router sets up an automatic reply when a &%mail%& or
&%vacation%& command is used in a filter file. The transport used is specified
by this option, which, after expansion, must be the name of a configured
transport. This should normally be an &(autoreply)& transport. Other transports
are unlikely to do anything sensible or useful.


.option rewrite redirect boolean true
.cindex "address redirection" "disabling rewriting"
If this option is set false, addresses generated by the router are not
subject to address rewriting. Otherwise, they are treated like new addresses
and are rewritten according to the global rewriting rules.


.option sieve_inbox redirect string&!! inbox
.new
The value of this option is passed to a Sieve filter to specify the
name of the mailbox used for "keep" operations (explicit or implicit).
.wen

.option sieve_subaddress redirect string&!! unset
The value of this option is passed to a Sieve filter to specify the
:subaddress part of an address.

.option sieve_useraddress redirect string&!! unset
The value of this option is passed to a Sieve filter to specify the :user part
of an address. However, if it is unset, the entire original local part
(including any prefix or suffix) is used for :user.


.option sieve_vacation_directory redirect string&!! unset
.cindex "Sieve filter" "vacation directory"
To enable the &"vacation"& extension for Sieve filters, you must set
&%sieve_vacation_directory%& to the directory where vacation databases are held
(do not put anything else in that directory), and ensure that the
&%reply_transport%& option refers to an &(autoreply)& transport. Each user
needs their own directory; Exim will create it if necessary.



.option skip_syntax_errors redirect boolean false
.cindex "forward file" "broken"
.cindex "address redirection" "broken files"
.cindex "alias file" "broken"
.cindex "broken alias or forward files"
.cindex "ignoring faulty addresses"
.cindex "skipping faulty addresses"
.cindex "error" "skipping bad syntax"
If &%skip_syntax_errors%& is set, syntactically malformed addresses in
non-filter redirection data are skipped, and each failing address is logged. If
&%syntax_errors_to%& is set, a message is sent to the address it defines,
giving details of the failures. If &%syntax_errors_text%& is set, its contents
are expanded and placed at the head of the error message generated by
&%syntax_errors_to%&. Usually it is appropriate to set &%syntax_errors_to%& to
be the same address as the generic &%errors_to%& option. The
&%skip_syntax_errors%& option is often used when handling mailing lists.

If all the addresses in a redirection list are skipped because of syntax
errors, the router declines to handle the original address, and it is passed to
the following routers.

If &%skip_syntax_errors%& is set when an Exim filter is interpreted, any syntax
error in the filter causes filtering to be abandoned without any action being
taken. The incident is logged, and the router declines to handle the address,
so it is passed to the following routers.

.cindex "Sieve filter" "syntax errors in"
Syntax errors in a Sieve filter file cause the &"keep"& action to occur. This
action is specified by
&url(https://www.rfc-editor.org/rfc/rfc3028.html,RFC 3028).
The values of &%skip_syntax_errors%&,
&%syntax_errors_to%&, and &%syntax_errors_text%& are not used.

&%skip_syntax_errors%& can be used to specify that errors in users' forward
lists or filter files should not prevent delivery. The &%syntax_errors_to%&
option, used with an address that does not get redirected, can be used to
notify users of these errors, by means of a router like this:
.code
userforward:
  driver = redirect
  allow_filter
  check_local_user
  file = $home/.forward
  file_transport = address_file
  pipe_transport = address_pipe
  reply_transport = address_reply
  no_verify
  skip_syntax_errors
  syntax_errors_to = real-$local_part@$domain
  syntax_errors_text = \
   This is an automatically generated message. An error has\n\
   been found in your .forward file. Details of the error are\n\
   reported below. While this error persists, you will receive\n\
   a copy of this message for every message that is addressed\n\
   to you. If your .forward file is a filter file, or if it is\n\
   a non-filter file containing no valid forwarding addresses,\n\
   a copy of each incoming message will be put in your normal\n\
   mailbox. If a non-filter file contains at least one valid\n\
   forwarding address, forwarding to the valid addresses will\n\
   happen, and those will be the only deliveries that occur.
.endd
You also need a router to ensure that local addresses that are prefixed by
&`real-`& are recognized, but not forwarded or filtered. For example, you could
put this immediately before the &(userforward)& router:
.code
real_localuser:
  driver = accept
  check_local_user
  local_part_prefix = real-
  transport = local_delivery
.endd
For security, it would probably be a good idea to restrict the use of this
router to locally-generated messages, using a condition such as this:
.code
  condition = ${if match_ip {$sender_host_address} \
                        {<; ; 127.0.0.1 ; ::1}}
.endd


.option syntax_errors_text redirect string&!! unset
See &%skip_syntax_errors%& above.


.option syntax_errors_to redirect string unset
See &%skip_syntax_errors%& above.
.ecindex IIDredrou1
.ecindex IIDredrou2






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Environment for running local transports" "CHAPenvironment" &&&
         "Environment for local transports"
.scindex IIDenvlotra1 "local transports" "environment for"
.scindex IIDenvlotra2 "environment" "local transports"
.scindex IIDenvlotra3 "transport" "local; environment for"
Local transports handle deliveries to files and pipes. (The &(autoreply)&
transport can be thought of as similar to a pipe.) Exim always runs transports
in subprocesses, under specified uids and gids. Typical deliveries to local
mailboxes run under the uid and gid of the local user.

Exim also sets a specific current directory while running the transport; for
some transports a home directory setting is also relevant. The &(pipe)&
transport is the only one that sets up environment variables; see section
&<<SECTpipeenv>>& for details.

The values used for the uid, gid, and the directories may come from several
different places. In many cases, the router that handles the address associates
settings with that address as a result of its &%check_local_user%&, &%group%&,
or &%user%& options. However, values may also be given in the transport's own
configuration, and these override anything that comes from the router.



.section "Concurrent deliveries" "SECID131"
.cindex "concurrent deliveries"
.cindex "simultaneous deliveries"
If two different messages for the same local recipient arrive more or less
simultaneously, the two delivery processes are likely to run concurrently. When
the &(appendfile)& transport is used to write to a file, Exim applies locking
rules to stop concurrent processes from writing to the same file at the same
time.

However, when you use a &(pipe)& transport, it is up to you to arrange any
locking that is needed. Here is a silly example:
.code
my_transport:
  driver = pipe
  command = /bin/sh -c 'cat >>/some/file'
.endd
This is supposed to write the message at the end of the file. However, if two
messages arrive at the same time, the file will be scrambled. You can use the
&%exim_lock%& utility program (see section &<<SECTmailboxmaint>>&) to lock a
file using the same algorithm that Exim itself uses.




.section "Uids and gids" "SECTenvuidgid"
.cindex "local transports" "uid and gid"
.cindex "transport" "local; uid and gid"
All transports have the options &%group%& and &%user%&. If &%group%& is set, it
overrides any group that the router set in the address, even if &%user%& is not
set for the transport. This makes it possible, for example, to run local mail
delivery under the uid of the recipient (set by the router), but in a special
group (set by the transport). For example:
.code
# Routers ...
# User/group are set by check_local_user in this router
local_users:
  driver = accept
  check_local_user
  transport = group_delivery

# Transports ...
# This transport overrides the group
group_delivery:
  driver = appendfile
  file = /var/spool/mail/$local_part_data
  group = mail
.endd
If &%user%& is set for a transport, its value overrides what is set in the
address by the router. If &%user%& is non-numeric and &%group%& is not set, the
gid associated with the user is used. If &%user%& is numeric, &%group%& must be
set.

.oindex "&%initgroups%&"
When the uid is taken from the transport's configuration, the &[initgroups()]&
function is called for the groups associated with that uid if the
&%initgroups%& option is set for the transport. When the uid is not specified
by the transport, but is associated with the address by a router, the option
for calling &[initgroups()]& is taken from the router configuration.

.cindex "&(pipe)& transport" "uid for"
The &(pipe)& transport contains the special option &%pipe_as_creator%&. If this
is set and &%user%& is not set, the uid of the process that called Exim to
receive the message is used, and if &%group%& is not set, the corresponding
original gid is also used.

This is the detailed preference order for obtaining a gid; the first of the
following that is set is used:

.ilist
A &%group%& setting of the transport;
.next
A &%group%& setting of the router;
.next
A gid associated with a user setting of the router, either as a result of
&%check_local_user%& or an explicit non-numeric &%user%& setting;
.next
The group associated with a non-numeric &%user%& setting of the transport;
.next
In a &(pipe)& transport, the creator's gid if &%deliver_as_creator%& is set and
the uid is the creator's uid;
.next
The Exim gid if the Exim uid is being used as a default.
.endlist

If, for example, the user is specified numerically on the router and there are
no group settings, no gid is available. In this situation, an error occurs.
This is different for the uid, for which there always is an ultimate default.
The first of the following that is set is used:

.ilist
A &%user%& setting of the transport;
.next
In a &(pipe)& transport, the creator's uid if &%deliver_as_creator%& is set;
.next
A &%user%& setting of the router;
.next
A &%check_local_user%& setting of the router;
.next
The Exim uid.
.endlist

Of course, an error will still occur if the uid that is chosen is on the
&%never_users%& list.





.section "Current and home directories" "SECID132"
.cindex "current directory for local transport"
.cindex "home directory" "for local transport"
.cindex "transport" "local; home directory for"
.cindex "transport" "local; current directory for"
Routers may set current and home directories for local transports by means of
the &%transport_current_directory%& and &%transport_home_directory%& options.
However, if the transport's &%current_directory%& or &%home_directory%& options
are set, they override the router's values. In detail, the home directory
for a local transport is taken from the first of these values that is set:

.ilist
The &%home_directory%& option on the transport;
.next
The &%transport_home_directory%& option on the router;
.next
The password data if &%check_local_user%& is set on the router;
.next
The &%router_home_directory%& option on the router.
.endlist

The current directory is taken from the first of these values that is set:

.ilist
The &%current_directory%& option on the transport;
.next
The &%transport_current_directory%& option on the router.
.endlist


If neither the router nor the transport sets a current directory, Exim uses the
value of the home directory, if it is set. Otherwise it sets the current
directory to &_/_& before running a local transport.



.section "Expansion variables derived from the address" "SECID133"
.vindex "&$domain$&"
.vindex "&$local_part$&"
.vindex "&$original_domain$&"
Normally a local delivery is handling a single address, and in that case the
variables such as &$domain$& and &$local_part$& are set during local
deliveries. However, in some circumstances more than one address may be handled
at once (for example, while writing batch SMTP for onward transmission by some
other means). In this case, the variables associated with the local part are
never set, &$domain$& is set only if all the addresses have the same domain,
and &$original_domain$& is never set.
.ecindex IIDenvlotra1
.ecindex IIDenvlotra2
.ecindex IIDenvlotra3







. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Generic options for transports" "CHAPtransportgeneric"
.scindex IIDgenoptra1 "generic options" "transport"
.scindex IIDgenoptra2 "options" "generic; for transports"
.scindex IIDgenoptra3 "transport" "generic options for"
The name of a transport is limited to be &drivernamemax; ASCII characters long;
prior to Exim 4.95 names would be silently truncated at this length, but now
it is enforced.

The following generic options apply to all transports:


.option body_only transports boolean false
.cindex "transport" "body only"
.cindex "message" "transporting body only"
.cindex "body of message" "transporting"
If this option is set, the message's headers are not transported. It is
mutually exclusive with &%headers_only%&. If it is used with the &(appendfile)&
or &(pipe)& transports, the settings of &%message_prefix%& and
&%message_suffix%& should be checked, because this option does not
automatically suppress them.


.option current_directory transports string&!! unset
.cindex "transport" "current directory for"
This specifies the current directory that is to be set while running the
transport, overriding any value that may have been set by the router.
If the expansion fails for any reason, including forced failure, an error is
logged, and delivery is deferred.


.option disable_logging transports boolean false
If this option is set true, nothing is logged for any
deliveries by the transport or for any
transport errors. You should not set this option unless you really, really know
what you are doing.


.option debug_print transports string&!! unset
.cindex "testing" "variables in drivers"
If this option is set and debugging is enabled (see the &%-d%& command line
option), the string is expanded and included in the debugging output when the
transport is run.
If expansion of the string fails, the error message is written to the debugging
output, and Exim carries on processing.
This facility is provided to help with checking out the values of variables and
so on when debugging driver configurations. For example, if a &%headers_add%&
option is not working properly, &%debug_print%& could be used to output the
variables it references. A newline is added to the text if it does not end with
one.
The variables &$transport_name$& and &$router_name$& contain the name of the
transport and the router that called it.

.option delivery_date_add transports boolean false
.cindex "&'Delivery-date:'& header line"
If this option is true, a &'Delivery-date:'& header is added to the message.
This gives the actual time the delivery was made. As this is not a standard
header, Exim has a configuration option (&%delivery_date_remove%&) which
requests its removal from incoming messages, so that delivered messages can
safely be resent to other recipients.


.option driver transports string unset
This specifies which of the available transport drivers is to be used.
There is no default, and this option must be set for every transport.


.option envelope_to_add transports boolean false
.cindex "&'Envelope-to:'& header line"
If this option is true, an &'Envelope-to:'& header is added to the message.
This gives the original address(es) in the incoming envelope that caused this
delivery to happen. More than one address may be present if the transport is
configured to handle several addresses at once, or if more than one original
address was redirected to the same final address. As this is not a standard
header, Exim has a configuration option (&%envelope_to_remove%&) which requests
its removal from incoming messages, so that delivered messages can safely be
resent to other recipients.

&*Note:*& If used on a transport handling multiple recipients
(the smtp transport unless &%max_rcpt%& is 1, the appendfile, pipe or lmtp
transport if &%batch_max%& is greater than 1)
then information about Bcc recipients will be leaked.
Doing so is generally not advised.


.option event_action transports string&!! unset
.cindex events
This option declares a string to be expanded for Exim's events mechanism.
For details see chapter &<<CHAPevents>>&.


.option group transports string&!! "Exim group"
.cindex "transport" "group; specifying"
This option specifies a gid for running the transport process, overriding any
value that the router supplies, and also overriding any value associated with
&%user%& (see below).


.option headers_add transports list&!! unset
.cindex "header lines" "adding in transport"
.cindex "transport" "header lines; adding"
This option specifies a list of text headers,
newline-separated (by default, changeable in the usual way &<<SECTlistsepchange>>&),
which are (separately) expanded and added to the header
portion of a message as it is transported, as described in section
&<<SECTheadersaddrem>>&. Additional header lines can also be specified by
routers. If the result of the expansion is an empty string, or if the expansion
is forced to fail, no action is taken. Other expansion failures are treated as
errors and cause the delivery to be deferred.

Unlike most options, &%headers_add%& can be specified multiple times
for a transport; all listed headers are added.


.option headers_only transports boolean false
.cindex "transport" "header lines only"
.cindex "message" "transporting headers only"
.cindex "header lines" "transporting"
If this option is set, the message's body is not transported. It is mutually
exclusive with &%body_only%&. If it is used with the &(appendfile)& or &(pipe)&
transports, the settings of &%message_prefix%& and &%message_suffix%& should be
checked, since this option does not automatically suppress them.


.option headers_remove transports list&!! unset
.cindex "header lines" "removing"
.cindex "transport" "header lines; removing"
This option specifies a list of text headers,
colon-separated (by default, changeable in the usual way &<<SECTlistsepchange>>&),
to be removed from the message.
However, the option has no effect when an address is just being verified.
Each list item is separately expanded.
If the result of the expansion is an empty string, or if the expansion
is forced to fail, no action is taken. Other expansion failures are treated as
errors and cause the delivery to be deferred.
If an item ends in *, it will match any header with the given prefix.

Matching headers are omitted from the message as it is transported, as described
in section &<<SECTheadersaddrem>>&. Header removal can also be specified by
routers.

Unlike most options, &%headers_remove%& can be specified multiple times
for a transport; all listed headers are removed.

&*Warning*&: Because of the separate expansion of the list items,
items that contain a list separator must have it doubled.
To avoid this, change the list separator (&<<SECTlistsepchange>>&).



.option headers_rewrite transports string unset
.cindex "transport" "header lines; rewriting"
.cindex "rewriting" "at transport time"
This option allows addresses in header lines to be rewritten at transport time,
that is, as the message is being copied to its destination. The contents of the
option are a colon-separated list of rewriting rules. Each rule is in exactly
the same form as one of the general rewriting rules that are applied when a
message is received. These are described in chapter &<<CHAPrewrite>>&. For
example,
.code
headers_rewrite = a@b c@d f : \
                  x@y w@z
.endd
changes &'a@b'& into &'c@d'& in &'From:'& header lines, and &'x@y'& into
&'w@z'& in all address-bearing header lines. The rules are applied to the
header lines just before they are written out at transport time, so they affect
only those copies of the message that pass through the transport. However, only
the message's original header lines, and any that were added by a system
filter, are rewritten. If a router or transport adds header lines, they are not
affected by this option. These rewriting rules are &'not'& applied to the
envelope. You can change the return path using &%return_path%&, but you cannot
change envelope recipients at this time.


.option home_directory transports string&!! unset
.cindex "transport" "home directory for"
.vindex "&$home$&"
This option specifies a home directory setting for a local transport,
overriding any value that may be set by the router. The home directory is
placed in &$home$& while expanding the transport's private options. It is also
used as the current directory if no current directory is set by the
&%current_directory%& option on the transport or the
&%transport_current_directory%& option on the router. If the expansion fails
for any reason, including forced failure, an error is logged, and delivery is
deferred.


.option initgroups transports boolean false
.cindex "additional groups"
.cindex "groups" "additional"
.cindex "transport" "group; additional"
If this option is true and the uid for the delivery process is provided by the
transport, the &[initgroups()]& function is called when running the transport
to ensure that any additional groups associated with the uid are set up.


.option max_parallel transports integer&!! unset
.cindex limit "transport parallelism"
.cindex transport "parallel processes"
.cindex transport "concurrency limit"
.cindex "delivery" "parallelism for transport"
If this option is set and expands to an integer greater than zero
it limits the number of concurrent runs of the transport.
The control does not apply to shadow transports.

.cindex "hints database" "transport concurrency control"
Exim implements this control by means of a hints database in which a record is
incremented whenever a transport process is being created. The record
is decremented and possibly removed when the process terminates.
Obviously there is scope for
records to get left lying around if there is a system or program crash. To
guard against this, Exim ignores any records that are more than six hours old.

If you use this option, you should also arrange to delete the
relevant hints database whenever your system reboots. The names of the files
start with &_misc_& and they are kept in the &_spool/db_& directory. There
may be one or two files, depending on the type of DBM in use. The same files
are used for ETRN and smtp transport serialization.


.option message_size_limit transports string&!! 0
.cindex "limit" "message size per transport"
.cindex "size" "of message, limit"
.cindex "transport" "message size; limiting"
This option controls the size of messages passed through the transport. It is
expanded before use; the result of the expansion must be a sequence of decimal
digits, optionally followed by K or M. If the expansion fails for any reason,
including forced failure, or if the result is not of the required form,
delivery is deferred. If the value is greater than zero and the size of a
message exceeds this limit, the address is failed. If there is any chance that
the resulting bounce message could be routed to the same transport, you should
ensure that &%return_size_limit%& is less than the transport's
&%message_size_limit%&, as otherwise the bounce message will fail to get
delivered.



.option rcpt_include_affixes transports boolean false
.cindex "prefix" "for local part, including in envelope"
.cindex "suffix for local part" "including in envelope"
.cindex "local part" "prefix"
.cindex "local part" "suffix"
When this option is false (the default), and an address that has had any
affixes (prefixes or suffixes) removed from the local part is delivered by any
form of SMTP or LMTP, the affixes are not included. For example, if a router
that contains
.code
local_part_prefix = *-
.endd
routes the address &'abc-xyz@some.domain'& to an SMTP transport, the envelope
is delivered with
.code
RCPT TO:<xyz@some.domain>
.endd
This is also the case when an ACL-time callout is being used to verify a
recipient address. However, if &%rcpt_include_affixes%& is set true, the
whole local part is included in the RCPT command. This option applies to BSMTP
deliveries by the &(appendfile)& and &(pipe)& transports as well as to the
&(lmtp)& and &(smtp)& transports.


.option retry_use_local_part transports boolean "see below"
.cindex "hints database" "retry keys"
When a delivery suffers a temporary failure, a retry record is created
in Exim's hints database. For remote deliveries, the key for the retry record
is based on the name and/or IP address of the failing remote host. For local
deliveries, the key is normally the entire address, including both the local
part and the domain. This is suitable for most common cases of local delivery
temporary failure &-- for example, exceeding a mailbox quota should delay only
deliveries to that mailbox, not to the whole domain.

However, in some special cases you may want to treat a temporary local delivery
as a failure associated with the domain, and not with a particular local part.
(For example, if you are storing all mail for some domain in files.) You can do
this by setting &%retry_use_local_part%& false.

For all the local transports, its default value is true. For remote transports,
the default value is false for tidiness, but changing the value has no effect
on a remote transport in the current implementation.


.option return_path transports string&!! unset
.cindex "envelope sender"
.cindex "envelope from"
.cindex "transport" "return path; changing"
.cindex "return path" "changing in transport"
If this option is set, the string is expanded at transport time and replaces
the existing return path (envelope sender) value in the copy of the message
that is being delivered. An empty return path is permitted. This feature is
designed for remote deliveries, where the value of this option is used in the
SMTP MAIL command. If you set &%return_path%& for a local transport, the
only effect is to change the address that is placed in the &'Return-path:'&
header line, if one is added to the message (see the next option).

&*Note:*& A changed return path is not logged unless you add
&%return_path_on_delivery%& to the log selector.

.vindex "&$return_path$&"
The expansion can refer to the existing value via &$return_path$&. This is
either the message's envelope sender, or an address set by the
&%errors_to%& option on a router. If the expansion is forced to fail, no
replacement occurs; if it fails for another reason, delivery is deferred. This
option can be used to support VERP (Variable Envelope Return Paths) &-- see
section &<<SECTverp>>&.

&*Note*&: If a delivery error is detected locally, including the case when a
remote server rejects a message at SMTP time, the bounce message is not sent to
the value of this option. It is sent to the previously set errors address.
This defaults to the incoming sender address, but can be changed by setting
&%errors_to%& in a router.



.option return_path_add transports boolean false
.chindex Return-path:
If this option is true, a &'Return-path:'& header is added to the message.
Although the return path is normally available in the prefix line of BSD
mailboxes, this is commonly not displayed by MUAs, and so the user does not
have easy access to it.

&url(https://www.rfc-editor.org/rfc/rfc2821,RFC 2821)
states that the &'Return-path:'& header is added to a message &"when
the delivery SMTP server makes the final delivery"&. This implies that this
header should not be present in incoming messages. Exim has a configuration
option, &%return_path_remove%&, which requests removal of this header from
incoming messages, so that delivered messages can safely be resent to other
recipients.


.option shadow_condition transports string&!! unset
See &%shadow_transport%& below.


.option shadow_transport transports string unset
.cindex "shadow transport"
.cindex "transport" "shadow"
A local transport may set the &%shadow_transport%& option to the name of
another local transport. Shadow remote transports are not supported.

Whenever a delivery to the main transport succeeds, and either
&%shadow_condition%& is unset, or its expansion does not result in the empty
string or one of the strings &"0"& or &"no"& or &"false"&, the message is also
passed to the shadow transport, with the same delivery address or addresses. If
expansion fails, no action is taken except that non-forced expansion failures
cause a log line to be written.

The result of the shadow transport is discarded and does not affect the
subsequent processing of the message. Only a single level of shadowing is
provided; the &%shadow_transport%& option is ignored on any transport when it
is running as a shadow. Options concerned with output from pipes are also
ignored. The log line for the successful delivery has an item added on the end,
of the form
.code
ST=<shadow transport name>
.endd
If the shadow transport did not succeed, the error message is put in
parentheses afterwards. Shadow transports can be used for a number of different
purposes, including keeping more detailed log information than Exim normally
provides, and implementing automatic acknowledgment policies based on message
headers that some sites insist on.


.option transport_filter transports string&!! unset
.cindex "transport" "filter"
.cindex "filter" "transport filter"
This option sets up a filtering (in the Unix shell sense) process for messages
at transport time. It should not be confused with mail filtering as set up by
individual users or via a system filter.
If unset, or expanding to an empty string, no filtering is done.

When the message is about to be written out, the command specified by
&%transport_filter%& is started up in a separate, parallel process, and
the entire message, including the header lines, is passed to it on its standard
input (this in fact is done from a third process, to avoid deadlock).
The command must be specified as an absolute path.

The process run by the command must use its standard input as the message
data to be transformed, and write the results on its standard output.

The lines of the message that are written to the transport filter are
terminated by newline (&"\n"&). The message is passed to the filter before any
SMTP-specific processing, such as turning &"\n"& into &"\r\n"& and escaping
lines beginning with a dot, and also before any processing implied by the
settings of &%check_string%& and &%escape_string%& in the &(appendfile)& or
&(pipe)& transports.

The standard error for the filter process is set to the same destination as its
standard output; this is read and written to the message's ultimate
destination. The process that writes the message to the filter, the
filter itself, and the original process that reads the result and delivers it
are all run in parallel, like a shell pipeline.

The filter can perform any transformations it likes, but of course should take
care not to break
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
syntax. Exim does not check the result, except to
test for a final newline when SMTP is in use. All messages transmitted over
SMTP must end with a newline, so Exim supplies one if it is missing.

.cindex "content scanning" "per user"
A transport filter can be used to provide content-scanning on a per-user basis
at delivery time if the only required effect of the scan is to modify the
message. For example, a content scan could insert a new header line containing
a spam score. This could be interpreted by a filter in the user's MUA. It is
not possible to discard a message at this stage.

.cindex "SIZE" "ESMTP extension"
A problem might arise if the filter increases the size of a message that is
being sent down an SMTP connection. If the receiving SMTP server has indicated
support for the SIZE parameter, Exim will have sent the size of the message
at the start of the SMTP session. If what is actually sent is substantially
more, the server might reject the message. This can be worked round by setting
the &%size_addition%& option on the &(smtp)& transport, either to allow for
additions to the message, or to disable the use of SIZE altogether.

.vindex "&$pipe_addresses$&"
The value of the &%transport_filter%& option is the command string for starting
the filter, which is run directly from Exim, not under a shell. The string is
parsed by Exim in the same way as a command string for the &(pipe)& transport:
Exim breaks it up into arguments and then expands each argument separately (see
section &<<SECThowcommandrun>>&). Any kind of expansion failure causes delivery
to be deferred. The special argument &$pipe_addresses$& is replaced by a number
of arguments, one for each address that applies to this delivery. (This isn't
an ideal name for this feature here, but as it was already implemented for the
&(pipe)& transport, it seemed sensible not to change it.)

.vindex "&$host$&"
.vindex "&$host_address$&"
The expansion variables &$host$& and &$host_address$& are available when the
transport is a remote one. They contain the name and IP address of the host to
which the message is being sent. For example:
. used to have $sender_address in this cmdline, but it's tainted
.code
transport_filter = /some/directory/transport-filter.pl \
  $host $host_address $pipe_addresses
.endd

Two problems arise if you want to use more complicated expansion items to
generate transport filter commands, both of which due to the fact that the
command is split up &'before'& expansion.
.ilist
If an expansion item contains white space, you must quote it, so that it is all
part of the same command item. If the entire option setting is one such
expansion item, you have to take care what kind of quoting you use. For
example:
.code
transport_filter = '/bin/cmd${if eq{$host}{a.b.c}{1}{2}}'
.endd
This runs the command &(/bin/cmd1)& if the host name is &'a.b.c'&, and
&(/bin/cmd2)& otherwise.

Option strings in general have any fully-surrounding double quote wrapping
removed early in parsing (see &<<SECTstrings>>&).
Then, for this option, quotes protect against whitespace being
regarded as a separator while splitting into the command argument vector.
Either double or single quotes can be used here;
the former interprets backlash-quoted charachters
and the latter does not.

If double quotes had been used in this example, they would have been
stripped by Exim when it read the option's value. When the value is used, if
the single quotes were missing, the line would be split into two items,
&`/bin/cmd${if`& and &`eq{$host}{a.b.c}{1}{2}`&, and an error would occur when
Exim tried to expand the first one.
.next
Except for the special case of &$pipe_addresses$& that is mentioned above, an
expansion cannot generate multiple arguments, or a command name followed by
arguments. Consider this example:
.code
transport_filter = ${lookup{$host}lsearch{/a/file}\
                    {$value}{/bin/cat}}
.endd
The result of the lookup is interpreted as the name of the command, even
if it contains white space. The simplest way round this is to use a shell:
.code
transport_filter = /bin/sh -c ${lookup{$host}lsearch{/a/file}\
                               {$value}{/bin/cat}}
.endd
.endlist

The filter process is run under the same uid and gid as the normal delivery.
For remote deliveries this is the Exim uid/gid by default. The command should
normally yield a zero return code. Transport filters are not supposed to fail.
A non-zero code is taken to mean that the transport filter encountered some
serious problem. Delivery of the message is deferred; the message remains on
the queue and is tried again later. It is not possible to cause a message to be
bounced from a transport filter.

If a transport filter is set on an autoreply transport, the original message is
passed through the filter as it is being copied into the newly generated
message, which happens if the &%return_message%& option is set.


.option transport_filter_timeout transports time 5m
.cindex "transport" "filter, timeout"
When Exim is reading the output of a transport filter, it applies a timeout
that can be set by this option. Exceeding the timeout is normally treated as a
temporary delivery failure. However, if a transport filter is used with a
&(pipe)& transport, a timeout in the transport filter is treated in the same
way as a timeout in the pipe command itself. By default, a timeout is a hard
error, but if the &(pipe)& transport's &%timeout_defer%& option is set true, it
becomes a temporary error.


.option user transports string&!! "Exim user"
.cindex "uid (user id)" "local delivery"
.cindex "transport" "user, specifying"
This option specifies the user under whose uid the delivery process is to be
run, overriding any uid that may have been set by the router. If the user is
given as a name, the uid is looked up from the password data, and the
associated group is taken as the value of the gid to be used if the &%group%&
option is not set.

For deliveries that use local transports, a user and group are normally
specified explicitly or implicitly (for example, as a result of
&%check_local_user%&) by the router or transport.

.cindex "hints database" "access by remote transport"
For remote transports, you should leave this option unset unless you really are
sure you know what you are doing. When a remote transport is running, it needs
to be able to access Exim's hints databases, because each host may have its own
retry data.
.ecindex IIDgenoptra1
.ecindex IIDgenoptra2
.ecindex IIDgenoptra3






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Address batching in local transports" "CHAPbatching" &&&
         "Address batching"
.cindex "transport" "local; address batching in"
The only remote transport (&(smtp)&) is normally configured to handle more than
one address at a time, so that when several addresses are routed to the same
remote host, just one copy of the message is sent. Local transports, however,
normally handle one address at a time. That is, a separate instance of the
transport is run for each address that is routed to the transport. A separate
copy of the message is delivered each time.

.cindex "batched local delivery"
.oindex "&%batch_max%&"
.oindex "&%batch_id%&"
In special cases, it may be desirable to handle several addresses at once in a
local transport, for example:

.ilist
In an &(appendfile)& transport, when storing messages in files for later
delivery by some other means, a single copy of the message with multiple
recipients saves space.
.next
In an &(lmtp)& transport, when delivering over &"local SMTP"& to some process,
a single copy saves time, and is the normal way LMTP is expected to work.
.next
In a &(pipe)& transport, when passing the message
to a scanner program or
to some other delivery mechanism such as UUCP, multiple recipients may be
acceptable.
.endlist

These three local transports all have the same options for controlling multiple
(&"batched"&) deliveries, namely &%batch_max%& and &%batch_id%&. To save
repeating the information for each transport, these options are described here.

The &%batch_max%& option specifies the maximum number of addresses that can be
delivered together in a single run of the transport. Its default value is one
(no batching). When more than one address is routed to a transport that has a
&%batch_max%& value greater than one, the addresses are delivered in a batch
(that is, in a single run of the transport with multiple recipients), subject
to certain conditions:

.ilist
.vindex "&$local_part$&"
If any of the transport's options contain a reference to &$local_part$&, no
batching is possible.
.next
.vindex "&$domain$&"
If any of the transport's options contain a reference to &$domain$&, only
addresses with the same domain are batched.
.next
.cindex "customizing" "batching condition"
If &%batch_id%& is set, it is expanded for each address, and only those
addresses with the same expanded value are batched. This allows you to specify
customized batching conditions. Failure of the expansion for any reason,
including forced failure, disables batching, but it does not stop the delivery
from taking place.
.next
Batched addresses must also have the same errors address (where to send
delivery errors), the same header additions and removals, the same user and
group for the transport, and if a host list is present, the first host must
be the same.
.endlist

In the case of the &(appendfile)& and &(pipe)& transports, batching applies
both when the file or pipe command is specified in the transport, and when it
is specified by a &(redirect)& router, but all the batched addresses must of
course be routed to the same file or pipe command. These two transports have an
option called &%use_bsmtp%&, which causes them to deliver the message in
&"batched SMTP"& format, with the envelope represented as SMTP commands. The
&%check_string%& and &%escape_string%& options are forced to the values
.code
check_string = "."
escape_string = ".."
.endd
when batched SMTP is in use. A full description of the batch SMTP mechanism is
given in section &<<SECTbatchSMTP>>&. The &(lmtp)& transport does not have a
&%use_bsmtp%& option, because it always delivers using the SMTP protocol.

.cindex "&'Envelope-to:'& header line"
If the generic &%envelope_to_add%& option is set for a batching transport, the
&'Envelope-to:'& header that is added to the message contains all the addresses
that are being processed together. If you are using a batching &(appendfile)&
transport without &%use_bsmtp%&, the only way to preserve the recipient
addresses is to set the &%envelope_to_add%& option.

.cindex "&(pipe)& transport" "with multiple addresses"
.vindex "&$pipe_addresses$&"
If you are using a &(pipe)& transport without BSMTP, and setting the
transport's &%command%& option, you can include &$pipe_addresses$& as part of
the command. This is not a true variable; it is a bit of magic that causes each
of the recipient addresses to be inserted into the command as a separate
argument. This provides a way of accessing all the addresses that are being
delivered in the batch. &*Note:*& This is not possible for pipe commands that
are specified by a &(redirect)& router.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The appendfile transport" "CHAPappendfile"
.scindex IIDapptra1 "&(appendfile)& transport"
.scindex IIDapptra2 "transports" "&(appendfile)&"
.cindex "directory creation"
.cindex "creating directories"
The &(appendfile)& transport delivers a message by appending it to an existing
file, or by creating an entirely new file in a specified directory. Single
files to which messages are appended can be in the traditional Unix mailbox
format, or optionally in the MBX format supported by the Pine MUA and
University of Washington IMAP daemon, &'inter alia'&. When each message is
being delivered as a separate file, &"maildir"& format can optionally be used
to give added protection against failures that happen part-way through the
delivery. A third form of separate-file delivery known as &"mailstore"& is also
supported. For all file formats, Exim attempts to create as many levels of
directory as necessary, provided that &%create_directory%& is set.

The code for the optional formats is not included in the Exim binary by
default. It is necessary to set SUPPORT_MBX, SUPPORT_MAILDIR and/or
SUPPORT_MAILSTORE in &_Local/Makefile_& to have the appropriate code
included.

.cindex "quota" "system"
Exim recognizes system quota errors, and generates an appropriate message. Exim
also supports its own quota control within the transport, for use when the
system facility is unavailable or cannot be used for some reason.

If there is an error while appending to a file (for example, quota exceeded or
partition filled), Exim attempts to reset the file's length and last
modification time back to what they were before. If there is an error while
creating an entirely new file, the new file is removed.

Before appending to a file, a number of security checks are made, and the
file is locked. A detailed description is given below, after the list of
private options.

The &(appendfile)& transport is most commonly used for local deliveries to
users' mailboxes. However, it can also be used as a pseudo-remote transport for
putting messages into files for remote delivery by some means other than Exim.
&"Batch SMTP"& format is often used in this case (see the &%use_bsmtp%&
option).



.section "The file and directory options" "SECTfildiropt"
The &%file%& option specifies a single file, to which the message is appended;
the &%directory%& option specifies a directory, in which a new file containing
the message is created. Only one of these two options can be set, and for
normal deliveries to mailboxes, one of them &'must'& be set.

.vindex "&$address_file$&"
.vindex "&$local_part$&"
However, &(appendfile)& is also used for delivering messages to files or
directories whose names (or parts of names) are obtained from alias,
forwarding, or filtering operations (for example, a &%save%& command in a
user's Exim filter). When such a transport is running, &$local_part$& contains
the local part that was aliased or forwarded, and &$address_file$& contains the
name (or partial name) of the file or directory generated by the redirection
operation. There are two cases:

.ilist
If neither &%file%& nor &%directory%& is set, the redirection operation
must specify an absolute path (one that begins with &`/`&). This is the most
common case when users with local accounts use filtering to sort mail into
different folders. See for example, the &(address_file)& transport in the
default configuration. If the path ends with a slash, it is assumed to be the
name of a directory. A delivery to a directory can also be forced by setting
&%maildir_format%& or &%mailstore_format%&.
.next
If &%file%& or &%directory%& is set for a delivery from a redirection, it is
used to determine the file or directory name for the delivery. Normally, the
contents of &$address_file$& are used in some way in the string expansion.
.endlist
If the &%create_file%& option is set to a path which
matches (see the option definition below for details)
a file or directory name
for the delivery, that name becomes de-tainted.

.cindex "tainted data" "in filenames"
.cindex appendfile "tainted data"
Tainted data may not be used for a file or directory name.
This means that, for instance, &$local_part$& cannot be used directly
as a component of a path.  It can however be used as the key for a lookup
which returns a path (or component).


.cindex "Sieve filter" "configuring &(appendfile)&"
.cindex "Sieve filter" "relative mailbox path handling"
As an example of the second case, consider an environment where users do not
have home directories. They may be permitted to use Exim filter commands of the
form:
.code
save folder23
.endd
or Sieve filter commands of the form:
.code
require "fileinto";
fileinto "folder23";
.endd
In this situation, the expansion of &%file%& or &%directory%& in the transport
must transform the relative path into an appropriate absolute filename. In the
case of Sieve filters, the name &'inbox'& must be handled. It is the
.new
default
.wen
name that
is used as a result of a &"keep"& action in the filter. This example shows one
way of handling this requirement:
.code
file = ${if eq{$address_file}{inbox} \
            {/var/mail/$local_part_data} \
            {${if eq{${substr_0_1:$address_file}}{/} \
                  {$address_file} \
                  {$home/mail/$address_file} \
            }} \
       }
.endd
With this setting of &%file%&, &'inbox'& refers to the standard mailbox
location, absolute paths are used without change, and other folders are in the
&_mail_& directory within the home directory.

.new
An alternative for the &"keep"& aspect is to use the &%sieve_inbox%& option
on the redirect router that calls the Sieve filter,
to explicitly set the filename used.
.wen

&*Note 1*&: While processing an Exim filter, a relative path such as
&_folder23_& is turned into an absolute path if a home directory is known to
the router. In particular, this is the case if &%check_local_user%& is set. If
you want to prevent this happening at routing time, you can set
&%router_home_directory%& empty. This forces the router to pass the relative
path to the transport.

&*Note 2*&: An absolute path in &$address_file$& is not treated specially;
the &%file%& or &%directory%& option is still used if it is set.

.new
&*Note 3*&: Permitting a user to enable writes to an absolute path
may be a security issue.
.wen



.section "Private options for appendfile" "SECID134"
.cindex "options" "&(appendfile)& transport"



.option allow_fifo appendfile boolean false
.cindex "fifo (named pipe)"
.cindex "named pipe (fifo)"
.cindex "pipe" "named (fifo)"
Setting this option permits delivery to named pipes (FIFOs) as well as to
regular files. If no process is reading the named pipe at delivery time, the
delivery is deferred.


.option allow_symlink appendfile boolean false
.cindex "symbolic link" "to mailbox"
.cindex "mailbox" "symbolic link"
By default, &(appendfile)& will not deliver if the path name for the file is
that of a symbolic link. Setting this option relaxes that constraint, but there
are security issues involved in the use of symbolic links. Be sure you know
what you are doing if you set this. Details of exactly what this option affects
are included in the discussion which follows this list of options.


.option batch_id appendfile string&!! unset
See the description of local delivery batching in chapter &<<CHAPbatching>>&.
However, batching is automatically disabled for &(appendfile)& deliveries that
happen as a result of forwarding or aliasing or other redirection directly to a
file.


.option batch_max appendfile integer 1
See the description of local delivery batching in chapter &<<CHAPbatching>>&.


.option check_group appendfile boolean false
When this option is set, the group owner of the file defined by the &%file%&
option is checked to see that it is the same as the group under which the
delivery process is running. The default setting is false because the default
file mode is 0600, which means that the group is irrelevant.


.option check_owner appendfile boolean true
When this option is set, the owner of the file defined by the &%file%& option
is checked to ensure that it is the same as the user under which the delivery
process is running.


.option check_string appendfile string "see below"
.cindex "&""From""& line"
As &(appendfile)& writes the message, the start of each line is tested for
matching &%check_string%&, and if it does, the initial matching characters are
replaced by the contents of &%escape_string%&. The value of &%check_string%& is
a literal string, not a regular expression, and the case of any letters it
contains is significant.

If &%use_bsmtp%& is set the values of &%check_string%& and &%escape_string%&
are forced to &"."& and &".."& respectively, and any settings in the
configuration are ignored. Otherwise, they default to &"From&~"& and
&">From&~"& when the &%file%& option is set, and unset when any of the
&%directory%&, &%maildir%&, or &%mailstore%& options are set.

The default settings, along with &%message_prefix%& and &%message_suffix%&, are
suitable for traditional &"BSD"& mailboxes, where a line beginning with
&"From&~"& indicates the start of a new message. All four options need changing
if another format is used. For example, to deliver to mailboxes in MMDF format:
.cindex "MMDF format mailbox"
.cindex "mailbox" "MMDF format"
.code
check_string = "\1\1\1\1\n"
escape_string = "\1\1\1\1 \n"
message_prefix = "\1\1\1\1\n"
message_suffix = "\1\1\1\1\n"
.endd
.option create_directory appendfile boolean true
.cindex "directory creation"
When this option is true, Exim attempts to create any missing superior
directories for the file that it is about to write. A created directory's mode
is given by the &%directory_mode%& option.

The group ownership of a newly created directory is highly dependent on the
operating system (and possibly the file system) that is being used. For
example, in Solaris, if the parent directory has the setgid bit set, its group
is propagated to the child; if not, the currently set group is used. However,
in FreeBSD, the parent's group is always used.



.option create_file appendfile string anywhere
This option constrains the location of files and directories that are created
by this transport. It applies to files defined by the &%file%& option and
directories defined by the &%directory%& option. In the case of maildir
delivery, it applies to the top level directory, not the maildir directories
beneath.

The option must be set to one of the words &"anywhere"&, &"inhome"&, or
&"belowhome"&, or to an absolute path.

In the second and third cases, a home directory must have been
set for the transport, and the file or directory being created must
reside within it.
The "belowhome" checking additionally checks for attempts to use "../"
to evade the testing.
This option is not useful when an explicit filename is
given for normal mailbox deliveries. It is intended for the case when filenames
are generated from users' &_.forward_& files. These are usually handled
by an &(appendfile)& transport called &%address_file%&. See also
&%file_must_exist%&.

In the fourth case,
the value given for this option must be an absolute path for an
existing directory.
The value is used for checking instead of a home directory;
checking is done in "belowhome" mode.

.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using appendfile create_file option"
If "belowhome" checking is used, the file or directory path
becomes de-tainted.


.option directory appendfile string&!! unset
This option is mutually exclusive with the &%file%& option, but one of &%file%&
or &%directory%& must be set, unless the delivery is the direct result of a
redirection (see section &<<SECTfildiropt>>&).

When &%directory%& is set, the string is expanded, and the message is delivered
into a new file or files in or below the given directory, instead of being
appended to a single mailbox file. A number of different formats are provided
(see &%maildir_format%& and &%mailstore_format%&), and see section
&<<SECTopdir>>& for further details of this form of delivery.

The result of expansion must not be tainted, unless the &%create_file%& option
specifies a path.


.option directory_file appendfile string&!! "see below"
.cindex "base62"
.vindex "&$inode$&"
When &%directory%& is set, but neither &%maildir_format%& nor
&%mailstore_format%& is set, &(appendfile)& delivers each message into a file
whose name is obtained by expanding this string. The default value is:
.code
q${base62:$tod_epoch}-$inode
.endd
This generates a unique name from the current time, in base 62 form, and the
inode of the file. The variable &$inode$& is available only when expanding this
option.


.option directory_mode appendfile "octal integer" 0700
If &(appendfile)& creates any directories as a result of the
&%create_directory%& option, their mode is specified by this option.


.option escape_string appendfile string "see description"
See &%check_string%& above.


.option file appendfile string&!! unset
This option is mutually exclusive with the &%directory%& option, but one of
&%file%& or &%directory%& must be set, unless the delivery is the direct result
of a redirection (see section &<<SECTfildiropt>>&). The &%file%& option
specifies a single file, to which the message is appended. One or more of
&%use_fcntl_lock%&, &%use_flock_lock%&, or &%use_lockfile%& must be set with
&%file%&.

The result of expansion must not be tainted, unless the &%create_file%& option
specifies a path.

.cindex "NFS" "lock file"
.cindex "locking files"
.cindex "lock files"
If you are using more than one host to deliver over NFS into the same
mailboxes, you should always use lock files.

The string value is expanded for each delivery, and must yield an absolute
path. The most common settings of this option are variations on one of these
examples:
.code
file = /var/spool/mail/$local_part_data
file = /home/$local_part_data/inbox
file = $home/inbox
.endd
.cindex "&""sticky""& bit"
In the first example, all deliveries are done into the same directory. If Exim
is configured to use lock files (see &%use_lockfile%& below) it must be able to
create a file in the directory, so the &"sticky"& bit must be turned on for
deliveries to be possible, or alternatively the &%group%& option can be used to
run the delivery under a group id which has write access to the directory.



.option file_format appendfile string unset
.cindex "file" "mailbox; checking existing format"
This option requests the transport to check the format of an existing file
before adding to it. The check consists of matching a specific string at the
start of the file. The value of the option consists of an even number of
colon-separated strings. The first of each pair is the test string, and the
second is the name of a transport. If the transport associated with a matched
string is not the current transport, control is passed over to the other
transport. For example, suppose the standard &(local_delivery)& transport has
this added to it:
.code
file_format = "From       : local_delivery :\
               \1\1\1\1\n : local_mmdf_delivery"
.endd
Mailboxes that begin with &"From"& are still handled by this transport, but if
a mailbox begins with four binary ones followed by a newline, control is passed
to a transport called &%local_mmdf_delivery%&, which presumably is configured
to do the delivery in MMDF format. If a mailbox does not exist or is empty, it
is assumed to match the current transport. If the start of a mailbox doesn't
match any string, or if the transport named for a given string is not defined,
delivery is deferred.


.option file_must_exist appendfile boolean false
If this option is true, the file specified by the &%file%& option must exist.
A temporary error occurs if it does not, causing delivery to be deferred.
If this option is false, the file is created if it does not exist.


.option lock_fcntl_timeout appendfile time 0s
.cindex "timeout" "mailbox locking"
.cindex "mailbox" "locking, blocking and non-blocking"
.cindex "locking files"
By default, the &(appendfile)& transport uses non-blocking calls to &[fcntl()]&
when locking an open mailbox file. If the call fails, the delivery process
sleeps for &%lock_interval%& and tries again, up to &%lock_retries%& times.
Non-blocking calls are used so that the file is not kept open during the wait
for the lock; the reason for this is to make it as safe as possible for
deliveries over NFS in the case when processes might be accessing an NFS
mailbox without using a lock file. This should not be done, but
misunderstandings and hence misconfigurations are not unknown.

On a busy system, however, the performance of a non-blocking lock approach is
not as good as using a blocking lock with a timeout. In this case, the waiting
is done inside the system call, and Exim's delivery process acquires the lock
and can proceed as soon as the previous lock holder releases it.

If &%lock_fcntl_timeout%& is set to a non-zero time, blocking locks, with that
timeout, are used. There may still be some retrying: the maximum number of
retries is
.code
(lock_retries * lock_interval) / lock_fcntl_timeout
.endd
rounded up to the next whole number. In other words, the total time during
which &(appendfile)& is trying to get a lock is roughly the same, unless
&%lock_fcntl_timeout%& is set very large.

You should consider setting this option if you are getting a lot of delayed
local deliveries because of errors of the form
.code
failed to lock mailbox /some/file (fcntl)
.endd

.option lock_flock_timeout appendfile time 0s
This timeout applies to file locking when using &[flock()]& (see
&%use_flock%&); the timeout operates in a similar manner to
&%lock_fcntl_timeout%&.


.option lock_interval appendfile time 3s
This specifies the time to wait between attempts to lock the file. See below
for details of locking.


.option lock_retries appendfile integer 10
This specifies the maximum number of attempts to lock the file. A value of zero
is treated as 1. See below for details of locking.


.option lockfile_mode appendfile "octal integer" 0600
This specifies the mode of the created lock file, when a lock file is being
used (see &%use_lockfile%& and &%use_mbx_lock%&).


.option lockfile_timeout appendfile time 30m
.cindex "timeout" "mailbox locking"
When a lock file is being used (see &%use_lockfile%&), if a lock file already
exists and is older than this value, it is assumed to have been left behind by
accident, and Exim attempts to remove it.


.option mailbox_filecount appendfile string&!! unset
.cindex "mailbox" "specifying size of"
.cindex "size" "of mailbox"
If this option is set, it is expanded, and the result is taken as the current
number of files in the mailbox. It must be a decimal number, optionally
followed by K or M. This provides a way of obtaining this information from an
external source that maintains the data.


.option mailbox_size appendfile string&!! unset
.cindex "mailbox" "specifying size of"
.cindex "size" "of mailbox"
If this option is set, it is expanded, and the result is taken as the current
size the mailbox. It must be a decimal number, optionally followed by K or M.
This provides a way of obtaining this information from an external source that
maintains the data. This is likely to be helpful for maildir deliveries where
it is computationally expensive to compute the size of a mailbox.



.option maildir_format appendfile boolean false
.cindex "maildir format" "specifying"
If this option is set with the &%directory%& option, the delivery is into a new
file, in the &"maildir"& format that is used by other mail software. When the
transport is activated directly from a &(redirect)& router (for example, the
&(address_file)& transport in the default configuration), setting
&%maildir_format%& causes the path received from the router to be treated as a
directory, whether or not it ends with &`/`&. This option is available only if
SUPPORT_MAILDIR is present in &_Local/Makefile_&. See section
&<<SECTmaildirdelivery>>& below for further details.


.option maildir_quota_directory_regex appendfile string "See below"
.cindex "maildir format" "quota; directories included in"
.cindex "quota" "maildir; directories included in"
This option is relevant only when &%maildir_use_size_file%& is set. It defines
a regular expression for specifying directories, relative to the quota
directory (see &%quota_directory%&), that should be included in the quota
calculation. The default value is:
.code
maildir_quota_directory_regex = ^(?:cur|new|\..*)$
.endd
This includes the &_cur_& and &_new_& directories, and any maildir++ folders
(directories whose names begin with a dot). If you want to exclude the
&_Trash_&
folder from the count (as some sites do), you need to change this setting to
.code
maildir_quota_directory_regex = ^(?:cur|new|\.(?!Trash).*)$
.endd
This uses a negative lookahead in the regular expression to exclude the
directory whose name is &_.Trash_&. When a directory is excluded from quota
calculations, quota processing is bypassed for any messages that are delivered
directly into that directory.


.option maildir_retries appendfile integer 10
This option specifies the number of times to retry when writing a file in
&"maildir"& format. See section &<<SECTmaildirdelivery>>& below.


.option maildir_tag appendfile string&!! unset
This option applies only to deliveries in maildir format, and is described in
section &<<SECTmaildirdelivery>>& below.


.option maildir_use_size_file appendfile&!! boolean false
.cindex "maildir format" "&_maildirsize_& file"
The result of string expansion for this option must be a valid boolean value.
If it is true, it enables support for &_maildirsize_& files. Exim
creates a &_maildirsize_& file in a maildir if one does not exist, taking the
quota from the &%quota%& option of the transport. If &%quota%& is unset, the
value is zero. See &%maildir_quota_directory_regex%& above and section
&<<SECTmaildirdelivery>>& below for further details.

.option maildirfolder_create_regex appendfile string unset
.cindex "maildir format" "&_maildirfolder_& file"
.cindex "&_maildirfolder_&, creating"
The value of this option is a regular expression. If it is unset, it has no
effect. Otherwise, before a maildir delivery takes place, the pattern is
matched against the name of the maildir directory, that is, the directory
containing the &_new_& and &_tmp_& subdirectories that will be used for the
delivery. If there is a match, Exim checks for the existence of a file called
&_maildirfolder_& in the directory, and creates it if it does not exist.
See section &<<SECTmaildirdelivery>>& for more details.


.option mailstore_format appendfile boolean false
.cindex "mailstore format" "specifying"
If this option is set with the &%directory%& option, the delivery is into two
new files in  &"mailstore"& format. The option is available only if
SUPPORT_MAILSTORE is present in &_Local/Makefile_&. See section &<<SECTopdir>>&
below for further details.


.option mailstore_prefix appendfile string&!! unset
This option applies only to deliveries in mailstore format, and is described in
section &<<SECTopdir>>& below.


.option mailstore_suffix appendfile string&!! unset
This option applies only to deliveries in mailstore format, and is described in
section &<<SECTopdir>>& below.


.option mbx_format appendfile boolean false
.cindex "locking files"
.cindex "file" "locking"
.cindex "file" "MBX format"
.cindex "MBX format, specifying"
This option is available only if Exim has been compiled with SUPPORT_MBX
set in &_Local/Makefile_&. If &%mbx_format%& is set with the &%file%& option,
the message is appended to the mailbox file in MBX format instead of
traditional Unix format. This format is supported by Pine4 and its associated
IMAP and POP daemons, by means of the &'c-client'& library that they all use.

&*Note*&: The &%message_prefix%& and &%message_suffix%& options are not
automatically changed by the use of &%mbx_format%&. They should normally be set
empty when using MBX format, so this option almost always appears in this
combination:
.code
mbx_format = true
message_prefix =
message_suffix =
.endd
If none of the locking options are mentioned in the configuration,
&%use_mbx_lock%& is assumed and the other locking options default to false. It
is possible to specify the other kinds of locking with &%mbx_format%&, but
&%use_fcntl_lock%& and &%use_mbx_lock%& are mutually exclusive. MBX locking
interworks with &'c-client'&, providing for shared access to the mailbox. It
should not be used if any program that does not use this form of locking is
going to access the mailbox, nor should it be used if the mailbox file is NFS
mounted, because it works only when the mailbox is accessed from a single host.

If you set &%use_fcntl_lock%& with an MBX-format mailbox, you cannot use
the standard version of &'c-client'&, because as long as it has a mailbox open
(this means for the whole of a Pine or IMAP session), Exim will not be able to
append messages to it.


.option message_prefix appendfile string&!! "see below"
.cindex "&""From""& line"
The string specified here is expanded and output at the start of every message.
The default is unset unless &%file%& is specified and &%use_bsmtp%& is not set,
in which case it is:
.code
message_prefix = "From ${if def:return_path{$return_path}\
  {MAILER-DAEMON}} $tod_bsdinbox\n"
.endd
&*Note:*& If you set &%use_crlf%& true, you must change any occurrences of
&`\n`& to &`\r\n`& in &%message_prefix%&.

.option message_suffix appendfile string&!! "see below"
The string specified here is expanded and output at the end of every message.
The default is unset unless &%file%& is specified and &%use_bsmtp%& is not set,
in which case it is a single newline character. The suffix can be suppressed by
setting
.code
message_suffix =
.endd
&*Note:*& If you set &%use_crlf%& true, you must change any occurrences of
&`\n`& to &`\r\n`& in &%message_suffix%&.

.option mode appendfile "octal integer" 0600
If the output file is created, it is given this mode. If it already exists and
has wider permissions, they are reduced to this mode. If it has narrower
permissions, an error occurs unless &%mode_fail_narrower%& is false. However,
if the delivery is the result of a &%save%& command in a filter file specifying
a particular mode, the mode of the output file is always forced to take that
value, and this option is ignored.


.option mode_fail_narrower appendfile boolean true
This option applies in the case when an existing mailbox file has a narrower
mode than that specified by the &%mode%& option. If &%mode_fail_narrower%& is
true, the delivery is deferred (&"mailbox has the wrong mode"&); otherwise Exim
continues with the delivery attempt, using the existing mode of the file.


.option notify_comsat appendfile boolean false
If this option is true, the &'comsat'& daemon is notified after every
successful delivery to a user mailbox. This is the daemon that notifies logged
on users about incoming mail.


.option quota appendfile string&!! unset
.cindex "quota" "imposed by Exim"
This option imposes a limit on the size of the file to which Exim is appending,
or to the total space used in the directory tree when the &%directory%& option
is set. In the latter case, computation of the space used is expensive, because
all the files in the directory (and any sub-directories) have to be
individually inspected and their sizes summed. (See &%quota_size_regex%& and
&%maildir_use_size_file%& for ways to avoid this in environments where users
have no shell access to their mailboxes).

As there is no interlock against two simultaneous deliveries into a
multi-file mailbox, it is possible for the quota to be overrun in this case.
For single-file mailboxes, of course, an interlock is a necessity.

A file's size is taken as its &'used'& value. Because of blocking effects, this
may be a lot less than the actual amount of disk space allocated to the file.
If the sizes of a number of files are being added up, the rounding effect can
become quite noticeable, especially on systems that have large block sizes.
Nevertheless, it seems best to stick to the &'used'& figure, because this is
the obvious value which users understand most easily.

The value of the option is expanded, and must then be a numerical value
(decimal point allowed), optionally followed by one of the letters K, M, or G,
for kilobytes, megabytes, or gigabytes, optionally followed by a slash
and further option modifiers. If Exim is running on a system with
large file support (Linux and FreeBSD have this), mailboxes larger than 2G can
be handled.

The option modifier &%no_check%& can be used to force delivery even if the over
quota condition is met. The quota gets updated as usual.

&*Note*&: A value of zero is interpreted as &"no quota"&.

The expansion happens while Exim is running as root, before it changes uid for
the delivery. This means that files that are inaccessible to the end user can
be used to hold quota values that are looked up in the expansion. When delivery
fails because this quota is exceeded, the handling of the error is as for
system quota failures.

By default, Exim's quota checking mimics system quotas, and restricts the
mailbox to the specified maximum size, though the value is not accurate to the
last byte, owing to separator lines and additional headers that may get added
during message delivery. When a mailbox is nearly full, large messages may get
refused even though small ones are accepted, because the size of the current
message is added to the quota when the check is made. This behaviour can be
changed by setting &%quota_is_inclusive%& false. When this is done, the check
for exceeding the quota does not include the current message. Thus, deliveries
continue until the quota has been exceeded; thereafter, no further messages are
delivered. See also &%quota_warn_threshold%&.


.option quota_directory appendfile string&!! unset
This option defines the directory to check for quota purposes when delivering
into individual files. The default is the delivery directory, or, if a file
called &_maildirfolder_& exists in a maildir directory, the parent of the
delivery directory.


.option quota_filecount appendfile string&!! 0
This option applies when the &%directory%& option is set. It limits the total
number of files in the directory (compare the inode limit in system quotas). It
can only be used if &%quota%& is also set. The value is expanded; an expansion
failure causes delivery to be deferred. A value of zero is interpreted as
&"no quota"&.

The option modifier &%no_check%& can be used to force delivery even if the over
quota condition is met. The quota gets updated as usual.

.option quota_is_inclusive appendfile boolean true
See &%quota%& above.


.option quota_size_regex appendfile string unset
This option applies when one of the delivery modes that writes a separate file
for each message is being used. When Exim wants to find the size of one of
these files in order to test the quota, it first checks &%quota_size_regex%&.
If this is set to a regular expression that matches the filename, and it
captures one string, that string is interpreted as a representation of the
file's size. The value of &%quota_size_regex%& is not expanded.

This feature is useful only when users have no shell access to their mailboxes
&-- otherwise they could defeat the quota simply by renaming the files. This
facility can be used with maildir deliveries, by setting &%maildir_tag%& to add
the file length to the filename. For example:
.code
maildir_tag = ,S=$message_size
quota_size_regex = ,S=(\d+)
.endd
An alternative to &$message_size$& is &$message_linecount$&, which contains the
number of lines in the message.

The regular expression should not assume that the length is at the end of the
filename (even though &%maildir_tag%& puts it there) because maildir MUAs
sometimes add other information onto the ends of message filenames.

Section &<<SECID136>>& contains further information.

This option should not be used when other message-handling software
may duplicate messages by making hardlinks to the files.  When that is done Exim
will count the message size once for each filename, in contrast with the actual
disk usage.  When the option is not set, calculating total usage requires
a system-call per file to get the size; the number of links is then available also
as is used to adjust the effective size.


.option quota_warn_message appendfile string&!! "see below"
See below for the use of this option. If it is not set when
&%quota_warn_threshold%& is set, it defaults to
.code
quota_warn_message = "\
  To: $local_part@$domain\n\
  Subject: Your mailbox\n\n\
  This message is automatically created \
  by mail delivery software.\n\n\
  The size of your mailbox has exceeded \
  a warning threshold that is\n\
  set by the system administrator.\n"
.endd


.option quota_warn_threshold appendfile string&!! 0
.cindex "quota" "warning threshold"
.cindex "mailbox" "size warning"
.cindex "size" "of mailbox"
This option is expanded in the same way as &%quota%& (see above). If the
resulting value is greater than zero, and delivery of the message causes the
size of the file or total space in the directory tree to cross the given
threshold, a warning message is sent. If &%quota%& is also set, the threshold
may be specified as a percentage of it by following the value with a percent
sign. For example:
.code
quota = 10M
quota_warn_threshold = 75%
.endd
If &%quota%& is not set, a setting of &%quota_warn_threshold%& that ends with a
percent sign is ignored.

The warning message itself is specified by the &%quota_warn_message%& option,
and it must start with a &'To:'& header line containing the recipient(s) of the
warning message. These do not necessarily have to include the recipient(s) of
the original message. A &'Subject:'& line should also normally be supplied. You
can include any other header lines that you want. If you do not include a
&'From:'& line, the default is:
.code
From: Mail Delivery System <mailer-daemon@$qualify_domain_sender>
.endd
.oindex &%errors_reply_to%&
If you supply a &'Reply-To:'& line, it overrides the global &%errors_reply_to%&
option.

The &%quota%& option does not have to be set in order to use this option; they
are independent of one another except when the threshold is specified as a
percentage.


.option use_bsmtp appendfile boolean false
.cindex "envelope from"
.cindex "envelope sender"
If this option is set true, &(appendfile)& writes messages in &"batch SMTP"&
format, with the envelope sender and recipient(s) included as SMTP commands. If
you want to include a leading HELO command with such messages, you can do
so by setting the &%message_prefix%& option. See section &<<SECTbatchSMTP>>&
for details of batch SMTP.


.option use_crlf appendfile boolean false
.cindex "carriage return"
.cindex "linefeed"
This option causes lines to be terminated with the two-character CRLF sequence
(carriage return, linefeed) instead of just a linefeed character. In the case
of batched SMTP, the byte sequence written to the file is then an exact image
of what would be sent down a real SMTP connection.

&*Note:*& The contents of the &%message_prefix%& and &%message_suffix%& options
(which are used to supply the traditional &"From&~"& and blank line separators
in Berkeley-style mailboxes) are written verbatim, so must contain their own
carriage return characters if these are needed. In cases where these options
have non-empty defaults, the values end with a single linefeed, so they must be
changed to end with &`\r\n`& if &%use_crlf%& is set.


.option use_fcntl_lock appendfile boolean "see below"
This option controls the use of the &[fcntl()]& function to lock a file for
exclusive use when a message is being appended. It is set by default unless
&%use_flock_lock%& is set. Otherwise, it should be turned off only if you know
that all your MUAs use lock file locking. When both &%use_fcntl_lock%& and
&%use_flock_lock%& are unset, &%use_lockfile%& must be set.


.option use_flock_lock appendfile boolean false
This option is provided to support the use of &[flock()]& for file locking, for
the few situations where it is needed. Most modern operating systems support
&[fcntl()]& and &[lockf()]& locking, and these two functions interwork with
each other. Exim uses &[fcntl()]& locking by default.

This option is required only if you are using an operating system where
&[flock()]& is used by programs that access mailboxes (typically MUAs), and
where &[flock()]& does not correctly interwork with &[fcntl()]&. You can use
both &[fcntl()]& and &[flock()]& locking simultaneously if you want.

.cindex "Solaris" "&[flock()]& support"
Not all operating systems provide &[flock()]&. Some versions of Solaris do not
have it (and some, I think, provide a not quite right version built on top of
&[lockf()]&). If the OS does not have &[flock()]&, Exim will be built without
the ability to use it, and any attempt to do so will cause a configuration
error.

&*Warning*&: &[flock()]& locks do not work on NFS files (unless &[flock()]&
is just being mapped onto &[fcntl()]& by the OS).


.option use_lockfile appendfile boolean "see below"
If this option is turned off, Exim does not attempt to create a lock file when
appending to a mailbox file. In this situation, the only locking is by
&[fcntl()]&. You should only turn &%use_lockfile%& off if you are absolutely
sure that every MUA that is ever going to look at your users' mailboxes uses
&[fcntl()]& rather than a lock file, and even then only when you are not
delivering over NFS from more than one host.

.cindex "NFS" "lock file"
In order to append to an NFS file safely from more than one host, it is
necessary to take out a lock &'before'& opening the file, and the lock file
achieves this. Otherwise, even with &[fcntl()]& locking, there is a risk of
file corruption.

The &%use_lockfile%& option is set by default unless &%use_mbx_lock%& is set.
It is not possible to turn both &%use_lockfile%& and &%use_fcntl_lock%& off,
except when &%mbx_format%& is set.


.option use_mbx_lock appendfile boolean "see below"
This option is available only if Exim has been compiled with SUPPORT_MBX
set in &_Local/Makefile_&. Setting the option specifies that special MBX
locking rules be used. It is set by default if &%mbx_format%& is set and none
of the locking options are mentioned in the configuration. The locking rules
are the same as are used by the &'c-client'& library that underlies Pine and
the IMAP4 and POP daemons that come with it (see the discussion below). The
rules allow for shared access to the mailbox. However, this kind of locking
does not work when the mailbox is NFS mounted.

You can set &%use_mbx_lock%& with either (or both) of &%use_fcntl_lock%& and
&%use_flock_lock%& to control what kind of locking is used in implementing the
MBX locking rules. The default is to use &[fcntl()]& if &%use_mbx_lock%& is set
without &%use_fcntl_lock%& or &%use_flock_lock%&.




.section "Operational details for appending" "SECTopappend"
.cindex "appending to a file"
.cindex "file" "appending"
Before appending to a file, the following preparations are made:

.ilist
If the name of the file is &_/dev/null_&, no action is taken, and a success
return is given.

.next
.cindex "directory creation"
If any directories on the file's path are missing, Exim creates them if the
&%create_directory%& option is set. A created directory's mode is given by the
&%directory_mode%& option.

.next
If &%file_format%& is set, the format of an existing file is checked. If this
indicates that a different transport should be used, control is passed to that
transport.

.next
.cindex "file" "locking"
.cindex "locking files"
.cindex "NFS" "lock file"
If &%use_lockfile%& is set, a lock file is built in a way that will work
reliably over NFS, as follows:

.olist
Create a &"hitching post"& file whose name is that of the lock file with the
current time, primary host name, and process id added, by opening for writing
as a new file. If this fails with an access error, delivery is deferred.
.next
Close the hitching post file, and hard link it to the lock filename.
.next
If the call to &[link()]& succeeds, creation of the lock file has succeeded.
Unlink the hitching post name.
.next
Otherwise, use &[stat()]& to get information about the hitching post file, and
then unlink hitching post name. If the number of links is exactly two, creation
of the lock file succeeded but something (for example, an NFS server crash and
restart) caused this fact not to be communicated to the &[link()]& call.
.next
If creation of the lock file failed, wait for &%lock_interval%& and try again,
up to &%lock_retries%& times. However, since any program that writes to a
mailbox should complete its task very quickly, it is reasonable to time out old
lock files that are normally the result of user agent and system crashes. If an
existing lock file is older than &%lockfile_timeout%& Exim attempts to unlink
it before trying again.
.endlist olist

.next
A call is made to &[lstat()]& to discover whether the main file exists, and if
so, what its characteristics are. If &[lstat()]& fails for any reason other
than non-existence, delivery is deferred.

.next
.cindex "symbolic link" "to mailbox"
.cindex "mailbox" "symbolic link"
If the file does exist and is a symbolic link, delivery is deferred, unless the
&%allow_symlink%& option is set, in which case the ownership of the link is
checked, and then &[stat()]& is called to find out about the real file, which
is then subjected to the checks below. The check on the top-level link
ownership prevents one user creating a link for another's mailbox in a sticky
directory, though allowing symbolic links in this case is definitely not a good
idea. If there is a chain of symbolic links, the intermediate ones are not
checked.

.next
If the file already exists but is not a regular file, or if the file's owner
and group (if the group is being checked &-- see &%check_group%& above) are
different from the user and group under which the delivery is running,
delivery is deferred.

.next
If the file's permissions are more generous than specified, they are reduced.
If they are insufficient, delivery is deferred, unless &%mode_fail_narrower%&
is set false, in which case the delivery is tried using the existing
permissions.

.next
The file's inode number is saved, and the file is then opened for appending.
If this fails because the file has vanished, &(appendfile)& behaves as if it
hadn't existed (see below). For any other failures, delivery is deferred.

.next
If the file is opened successfully, check that the inode number hasn't
changed, that it is still a regular file, and that the owner and permissions
have not changed. If anything is wrong, defer delivery and freeze the message.

.next
If the file did not exist originally, defer delivery if the &%file_must_exist%&
option is set. Otherwise, check that the file is being created in a permitted
directory if the &%create_file%& option is set (deferring on failure), and then
open for writing as a new file, with the O_EXCL and O_CREAT options,
except when dealing with a symbolic link (the &%allow_symlink%& option must be
set). In this case, which can happen if the link points to a non-existent file,
the file is opened for writing using O_CREAT but not O_EXCL, because
that prevents link following.

.next
.cindex "loop" "while file testing"
If opening fails because the file exists, obey the tests given above for
existing files. However, to avoid looping in a situation where the file is
being continuously created and destroyed, the exists/not-exists loop is broken
after 10 repetitions, and the message is then frozen.

.next
If opening fails with any other error, defer delivery.

.next
.cindex "file" "locking"
.cindex "locking files"
Once the file is open, unless both &%use_fcntl_lock%& and &%use_flock_lock%&
are false, it is locked using &[fcntl()]& or &[flock()]& or both. If
&%use_mbx_lock%& is false, an exclusive lock is requested in each case.
However, if &%use_mbx_lock%& is true, Exim takes out a shared lock on the open
file, and an exclusive lock on the file whose name is
.code
/tmp/.<device-number>.<inode-number>
.endd
using the device and inode numbers of the open mailbox file, in accordance with
the MBX locking rules. This file is created with a mode that is specified by
the &%lockfile_mode%& option.

If Exim fails to lock the file, there are two possible courses of action,
depending on the value of the locking timeout. This is obtained from
&%lock_fcntl_timeout%& or &%lock_flock_timeout%&, as appropriate.

If the timeout value is zero, the file is closed, Exim waits for
&%lock_interval%&, and then goes back and re-opens the file as above and tries
to lock it again. This happens up to &%lock_retries%& times, after which the
delivery is deferred.

If the timeout has a value greater than zero, blocking calls to &[fcntl()]& or
&[flock()]& are used (with the given timeout), so there has already been some
waiting involved by the time locking fails. Nevertheless, Exim does not give up
immediately. It retries up to
.code
(lock_retries * lock_interval) / <timeout>
.endd
times (rounded up).
.endlist

At the end of delivery, Exim closes the file (which releases the &[fcntl()]&
and/or &[flock()]& locks) and then deletes the lock file if one was created.


.section "Operational details for delivery to a new file" "SECTopdir"
.cindex "delivery" "to single file"
.cindex "&""From""& line"
When the &%directory%& option is set instead of &%file%&, each message is
delivered into a newly-created file or set of files. When &(appendfile)& is
activated directly from a &(redirect)& router, neither &%file%& nor
&%directory%& is normally set, because the path for delivery is supplied by the
router. (See for example, the &(address_file)& transport in the default
configuration.) In this case, delivery is to a new file if either the path name
ends in &`/`&, or the &%maildir_format%& or &%mailstore_format%& option is set.

No locking is required while writing the message to a new file, so the various
locking options of the transport are ignored. The &"From"& line that by default
separates messages in a single file is not normally needed, nor is the escaping
of message lines that start with &"From"&, and there is no need to ensure a
newline at the end of each message. Consequently, the default values for
&%check_string%&, &%message_prefix%&, and &%message_suffix%& are all unset when
any of &%directory%&, &%maildir_format%&, or &%mailstore_format%& is set.

If Exim is required to check a &%quota%& setting, it adds up the sizes of all
the files in the delivery directory by default. However, you can specify a
different directory by setting &%quota_directory%&. Also, for maildir
deliveries (see below) the &_maildirfolder_& convention is honoured.


.cindex "maildir format"
.cindex "mailstore format"
There are three different ways in which delivery to individual files can be
done, controlled by the settings of the &%maildir_format%& and
&%mailstore_format%& options. Note that code to support maildir or mailstore
formats is not included in the binary unless SUPPORT_MAILDIR or
SUPPORT_MAILSTORE, respectively, is set in &_Local/Makefile_&.

.cindex "directory creation"
In all three cases an attempt is made to create the directory and any necessary
sub-directories if they do not exist, provided that the &%create_directory%&
option is set (the default). The location of a created directory can be
constrained by setting &%create_file%&. A created directory's mode is given by
the &%directory_mode%& option. If creation fails, or if the
&%create_directory%& option is not set when creation is required, delivery is
deferred.



.section "Maildir delivery" "SECTmaildirdelivery"
.cindex "maildir format" "description of"
If the &%maildir_format%& option is true, Exim delivers each message by writing
it to a file whose name is &_tmp/<stime>.H<mtime>P<pid>.<host>_& in the
directory that is defined by the &%directory%& option (the &"delivery
directory"&). If the delivery is successful, the file is renamed into the
&_new_& subdirectory.

In the filename, <&'stime'&> is the current time of day in seconds, and
<&'mtime'&> is the microsecond fraction of the time. After a maildir delivery,
Exim checks that the time-of-day clock has moved on by at least one microsecond
before terminating the delivery process. This guarantees uniqueness for the
filename. However, as a precaution, Exim calls &[stat()]& for the file before
opening it. If any response other than ENOENT (does not exist) is given,
Exim waits 2 seconds and tries again, up to &%maildir_retries%& times.

Before Exim carries out a maildir delivery, it ensures that subdirectories
called &_new_&, &_cur_&, and &_tmp_& exist in the delivery directory. If they
do not exist, Exim tries to create them and any superior directories in their
path, subject to the &%create_directory%& and &%create_file%& options. If the
&%maildirfolder_create_regex%& option is set, and the regular expression it
contains matches the delivery directory, Exim also ensures that a file called
&_maildirfolder_& exists in the delivery directory. If a missing directory or
&_maildirfolder_& file cannot be created, delivery is deferred.

These features make it possible to use Exim to create all the necessary files
and directories in a maildir mailbox, including subdirectories for maildir++
folders. Consider this example:
.code
maildir_format = true
directory = /var/mail/$local_part_data\
           ${if eq{$local_part_suffix}{}{}\
           {/.${substr_1:$local_part_suffix}}}
maildirfolder_create_regex = /\.[^/]+$
.endd
If &$local_part_suffix$& is empty (there was no suffix for the local part),
delivery is into a toplevel maildir with a name like &_/var/mail/pimbo_& (for
the user called &'pimbo'&). The pattern in &%maildirfolder_create_regex%& does
not match this name, so Exim will not look for or create the file
&_/var/mail/pimbo/maildirfolder_&, though it will create
&_/var/mail/pimbo/{cur,new,tmp}_& if necessary.

However, if &$local_part_suffix$& contains &`-eximusers`& (for example),
delivery is into the maildir++ folder &_/var/mail/pimbo/.eximusers_&, which
does match &%maildirfolder_create_regex%&. In this case, Exim will create
&_/var/mail/pimbo/.eximusers/maildirfolder_& as well as the three maildir
directories &_/var/mail/pimbo/.eximusers/{cur,new,tmp}_&.

&*Warning:*& Take care when setting &%maildirfolder_create_regex%& that it does
not inadvertently match the toplevel maildir directory, because a
&_maildirfolder_& file at top level would completely break quota calculations.

.cindex "quota" "in maildir delivery"
.cindex "maildir++"
If Exim is required to check a &%quota%& setting before a maildir delivery, and
&%quota_directory%& is not set, it looks for a file called &_maildirfolder_& in
the maildir directory (alongside &_new_&, &_cur_&, &_tmp_&). If this exists,
Exim assumes the directory is a maildir++ folder directory, which is one level
down from the user's top level mailbox directory. This causes it to start at
the parent directory instead of the current directory when calculating the
amount of space used.

One problem with delivering into a multi-file mailbox is that it is
computationally expensive to compute the size of the mailbox for quota
checking. Various approaches have been taken to reduce the amount of work
needed. The next two sections describe two of them. A third alternative is to
use some external process for maintaining the size data, and use the expansion
of the &%mailbox_size%& option as a way of importing it into Exim.




.section "Using tags to record message sizes" "SECID135"
If &%maildir_tag%& is set, the string is expanded for each delivery.
When the maildir file is renamed into the &_new_& sub-directory, the
tag is added to its name. However, if adding the tag takes the length of the
name to the point where the test &[stat()]& call fails with ENAMETOOLONG,
the tag is dropped and the maildir file is created with no tag.


.vindex "&$message_size$&"
Tags can be used to encode the size of files in their names; see
&%quota_size_regex%& above for an example. The expansion of &%maildir_tag%&
happens after the message has been written. The value of the &$message_size$&
variable is set to the number of bytes actually written. If the expansion is
forced to fail, the tag is ignored, but a non-forced failure causes delivery to
be deferred. The expanded tag may contain any printing characters except &"/"&.
Non-printing characters in the string are ignored; if the resulting string is
empty, it is ignored. If it starts with an alphanumeric character, a leading
colon is inserted; this default has not proven to be the path that popular
maildir implementations have chosen (but changing it in Exim would break
backwards compatibility).

For one common implementation, you might set:
.code
maildir_tag = ,S=${message_size}
.endd
but you should check the documentation of the other software to be sure.

It is advisable to also set &%quota_size_regex%& when setting &%maildir_tag%&
as this allows Exim to extract the size from your tag, instead of having to
&[stat()]& each message file.


.section "Using a maildirsize file" "SECID136"
.cindex "quota" "in maildir delivery"
.cindex "maildir format" "&_maildirsize_& file"
If &%maildir_use_size_file%& is true, Exim implements the maildir++ rules for
storing quota and message size information in a file called &_maildirsize_&
within the toplevel maildir directory. If this file does not exist, Exim
creates it, setting the quota from the &%quota%& option of the transport. If
the maildir directory itself does not exist, it is created before any attempt
to write a &_maildirsize_& file.

The &_maildirsize_& file is used to hold information about the sizes of
messages in the maildir, thus speeding up quota calculations. The quota value
in the file is just a cache; if the quota is changed in the transport, the new
value overrides the cached value when the next message is delivered. The cache
is maintained for the benefit of other programs that access the maildir and
need to know the quota.

If the &%quota%& option in the transport is unset or zero, the &_maildirsize_&
file is maintained (with a zero quota setting), but no quota is imposed.

A regular expression is available for controlling which directories in the
maildir participate in quota calculations when a &_maildirsizefile_& is in use.
See the description of the &%maildir_quota_directory_regex%& option above for
details.


.section "Mailstore delivery" "SECID137"
.cindex "mailstore format" "description of"
If the &%mailstore_format%& option is true, each message is written as two
files in the given directory. A unique base name is constructed from the
message id and the current delivery process, and the files that are written use
this base name plus the suffixes &_.env_& and &_.msg_&. The &_.env_& file
contains the message's envelope, and the &_.msg_& file contains the message
itself. The base name is placed in the variable &$mailstore_basename$&.

During delivery, the envelope is first written to a file with the suffix
&_.tmp_&. The &_.msg_& file is then written, and when it is complete, the
&_.tmp_& file is renamed as the &_.env_& file. Programs that access messages in
mailstore format should wait for the presence of both a &_.msg_& and a &_.env_&
file before accessing either of them. An alternative approach is to wait for
the absence of a &_.tmp_& file.

The envelope file starts with any text defined by the &%mailstore_prefix%&
option, expanded and terminated by a newline if there isn't one. Then follows
the sender address on one line, then all the recipient addresses, one per line.
There can be more than one recipient only if the &%batch_max%& option is set
greater than one. Finally, &%mailstore_suffix%& is expanded and the result
appended to the file, followed by a newline if it does not end with one.

If expansion of &%mailstore_prefix%& or &%mailstore_suffix%& ends with a forced
failure, it is ignored. Other expansion errors are treated as serious
configuration errors, and delivery is deferred. The variable
&$mailstore_basename$& is available for use during these expansions.


.section "Non-special new file delivery" "SECID138"
If neither &%maildir_format%& nor &%mailstore_format%& is set, a single new
file is created directly in the named directory. For example, when delivering
messages into files in batched SMTP format for later delivery to some host (see
section &<<SECTbatchSMTP>>&), a setting such as
.code
directory = /var/bsmtp/$host
.endd
might be used. A message is written to a file with a temporary name, which is
then renamed when the delivery is complete. The final name is obtained by
expanding the contents of the &%directory_file%& option.
.ecindex IIDapptra1
.ecindex IIDapptra2






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The autoreply transport" "CHID8"
.scindex IIDauttra1 "transports" "&(autoreply)&"
.scindex IIDauttra2 "&(autoreply)& transport"
The &(autoreply)& transport is not a true transport in that it does not cause
the message to be transmitted. Instead, it generates a new mail message as an
automatic reply to the incoming message. &'References:'& and
&'Auto-Submitted:'& header lines are included. These are constructed according
to the rules in RFCs
&url(https://www.rfc-editor.org/rfc/rfc2822,2822) and
&url(https://datatracker.ietf.org/doc/html/rfc3834,3834), respectively.

If the router that passes the message to this transport does not have the
&%unseen%& option set, the original message (for the current recipient) is not
delivered anywhere. However, when the &%unseen%& option is set on the router
that passes the message to this transport, routing of the address continues, so
another router can set up a normal message delivery.


The &(autoreply)& transport is usually run as the result of mail filtering, a
&"vacation"& message being the standard example. However, it can also be run
directly from a router like any other transport. To reduce the possibility of
message cascades, messages created by the &(autoreply)& transport always have
empty envelope sender addresses, like bounce messages.

The parameters of the message to be sent can be specified in the configuration
by options described below. However, these are used only when the address
passed to the transport does not contain its own reply information. When the
transport is run as a consequence of a
&%mail%&
or &%vacation%& command in a filter file, the parameters of the message are
supplied by the filter, and passed with the address. The transport's options
that define the message are then ignored (so they are not usually set in this
case). The message is specified entirely by the filter or by the transport; it
is never built from a mixture of options. However, the &%file_optional%&,
&%mode%&, and &%return_message%& options apply in all cases.

&(Autoreply)& is implemented as a local transport. When used as a result of a
command in a user's filter file, &(autoreply)& normally runs under the uid and
gid of the user, and with appropriate current and home directories (see chapter
&<<CHAPenvironment>>&).

There is a subtle difference between routing a message to a &(pipe)& transport
that generates some text to be returned to the sender, and routing it to an
&(autoreply)& transport. This difference is noticeable only if more than one
address from the same message is so handled. In the case of a pipe, the
separate outputs from the different addresses are gathered up and returned to
the sender in a single message, whereas if &(autoreply)& is used, a separate
message is generated for each address that is passed to it.

Non-printing characters are not permitted in the header lines generated for the
message that &(autoreply)& creates, with the exception of newlines that are
immediately followed by white space. If any non-printing characters are found,
the transport defers.
Whether characters with the top bit set count as printing characters or not is
controlled by the &%print_topbitchars%& global option.

If any of the generic options for manipulating headers (for example,
&%headers_add%&) are set on an &(autoreply)& transport, they apply to the copy
of the original message that is included in the generated message when
&%return_message%& is set. They do not apply to the generated message itself.

.vindex "&$sender_address$&"
If the &(autoreply)& transport receives return code 2 from Exim when it submits
the message, indicating that there were no recipients, it does not treat this
as an error. This means that autoreplies sent to &$sender_address$& when this
is empty (because the incoming message is a bounce message) do not cause
problems. They are just discarded.



.section "Private options for autoreply" "SECID139"
.cindex "options" "&(autoreply)& transport"

.option bcc autoreply string&!! unset
This specifies the addresses that are to receive &"blind carbon copies"& of the
message when the message is specified by the transport.


.option cc autoreply string&!! unset
This specifies recipients of the message and the contents of the &'Cc:'& header
when the message is specified by the transport.


.option file autoreply string&!! unset
The contents of the file are sent as the body of the message when the message
is specified by the transport. If both &%file%& and &%text%& are set, the text
string comes first.


.option file_expand autoreply boolean false
If this is set, the contents of the file named by the &%file%& option are
subjected to string expansion as they are added to the message.


.option file_optional autoreply boolean false
If this option is true, no error is generated if the file named by the &%file%&
option or passed with the address does not exist or cannot be read.


.option from autoreply string&!! unset
This specifies the contents of the &'From:'& header when the message is
specified by the transport.


.option headers autoreply string&!! unset
This specifies additional
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
headers that are to be added to the message
when the message is specified by the transport. Several can be given by using
&"\n"& to separate them. There is no check on the format.


.option log autoreply string&!! unset
This option names a file in which a record of every message sent is logged when
the message is specified by the transport.


.option mode autoreply "octal integer" 0600
If either the log file or the &"once"& file has to be created, this mode is
used.


.option never_mail autoreply "address list&!!" unset
If any run of the transport creates a message with a recipient that matches any
item in the list, that recipient is quietly discarded. If all recipients are
discarded, no message is created. This applies both when the recipients are
generated by a filter and when they are specified in the transport.



.option once autoreply string&!! unset
This option names a file or DBM database in which a record of each &'To:'&
recipient is kept when the message is specified by the transport. &*Note*&:
This does not apply to &'Cc:'& or &'Bcc:'& recipients.

If &%once%& is unset, or is set to an empty string, the message is always sent.
By default, if &%once%& is set to a non-empty filename, the message
is not sent if a potential recipient is already listed in the database.
However, if the &%once_repeat%& option specifies a time greater than zero, the
message is sent if that much time has elapsed since a message was last sent to
this recipient. A setting of zero time for &%once_repeat%& (the default)
prevents a message from being sent a second time &-- in this case, zero means
infinity.

If &%once_file_size%& is zero, a DBM database is used to remember recipients,
and it is allowed to grow as large as necessary. If &%once_file_size%& is set
greater than zero, it changes the way Exim implements the &%once%& option.
Instead of using a DBM file to record every recipient it sends to, it uses a
regular file, whose size will never get larger than the given value.

In the file, Exim keeps a linear list of recipient addresses and the times at
which they were sent messages. If the file is full when a new address needs to
be added, the oldest address is dropped. If &%once_repeat%& is not set, this
means that a given recipient may receive multiple messages, but at
unpredictable intervals that depend on the rate of turnover of addresses in the
file. If &%once_repeat%& is set, it specifies a maximum time between repeats.


.option once_file_size autoreply integer 0
See &%once%& above.


.option once_repeat autoreply time&!! 0s
See &%once%& above.
After expansion, the value of this option must be a valid time value.


.option reply_to autoreply string&!! unset
This specifies the contents of the &'Reply-To:'& header when the message is
specified by the transport.


.option return_message autoreply boolean false
If this is set, a copy of the original message is returned with the new
message, subject to the maximum size set in the &%return_size_limit%& global
configuration option.


.option subject autoreply string&!! unset
This specifies the contents of the &'Subject:'& header when the message is
specified by the transport. It is tempting to quote the original subject in
automatic responses. For example:
.code
subject = Re: $h_subject:
.endd
There is a danger in doing this, however. It may allow a third party to
subscribe your users to an opt-in mailing list, provided that the list accepts
bounce messages as subscription confirmations. Well-managed lists require a
non-bounce message to confirm a subscription, so the danger is relatively
small.



.option text autoreply string&!! unset
This specifies a single string to be used as the body of the message when the
message is specified by the transport. If both &%text%& and &%file%& are set,
the text comes first.


.option to autoreply string&!! unset
This specifies recipients of the message and the contents of the &'To:'& header
when the message is specified by the transport.
.ecindex IIDauttra1
.ecindex IIDauttra2




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The lmtp transport" "CHAPLMTP"
.cindex "transports" "&(lmtp)&"
.cindex "&(lmtp)& transport"
.cindex "LMTP" "over a pipe"
.cindex "LMTP" "over a unix-domain socket"
The &(lmtp)& transport runs the LMTP protocol
(&url(https://www.rfc-editor.org/rfc/rfc2033,RFC 2033)) over a pipe to a
specified command
or by interacting with a Unix domain socket.
This transport is something of a cross between the &(pipe)& and &(smtp)&
transports. Exim also has support for using LMTP over TCP/IP; this is
implemented as an option for the &(smtp)& transport. Because LMTP is expected
to be of minority interest, the default build-time configure in &_src/EDITME_&
has it commented out. You need to ensure that
.code
TRANSPORT_LMTP=yes
.endd
.cindex "options" "&(lmtp)& transport"
is present in your &_Local/Makefile_& in order to have the &(lmtp)& transport
included in the Exim binary. The private options of the &(lmtp)& transport are
as follows:

.option batch_id lmtp string&!! unset
See the description of local delivery batching in chapter &<<CHAPbatching>>&.


.option batch_max lmtp integer 1
This limits the number of addresses that can be handled in a single delivery.
Most LMTP servers can handle several addresses at once, so it is normally a
good idea to increase this value. See the description of local delivery
batching in chapter &<<CHAPbatching>>&.


.option command lmtp string&!! unset
This option must be set if &%socket%& is not set. The string is a command which
is run in a separate process. It is split up into a command name and list of
arguments, each of which is separately expanded (so expansion cannot change the
number of arguments). The command is run directly, not via a shell. The message
is passed to the new process using the standard input and output to operate the
LMTP protocol.

.option ignore_quota lmtp boolean false
.cindex "LMTP" "ignoring quota errors"
If this option is set true, the string &`IGNOREQUOTA`& is added to RCPT
commands, provided that the LMTP server has advertised support for IGNOREQUOTA
in its response to the LHLO command.

.option socket lmtp string&!! unset
This option must be set if &%command%& is not set. The result of expansion must
be the name of a Unix domain socket. The transport connects to the socket and
delivers the message to it using the LMTP protocol.


.option timeout lmtp time 5m
The transport is aborted if the created process or Unix domain socket does not
respond to LMTP commands or message input within this timeout. Delivery
is deferred, and will be tried again later. Here is an example of a typical
LMTP transport:
.code
lmtp:
  driver = lmtp
  command = /some/local/lmtp/delivery/program
  batch_max = 20
  user = exim
.endd
This delivers up to 20 addresses at a time, in a mixture of domains if
necessary, running as the user &'exim'&.



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The pipe transport" "CHAPpipetransport"
.scindex IIDpiptra1 "transports" "&(pipe)&"
.scindex IIDpiptra2 "&(pipe)& transport"
The &(pipe)& transport is used to deliver messages via a pipe to a command
running in another process. One example is the use of &(pipe)& as a
pseudo-remote transport for passing messages to some other delivery mechanism
(such as UUCP). Another is the use by individual users to automatically process
their incoming messages. The &(pipe)& transport can be used in one of the
following ways:

.ilist
.vindex "&$local_part$&"
A router routes one address to a transport in the normal way, and the
transport is configured as a &(pipe)& transport. In this case, &$local_part$&
contains the local part of the address (as usual), and the command that is run
is specified by the &%command%& option on the transport.
.next
.vindex "&$pipe_addresses$&"
If the &%batch_max%& option is set greater than 1 (the default is 1), the
transport can handle more than one address in a single run. In this case, when
more than one address is routed to the transport, &$local_part$& is not set
(because it is not unique). However, the pseudo-variable &$pipe_addresses$&
(described in section &<<SECThowcommandrun>>& below) contains all the addresses
that are routed to the transport.
.next
.vindex "&$address_pipe$&"
A router redirects an address directly to a pipe command (for example, from an
alias or forward file). In this case, &$address_pipe$& contains the text of the
pipe command, and the &%command%& option on the transport is ignored unless
&%force_command%& is set. If only one address is being transported
(&%batch_max%& is not greater than one, or only one address was redirected to
this pipe command), &$local_part$& contains the local part that was redirected.
.endlist


The &(pipe)& transport is a non-interactive delivery method. Exim can also
deliver messages over pipes using the LMTP interactive protocol. This is
implemented by the &(lmtp)& transport.

In the case when &(pipe)& is run as a consequence of an entry in a local user's
&_.forward_& file, the command runs under the uid and gid of that user. In
other cases, the uid and gid have to be specified explicitly, either on the
transport or on the router that handles the address. Current and &"home"&
directories are also controllable. See chapter &<<CHAPenvironment>>& for
details of the local delivery environment and chapter &<<CHAPbatching>>&
for a discussion of local delivery batching.

.cindex "tainted data" "in pipe command"
.cindex pipe "tainted data"
Tainted data may not be used for the command name.


.section "Concurrent delivery" "SECID140"
If two messages arrive at almost the same time, and both are routed to a pipe
delivery, the two pipe transports may be run concurrently. You must ensure that
any pipe commands you set up are robust against this happening. If the commands
write to a file, the &%exim_lock%& utility might be of use.
Alternatively the &%max_parallel%& option could be used with a value
of "1" to enforce serialization.




.section "Returned status and data" "SECID141"
.cindex "&(pipe)& transport" "returned data"
If the command exits with a non-zero return code, the delivery is deemed to
have failed, unless either the &%ignore_status%& option is set (in which case
the return code is treated as zero), or the return code is one of those listed
in the &%temp_errors%& option, which are interpreted as meaning &"try again
later"&. In this case, delivery is deferred. Details of a permanent failure are
logged, but are not included in the bounce message, which merely contains
&"local delivery failed"&.

If the command exits on a signal and the &%freeze_signal%& option is set then
the message will be frozen in the queue. If that option is not set, a bounce
will be sent as normal.

If the return code is greater than 128 and the command being run is a shell
script, it normally means that the script was terminated by a signal whose
value is the return code minus 128. The &%freeze_signal%& option does not
apply in this case.

If Exim is unable to run the command (that is, if &[execve()]& fails), the
return code is set to 127. This is the value that a shell returns if it is
asked to run a non-existent command. The wording for the log line suggests that
a non-existent command may be the problem.

The &%return_output%& option can affect the result of a pipe delivery. If it is
set and the command produces any output on its standard output or standard
error streams, the command is considered to have failed, even if it gave a zero
return code or if &%ignore_status%& is set. The output from the command is
included as part of the bounce message. The &%return_fail_output%& option is
similar, except that output is returned only when the command exits with a
failure return code, that is, a value other than zero or a code that matches
&%temp_errors%&.



.section "How the command is run" "SECThowcommandrun"
.cindex "&(pipe)& transport" "path for command"
The command line is (by default) broken down into a command name and arguments
by the &(pipe)& transport itself. The &%allow_commands%& and
&%restrict_to_path%& options can be used to restrict the commands that may be
run.

.cindex "quoting" "in pipe command"
Unquoted arguments are delimited by white space. If an argument appears in
double quotes, backslash is interpreted as an escape character in the usual
way. If an argument appears in single quotes, no escaping is done.

String expansion is applied to the command line except when it comes from a
traditional &_.forward_& file (commands from a filter file are expanded). The
expansion is applied to each argument in turn rather than to the whole line.
For this reason, any string expansion item that contains white space must be
quoted so as to be contained within a single argument. A setting such as
.code
command = /some/path ${if eq{$local_part}{postmaster}{xx}{yy}}
.endd
will not work, because the expansion item gets split between several
arguments. You have to write
.code
command = /some/path "${if eq{$local_part}{postmaster}{xx}{yy}}"
.endd
to ensure that it is all in one argument. The expansion is done in this way,
argument by argument, so that the number of arguments cannot be changed as a
result of expansion, and quotes or backslashes in inserted variables do not
interact with external quoting. However, this leads to problems if you want to
generate multiple arguments (or the command name plus arguments) from a single
expansion. In this situation, the simplest solution is to use a shell. For
example:
.code
command = /bin/sh -c ${lookup{$local_part}lsearch{/some/file}}
.endd

.cindex "transport" "filter"
.cindex "filter" "transport filter"
.vindex "&$pipe_addresses$&"
Special handling takes place when an argument consists of precisely the text
&`$pipe_addresses`& (no quotes).
This is not a general expansion variable; the only
place this string is recognized is when it appears as an argument for a pipe or
transport filter command. It causes each address that is being handled to be
inserted in the argument list at that point &'as a separate argument'&. This
avoids any problems with spaces or shell metacharacters, and is of use when a
&(pipe)& transport is handling groups of addresses in a batch.

If &%force_command%& is enabled on the transport, special handling takes place
for an argument that consists of precisely the text &`$address_pipe`&.  It
is handled similarly to &$pipe_addresses$& above.  It is expanded and each
argument is inserted in the argument list at that point
&'as a separate argument'&.  The &`$address_pipe`& item does not need to be
the only item in the argument; in fact, if it were then &%force_command%&
should behave as a no-op.  Rather, it should be used to adjust the command
run while preserving the argument vector separation.

After splitting up into arguments and expansion, the resulting command is run
in a subprocess directly from the transport, &'not'& under a shell. The
message that is being delivered is supplied on the standard input, and the
standard output and standard error are both connected to a single pipe that is
read by Exim. The &%max_output%& option controls how much output the command
may produce, and the &%return_output%& and &%return_fail_output%& options
control what is done with it.

Not running the command under a shell (by default) lessens the security risks
in cases when a command from a user's filter file is built out of data that was
taken from an incoming message. If a shell is required, it can of course be
explicitly specified as the command to be run. However, there are circumstances
where existing commands (for example, in &_.forward_& files) expect to be run
under a shell and cannot easily be modified. To allow for these cases, there is
an option called &%use_shell%&, which changes the way the &(pipe)& transport
works. Instead of breaking up the command line as just described, it expands it
as a single string and passes the result to &_/bin/sh_&. The
&%restrict_to_path%& option and the &$pipe_addresses$& facility cannot be used
with &%use_shell%&, and the whole mechanism is inherently less secure.



.section "Environment variables" "SECTpipeenv"
.cindex "&(pipe)& transport" "environment for command"
.cindex "environment" "&(pipe)& transport"
The environment variables listed below are set up when the command is invoked.
This list is a compromise for maximum compatibility with other MTAs. Note that
the &%environment%& option can be used to add additional variables to this
environment. The environment for the &(pipe)& transport is not subject
to the &%add_environment%& and &%keep_environment%& main config options.
&*Note*&: Using enviroment variables loses track of tainted data.
Writers of &(pipe)& transport commands should be wary of data supplied
by potential attackers.
.display
&`DOMAIN            `&   the domain of the address
&`HOME              `&   the home directory, if set
&`HOST              `&   the host name when called from a router (see below)
&`LOCAL_PART        `&   see below
&`LOCAL_PART_PREFIX `&   see below
&`LOCAL_PART_SUFFIX `&   see below
&`LOGNAME           `&   see below
&`MESSAGE_ID        `&   Exim's local ID for the message
&`PATH              `&   as specified by the &%path%& option below
&`QUALIFY_DOMAIN    `&   the sender qualification domain
&`RECIPIENT         `&   the complete recipient address
&`SENDER            `&   the sender of the message (empty if a bounce)
&`SHELL             `&   &`/bin/sh`&
&`TZ                `&   the value of the &%timezone%& option, if set
&`USER              `&   see below
.endd
When a &(pipe)& transport is called directly from (for example) an &(accept)&
router, LOCAL_PART is set to the local part of the address. When it is
called as a result of a forward or alias expansion, LOCAL_PART is set to
the local part of the address that was expanded. In both cases, any affixes are
removed from the local part, and made available in LOCAL_PART_PREFIX and
LOCAL_PART_SUFFIX, respectively. LOGNAME and USER are set to the
same value as LOCAL_PART for compatibility with other MTAs.

.cindex "HOST"
HOST is set only when a &(pipe)& transport is called from a router that
associates hosts with an address, typically when using &(pipe)& as a
pseudo-remote transport. HOST is set to the first host name specified by
the router.

.cindex "HOME"
If the transport's generic &%home_directory%& option is set, its value is used
for the HOME environment variable. Otherwise, a home directory may be set
by the router's &%transport_home_directory%& option, which defaults to the
user's home directory if &%check_local_user%& is set.


.section "Private options for pipe" "SECID142"
.cindex "options" "&(pipe)& transport"



.option allow_commands pipe "string list&!!" unset
.cindex "&(pipe)& transport" "permitted commands"
The string is expanded, and is then interpreted as a colon-separated list of
permitted commands. If &%restrict_to_path%& is not set, the only commands
permitted are those in the &%allow_commands%& list. They need not be absolute
paths; the &%path%& option is still used for relative paths. If
&%restrict_to_path%& is set with &%allow_commands%&, the command must either be
in the &%allow_commands%& list, or a name without any slashes that is found on
the path. In other words, if neither &%allow_commands%& nor
&%restrict_to_path%& is set, there is no restriction on the command, but
otherwise only commands that are permitted by one or the other are allowed. For
example, if
.code
allow_commands = /usr/bin/vacation
.endd
and &%restrict_to_path%& is not set, the only permitted command is
&_/usr/bin/vacation_&. The &%allow_commands%& option may not be set if
&%use_shell%& is set.


.option batch_id pipe string&!! unset
See the description of local delivery batching in chapter &<<CHAPbatching>>&.


.option batch_max pipe integer 1
This limits the number of addresses that can be handled in a single delivery.
See the description of local delivery batching in chapter &<<CHAPbatching>>&.


.option check_string pipe string unset
As &(pipe)& writes the message, the start of each line is tested for matching
&%check_string%&, and if it does, the initial matching characters are replaced
by the contents of &%escape_string%&, provided both are set. The value of
&%check_string%& is a literal string, not a regular expression, and the case of
any letters it contains is significant. When &%use_bsmtp%& is set, the contents
of &%check_string%& and &%escape_string%& are forced to values that implement
the SMTP escaping protocol. Any settings made in the configuration file are
ignored.


.option command pipe string&!! unset
This option need not be set when &(pipe)& is being used to deliver to pipes
obtained directly from address redirections. In other cases, the option must be
set, to provide a command to be run. It need not yield an absolute path (see
the &%path%& option below). The command is split up into separate arguments by
Exim, and each argument is separately expanded, as described in section
&<<SECThowcommandrun>>& above.

.cindex "tainted data"
No part of the resulting command may be tainted.


.option environment pipe "string list&!!" unset
.cindex "&(pipe)& transport" "environment for command"
.cindex "environment" "&(pipe)& transport"
This option is used to add additional variables to the environment in which the
command runs (see section &<<SECTpipeenv>>& for the default list). Its value is
a string which is expanded, and then interpreted as a colon-separated list of
environment settings of the form <&'name'&>=<&'value'&>.


.option escape_string pipe string unset
See &%check_string%& above.


.option freeze_exec_fail pipe boolean false
.cindex "exec failure"
.cindex "failure of exec"
.cindex "&(pipe)& transport" "failure of exec"
Failure to exec the command in a pipe transport is by default treated like
any other failure while running the command. However, if &%freeze_exec_fail%&
is set, failure to exec is treated specially, and causes the message to be
frozen, whatever the setting of &%ignore_status%&.


.option freeze_signal pipe boolean false
.cindex "signal exit"
.cindex "&(pipe)& transport", "signal exit"
Normally if the process run by a command in a pipe transport exits on a signal,
a bounce message is sent. If &%freeze_signal%& is set, the message will be
frozen in Exim's queue instead.


.option force_command pipe boolean false
.cindex "force command"
.cindex "&(pipe)& transport", "force command"
Normally when a router redirects an address directly to a pipe command
the &%command%& option on the transport is ignored.  If &%force_command%&
is set, the &%command%& option will used. This is especially
useful for forcing a wrapper or additional argument to be added to the
command. For example:
.code
command = /usr/bin/remote_exec myhost -- $address_pipe
force_command
.endd

Note that &$address_pipe$& is handled specially in &%command%& when
&%force_command%& is set, expanding out to the original argument vector as
separate items, similarly to a Unix shell &`"$@"`& construct.


.option ignore_status pipe boolean false
If this option is true, the status returned by the subprocess that is set up to
run the command is ignored, and Exim behaves as if zero had been returned.
Otherwise, a non-zero status or termination by signal causes an error return
from the transport unless the status value is one of those listed in
&%temp_errors%&; these cause the delivery to be deferred and tried again later.

&*Note*&: This option does not apply to timeouts, which do not return a status.
See the &%timeout_defer%& option for how timeouts are handled.


.option log_defer_output pipe boolean false
.cindex "&(pipe)& transport" "logging output"
If this option is set, and the status returned by the command is
one of the codes listed in &%temp_errors%& (that is, delivery was deferred),
and any output was produced on stdout or stderr, the first line of it is
written to the main log.


.option log_fail_output pipe boolean false
If this option is set, and the command returns any output on stdout or
stderr, and also ends with a return code that is neither zero nor one of
the return codes listed in &%temp_errors%& (that is, the delivery
failed), the first line of output is written to the main log. This
option and &%log_output%& are mutually exclusive.  Only one of them may
be set.


.option log_output pipe boolean false
If this option is set and the command returns any output on stdout or
stderr, the first line of output is written to the main log, whatever
the return code. This option and &%log_fail_output%& are mutually
exclusive. Only one of them may be set.


.option max_output pipe integer 20K
This specifies the maximum amount of output that the command may produce on its
standard output and standard error file combined. If the limit is exceeded, the
process running the command is killed. This is intended as a safety measure to
catch runaway processes. The limit is applied independently of the settings of
the options that control what is done with such output (for example,
&%return_output%&). Because of buffering effects, the amount of output may
exceed the limit by a small amount before Exim notices.


.option message_prefix pipe string&!! "see below"
The string specified here is expanded and output at the start of every message.
The default is unset if &%use_bsmtp%& is set. Otherwise it is
.code
message_prefix = \
  From ${if def:return_path{$return_path}{MAILER-DAEMON}}\
  ${tod_bsdinbox}\n
.endd
.cindex "Cyrus"
.cindex "&%tmail%&"
.cindex "&""From""& line"
This is required by the commonly used &_/usr/bin/vacation_& program.
However, it must &'not'& be present if delivery is to the Cyrus IMAP server,
or to the &%tmail%& local delivery agent. The prefix can be suppressed by
setting
.code
message_prefix =
.endd
&*Note:*& If you set &%use_crlf%& true, you must change any occurrences of
&`\n`& to &`\r\n`& in &%message_prefix%&.


.option message_suffix pipe string&!! "see below"
The string specified here is expanded and output at the end of every message.
The default is unset if &%use_bsmtp%& is set. Otherwise it is a single newline.
The suffix can be suppressed by setting
.code
message_suffix =
.endd
&*Note:*& If you set &%use_crlf%& true, you must change any occurrences of
&`\n`& to &`\r\n`& in &%message_suffix%&.


.option path pipe string&!! "/bin:/usr/bin"
This option is expanded and
specifies the string that is set up in the PATH environment
variable of the subprocess.
If the &%command%& option does not yield an absolute path name, the command is
sought in the PATH directories, in the usual way. &*Warning*&: This does not
apply to a command specified as a transport filter.


.option permit_coredump pipe boolean false
Normally Exim inhibits core-dumps during delivery.  If you have a need to get
a core-dump of a pipe command, enable this command.  This enables core-dumps
during delivery and affects both the Exim binary and the pipe command run.
It is recommended that this option remain off unless and until you have a need
for it and that this only be enabled when needed, as the risk of excessive
resource consumption can be quite high.  Note also that Exim is typically
installed as a setuid binary and most operating systems will inhibit coredumps
of these by default, so further OS-specific action may be required.


.option pipe_as_creator pipe boolean false
.cindex "uid (user id)" "local delivery"
If the generic &%user%& option is not set and this option is true, the delivery
process is run under the uid that was in force when Exim was originally called
to accept the message. If the group id is not otherwise set (via the generic
&%group%& option), the gid that was in force when Exim was originally called to
accept the message is used.


.option restrict_to_path pipe boolean false
When this option is set, any command name not listed in &%allow_commands%& must
contain no slashes. The command is searched for only in the directories listed
in the &%path%& option. This option is intended for use in the case when a pipe
command has been generated from a user's &_.forward_& file. This is usually
handled by a &(pipe)& transport called &%address_pipe%&.


.option return_fail_output pipe boolean false
If this option is true, and the command produced any output and ended with a
return code other than zero or one of the codes listed in &%temp_errors%& (that
is, the delivery failed), the output is returned in the bounce message.
However, if the message has a null sender (that is, it is itself a bounce
message), output from the command is discarded. This option and
&%return_output%& are mutually exclusive. Only one of them may be set.



.option return_output pipe boolean false
If this option is true, and the command produced any output, the delivery is
deemed to have failed whatever the return code from the command, and the output
is returned in the bounce message. Otherwise, the output is just discarded.
However, if the message has a null sender (that is, it is a bounce message),
output from the command is always discarded, whatever the setting of this
option. This option and &%return_fail_output%& are mutually exclusive. Only one
of them may be set.



.option temp_errors pipe "string list" "see below"
.cindex "&(pipe)& transport" "temporary failure"
This option contains either a colon-separated list of numbers, or a single
asterisk. If &%ignore_status%& is false
and &%return_output%& is not set,
and the command exits with a non-zero return code, the failure is treated as
temporary and the delivery is deferred if the return code matches one of the
numbers, or if the setting is a single asterisk. Otherwise, non-zero return
codes are treated as permanent errors. The default setting contains the codes
defined by EX_TEMPFAIL and EX_CANTCREAT in &_sysexits.h_&. If Exim is
compiled on a system that does not define these macros, it assumes values of 75
and 73, respectively.


.option timeout pipe time 1h
If the command fails to complete within this time, it is killed. This normally
causes the delivery to fail (but see &%timeout_defer%&). A zero time interval
specifies no timeout. In order to ensure that any subprocesses created by the
command are also killed, Exim makes the initial process a process group leader,
and kills the whole process group on a timeout. However, this can be defeated
if one of the processes starts a new process group.

.option timeout_defer pipe boolean false
A timeout in a &(pipe)& transport, either in the command that the transport
runs, or in a transport filter that is associated with it, is by default
treated as a hard error, and the delivery fails. However, if &%timeout_defer%&
is set true, both kinds of timeout become temporary errors, causing the
delivery to be deferred.

.option umask pipe "octal integer" 022
This specifies the umask setting for the subprocess that runs the command.


.option use_bsmtp pipe boolean false
.cindex "envelope sender"
If this option is set true, the &(pipe)& transport writes messages in &"batch
SMTP"& format, with the envelope sender and recipient(s) included as SMTP
commands. If you want to include a leading HELO command with such messages,
you can do so by setting the &%message_prefix%& option. See section
&<<SECTbatchSMTP>>& for details of batch SMTP.

.option use_classresources pipe boolean false
.cindex "class resources (BSD)"
This option is available only when Exim is running on FreeBSD, NetBSD, or
BSD/OS. If it is set true, the &[setclassresources()]& function is used to set
resource limits when a &(pipe)& transport is run to perform a delivery. The
limits for the uid under which the pipe is to run are obtained from the login
class database.


.option use_crlf pipe boolean false
.cindex "carriage return"
.cindex "linefeed"
This option causes lines to be terminated with the two-character CRLF sequence
(carriage return, linefeed) instead of just a linefeed character. In the case
of batched SMTP, the byte sequence written to the pipe is then an exact image
of what would be sent down a real SMTP connection.

The contents of the &%message_prefix%& and &%message_suffix%& options are
written verbatim, so must contain their own carriage return characters if these
are needed. When &%use_bsmtp%& is not set, the default values for both
&%message_prefix%& and &%message_suffix%& end with a single linefeed, so their
values must be changed to end with &`\r\n`& if &%use_crlf%& is set.


.option use_shell pipe boolean false
.vindex "&$pipe_addresses$&"
If this option is set, it causes the command to be passed to &_/bin/sh_&
instead of being run directly from the transport, as described in section
&<<SECThowcommandrun>>&. This is less secure, but is needed in some situations
where the command is expected to be run under a shell and cannot easily be
modified. The &%allow_commands%& and &%restrict_to_path%& options, and the
&`$pipe_addresses`& facility are incompatible with &%use_shell%&. The
command is expanded as a single string, and handed to &_/bin/sh_& as data for
its &%-c%& option.



.section "Using an external local delivery agent" "SECID143"
.cindex "local delivery" "using an external agent"
.cindex "&'procmail'&"
.cindex "external local delivery"
.cindex "delivery" "&'procmail'&"
.cindex "delivery" "by external agent"
The &(pipe)& transport can be used to pass all messages that require local
delivery to a separate local delivery agent such as &%procmail%&. When doing
this, care must be taken to ensure that the pipe is run under an appropriate
uid and gid. In some configurations one wants this to be a uid that is trusted
by the delivery agent to supply the correct sender of the message. It may be
necessary to recompile or reconfigure the delivery agent so that it trusts an
appropriate user. The following is an example transport and router
configuration for &%procmail%&:
.code
# transport
procmail_pipe:
  driver = pipe
  command = /usr/local/bin/procmail -d $local_part_data
  return_path_add
  delivery_date_add
  envelope_to_add
  check_string = "From "
  escape_string = ">From "
  umask = 077
  user = $local_part_data
  group = mail

# router
procmail:
  driver = accept
  check_local_user
  transport = procmail_pipe
.endd
In this example, the pipe is run as the local user, but with the group set to
&'mail'&. An alternative is to run the pipe as a specific user such as &'mail'&
or &'exim'&, but in this case you must arrange for &%procmail%& to trust that
user to supply a correct sender address. If you do not specify either a
&%group%& or a &%user%& option, the pipe command is run as the local user. The
home directory is the user's home directory by default.

&*Note*&: The command that the pipe transport runs does &'not'& begin with
.code
IFS=" "
.endd
as shown in some &%procmail%& documentation, because Exim does not by default
use a shell to run pipe commands.

.cindex "Cyrus"
The next example shows a transport and a router for a system where local
deliveries are handled by the Cyrus IMAP server.
. Used to have R: local_part_suffix = .* +  T:   -m $local_part_suffix_v
. but that suffix is tainted so cannot be used in a command arg
. Really, you'd want to use a lookup for acceptable suffixes to do real detainting
.code
# transport
local_delivery_cyrus:
  driver = pipe
  command = /usr/cyrus/bin/deliver \
            -- $local_part_data
  user = cyrus
  group = mail
  return_output
  log_output
  message_prefix =
  message_suffix =

# router
local_user_cyrus:
  driver = accept
  check_local_user
  transport = local_delivery_cyrus
.endd
Note the unsetting of &%message_prefix%& and &%message_suffix%&, and the use of
&%return_output%& to cause any text written by Cyrus to be returned to the
sender.
.ecindex IIDpiptra1
.ecindex IIDpiptra2


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The smtp transport" "CHAPsmtptrans"
.scindex IIDsmttra1 "transports" "&(smtp)&"
.scindex IIDsmttra2 "&(smtp)& transport"
The &(smtp)& transport delivers messages over TCP/IP connections using the SMTP
or LMTP protocol. The list of hosts to try can either be taken from the address
that is being processed (having been set up by the router), or specified
explicitly for the transport. Timeout and retry processing (see chapter
&<<CHAPretry>>&) is applied to each IP address independently.


.section "Multiple messages on a single connection" "SECID144"
The sending of multiple messages over a single TCP/IP connection can arise in
two ways:

.ilist
If a message contains more than &%max_rcpt%& (see below) addresses that are
routed to the same host, more than one copy of the message has to be sent to
that host. In this situation, multiple copies may be sent in a single run of
the &(smtp)& transport over a single TCP/IP connection. (What Exim actually
does when it has too many addresses to send in one message also depends on the
value of the global &%remote_max_parallel%& option. Details are given in
section &<<SECToutSMTPTCP>>&.)
.next
.cindex "hints database" "remembering routing"
When a message has been successfully delivered over a TCP/IP connection, Exim
looks in its hints database to see if there are any other messages awaiting a
connection to the same host. If there are, a new delivery process is started
for one of them, and the current TCP/IP connection is passed on to it. The new
process may in turn send multiple copies and possibly create yet another
process.
.endlist


For each copy sent over the same TCP/IP connection, a sequence counter is
incremented, and if it ever gets to the value of &%connection_max_messages%&,
no further messages are sent over that connection.



.section "Use of the $host and $host_address variables" "SECID145"
.vindex "&$host$&"
.vindex "&$host_address$&"
At the start of a run of the &(smtp)& transport, the values of &$host$& and
&$host_address$& are the name and IP address of the first host on the host list
passed by the router. However, when the transport is about to connect to a
specific host, and while it is connected to that host, &$host$& and
&$host_address$& are set to the values for that host. These are the values
that are in force when the &%helo_data%&, &%hosts_try_auth%&, &%interface%&,
&%serialize_hosts%&, and the various TLS options are expanded.


.section "Use of $tls_cipher and $tls_peerdn" "usecippeer"
.vindex &$tls_bits$&
.vindex &$tls_cipher$&
.vindex &$tls_peerdn$&
.vindex &$tls_sni$&
At the start of a run of the &(smtp)& transport, the values of &$tls_bits$&,
&$tls_cipher$&, &$tls_peerdn$& and &$tls_sni$&
are the values that were set when the message was received.
These are the values that are used for options that are expanded before any
SMTP connections are made. Just before each connection is made, these four
variables are emptied. If TLS is subsequently started, they are set to the
appropriate values for the outgoing connection, and these are the values that
are in force when any authenticators are run and when the
&%authenticated_sender%& option is expanded.

These variables are deprecated in favour of &$tls_in_cipher$& et. al.
and will be removed in a future release.


.section "Private options for smtp" "SECID146"
.cindex "options" "&(smtp)& transport"
The private options of the &(smtp)& transport are as follows:


.option address_retry_include_sender smtp boolean true
.cindex "4&'xx'& responses" "retrying after"
When an address is delayed because of a 4&'xx'& response to a RCPT command, it
is the combination of sender and recipient that is delayed in subsequent queue
runs until the retry time is reached. You can delay the recipient without
reference to the sender (which is what earlier versions of Exim did), by
setting &%address_retry_include_sender%& false. However, this can lead to
problems with servers that regularly issue 4&'xx'& responses to RCPT commands.

.option allow_localhost smtp boolean false
.cindex "local host" "sending to"
.cindex "fallback" "hosts specified on transport"
When a host specified in &%hosts%& or &%fallback_hosts%& (see below) turns out
to be the local host, or is listed in &%hosts_treat_as_local%&, delivery is
deferred by default. However, if &%allow_localhost%& is set, Exim goes on to do
the delivery anyway. This should be used only in special cases when the
configuration ensures that no looping will result (for example, a differently
configured Exim is listening on the port to which the message is sent).


.option authenticated_sender smtp string&!! unset
.cindex "Cyrus"
When Exim has authenticated as a client, or if &%authenticated_sender_force%&
is true, this option sets a value for the AUTH= item on outgoing MAIL commands,
overriding any existing authenticated sender value. If the string expansion is
forced to fail, the option is ignored. Other expansion failures cause delivery
to be deferred. If the result of expansion is an empty string, that is also
ignored.

The expansion happens after the outgoing connection has been made and TLS
started, if required. This means that the &$host$&, &$host_address$&,
&$tls_out_cipher$&, and &$tls_out_peerdn$& variables are set according to the
particular connection.

If the SMTP session is not authenticated, the expansion of
&%authenticated_sender%& still happens (and can cause the delivery to be
deferred if it fails), but no AUTH= item is added to MAIL commands
unless &%authenticated_sender_force%& is true.

This option allows you to use the &(smtp)& transport in LMTP mode to
deliver mail to Cyrus IMAP and provide the proper local part as the
&"authenticated sender"&, via a setting such as:
.code
authenticated_sender = $local_part
.endd
This removes the need for IMAP subfolders to be assigned special ACLs to
allow direct delivery to those subfolders.

Because of expected uses such as that just described for Cyrus (when no
domain is involved), there is no checking on the syntax of the provided
value.


.option authenticated_sender_force smtp boolean false
If this option is set true, the &%authenticated_sender%& option's value
is used for the AUTH= item on outgoing MAIL commands, even if Exim has not
authenticated as a client.


.option command_timeout smtp time 5m
.cindex timeout "smtp transport command"
This sets a timeout for receiving a response to an SMTP command that has been
sent out. It is also used when waiting for the initial banner line from the
remote host. Its value must not be zero.


.option connect_timeout smtp time 5m
.cindex timeout "smtp transport connect"
This sets a timeout for the &[connect()]& function, which sets up a TCP/IP call
to a remote host. A setting of zero allows the system timeout (typically
several minutes) to act. To have any effect, the value of this option must be
less than the system timeout. However, it has been observed that on some
systems there is no system timeout, which is why the default value for this
option is 5 minutes, a value recommended by
&url(https://www.rfc-editor.org/rfc/rfc1123,RFC 1123).


.option connection_max_messages smtp integer 500
.cindex "SMTP" "passed connection"
.cindex "SMTP" "multiple deliveries"
.cindex "multiple SMTP deliveries"
This controls the maximum number of separate message deliveries that are sent
over a single TCP/IP connection. If the value is zero, there is no limit.
For testing purposes, this value can be overridden by the &%-oB%& command line
option.

.cindex "ESMTP extensions" LIMITS
If the peer advertises a LIMITS extension with a MAILMAX value,
and either TLSS is in use or was not advertised,
that value also constrains the result of this option.


.option dane_require_tls_ciphers smtp string&!! unset
.cindex "TLS" "requiring specific ciphers for DANE"
.cindex "cipher" "requiring specific"
.cindex DANE "TLS ciphers"
This option may be used to override &%tls_require_ciphers%& for connections
where DANE has been determined to be in effect.
If not set, then &%tls_require_ciphers%& will be used.
Normal SMTP delivery is not able to make strong demands of TLS cipher
configuration, because delivery will fall back to plaintext.  Once DANE has
been determined to be in effect, there is no plaintext fallback and making the
TLS cipherlist configuration stronger will increase security, rather than
counter-intuitively decreasing it.
If the option expands to be empty or is forced to fail, then it will
be treated as unset and &%tls_require_ciphers%& will be used instead.


.option data_timeout smtp time 5m
.cindex timeout "for transmitted SMTP data blocks"
This sets a timeout for the transmission of each block in the data portion of
the message. As a result, the overall timeout for a message depends on the size
of the message. Its value must not be zero. See also &%final_timeout%&.


.option dkim_canon smtp string&!! unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_domain smtp "string list&!!" unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_hash smtp string&!! sha256
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_identity smtp string&!! unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_private_key smtp string&!! unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_selector smtp string&!! unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_strict smtp string&!! unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_sign_headers smtp string&!! "per RFC"
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.
.option dkim_timestamps smtp string&!! unset
DKIM signing option.  For details see section &<<SECDKIMSIGN>>&.


.option delay_after_cutoff smtp boolean true
.cindex "final cutoff" "retries, controlling"
.cindex retry "final cutoff"
This option controls what happens when all remote IP addresses for a given
domain have been inaccessible for so long that they have passed their retry
cutoff times.

In the default state, if the next retry time has not been reached for any of
them, the address is bounced without trying any deliveries. In other words,
Exim delays retrying an IP address after the final cutoff time until a new
retry time is reached, and can therefore bounce an address without ever trying
a delivery, when machines have been down for a long time. Some people are
unhappy at this prospect, so...

If &%delay_after_cutoff%& is set false, Exim behaves differently. If all IP
addresses are past their final cutoff time, Exim tries to deliver to those
IP addresses that have not been tried since the message arrived. If there are
none, of if they all fail, the address is bounced. In other words, it does not
delay when a new message arrives, but immediately tries those expired IP
addresses that haven't been tried since the message arrived. If there is a
continuous stream of messages for the dead hosts, unsetting
&%delay_after_cutoff%& means that there will be many more attempts to deliver
to them.


.option dns_qualify_single smtp boolean true
If the &%hosts%& or &%fallback_hosts%& option is being used,
and the &%gethostbyname%& option is false,
the RES_DEFNAMES resolver option is set. See the &%qualify_single%& option
in chapter &<<CHAPdnslookup>>& for more details.


.option dns_search_parents smtp boolean false
If the &%hosts%& or &%fallback_hosts%& option is being used, and the
&%gethostbyname%& option is false, the RES_DNSRCH resolver option is set.
See the &%search_parents%& option in chapter &<<CHAPdnslookup>>& for more
details.


.option dnssec_request_domains smtp "domain list&!!" *
.cindex "MX record" "security"
.cindex "DNSSEC" "MX lookup"
.cindex "security" "MX lookup"
.cindex "DNS" "DNSSEC"
DNS lookups for domains matching &%dnssec_request_domains%& will be done with
the DNSSEC request bit set. Setting this transport option is only useful if the
transport overrides or sets the host names. See the &%dnssec_request_domains%&
router option.



.option dnssec_require_domains smtp "domain list&!!" unset
.cindex "MX record" "security"
.cindex "DNSSEC" "MX lookup"
.cindex "security" "MX lookup"
.cindex "DNS" "DNSSEC"
DNS lookups for domains matching &%dnssec_require_domains%& will be done with
the DNSSEC request bit set.  Setting this transport option is only
useful if the transport overrides or sets the host names. See the
&%dnssec_require_domains%& router option.



.option dscp smtp string&!! unset
.cindex "DCSP" "outbound"
This option causes the DSCP value associated with a socket to be set to one
of a number of fixed strings or to numeric value.
The &%-bI:dscp%& option may be used to ask Exim which names it knows of.
Common values include &`throughput`&, &`mincost`&, and on newer systems
&`ef`&, &`af41`&, etc.  Numeric values may be in the range 0 to 0x3F.

The outbound packets from Exim will be marked with this value in the header
(for IPv4, the TOS field; for IPv6, the TCLASS field); there is no guarantee
that these values will have any effect, not be stripped by networking
equipment, or do much of anything without cooperation with your Network
Engineer and those of all network operators between the source and destination.


.option fallback_hosts smtp "string list" unset
.cindex "fallback" "hosts specified on transport"
String expansion is not applied to this option. The argument must be a
colon-separated list of host names or IP addresses, optionally also including
port numbers, though the separator can be changed, as described in section
&<<SECTlistconstruct>>&. Each individual item in the list is the same as an
item in a &%route_list%& setting for the &(manualroute)& router, as described
in section &<<SECTformatonehostitem>>&.

Fallback hosts can also be specified on routers, which associate them with the
addresses they process. As for the &%hosts%& option without &%hosts_override%&,
&%fallback_hosts%& specified on the transport is used only if the address does
not have its own associated fallback host list. Unlike &%hosts%&, a setting of
&%fallback_hosts%& on an address is not overridden by &%hosts_override%&.
However, &%hosts_randomize%& does apply to fallback host lists.

If Exim is unable to deliver to any of the hosts for a particular address, and
the errors are not permanent rejections, the address is put on a separate
transport queue with its host list replaced by the fallback hosts, unless the
address was routed via MX records and the current host was in the original MX
list. In that situation, the fallback host list is not used.

Once normal deliveries are complete, the fallback queue is delivered by
re-running the same transports with the new host lists. If several failing
addresses have the same fallback hosts (and &%max_rcpt%& permits it), a single
copy of the message is sent.

The resolution of the host names on the fallback list is controlled by the
&%gethostbyname%& option, as for the &%hosts%& option. Fallback hosts apply
both to cases when the host list comes with the address and when it is taken
from &%hosts%&. This option provides a &"use a smart host only if delivery
fails"& facility.


.option final_timeout smtp time 10m
.cindex timeout "for transmitted SMTP data accept"
This is the timeout that applies while waiting for the response to the final
line containing just &"."& that terminates a message. Its value must not be
zero.

.option gethostbyname smtp boolean false
If this option is true when the &%hosts%& and/or &%fallback_hosts%& options are
being used, names are looked up using &[gethostbyname()]&
(or &[getipnodebyname()]& when available)
instead of using the DNS. Of course, that function may in fact use the DNS, but
it may also consult other sources of information such as &_/etc/hosts_&.

.option gnutls_compat_mode smtp boolean unset
This option controls whether GnuTLS is used in compatibility mode in an Exim
server. This reduces security slightly, but improves interworking with older
implementations of TLS.

.option helo_data smtp string&!! "see below"
.cindex "HELO" "argument, setting"
.cindex "EHLO" "argument, setting"
.cindex "LHLO argument setting"
The value of this option is expanded after a connection to a another host has
been set up. The result is used as the argument for the EHLO, HELO, or LHLO
command that starts the outgoing SMTP or LMTP session. The default value of the
option is:
.code
$primary_hostname
.endd
During the expansion, the variables &$host$& and &$host_address$& are set to
the identity of the remote host, and the variables &$sending_ip_address$& and
&$sending_port$& are set to the local IP address and port number that are being
used. These variables can be used to generate different values for different
servers or different local IP addresses. For example, if you want the string
that is used for &%helo_data%& to be obtained by a DNS lookup of the outgoing
interface address, you could use this:
.code
helo_data = ${lookup dnsdb{ptr=$sending_ip_address} \
  {${listextract{1}{<\n $value}}} \
  {$primary_hostname}}
.endd
The use of &%helo_data%& applies both to sending messages and when doing
callouts.

.option host_name_extract smtp "string list&!!" "see below"
.cindex "load balancer" "hosts behind"
.cindex TLS resumption
Some mail-accepting sites
(notably Microsoft)
operate many servers behind a network load-balancer.  When this is done,
with separated TLS session caches, TLS session resuption becomes problematic.
It will only succeed when the same server happens to be selected by the
load-balancer, matching the session stored in the client's cache.

Exim can pull out a server name, if there is one, from the response to the
client's SMTP EHLO command.
For normal STARTTLS use, the default value of this option:
.code
    ${if and { {match {$host} {.outlook.com\$}} \
               {match {$item} {\N^250-([\w.]+)\s\N}} \
             } {$1}}
.endd
suffices for one known case.

During the expansion of this option the &$item$& variable will have the
server's EHLO response.

For TLS-on-connect connections we do not have an EHLO
response to use. Because of this the default value of this option is
set to a static string for those cases, meaning that resumption will
always be attempted if permitted by the &%tls_resumption_hosts%& option.

The result of the option expansion is included in the key used to store and
retrieve the TLS session, for session resumption.

Operators of high-load sites may wish to evaluate their logs for indications
of other destination sites operating load-balancers, and develop a suitable
expression for this option.
The smtp:ehlo event and the &$tls_out_resumption$& variable
will be useful for such work.

.option hosts smtp "string list&!!" unset
Hosts are associated with an address by a router such as &(dnslookup)&, which
finds the hosts by looking up the address domain in the DNS, or by
&(manualroute)&, which has lists of hosts in its configuration. However,
email addresses can be passed to the &(smtp)& transport by any router, and not
all of them can provide an associated list of hosts.

The &%hosts%& option specifies a list of hosts to be used if the address being
processed does not have any hosts associated with it. The hosts specified by
&%hosts%& are also used, whether or not the address has its own hosts, if
&%hosts_override%& is set.

The string is first expanded, before being interpreted as a colon-separated
list of host names or IP addresses, possibly including port numbers. The
separator may be changed to something other than colon, as described in section
&<<SECTlistconstruct>>&. Each individual item in the list is the same as an
item in a &%route_list%& setting for the &(manualroute)& router, as described
in section &<<SECTformatonehostitem>>&. However, note that the &`/MX`& facility
of the &(manualroute)& router is not available here.

If the expansion fails, delivery is deferred. Unless the failure was caused by
the inability to complete a lookup, the error is logged to the panic log as
well as the main log. Host names are looked up either by searching directly for
address records in the DNS or by calling &[gethostbyname()]& (or
&[getipnodebyname()]& when available), depending on the setting of the
&%gethostbyname%& option. When Exim is compiled with IPv6 support, if a host
that is looked up in the DNS has both IPv4 and IPv6 addresses, both types of
address are used.

During delivery, the hosts are tried in order, subject to their retry status,
unless &%hosts_randomize%& is set.


.option hosts_avoid_esmtp smtp "host list&!!" unset
.cindex "ESMTP, avoiding use of"
.cindex "HELO" "forcing use of"
.cindex "EHLO" "avoiding use of"
.cindex "PIPELINING" "avoiding the use of"
This option is for use with broken hosts that announce ESMTP facilities (for
example, PIPELINING) and then fail to implement them properly. When a host
matches &%hosts_avoid_esmtp%&, Exim sends HELO rather than EHLO at the
start of the SMTP session. This means that it cannot use any of the ESMTP
facilities such as AUTH, PIPELINING, SIZE, and STARTTLS.


.option hosts_avoid_pipelining smtp "host list&!!" unset
.cindex "PIPELINING" "avoiding the use of"
.cindex "ESMTP extensions" PIPELINING
Exim will not use the ESMTP PIPELINING extension when delivering to any host
that matches this list, even if the server host advertises PIPELINING support.

.option hosts_pipe_connect smtp "host list&!!" unset
.cindex "pipelining" "early connection"
.cindex "pipelining" PIPECONNECT
If Exim is built with the SUPPORT_PIPE_CONNECT build option
this option controls which to hosts the facility watched for
and recorded, and used for subsequent connections.

The retry hints database is used for the record,
and records are subject to the &%retry_data_expire%& option.
When used, the pipelining saves on roundtrip times.
It also turns SMTP into a client-first protocol
so combines well with TCP Fast Open.

See also the &%pipelining_connect_advertise_hosts%& main option.

Note:
When the facility is used, if the transport &%interface%& option is unset
the &%helo_data%& option
will be expanded before the &$sending_ip_address$& variable
is filled in.
A check is made for the use of that variable, without the
presence of a &"def:"& test on it, but suitably complex coding
can avoid the check and produce unexpected results.
You have been warned.


.option hosts_avoid_tls smtp "host list&!!" unset
.cindex "TLS" "avoiding for certain hosts"
Exim will not try to start a TLS session when delivering to any host that
matches this list. See chapter &<<CHAPTLS>>& for details of TLS.

.option hosts_verify_avoid_tls smtp "host list&!!" unset
.cindex "TLS" "avoiding for certain hosts"
Exim will not try to start a TLS session for a verify callout,
or when delivering in cutthrough mode,
to any host that matches this list.


.option hosts_max_try smtp integer 5
.cindex "host" "maximum number to try"
.cindex "limit" "number of hosts tried"
.cindex "limit" "number of MX tried"
.cindex "MX record" "maximum tried"
This option limits the number of IP addresses that are tried for any one
delivery in cases where there are temporary delivery errors. Section
&<<SECTvalhosmax>>& describes in detail how the value of this option is used.


.option hosts_max_try_hardlimit smtp integer 50
This is an additional check on the maximum number of IP addresses that Exim
tries for any one delivery. Section &<<SECTvalhosmax>>& describes its use and
why it exists.



.option hosts_nopass_tls smtp "host list&!!" unset
.cindex "TLS" "passing connection"
.cindex "multiple SMTP deliveries"
.cindex "TLS" "multiple message deliveries"
For any host that matches this list, a connection on which a TLS session has
been started will not be passed to a new delivery process for sending another
message on the same connection. See section &<<SECTmulmessam>>& for an
explanation of when this might be needed.

.option hosts_noproxy_tls smtp "host list&!!" unset
.cindex "TLS" "passing connection"
.cindex "multiple SMTP deliveries"
.cindex "TLS" "multiple message deliveries"
For any host that matches this list, a TLS session which has
been started will not be passed to a new delivery process for sending another
message on the same session.

The traditional implementation closes down TLS and re-starts it in the new
process, on the same open TCP connection, for each successive message
sent.  If permitted by this option a pipe to to the new process is set up
instead, and the original process maintains the TLS connection and proxies
the SMTP connection from and to the new process and any subsequents.
The new process has no access to TLS information, so cannot include it in
logging.



.option hosts_override smtp boolean false
If this option is set and the &%hosts%& option is also set, any hosts that are
attached to the address are ignored, and instead the hosts specified by the
&%hosts%& option are always used. This option does not apply to
&%fallback_hosts%&.


.option hosts_randomize smtp boolean false
.cindex "randomized host list"
.cindex "host" "list of; randomized"
.cindex "fallback" "randomized hosts"
If this option is set, and either the list of hosts is taken from the
&%hosts%& or the &%fallback_hosts%& option, or the hosts supplied by the router
were not obtained from MX records (this includes fallback hosts from the
router), and were not randomized by the router, the order of trying the hosts
is randomized each time the transport runs. Randomizing the order of a host
list can be used to do crude load sharing.

When &%hosts_randomize%& is true, a host list may be split into groups whose
order is separately randomized. This makes it possible to set up MX-like
behaviour. The boundaries between groups are indicated by an item that is just
&`+`& in the host list. For example:
.code
hosts = host1:host2:host3:+:host4:host5
.endd
The order of the first three hosts and the order of the last two hosts is
randomized for each use, but the first three always end up before the last two.
If &%hosts_randomize%& is not set, a &`+`& item in the list is ignored.

.option hosts_require_auth smtp "host list&!!" unset
.cindex "authentication" "required by client"
This option provides a list of servers for which authentication must succeed
before Exim will try to transfer a message. If authentication fails for
servers which are not in this list, Exim tries to send unauthenticated. If
authentication fails for one of these servers, delivery is deferred. This
temporary error is detectable in the retry rules, so it can be turned into a
hard failure if required. See also &%hosts_try_auth%&, and chapter
&<<CHAPSMTPAUTH>>& for details of authentication.


.option hosts_request_ocsp smtp "host list&!!" "see below"
.cindex "TLS" "requiring for certain servers"
Exim will request a Certificate Status on a
TLS session for any host that matches this list.
&%tls_verify_certificates%& should also be set for the transport.

The default is &"**"& if DANE is not in use for the connection,
or if DANE-TA us used.
It is empty if DANE-EE is used.

.option hosts_require_alpn smtp "host list&!!" unset
.cindex ALPN "require negotiation in client"
.cindex TLS ALPN
.cindex TLS "Application Layer Protocol Names"
If the TLS library supports ALPN
then a successful negotiation of ALPN will be required for any host
matching the list, for TLS to be used.
See also the &%tls_alpn%& option.

&*Note*&: prevention of fallback to in-clear connection is not
managed by this option; see &%hosts_require_tls%&.

.option hosts_require_dane smtp "host list&!!" unset
.cindex DANE "transport options"
.cindex DANE "requiring for certain servers"
If built with DANE support, Exim  will require that a DNSSEC-validated
TLSA record is present for any host matching the list,
and that a DANE-verified TLS connection is made.
There will be no fallback to in-clear communication.
See the &%dnssec_request_domains%& router and transport options.
See section &<<SECDANE>>&.

.option hosts_require_ocsp smtp "host list&!!" unset
.cindex "TLS" "requiring for certain servers"
Exim will request, and check for a valid Certificate Status being given, on a
TLS session for any host that matches this list.
&%tls_verify_certificates%& should also be set for the transport.

.option hosts_require_tls smtp "host list&!!" unset
.cindex "TLS" "requiring for certain servers"
Exim will insist on using a TLS session when delivering to any host that
matches this list. See chapter &<<CHAPTLS>>& for details of TLS.
&*Note*&: This option affects outgoing mail only. To insist on TLS for
incoming messages, use an appropriate ACL.

.option hosts_try_auth smtp "host list&!!" unset
.cindex "authentication" "optional in client"
This option provides a list of servers to which, provided they announce
authentication support, Exim will attempt to authenticate as a client when it
connects. If authentication fails
and &%hosts_require_auth%& permits,
Exim will try to transfer the message unauthenticated.
See also chapter &<<CHAPSMTPAUTH>>& for details of authentication.

.option hosts_try_chunking smtp "host list&!!" *
.cindex CHUNKING "enabling, in client"
.cindex BDAT "SMTP command"
.cindex "RFC 3030" "CHUNKING"
This option provides a list of servers to which, provided they announce
CHUNKING support, Exim will attempt to use BDAT commands rather than DATA.
Unless DKIM signing is being done,
BDAT will not be used in conjunction with a transport filter.

.option hosts_try_dane smtp "host list&!!" *
.cindex DANE "transport options"
.cindex DANE "attempting for certain servers"
If built with DANE support, Exim  will look up a
TLSA record for any host matching the list,
If one is found and that lookup was DNSSEC-validated,
then Exim requires that a DANE-verified TLS connection is made for that host;
there will be no fallback to in-clear communication.
See the &%dnssec_request_domains%& router and transport options.
See section &<<SECDANE>>&.

.option hosts_try_fastopen smtp "host list&!!" *
.cindex "fast open, TCP" "enabling, in client"
.cindex "TCP Fast Open" "enabling, in client"
.cindex "RFC 7413" "TCP Fast Open"
This option provides a list of servers to which, provided
the facility is supported by this system, Exim will attempt to
perform a TCP Fast Open (&url(https://www.rfc-editor.org/rfc/rfc7413,RFC 7413).
No data is sent on the SYN segment but, if the remote server also
supports the facility, it can send its SMTP banner immediately after
the SYN,ACK segment.  This can save up to one round-trip time.

The facility is only active for previously-contacted servers,
as the initiator must present a cookie in the SYN segment.

On (at least some) current Linux distributions the facility must be enabled
in the kernel by the sysadmin before the support is usable.
There is no option for control of the server side; if the system supports
it it is always enabled.  Note that lengthy operations in the connect ACL,
such as DNSBL lookups, will still delay the emission of the SMTP banner.

.option hosts_try_prdr smtp "host list&!!" *
.cindex "PRDR" "enabling, optional in client"
.cindex "ESMTP extensions" PRDR
This option provides a list of servers to which, provided they announce
PRDR support, Exim will attempt to negotiate PRDR
for multi-recipient messages.
The option can usually be left as default.

.option interface smtp "string list&!!" unset
.cindex "bind IP address"
.cindex "IP address" "binding"
.vindex "&$host$&"
.vindex "&$host_address$&"
This option specifies which interface to bind to when making an outgoing SMTP
call. The value is an IP address, not an interface name such as
&`eth0`&. Do not confuse this with the interface address that was used when a
message was received, which is in &$received_ip_address$&, formerly known as
&$interface_address$&. The name was changed to minimize confusion with the
outgoing interface address. There is no variable that contains an outgoing
interface address because, unless it is set by this option, its value is
unknown.

During the expansion of the &%interface%& option the variables &$host$& and
&$host_address$& refer to the host to which a connection is about to be made
during the expansion of the string. Forced expansion failure, or an empty
string result causes the option to be ignored. Otherwise, after expansion, the
string must be a list of IP addresses, colon-separated by default, but the
separator can be changed in the usual way (&<<SECTlistsepchange>>&).
For example:
.code
interface = <; 192.168.123.123 ; 3ffe:ffff:836f::fe86:a061
.endd
The first interface of the correct type (IPv4 or IPv6) is used for the outgoing
connection. If none of them are the correct type, the option is ignored. If
&%interface%& is not set, or is ignored, the system's IP functions choose which
interface to use if the host has more than one.


.option keepalive smtp boolean true
.cindex "keepalive" "on outgoing connection"
This option controls the setting of SO_KEEPALIVE on outgoing TCP/IP socket
connections. When set, it causes the kernel to probe idle connections
periodically, by sending packets with &"old"& sequence numbers. The other end
of the connection should send a acknowledgment if the connection is still okay
or a reset if the connection has been aborted. The reason for doing this is
that it has the beneficial effect of freeing up certain types of connection
that can get stuck when the remote host is disconnected without tidying up the
TCP/IP call properly. The keepalive mechanism takes several hours to detect
unreachable hosts.


.option lmtp_ignore_quota smtp boolean false
.cindex "LMTP" "ignoring quota errors"
If this option is set true when the &%protocol%& option is set to &"lmtp"&, the
string &`IGNOREQUOTA`& is added to RCPT commands, provided that the LMTP server
has advertised support for IGNOREQUOTA in its response to the LHLO command.

.option max_rcpt smtp integer&!! 100
.cindex "RCPT" "maximum number of outgoing"
This option,
after expansion,
limits the number of RCPT commands that are sent in a single
SMTP message transaction.
A value setting of zero disables the limit.

If a constant is given,
each set of addresses is treated independently, and
so can cause parallel connections to the same host if &%remote_max_parallel%&
permits this.

.cindex "ESMTP extensions" LIMITS
If the peer advertises a LIMITS extension with a RCPTMAX value,
and either TLSS is in use or was not advertised,
that value also constrains the result of this option
and no parallel connections will be caused on meeting the RCPTMAX limit.


.option message_linelength_limit smtp integer 998
.cindex "line length" limit
This option sets the maximum line length, in bytes, that the transport
will send.  Any messages with lines exceeding the given value
(before a transport filter, if any)
will fail and a failure-DSN ("bounce") message will if possible be returned
to the sender.
The default value is that defined by the SMTP standards.

It is generally wise to also check in the data ACL so that messages
received via SMTP can be refused without producing a bounce.


.option multi_domain smtp boolean&!! true
.vindex "&$domain$&"
When this option is set, the &(smtp)& transport can handle a number of
addresses containing a mixture of different domains provided they all resolve
to the same list of hosts. Turning the option off restricts the transport to
handling only one domain at a time. This is useful if you want to use
&$domain$& in an expansion for the transport, because it is set only when there
is a single domain involved in a remote delivery.

It is expanded per-address and can depend on any of
&$address_data$&, &$domain_data$&, &$local_part_data$&,
&$host$&, &$host_address$& and &$host_port$&.

If the connection is DANE-enabled then this option is ignored;
only messages having the domain used for the DANE TLSA lookup are
sent on the connection.

.cindex "ESMTP extensions" LIMITS
If the peer advertises a LIMITS extension with a RCPTDOMAINMAX value,
and either TLSS is in use or was not advertised,
this option is regarded as being false.


.option port smtp string&!! "see below"
.cindex "port" "sending TCP/IP"
.cindex "TCP/IP" "setting outgoing port"
This option specifies the TCP/IP port on the server to which Exim connects.
&*Note:*& Do not confuse this with the port that was used when a message was
received, which is in &$received_port$&, formerly known as &$interface_port$&.
The name was changed to minimize confusion with the outgoing port. There is no
variable that contains an outgoing port.

If the value of this option begins with a digit it is taken as a port number;
otherwise it is looked up using &[getservbyname()]&. The default value is
normally &"smtp"&,
but if &%protocol%& is set to &"lmtp"& the default is &"lmtp"&
and if &%protocol%& is set to &"smtps"& the default is &"smtps"&.
If the expansion fails, or if a port number cannot be found, delivery
is deferred.

Note that at least one Linux distribution has been seen failing
to put &"smtps"& in its &"/etc/services"& file, resulting is such deferrals.



.option protocol smtp string smtp
.cindex "LMTP" "over TCP/IP"
.cindex "ssmtp protocol" "outbound"
.cindex "TLS" "SSL-on-connect outbound"
.vindex "&$port$&"
If this option is set to &"lmtp"& instead of &"smtp"&, the default value for
the &%port%& option changes to &"lmtp"&, and the transport operates the LMTP
protocol
(&url(https://www.rfc-editor.org/rfc/rfc2033,RFC 2033)) instead of SMTP.
This protocol is sometimes used for local
deliveries into closed message stores. Exim also has support for running LMTP
over a pipe to a local process &-- see chapter &<<CHAPLMTP>>&.

&*Note*&: When using LMTP it should be considered whether the default values
for some other features, such as DANE, are appropriate.

If this option is set to &"smtps"&, the default value for the &%port%& option
changes to &"smtps"&, and the transport initiates TLS immediately after
connecting, as an outbound SSL-on-connect, instead of using STARTTLS to upgrade.
The Internet standards bodies used to strongly discourage use of this mode,
but as of &url(https://www.rfc-editor.org/rfc/rfc8314,RFC 8314) it is preferred
over STARTTLS for message submission
(as distinct from MTA-MTA communication).


.option retry_include_ip_address smtp boolean&!! true
Exim normally includes both the host name and the IP address in the key it
constructs for indexing retry data after a temporary delivery failure. This
means that when one of several IP addresses for a host is failing, it gets
tried periodically (controlled by the retry rules), but use of the other IP
addresses is not affected.

However, in some dialup environments hosts are assigned a different IP address
each time they connect. In this situation the use of the IP address as part of
the retry key leads to undesirable behaviour. Setting this option false causes
Exim to use only the host name.
Since it is expanded it can be made to depend on the host or domain.


.option serialize_hosts smtp "host list&!!" unset
.cindex "serializing connections"
.cindex "host" "serializing connections"
Because Exim operates in a distributed manner, if several messages for the same
host arrive at around the same time, more than one simultaneous connection to
the remote host can occur. This is not usually a problem except when there is a
slow link between the hosts. In that situation it may be helpful to restrict
Exim to one connection at a time. This can be done by setting
&%serialize_hosts%& to match the relevant hosts.

.cindex "hints database" "serializing deliveries to a host"
Exim implements serialization by means of a hints database in which a record is
written whenever a process connects to one of the restricted hosts. The record
is deleted when the connection is completed. Obviously there is scope for
records to get left lying around if there is a system or program crash. To
guard against this, Exim ignores any records that are more than six hours old.

If you set up this kind of serialization, you should also arrange to delete the
relevant hints database whenever your system reboots. The names of the files
start with &_misc_& and they are kept in the &_spool/db_& directory. There
may be one or two files, depending on the type of DBM in use. The same files
are used for ETRN serialization.

See also the &%max_parallel%& generic transport option.


.option size_addition smtp integer 1024
.cindex "SIZE" "ESMTP extension"
.cindex "message" "size issue for transport filter"
.cindex "size" "of message"
.cindex "transport" "filter"
.cindex "filter" "transport filter"
If a remote SMTP server indicates that it supports the SIZE option of the
MAIL command, Exim uses this to pass over the message size at the start of
an SMTP transaction. It adds the value of &%size_addition%& to the value it
sends, to allow for headers and other text that may be added during delivery by
configuration options or in a transport filter. It may be necessary to increase
this if a lot of text is added to messages.

Alternatively, if the value of &%size_addition%& is set negative, it disables
the use of the SIZE option altogether.


.option socks_proxy smtp string&!! unset
.cindex proxy SOCKS
This option enables use of SOCKS proxies for connections made by the
transport.  For details see section &<<SECTproxySOCKS>>&.


.option tls_alpn smtp string&!! unset
.cindex TLS "Application Layer Protocol Names"
.cindex TLS ALPN
.cindex ALPN "set name in client"
If this option is set
and the TLS library supports ALPN,
the value given is used.

As of writing no value has been standardised for email use.
The authors suggest using &"smtp"&.



.option tls_certificate smtp string&!! unset
.cindex "TLS" "client certificate, location of"
.cindex "certificate" "client, location of"
.vindex "&$host$&"
.vindex "&$host_address$&"
The value of this option must be the absolute path to a file which contains the
client's certificate, for possible use when sending a message over an encrypted
connection. The values of &$host$& and &$host_address$& are set to the name and
address of the server during the expansion. See chapter &<<CHAPTLS>>& for
details of TLS.

&*Note*&: This option must be set if you want Exim to be able to use a TLS
certificate when sending messages as a client. The global option of the same
name specifies the certificate for Exim as a server; it is not automatically
assumed that the same certificate should be used when Exim is operating as a
client.


.option tls_crl smtp string&!! unset
.cindex "TLS" "client certificate revocation list"
.cindex "certificate" "revocation list for client"
This option specifies a certificate revocation list. The expanded value must
be the name of a file that contains a CRL in PEM format.


.option tls_dh_min_bits smtp integer 1024
.cindex "TLS" "Diffie-Hellman minimum acceptable size"
When establishing a TLS session, if a ciphersuite which uses Diffie-Hellman
key agreement is negotiated, the server will provide a large prime number
for use.  This option establishes the minimum acceptable size of that number.
If the parameter offered by the server is too small, then the TLS handshake
will fail.

Only supported when using GnuTLS.


.option tls_privatekey smtp string&!! unset
.cindex "TLS" "client private key, location of"
.vindex "&$host$&"
.vindex "&$host_address$&"
The value of this option must be the absolute path to a file which contains the
client's private key. This is used when sending a message over an encrypted
connection using a client certificate. The values of &$host$& and
&$host_address$& are set to the name and address of the server during the
expansion. If this option is unset, or the expansion is forced to fail, or the
result is an empty string, the private key is assumed to be in the same file as
the certificate. See chapter &<<CHAPTLS>>& for details of TLS.


.option tls_require_ciphers smtp string&!! unset
.cindex "TLS" "requiring specific ciphers"
.cindex "cipher" "requiring specific"
.vindex "&$host$&"
.vindex "&$host_address$&"
The value of this option must be a list of permitted cipher suites, for use
when setting up an outgoing encrypted connection. (There is a global option of
the same name for controlling incoming connections.) The values of &$host$& and
&$host_address$& are set to the name and address of the server during the
expansion. See chapter &<<CHAPTLS>>& for details of TLS; note that this option
is used in different ways by OpenSSL and GnuTLS (see sections
&<<SECTreqciphssl>>& and &<<SECTreqciphgnu>>&). For GnuTLS, the order of the
ciphers is a preference order.


.option tls_resumption_hosts smtp "host list&!!" unset
.cindex TLS resumption
This option controls which connections to use the TLS resumption feature.
See &<<SECTresumption>>& for details.



.option tls_sni smtp string&!! unset
.cindex "TLS" "Server Name Indication"
.cindex "TLS" SNI
.cindex SNI "setting in client"
.vindex "&$tls_sni$&"
If this option is set
and the connection is not DANE-validated
then it sets the $tls_out_sni variable and causes any
TLS session to pass this value as the Server Name Indication extension to
the remote side, which can be used by the remote side to select an appropriate
certificate and private key for the session.

See &<<SECTtlssni>>& for more information.

Note that for OpenSSL, this feature requires a build of OpenSSL that supports
TLS extensions.




.option tls_tempfail_tryclear smtp boolean true
.cindex "4&'xx'& responses" "to STARTTLS"
When the server host is not in &%hosts_require_tls%&, and there is a problem in
setting up a TLS session, this option determines whether or not Exim should try
to deliver the message unencrypted. If it is set false, delivery to the
current host is deferred; if there are other hosts, they are tried. If this
option is set true, Exim attempts to deliver unencrypted after a 4&'xx'&
response to STARTTLS. Also, if STARTTLS is accepted, but the subsequent
TLS negotiation fails, Exim closes the current connection (because it is in an
unknown state), opens a new one to the same host, and then tries the delivery
in clear.


.option tls_try_verify_hosts smtp "host list&!!" *
.cindex "TLS" "server certificate verification"
.cindex "certificate" "verification of server"
This option gives a list of hosts for which, on encrypted connections,
certificate verification will be tried but need not succeed.
The &%tls_verify_certificates%& option must also be set.
Note that unless the host is in this list
TLS connections will be denied to hosts using self-signed certificates
when &%tls_verify_certificates%& is matched.
The &$tls_out_certificate_verified$& variable is set when
certificate verification succeeds.


.option tls_verify_cert_hostnames smtp "host list&!!" *
.cindex "TLS" "server certificate hostname verification"
.cindex "certificate" "verification of server"
This option give a list of hosts for which,
while verifying the server certificate,
checks will be included on the host name
(note that this will generally be the result of a DNS MX lookup)
versus the Subject-Alternate-Name (or, if none, Subject-Name) fields.
Wildcard names are permitted,
limited to being the initial component of a 3-or-more component FQDN.

There is no equivalent checking on client certificates.


.option tls_verify_certificates smtp string&!! system
.cindex "TLS" "server certificate verification"
.cindex "certificate" "verification of server"
.vindex "&$host$&"
.vindex "&$host_address$&"
The value of this option must be either the
word "system"
or the absolute path to
a file or directory containing permitted certificates for servers,
for use when setting up an encrypted connection.

The "system" value for the option will use a location compiled into the SSL library.
This is not available for GnuTLS versions preceding 3.0.20; a value of "system"
is taken as empty and an explicit location
must be specified.

The use of a directory for the option value is not available for GnuTLS versions
preceding 3.3.6 and a single file must be used.

With OpenSSL the certificates specified
explicitly
either by file or directory
are added to those given by the system default location.

The values of &$host$& and
&$host_address$& are set to the name and address of the server during the
expansion of this option. See chapter &<<CHAPTLS>>& for details of TLS.

For back-compatibility,
if neither tls_verify_hosts nor tls_try_verify_hosts are set
(a single-colon empty list counts as being set)
and certificate verification fails the TLS connection is closed.


.option tls_verify_hosts smtp "host list&!!" unset
.cindex "TLS" "server certificate verification"
.cindex "certificate" "verification of server"
This option gives a list of hosts for which, on encrypted connections,
certificate verification must succeed.
The &%tls_verify_certificates%& option must also be set.
If both this option and &%tls_try_verify_hosts%& are unset
operation is as if this option selected all hosts.
&*Warning*&: Including a host in &%tls_verify_hosts%& does not require
that connections use TLS.
Fallback to in-clear communication will be done unless restricted by
the &%hosts_require_tls%& option.

.option utf8_downconvert smtp integer&!! -1
.cindex utf8 "address downconversion"
.cindex i18n "utf8 address downconversion"
If built with internationalization support,
this option controls conversion of UTF-8 in message envelope addresses
to a-label form.
If, after expansion, the value is 1, 0, or -1 then this value overrides
any value previously set for the message.  Otherwise, any previously
set value is used.  To permit use of a previous value,
set this option to an empty string.
For details on the values see section &<<SECTi18nMTA>>&.




.section "How the limits for the number of hosts to try are used" &&&
         "SECTvalhosmax"
.cindex "host" "maximum number to try"
.cindex "limit" "hosts; maximum number tried"
There are two options that are concerned with the number of hosts that are
tried when an SMTP delivery takes place. They are &%hosts_max_try%& and
&%hosts_max_try_hardlimit%&.


The &%hosts_max_try%& option limits the number of hosts that are tried
for a single delivery. However, despite the term &"host"& in its name, the
option actually applies to each IP address independently. In other words, a
multihomed host is treated as several independent hosts, just as it is for
retrying.

Many of the larger ISPs have multiple MX records which often point to
multihomed hosts. As a result, a list of a dozen or more IP addresses may be
created as a result of routing one of these domains.

Trying every single IP address on such a long list does not seem sensible; if
several at the top of the list fail, it is reasonable to assume there is some
problem that is likely to affect all of them. Roughly speaking, the value of
&%hosts_max_try%& is the maximum number that are tried before deferring the
delivery. However, the logic cannot be quite that simple.

Firstly, IP addresses that are skipped because their retry times have not
arrived do not count, and in addition, addresses that are past their retry
limits are also not counted, even when they are tried. This means that when
some IP addresses are past their retry limits, more than the value of
&%hosts_max_retry%& may be tried. The reason for this behaviour is to ensure
that all IP addresses are considered before timing out an email address (but
see below for an exception).

Secondly, when the &%hosts_max_try%& limit is reached, Exim looks down the host
list to see if there is a subsequent host with a different (higher valued) MX.
If there is, that host is considered next, and the current IP address is used
but not counted. This behaviour helps in the case of a domain with a retry rule
that hardly ever delays any hosts, as is now explained:

Consider the case of a long list of hosts with one MX value, and a few with a
higher MX value. If &%hosts_max_try%& is small (the default is 5) only a few
hosts at the top of the list are tried at first. With the default retry rule,
which specifies increasing retry times, the higher MX hosts are eventually
tried when those at the top of the list are skipped because they have not
reached their retry times.

However, it is common practice to put a fixed short retry time on domains for
large ISPs, on the grounds that their servers are rarely down for very long.
Unfortunately, these are exactly the domains that tend to resolve to long lists
of hosts. The short retry time means that the lowest MX hosts are tried every
time. The attempts may be in a different order because of random sorting, but
without the special MX check, the higher MX hosts would never be tried until
all the lower MX hosts had timed out (which might be several days), because
there are always some lower MX hosts that have reached their retry times. With
the special check, Exim considers at least one IP address from each MX value at
every delivery attempt, even if the &%hosts_max_try%& limit has already been
reached.

The above logic means that &%hosts_max_try%& is not a hard limit, and in
particular, Exim normally eventually tries all the IP addresses before timing
out an email address. When &%hosts_max_try%& was implemented, this seemed a
reasonable thing to do. Recently, however, some lunatic DNS configurations have
been set up with hundreds of IP addresses for some domains. It can
take a very long time indeed for an address to time out in these cases.

The &%hosts_max_try_hardlimit%& option was added to help with this problem.
Exim never tries more than this number of IP addresses; if it hits this limit
and they are all timed out, the email address is bounced, even though not all
possible IP addresses have been tried.
.ecindex IIDsmttra1
.ecindex IIDsmttra2





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Address rewriting" "CHAPrewrite"
.scindex IIDaddrew "rewriting" "addresses"
There are some circumstances in which Exim automatically rewrites domains in
addresses. The two most common are when an address is given without a domain
(referred to as an &"unqualified address"&) or when an address contains an
abbreviated domain that is expanded by DNS lookup.

Unqualified envelope addresses are accepted only for locally submitted
messages, or for messages that are received from hosts matching
&%sender_unqualified_hosts%& or &%recipient_unqualified_hosts%&, as
appropriate. Unqualified addresses in header lines are qualified if they are in
locally submitted messages, or messages from hosts that are permitted to send
unqualified envelope addresses. Otherwise, unqualified addresses in header
lines are neither qualified nor rewritten.

One situation in which Exim does &'not'& automatically rewrite a domain is
when it is the name of a CNAME record in the DNS. The older RFCs suggest that
such a domain should be rewritten using the &"canonical"& name, and some MTAs
do this. The new RFCs do not contain this suggestion.


.section "Explicitly configured address rewriting" "SECID147"
This chapter describes the rewriting rules that can be used in the
main rewrite section of the configuration file, and also in the generic
&%headers_rewrite%& option that can be set on any transport.

Some people believe that configured address rewriting is a Mortal Sin.
Others believe that life is not possible without it. Exim provides the
facility; you do not have to use it.

The main rewriting rules that appear in the &"rewrite"& section of the
configuration file are applied to addresses in incoming messages, both envelope
addresses and addresses in header lines. Each rule specifies the types of
address to which it applies.

Whether or not addresses in header lines are rewritten depends on the origin of
the headers and the type of rewriting. Global rewriting, that is, rewriting
rules from the rewrite section of the configuration file, is applied only to
those headers that were received with the message. Header lines that are added
by ACLs or by a system filter or by individual routers or transports (which
are specific to individual recipient addresses) are not rewritten by the global
rules.

Rewriting at transport time, by means of the &%headers_rewrite%& option,
applies all headers except those added by routers and transports. That is, as
well as the headers that were received with the message, it also applies to
headers that were added by an ACL or a system filter.


In general, rewriting addresses from your own system or domain has some
legitimacy. Rewriting other addresses should be done only with great care and
in special circumstances. The author of Exim believes that rewriting should be
used sparingly, and mainly for &"regularizing"& addresses in your own domains.
Although it can sometimes be used as a routing tool, this is very strongly
discouraged.

There are two commonly encountered circumstances where rewriting is used, as
illustrated by these examples:

.ilist
The company whose domain is &'hitch.fict.example'& has a number of hosts that
exchange mail with each other behind a firewall, but there is only a single
gateway to the outer world. The gateway rewrites &'*.hitch.fict.example'& as
&'hitch.fict.example'& when sending mail off-site.
.next
A host rewrites the local parts of its own users so that, for example,
&'fp42@hitch.fict.example'& becomes &'Ford.Prefect@hitch.fict.example'&.
.endlist



.section "When does rewriting happen?" "SECID148"
.cindex "rewriting" "timing of"
.cindex "&ACL;" "rewriting addresses in"
Configured address rewriting can take place at several different stages of a
message's processing.

.vindex "&$sender_address$&"
At the start of an ACL for MAIL, the sender address may have been rewritten
by a special SMTP-time rewrite rule (see section &<<SSECTrewriteS>>&), but no
ordinary rewrite rules have yet been applied. If, however, the sender address
is verified in the ACL, it is rewritten before verification, and remains
rewritten thereafter. The subsequent value of &$sender_address$& is the
rewritten address. This also applies if sender verification happens in a
RCPT ACL. Otherwise, when the sender address is not verified, it is
rewritten as soon as a message's header lines have been received.

.vindex "&$domain$&"
.vindex "&$local_part$&"
Similarly, at the start of an ACL for RCPT, the current recipient's address
may have been rewritten by a special SMTP-time rewrite rule, but no ordinary
rewrite rules have yet been applied to it. However, the behaviour is different
from the sender address when a recipient is verified. The address is rewritten
for the verification, but the rewriting is not remembered at this stage. The
value of &$local_part$& and &$domain$& after verification are always the same
as they were before (that is, they contain the unrewritten &-- except for
SMTP-time rewriting &-- address).

As soon as a message's header lines have been received, all the envelope
recipient addresses are permanently rewritten, and rewriting is also applied to
the addresses in the header lines (if configured). This happens before adding
any header lines that were specified in MAIL or RCPT ACLs, and
.cindex "&[local_scan()]& function" "address rewriting; timing of"
before the DATA ACL and &[local_scan()]& functions are run.

When an address is being routed, either for delivery or for verification,
rewriting is applied immediately to child addresses that are generated by
redirection, unless &%no_rewrite%& is set on the router.

.cindex "envelope from"
.cindex "envelope sender" "rewriting at transport time"
.cindex "rewriting" "at transport time"
.cindex "header lines" "rewriting at transport time"
At transport time, additional rewriting of addresses in header lines can be
specified by setting the generic &%headers_rewrite%& option on a transport.
This option contains rules that are identical in form to those in the rewrite
section of the configuration file. They are applied to the original message
header lines and any that were added by ACLs or a system filter. They are not
applied to header lines that are added by routers or the transport.

The outgoing envelope sender can be rewritten by means of the &%return_path%&
transport option. However, it is not possible to rewrite envelope recipients at
transport time.




.section "Testing the rewriting rules that apply on input" "SECID149"
.cindex "rewriting" "testing"
.cindex "testing" "rewriting"
Exim's input rewriting configuration appears in a part of the runtime
configuration file headed by &"begin rewrite"&. It can be tested by the
&%-brw%& command line option. This takes an address (which can be a full
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address) as its argument. The output is a list of how the address would be
transformed by the rewriting rules for each of the different places it might
appear in an incoming message, that is, for each different header and for the
envelope sender and recipient fields. For example,
.code
exim -brw ph10@exim.workshop.example
.endd
might produce the output
.code
sender: Philip.Hazel@exim.workshop.example
from: Philip.Hazel@exim.workshop.example
to: ph10@exim.workshop.example
cc: ph10@exim.workshop.example
bcc: ph10@exim.workshop.example
reply-to: Philip.Hazel@exim.workshop.example
env-from: Philip.Hazel@exim.workshop.example
env-to: ph10@exim.workshop.example
.endd
which shows that rewriting has been set up for that address when used in any of
the source fields, but not when it appears as a recipient address. At the
present time, there is no equivalent way of testing rewriting rules that are
set for a particular transport.


.section "Rewriting rules" "SECID150"
.cindex "rewriting" "rules"
The rewrite section of the configuration file consists of lines of rewriting
rules in the form
.display
<&'source pattern'&>  <&'replacement'&>  <&'flags'&>
.endd
Rewriting rules that are specified for the &%headers_rewrite%& generic
transport option are given as a colon-separated list. Each item in the list
takes the same form as a line in the main rewriting configuration (except that
any colons must be doubled, of course).

The formats of source patterns and replacement strings are described below.
Each is terminated by white space, unless enclosed in double quotes, in which
case normal quoting conventions apply inside the quotes. The flags are single
characters which may appear in any order. Spaces and tabs between them are
ignored.

For each address that could potentially be rewritten, the rules are scanned in
order, and replacements for the address from earlier rules can themselves be
replaced by later rules (but see the &"q"& and &"R"& flags).

The order in which addresses are rewritten is undefined, may change between
releases, and must not be relied on, with one exception: when a message is
received, the envelope sender is always rewritten first, before any header
lines are rewritten. For example, the replacement string for a rewrite of an
address in &'To:'& must not assume that the message's address in &'From:'& has
(or has not) already been rewritten. However, a rewrite of &'From:'& may assume
that the envelope sender has already been rewritten.

.vindex "&$domain$&"
.vindex "&$local_part$&"
The variables &$local_part$& and &$domain$& can be used in the replacement
string to refer to the address that is being rewritten. Note that lookup-driven
rewriting can be done by a rule of the form
.code
*@*   ${lookup ...
.endd
where the lookup key uses &$1$& and &$2$& or &$local_part$& and &$domain$& to
refer to the address that is being rewritten.


.section "Rewriting patterns" "SECID151"
.cindex "rewriting" "patterns"
.cindex "address list" "in a rewriting pattern"
The source pattern in a rewriting rule is any item which may appear in an
address list (see section &<<SECTaddresslist>>&). It is in fact processed as a
single-item address list, which means that it is expanded before being tested
against the address. As always, if you use a regular expression as a pattern,
you must take care to escape dollar and backslash characters, or use the &`\N`&
facility to suppress string expansion within the regular expression.

Domains in patterns should be given in lower case. Local parts in patterns are
case-sensitive. If you want to do case-insensitive matching of local parts, you
can use a regular expression that starts with &`^(?i)`&.

.cindex "numerical variables (&$1$& &$2$& etc)" "in rewriting rules"
After matching, the numerical variables &$1$&, &$2$&, etc. may be set,
depending on the type of match which occurred. These can be used in the
replacement string to insert portions of the incoming address. &$0$& always
refers to the complete incoming address. When a regular expression is used, the
numerical variables are set from its capturing subexpressions. For other types
of pattern they are set as follows:

.ilist
If a local part or domain starts with an asterisk, the numerical variables
refer to the character strings matched by asterisks, with &$1$& associated with
the first asterisk, and &$2$& with the second, if present. For example, if the
pattern
.code
*queen@*.fict.example
.endd
is matched against the address &'hearts-queen@wonderland.fict.example'& then
.code
$0 = hearts-queen@wonderland.fict.example
$1 = hearts-
$2 = wonderland
.endd
Note that if the local part does not start with an asterisk, but the domain
does, it is &$1$& that contains the wild part of the domain.

.next
If the domain part of the pattern is a partial lookup, the wild and fixed parts
of the domain are placed in the next available numerical variables. Suppose,
for example, that the address &'foo@bar.baz.example'& is processed by a
rewriting rule of the form
.display
&`*@partial-dbm;/some/dbm/file`&    <&'replacement string'&>
.endd
and the key in the file that matches the domain is &`*.baz.example`&. Then
.code
$1 = foo
$2 = bar
$3 = baz.example
.endd
If the address &'foo@baz.example'& is looked up, this matches the same
wildcard file entry, and in this case &$2$& is set to the empty string, but
&$3$& is still set to &'baz.example'&. If a non-wild key is matched in a
partial lookup, &$2$& is again set to the empty string and &$3$& is set to the
whole domain. For non-partial domain lookups, no numerical variables are set.
.endlist


.section "Rewriting replacements" "SECID152"
.cindex "rewriting" "replacements"
If the replacement string for a rule is a single asterisk, addresses that
match the pattern and the flags are &'not'& rewritten, and no subsequent
rewriting rules are scanned. For example,
.code
hatta@lookingglass.fict.example  *  f
.endd
specifies that &'hatta@lookingglass.fict.example'& is never to be rewritten in
&'From:'& headers.

.vindex "&$domain$&"
.vindex "&$local_part$&"
If the replacement string is not a single asterisk, it is expanded, and must
yield a fully qualified address. Within the expansion, the variables
&$local_part$& and &$domain$& refer to the address that is being rewritten.
Any letters they contain retain their original case &-- they are not lower
cased. The numerical variables are set up according to the type of pattern that
matched the address, as described above. If the expansion is forced to fail by
the presence of &"fail"& in a conditional or lookup item, rewriting by the
current rule is abandoned, but subsequent rules may take effect. Any other
expansion failure causes the entire rewriting operation to be abandoned, and an
entry written to the panic log.



.subsection "Rewriting flags" "SSECID153"
There are three different kinds of flag that may appear on rewriting rules:

.ilist
Flags that specify which headers and envelope addresses to rewrite: E, F, T, b,
c, f, h, r, s, t.
.next
A flag that specifies rewriting at SMTP time: S.
.next
Flags that control the rewriting process: Q, q, R, w.
.endlist

For rules that are part of the &%headers_rewrite%& generic transport option,
E, F, T, and S are not permitted.



.subsection "Flags specifying which headers and envelope addresses to rewrite" &&&
         "SSECID154"
.cindex rewriting flags
If none of the following flag letters, nor the &"S"& flag (see section
&<<SSECTrewriteS>>&) are present, a main rewriting rule applies to all headers
and to both the sender and recipient fields of the envelope, whereas a
transport-time rewriting rule just applies to all headers. Otherwise, the
rewriting rule is skipped unless the relevant addresses are being processed.
.display
&`E`&       rewrite all envelope fields
&`F`&       rewrite the envelope From field
&`T`&       rewrite the envelope To field
&`b`&       rewrite the &'Bcc:'& header
&`c`&       rewrite the &'Cc:'& header
&`f`&       rewrite the &'From:'& header
&`h`&       rewrite all headers
&`r`&       rewrite the &'Reply-To:'& header
&`s`&       rewrite the &'Sender:'& header
&`t`&       rewrite the &'To:'& header
.endd
"All headers" means all of the headers listed above that can be selected
individually, plus their &'Resent-'& versions. It does not include
other headers such as &'Subject:'& etc.

You should be particularly careful about rewriting &'Sender:'& headers, and
restrict this to special known cases in your own domains.


.subsection "The SMTP-time rewriting flag" SSECTrewriteS
.cindex SMTP "rewriting malformed addresses"
.cindex RCPT "rewriting argument of"
.cindex MAIL "rewriting argument of"
The rewrite flag &"S"& specifies a rewrite of incoming envelope addresses at
SMTP time, as soon as an address is received in a MAIL or RCPT command, and
before any other processing; even before syntax checking. The pattern is
required to be a regular expression, and it is matched against the whole of the
data for the command, including any surrounding angle brackets.

.vindex "&$domain$&"
.vindex "&$local_part$&"
This form of rewrite rule allows for the handling of addresses that are not
compliant with RFCs
&url(https://www.rfc-editor.org/rfc/rfc2821,2821) and
&url(https://www.rfc-editor.org/rfc/rfc2822,2822)
(for example, &"bang paths"& in batched SMTP
input). Because the input is not required to be a syntactically valid address,
the variables &$local_part$& and &$domain$& are not available during the
expansion of the replacement string. The result of rewriting replaces the
original address in the MAIL or RCPT command.


.subsection "Flags controlling the rewriting process" SSECID155
There are four flags which control the way the rewriting process works. These
take effect only when a rule is invoked, that is, when the address is of the
correct type (matches the flags) and matches the pattern:

.ilist
If the &"Q"& flag is set on a rule, the rewritten address is permitted to be an
unqualified local part. It is qualified with &%qualify_recipient%&. In the
absence of &"Q"& the rewritten address must always include a domain.
.next
If the &"q"& flag is set on a rule, no further rewriting rules are considered,
even if no rewriting actually takes place because of a &"fail"& in the
expansion. The &"q"& flag is not effective if the address is of the wrong type
(does not match the flags) or does not match the pattern.
.next
The &"R"& flag causes a successful rewriting rule to be re-applied to the new
address, up to ten times. It can be combined with the &"q"& flag, to stop
rewriting once it fails to match (after at least one successful rewrite).
.next
.cindex "rewriting" "whole addresses"
When an address in a header is rewritten, the rewriting normally applies only
to the working part of the address, with any comments and
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
&"phrase"& left unchanged. For example, rewriting might change
.code
From: Ford Prefect <fp42@restaurant.hitch.fict.example>
.endd
into
.code
From: Ford Prefect <prefectf@hitch.fict.example>
.endd
.cindex "RFC 2047"
Sometimes there is a need to replace the whole address item, and this can be
done by adding the flag letter &"w"& to a rule. If this is set on a rule that
causes an address in a header line to be rewritten, the entire address is
replaced, not just the working part. The replacement must be a complete
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address, including the angle brackets if necessary. If text outside angle
brackets contains a character whose value is greater than 126 or less than 32
(except for tab), the text is encoded according to
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047). The character set
is taken from &%headers_charset%&, which gets its default at build time.

When the &"w"& flag is set on a rule that causes an envelope address to be
rewritten, all but the working part of the replacement address is discarded.
.endlist


.section "Rewriting examples" "SECID156"
Here is an example of the two common rewriting paradigms:
.code
*@*.hitch.fict.example  $1@hitch.fict.example
*@hitch.fict.example    ${lookup{$1}dbm{/etc/realnames}\
                     {$value}fail}@hitch.fict.example bctfrF
.endd
Note the use of &"fail"& in the lookup expansion in the second rule, forcing
the string expansion to fail if the lookup does not succeed. In this context it
has the effect of leaving the original address unchanged, but Exim goes on to
consider subsequent rewriting rules, if any, because the &"q"& flag is not
present in that rule. An alternative to &"fail"& would be to supply &$1$&
explicitly, which would cause the rewritten address to be the same as before,
at the cost of a small bit of processing. Not supplying either of these is an
error, since the rewritten address would then contain no local part.

The first example above replaces the domain with a superior, more general
domain. This may not be desirable for certain local parts. If the rule
.code
root@*.hitch.fict.example  *
.endd
were inserted before the first rule, rewriting would be suppressed for the
local part &'root'& at any domain ending in &'hitch.fict.example'&.

Rewriting can be made conditional on a number of tests, by making use of
&${if$& in the expansion item. For example, to apply a rewriting rule only to
messages that originate outside the local host:
.code
*@*.hitch.fict.example  "${if !eq {$sender_host_address}{}\
                         {$1@hitch.fict.example}fail}"
.endd
The replacement string is quoted in this example because it contains white
space.

.cindex "rewriting" "bang paths"
.cindex "bang paths" "rewriting"
Exim does not handle addresses in the form of &"bang paths"&. If it sees such
an address it treats it as an unqualified local part which it qualifies with
the local qualification domain (if the source of the message is local or if the
remote host is permitted to send unqualified addresses). Rewriting can
sometimes be used to handle simple bang paths with a fixed number of
components. For example, the rule
.code
\N^([^!]+)!(.*)@your.domain.example$\N   $2@$1
.endd
rewrites a two-component bang path &'host.name!user'& as the domain address
&'user@host.name'&. However, there is a security implication in using this as
a global rewriting rule for envelope addresses. It can provide a backdoor
method for using your system as a relay, because the incoming addresses appear
to be local. If the bang path addresses are received via SMTP, it is safer to
use the &"S"& flag to rewrite them as they are received, so that relay checking
can be done on the rewritten addresses.
.ecindex IIDaddrew





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Retry configuration" "CHAPretry"
.scindex IIDretconf1 "retry" "configuration, description of"
.scindex IIDregconf2 "configuration file" "retry section"
The &"retry"& section of the runtime configuration file contains a list of
retry rules that control how often Exim tries to deliver messages that cannot
be delivered at the first attempt. If there are no retry rules (the section is
empty or not present), there are no retries. In this situation, temporary
errors are treated as permanent. The default configuration contains a single,
general-purpose retry rule (see section &<<SECID57>>&). The &%-brt%& command
line option can be used to test which retry rule will be used for a given
address, domain and error.

The most common cause of retries is temporary failure to deliver to a remote
host because the host is down, or inaccessible because of a network problem.
Exim's retry processing in this case is applied on a per-host (strictly, per IP
address) basis, not on a per-message basis. Thus, if one message has recently
been delayed, delivery of a new message to the same host is not immediately
tried, but waits for the host's retry time to arrive. If the &%retry_defer%&
log selector is set, the message
.cindex "retry" "time not reached"
&"retry time not reached"& is written to the main log whenever a delivery is
skipped for this reason. Section &<<SECToutSMTPerr>>& contains more details of
the handling of errors during remote deliveries.

Retry processing applies to routing as well as to delivering, except as covered
in the next paragraph. The retry rules do not distinguish between these
actions. It is not possible, for example, to specify different behaviour for
failures to route the domain &'snark.fict.example'& and failures to deliver to
the host &'snark.fict.example'&. I didn't think anyone would ever need this
added complication, so did not implement it. However, although they share the
same retry rule, the actual retry times for routing and transporting a given
domain are maintained independently.

When a delivery is not part of a queue run (typically an immediate delivery on
receipt of a message), the routers are always run, and local deliveries are
always attempted, even if retry times are set for them. This makes for better
behaviour if one particular message is causing problems (for example, causing
quota overflow, or provoking an error in a filter file). If such a delivery
suffers a temporary failure, the retry data is updated as normal, and
subsequent delivery attempts from queue runs occur only when the retry time for
the local address is reached.

.section "Changing retry rules" "SECID157"
If you change the retry rules in your configuration, you should consider
whether or not to delete the retry data that is stored in Exim's spool area in
files with names like &_db/retry_&. Deleting any of Exim's hints files is
always safe; that is why they are called &"hints"&.

The hints retry data contains suggested retry times based on the previous
rules. In the case of a long-running problem with a remote host, it might
record the fact that the host has timed out. If your new rules increase the
timeout time for such a host, you should definitely remove the old retry data
and let Exim recreate it, based on the new rules. Otherwise Exim might bounce
messages that it should now be retaining.



.section "Format of retry rules" "SECID158"
.cindex "retry" "rules"
Each retry rule occupies one line and consists of three or four parts,
separated by white space: a pattern, an error name, an optional list of sender
addresses, and a list of retry parameters. The pattern and sender lists must be
enclosed in double quotes if they contain white space. The rules are searched
in order until one is found where the pattern, error name, and sender list (if
present) match the failing host or address, the error that occurred, and the
message's sender, respectively.


The pattern is any single item that may appear in an address list (see section
&<<SECTaddresslist>>&). It is in fact processed as a one-item address list,
which means that it is expanded before being tested against the address that
has been delayed. A negated address list item is permitted. Address
list processing treats a plain domain name as if it were preceded by &"*@"&,
which makes it possible for many retry rules to start with just a domain. For
example,
.code
lookingglass.fict.example        *  F,24h,30m;
.endd
provides a rule for any address in the &'lookingglass.fict.example'& domain,
whereas
.code
alice@lookingglass.fict.example  *  F,24h,30m;
.endd
applies only to temporary failures involving the local part &%alice%&.
In practice, almost all rules start with a domain name pattern without a local
part.

.cindex "regular expressions" "in retry rules"
&*Warning*&: If you use a regular expression in a retry rule pattern, it
must match a complete address, not just a domain, because that is how regular
expressions work in address lists.
.display
&`^\Nxyz\d+\.abc\.example$\N        *  G,1h,10m,2`&     &%Wrong%&
&`^\N[^@]+@xyz\d+\.abc\.example$\N  *  G,1h,10m,2`&     &%Right%&
.endd


.section "Choosing which retry rule to use for address errors" "SECID159"
When Exim is looking for a retry rule after a routing attempt has failed (for
example, after a DNS timeout), each line in the retry configuration is tested
against the complete address only if &%retry_use_local_part%& is set for the
router. Otherwise, only the domain is used, except when matching against a
regular expression, when the local part of the address is replaced with &"*"&.
A domain on its own can match a domain pattern, or a pattern that starts with
&"*@"&. By default, &%retry_use_local_part%& is true for routers where
&%check_local_user%& is true, and false for other routers.

Similarly, when Exim is looking for a retry rule after a local delivery has
failed (for example, after a mailbox full error), each line in the retry
configuration is tested against the complete address only if
&%retry_use_local_part%& is set for the transport (it defaults true for all
local transports).

.cindex "4&'xx'& responses" "retry rules for"
However, when Exim is looking for a retry rule after a remote delivery attempt
suffers an address error (a 4&'xx'& SMTP response for a recipient address), the
whole address is always used as the key when searching the retry rules. The
rule that is found is used to create a retry time for the combination of the
failing address and the message's sender. It is the combination of sender and
recipient that is delayed in subsequent queue runs until its retry time is
reached. You can delay the recipient without regard to the sender by setting
&%address_retry_include_sender%& false in the &(smtp)& transport but this can
lead to problems with servers that regularly issue 4&'xx'& responses to RCPT
commands.



.section "Choosing which retry rule to use for host and message errors" &&&
         "SECID160"
For a temporary error that is not related to an individual address (for
example, a connection timeout), each line in the retry configuration is checked
twice. First, the name of the remote host is used as a domain name (preceded by
&"*@"& when matching a regular expression). If this does not match the line,
the domain from the email address is tried in a similar fashion. For example,
suppose the MX records for &'a.b.c.example'& are
.code
a.b.c.example  MX  5  x.y.z.example
               MX  6  p.q.r.example
               MX  7  m.n.o.example
.endd
and the retry rules are
.code
p.q.r.example    *      F,24h,30m;
a.b.c.example    *      F,4d,45m;
.endd
and a delivery to the host &'x.y.z.example'& suffers a connection failure. The
first rule matches neither the host nor the domain, so Exim looks at the second
rule. This does not match the host, but it does match the domain, so it is used
to calculate the retry time for the host &'x.y.z.example'&. Meanwhile, Exim
tries to deliver to &'p.q.r.example'&. If this also suffers a host error, the
first retry rule is used, because it matches the host.

In other words, temporary failures to deliver to host &'p.q.r.example'& use the
first rule to determine retry times, but for all the other hosts for the domain
&'a.b.c.example'&, the second rule is used. The second rule is also used if
routing to &'a.b.c.example'& suffers a temporary failure.

&*Note*&: The host name is used when matching the patterns, not its IP address.
However, if a message is routed directly to an IP address without the use of a
host name, for example, if a &(manualroute)& router contains a setting such as:
.code
route_list = *.a.example  192.168.34.23
.endd
then the &"host name"& that is used when searching for a retry rule is the
textual form of the IP address.

.section "Retry rules for specific errors" "SECID161"
.cindex "retry" "specific errors; specifying"
The second field in a retry rule is the name of a particular error, or an
asterisk, which matches any error. The errors that can be tested for are:

.vlist
.vitem &%auth_failed%&
Authentication failed when trying to send to a host in the
&%hosts_require_auth%& list in an &(smtp)& transport.

.vitem &%data_4xx%&
A 4&'xx'& error was received for an outgoing DATA command, either immediately
after the command, or after sending the message's data.

.vitem &%mail_4xx%&
A 4&'xx'& error was received for an outgoing MAIL command.

.vitem &%rcpt_4xx%&
A 4&'xx'& error was received for an outgoing RCPT command.
.endlist

For the three 4&'xx'& errors, either the first or both of the x's can be given
as specific digits, for example: &`mail_45x`& or &`rcpt_436`&. For example, to
recognize 452 errors given to RCPT commands for addresses in a certain domain,
and have retries every ten minutes with a one-hour timeout, you could set up a
retry rule of this form:
.code
the.domain.name  rcpt_452   F,1h,10m
.endd
These errors apply to both outgoing SMTP (the &(smtp)& transport) and outgoing
LMTP (either the &(lmtp)& transport, or the &(smtp)& transport in LMTP mode).

.vlist
.vitem &%lost_connection%&
A server unexpectedly closed the SMTP connection. There may, of course,
legitimate reasons for this (host died, network died), but if it repeats a lot
for the same host, it indicates something odd.

.vitem &%lookup%&
A DNS lookup for a host failed.
Note that a &%dnslookup%& router will need to have matched
its &%fail_defer_domains%& option for this retry type to be usable.
Also note that a &%manualroute%& router will probably need
its &%host_find_failed%& option set to &%defer%&.

.vitem &%refused_MX%&
A connection to a host obtained from an MX record was refused.

.vitem &%refused_A%&
A connection to a host not obtained from an MX record was refused.

.vitem &%refused%&
A connection was refused.

.vitem &%timeout_connect_MX%&
A connection attempt to a host obtained from an MX record timed out.

.vitem &%timeout_connect_A%&
A connection attempt to a host not obtained from an MX record timed out.

.vitem &%timeout_connect%&
A connection attempt timed out.

.vitem &%timeout_MX%&
There was a timeout while connecting or during an SMTP session with a host
obtained from an MX record.

.vitem &%timeout_A%&
There was a timeout while connecting or during an SMTP session with a host not
obtained from an MX record.

.vitem &%timeout%&
There was a timeout while connecting or during an SMTP session.

.vitem &%tls_required%&
The server was required to use TLS (it matched &%hosts_require_tls%& in the
&(smtp)& transport), but either did not offer TLS, or it responded with 4&'xx'&
to STARTTLS, or there was a problem setting up the TLS connection.

.vitem &%quota%&
A mailbox quota was exceeded in a local delivery by the &(appendfile)&
transport.

.vitem &%quota_%&<&'time'&>
.cindex "quota" "error testing in retry rule"
.cindex "retry" "quota error testing"
A mailbox quota was exceeded in a local delivery by the &(appendfile)&
transport, and the mailbox has not been accessed for <&'time'&>. For example,
&'quota_4d'& applies to a quota error when the mailbox has not been accessed
for four days.
.endlist

.cindex "mailbox" "time of last read"
The idea of &%quota_%&<&'time'&> is to make it possible to have shorter
timeouts when the mailbox is full and is not being read by its owner. Ideally,
it should be based on the last time that the user accessed the mailbox.
However, it is not always possible to determine this. Exim uses the following
heuristic rules:

.ilist
If the mailbox is a single file, the time of last access (the &"atime"&) is
used. As no new messages are being delivered (because the mailbox is over
quota), Exim does not access the file, so this is the time of last user access.
.next
.cindex "maildir format" "time of last read"
For a maildir delivery, the time of last modification of the &_new_&
subdirectory is used. As the mailbox is over quota, no new files are created in
the &_new_& subdirectory, because no new messages are being delivered. Any
change to the &_new_& subdirectory is therefore assumed to be the result of an
MUA moving a new message to the &_cur_& directory when it is first read. The
time that is used is therefore the last time that the user read a new message.
.next
For other kinds of multi-file mailbox, the time of last access cannot be
obtained, so a retry rule that uses this type of error field is never matched.
.endlist

The quota errors apply both to system-enforced quotas and to Exim's own quota
mechanism in the &(appendfile)& transport. The &'quota'& error also applies
when a local delivery is deferred because a partition is full (the ENOSPC
error).



.section "Retry rules for specified senders" "SECID162"
.cindex "retry" "rules; sender-specific"
You can specify retry rules that apply only when the failing message has a
specific sender. In particular, this can be used to define retry rules that
apply only to bounce messages. The third item in a retry rule can be of this
form:
.display
&`senders=`&<&'address list'&>
.endd
The retry timings themselves are then the fourth item. For example:
.code
*   rcpt_4xx   senders=:   F,1h,30m
.endd
matches recipient 4&'xx'& errors for bounce messages sent to any address at any
host. If the address list contains white space, it must be enclosed in quotes.
For example:
.code
a.domain  rcpt_452  senders="xb.dom : yc.dom"  G,8h,10m,1.5
.endd
&*Warning*&: This facility can be unhelpful if it is used for host errors
(which do not depend on the recipient). The reason is that the sender is used
only to match the retry rule. Once the rule has been found for a host error,
its contents are used to set a retry time for the host, and this will apply to
all messages, not just those with specific senders.

When testing retry rules using &%-brt%&, you can supply a sender using the
&%-f%& command line option, like this:
.code
exim -f "" -brt user@dom.ain
.endd
If you do not set &%-f%& with &%-brt%&, a retry rule that contains a senders
list is never matched.





.section "Retry parameters" "SECID163"
.cindex "retry" "parameters in rules"
The third (or fourth, if a senders list is present) field in a retry rule is a
sequence of retry parameter sets, separated by semicolons. Each set consists of
.display
<&'letter'&>,<&'cutoff time'&>,<&'arguments'&>
.endd
The letter identifies the algorithm for computing a new retry time; the cutoff
time is the time beyond which this algorithm no longer applies, and the
arguments vary the algorithm's action. The cutoff time is measured from the
time that the first failure for the domain (combined with the local part if
relevant) was detected, not from the time the message was received.

.cindex "retry" "algorithms"
.cindex "retry" "fixed intervals"
.cindex "retry" "increasing intervals"
.cindex "retry" "random intervals"
The available algorithms are:

.ilist
&'F'&: retry at fixed intervals. There is a single time parameter specifying
the interval.
.next
&'G'&: retry at geometrically increasing intervals. The first argument
specifies a starting value for the interval, and the second a multiplier, which
is used to increase the size of the interval at each retry.
.next
&'H'&: retry at randomized intervals. The arguments are as for &'G'&. For each
retry, the previous interval is multiplied by the factor in order to get a
maximum for the next interval. The minimum interval is the first argument of
the parameter, and an actual interval is chosen randomly between them. Such a
rule has been found to be helpful in cluster configurations when all the
members of the cluster restart at once, and may therefore synchronize their
queue processing times.
.endlist

When computing the next retry time, the algorithm definitions are scanned in
order until one whose cutoff time has not yet passed is reached. This is then
used to compute a new retry time that is later than the current time. In the
case of fixed interval retries, this simply means adding the interval to the
current time. For geometrically increasing intervals, retry intervals are
computed from the rule's parameters until one that is greater than the previous
interval is found. The main configuration variable
.cindex "limit" "retry interval"
.cindex "retry" "interval, maximum"
.oindex "&%retry_interval_max%&"
&%retry_interval_max%& limits the maximum interval between retries. It
cannot be set greater than &`24h`&, which is its default value.

A single remote domain may have a number of hosts associated with it, and each
host may have more than one IP address. Retry algorithms are selected on the
basis of the domain name, but are applied to each IP address independently. If,
for example, a host has two IP addresses and one is unusable, Exim will
generate retry times for it and will not try to use it until its next retry
time comes. Thus the good IP address is likely to be tried first most of the
time.

.cindex "hints database" "use for retrying"
Retry times are hints rather than promises. Exim does not make any attempt to
run deliveries exactly at the computed times. Instead, a queue runner process
starts delivery processes for delayed messages periodically, and these attempt
new deliveries only for those addresses that have passed their next retry time.
If a new message arrives for a deferred address, an immediate delivery attempt
occurs only if the address has passed its retry time. In the absence of new
messages, the minimum time between retries is the interval between queue runner
processes. There is not much point in setting retry times of five minutes if
your queue runners happen only once an hour, unless there are a significant
number of incoming messages (which might be the case on a system that is
sending everything to a smart host, for example).

The data in the retry hints database can be inspected by using the
&'exim_dumpdb'& or &'exim_fixdb'& utility programs (see chapter
&<<CHAPutils>>&). The latter utility can also be used to change the data. The
&'exinext'& utility script can be used to find out what the next retry times
are for the hosts associated with a particular mail domain, and also for local
deliveries that have been deferred.


.section "Retry rule examples" "SECID164"
Here are some example retry rules:
.code
alice@wonderland.fict.example quota_5d  F,7d,3h
wonderland.fict.example       quota_5d
wonderland.fict.example       *         F,1h,15m; G,2d,1h,2;
lookingglass.fict.example     *         F,24h,30m;
*                 refused_A   F,2h,20m;
*                 *           F,2h,15m; G,16h,1h,1.5; F,5d,8h
.endd
The first rule sets up special handling for mail to
&'alice@wonderland.fict.example'& when there is an over-quota error and the
mailbox has not been read for at least 5 days. Retries continue every three
hours for 7 days. The second rule handles over-quota errors for all other local
parts at &'wonderland.fict.example'&; the absence of a local part has the same
effect as supplying &"*@"&. As no retry algorithms are supplied, messages that
fail are bounced immediately if the mailbox has not been read for at least 5
days.

The third rule handles all other errors at &'wonderland.fict.example'&; retries
happen every 15 minutes for an hour, then with geometrically increasing
intervals until two days have passed since a delivery first failed. After the
first hour there is a delay of one hour, then two hours, then four hours, and
so on (this is a rather extreme example).

The fourth rule controls retries for the domain &'lookingglass.fict.example'&.
They happen every 30 minutes for 24 hours only. The remaining two rules handle
all other domains, with special action for connection refusal from hosts that
were not obtained from an MX record.

The final rule in a retry configuration should always have asterisks in the
first two fields so as to provide a general catch-all for any addresses that do
not have their own special handling. This example tries every 15 minutes for 2
hours, then with intervals starting at one hour and increasing by a factor of
1.5 up to 16 hours, then every 8 hours up to 5 days.



.section "Timeout of retry data" "SECID165"
.cindex "timeout" "of retry data"
.oindex "&%retry_data_expire%&"
.cindex "hints database" "data expiry"
.cindex "retry" "timeout of data"
Exim timestamps the data that it writes to its retry hints database. When it
consults the data during a delivery it ignores any that is older than the value
set in &%retry_data_expire%& (default 7 days). If, for example, a host hasn't
been tried for 7 days, Exim will try to deliver to it immediately a message
arrives, and if that fails, it will calculate a retry time as if it were
failing for the first time.

This improves the behaviour for messages routed to rarely-used hosts such as MX
backups. If such a host was down at one time, and happens to be down again when
Exim tries a month later, using the old retry data would imply that it had been
down all the time, which is not a justified assumption.

If a host really is permanently dead, this behaviour causes a burst of retries
every now and again, but only if messages routed to it are rare. If there is a
message at least once every 7 days the retry data never expires.




.section "Long-term failures" "SECID166"
.cindex "delivery failure, long-term"
.cindex "retry" "after long-term failure"
Special processing happens when an email address has been failing for so long
that the cutoff time for the last algorithm is reached. For example, using the
default retry rule:
.code
* * F,2h,15m; G,16h,1h,1.5; F,4d,6h
.endd
the cutoff time is four days. Reaching the retry cutoff is independent of how
long any specific message has been failing; it is the length of continuous
failure for the recipient address that counts.

When the cutoff time is reached for a local delivery, or for all the IP
addresses associated with a remote delivery, a subsequent delivery failure
causes Exim to give up on the address, and a bounce message is generated.
In order to cater for new messages that use the failing address, a next retry
time is still computed from the final algorithm, and is used as follows:

For local deliveries, one delivery attempt is always made for any subsequent
messages. If this delivery fails, the address fails immediately. The
post-cutoff retry time is not used.

.cindex "final cutoff" "retries, controlling"
.cindex retry "final cutoff"
If the delivery is remote, there are two possibilities, controlled by the
.oindex "&%delay_after_cutoff%&"
&%delay_after_cutoff%& option of the &(smtp)& transport. The option is true by
default. Until the post-cutoff retry time for one of the IP addresses,
as set by the &%retry_data_expire%& option, is
reached, the failing email address is bounced immediately, without a delivery
attempt taking place. After that time, one new delivery attempt is made to
those IP addresses that are past their retry times, and if that still fails,
the address is bounced and new retry times are computed.

In other words, when all the hosts for a given email address have been failing
for a long time, Exim bounces rather then defers until one of the hosts' retry
times is reached. Then it tries once, and bounces if that attempt fails. This
behaviour ensures that few resources are wasted in repeatedly trying to deliver
to a broken destination, but if the host does recover, Exim will eventually
notice.

If &%delay_after_cutoff%& is set false, Exim behaves differently. If all IP
addresses are past their final cutoff time, Exim tries to deliver to those IP
addresses that have not been tried since the message arrived. If there are
no suitable IP addresses, or if they all fail, the address is bounced. In other
words, it does not delay when a new message arrives, but tries the expired
addresses immediately, unless they have been tried since the message arrived.
If there is a continuous stream of messages for the failing domains, setting
&%delay_after_cutoff%& false means that there will be many more attempts to
deliver to permanently failing IP addresses than when &%delay_after_cutoff%& is
true.

.section "Deliveries that work intermittently" "SECID167"
.cindex "retry" "intermittently working deliveries"
Some additional logic is needed to cope with cases where a host is
intermittently available, or when a message has some attribute that prevents
its delivery when others to the same address get through. In this situation,
because some messages are successfully delivered, the &"retry clock"& for the
host or address keeps getting reset by the successful deliveries, and so
failing messages remain in the queue for ever because the cutoff time is never
reached.

Two exceptional actions are applied to prevent this happening. The first
applies to errors that are related to a message rather than a remote host.
Section &<<SECToutSMTPerr>>& has a discussion of the different kinds of error;
examples of message-related errors are 4&'xx'& responses to MAIL or DATA
commands, and quota failures. For this type of error, if a message's arrival
time is earlier than the &"first failed"& time for the error, the earlier time
is used when scanning the retry rules to decide when to try next and when to
time out the address.

The exceptional second action applies in all cases. If a message has been on
the queue for longer than the cutoff time of any applicable retry rule for a
given address, a delivery is attempted for that address, even if it is not yet
time, and if this delivery fails, the address is timed out. A new retry time is
not computed in this case, so that other messages for the same address are
considered immediately.
.ecindex IIDretconf1
.ecindex IIDregconf2






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "SMTP authentication" "CHAPSMTPAUTH"
.scindex IIDauthconf1 "SMTP" "authentication configuration"
.scindex IIDauthconf2 "authentication"
The &"authenticators"& section of Exim's runtime configuration is concerned
with SMTP authentication. This facility is an extension to the SMTP protocol,
described in
&url(https://www.rfc-editor.org/rfc/rfc2554,RFC 2554),
which allows a client SMTP host to authenticate itself
to a server. This is a common way for a server to recognize clients that are
permitted to use it as a relay. SMTP authentication is not of relevance to the
transfer of mail between servers that have no managerial connection with each
other.

The name of an authenticator is limited to be &drivernamemax; ASCII characters long;
prior to Exim 4.95 names would be silently truncated at this length, but now
it is enforced.

.cindex "AUTH" "description of"
.cindex "ESMTP extensions" AUTH
Very briefly, the way SMTP authentication works is as follows:

.olist
The server advertises a number of authentication &'mechanisms'& in response to
the client's EHLO command.
.next
The client issues an AUTH command, naming a specific mechanism. The command
may, optionally, contain some authentication data.
.next
The server may issue one or more &'challenges'&, to which the client must send
appropriate responses. In simple authentication mechanisms, the challenges are
just prompts for user names and passwords. The server does not have to issue
any challenges &-- in some mechanisms the relevant data may all be transmitted
with the AUTH command.
.next
The server either accepts or denies authentication.
.next
If authentication succeeds, the client may optionally make use of the AUTH
option on the MAIL command to pass an authenticated sender in subsequent
mail transactions. Authentication lasts for the remainder of the SMTP
connection.
.next
If authentication fails, the client may give up, or it may try a different
authentication mechanism, or it may try transferring mail over the
unauthenticated connection.
.endlist

If you are setting up a client, and want to know which authentication
mechanisms the server supports, you can use Telnet to connect to port 25 (the
SMTP port) on the server, and issue an EHLO command. The response to this
includes the list of supported mechanisms. For example:
.display
&`$ `&&*&`telnet server.example 25`&*&
&`Trying 192.168.34.25...`&
&`Connected to server.example.`&
&`Escape character is &#x0027;^]&#x0027;.`&
&`220 server.example ESMTP Exim 4.20 ...`&
&*&`ehlo client.example`&*&
&`250-server.example Hello client.example [10.8.4.5]`&
&`250-SIZE 52428800`&
&`250-PIPELINING`&
&`250-AUTH PLAIN`&
&`250 HELP`&
.endd
The second-last line of this example output shows that the server supports
authentication using the PLAIN mechanism. In Exim, the different authentication
mechanisms are configured by specifying &'authenticator'& drivers. Like the
routers and transports, which authenticators are included in the binary is
controlled by build-time definitions. The following are currently available,
included by setting
.code
AUTH_CRAM_MD5=yes
AUTH_CYRUS_SASL=yes
AUTH_DOVECOT=yes
AUTH_EXTERNAL=yes
AUTH_GSASL=yes
AUTH_HEIMDAL_GSSAPI=yes
AUTH_PLAINTEXT=yes
AUTH_SPA=yes
AUTH_TLS=yes
.endd
in &_Local/Makefile_&, respectively. The first of these supports the CRAM-MD5
authentication mechanism
(&url(https://www.rfc-editor.org/rfc/rfc2195,RFC 2195)),
and the second provides an interface to the Cyrus SASL authentication library.
The third is an interface to Dovecot's authentication system, delegating the
work via a socket interface.
The fourth provides for negotiation of authentication done via non-SMTP means,
as defined by &url(https://www.rfc-editor.org/rfc/rfc4422,RFC 4422) Appendix A.
The fifth provides an interface to the GNU SASL authentication library, which
provides mechanisms but typically not data sources.
The sixth provides direct access to Heimdal GSSAPI, geared for Kerberos, but
supporting setting a server keytab.
The seventh can be configured to support
the PLAIN authentication mechanism
(&url(https://www.rfc-editor.org/rfc/rfc2595,RFC 2595))
or the LOGIN mechanism, which is
not formally documented, but used by several MUAs.
The eighth authenticator
supports Microsoft's &'Secure Password Authentication'& mechanism.
The last is an Exim authenticator but not an SMTP one;
instead it can use information from a TLS negotiation.

The authenticators are configured using the same syntax as other drivers (see
section &<<SECTfordricon>>&). If no authenticators are required, no
authentication section need be present in the configuration file. Each
authenticator can in principle have both server and client functions. When Exim
is receiving SMTP mail, it is acting as a server; when it is sending out
messages over SMTP, it is acting as a client. Authenticator configuration
options are provided for use in both these circumstances.

To make it clear which options apply to which situation, the prefixes
&%server_%& and &%client_%& are used on option names that are specific to
either the server or the client function, respectively. Server and client
functions are disabled if none of their options are set. If an authenticator is
to be used for both server and client functions, a single definition, using
both sets of options, is required. For example:
.code
cram:
  driver = cram_md5
  public_name = CRAM-MD5
  server_secret = ${if eq{$auth1}{ph10}{secret1}fail}
  client_name = ph10
  client_secret = secret2
.endd
The &%server_%& option is used when Exim is acting as a server, and the
&%client_%& options when it is acting as a client.

Descriptions of the individual authenticators are given in subsequent chapters.
The remainder of this chapter covers the generic options for the
authenticators, followed by general discussion of the way authentication works
in Exim.

&*Beware:*& the meaning of &$auth1$&, &$auth2$&, ... varies on a per-driver and
per-mechanism basis.  Please read carefully to determine which variables hold
account labels such as usercodes and which hold passwords or other
authenticating data.

Note that some mechanisms support two different identifiers for accounts: the
&'authentication id'& and the &'authorization id'&.  The contractions &'authn'&
and &'authz'& are commonly encountered.  The American spelling is standard here.
Conceptually, authentication data such as passwords are tied to the identifier
used to authenticate; servers may have rules to permit one user to act as a
second user, so that after login the session is treated as though that second
user had logged in.  That second user is the &'authorization id'&.  A robust
configuration might confirm that the &'authz'& field is empty or matches the
&'authn'& field.  Often this is just ignored.  The &'authn'& can be considered
as verified data, the &'authz'& as an unverified request which the server might
choose to honour.

A &'realm'& is a text string, typically a domain name, presented by a server
to a client to help it select an account and credentials to use.  In some
mechanisms, the client and server provably agree on the realm, but clients
typically can not treat the realm as secure data to be blindly trusted.



.section "Generic options for authenticators" "SECID168"
.cindex "authentication" "generic options"
.cindex "options" "generic; for authenticators"

.option client_condition authenticators string&!! unset
When Exim is authenticating as a client, it skips any authenticator whose
&%client_condition%& expansion yields &"0"&, &"no"&, or &"false"&. This can be
used, for example, to skip plain text authenticators when the connection is not
encrypted by a setting such as:
.code
client_condition = ${if !eq{$tls_out_cipher}{}}
.endd


.option client_set_id authenticators string&!! unset
When client authentication succeeds, this condition is expanded; the
result is used in the log lines for outbound messages.
Typically it will be the user name used for authentication.


.option driver authenticators string unset
This option must always be set. It specifies which of the available
authenticators is to be used.


.option public_name authenticators string unset
This option specifies the name of the authentication mechanism that the driver
implements, and by which it is known to the outside world. These names should
contain only upper case letters, digits, underscores, and hyphens
(&url(https://www.rfc-editor.org/rfc/rfc2222,RFC 2222)),
but Exim in fact matches them caselessly. If &%public_name%& is not set, it
defaults to the driver's instance name.


.option server_advertise_condition authenticators string&!! unset
When a server is about to advertise an authentication mechanism, the condition
is expanded. If it yields the empty string, &"0"&, &"no"&, or &"false"&, the
mechanism is not advertised.
If the expansion fails, the mechanism is not advertised. If the failure was not
forced, and was not caused by a lookup defer, the incident is logged.
See section &<<SECTauthexiser>>& below for further discussion.


.option server_condition authenticators string&!! unset
This option must be set for a &%plaintext%& server authenticator, where it
is used directly to control authentication. See section &<<SECTplainserver>>&
for details.

For the &(gsasl)& authenticator, this option is required for various
mechanisms; see chapter &<<CHAPgsasl>>& for details.

For the other authenticators, &%server_condition%& can be used as an additional
authentication or authorization mechanism that is applied after the other
authenticator conditions succeed. If it is set, it is expanded when the
authenticator would otherwise return a success code. If the expansion is forced
to fail, authentication fails. Any other expansion failure causes a temporary
error code to be returned. If the result of a successful expansion is an empty
string, &"0"&, &"no"&, or &"false"&, authentication fails. If the result of the
expansion is &"1"&, &"yes"&, or &"true"&, authentication succeeds. For any
other result, a temporary error code is returned, with the expanded string as
the error text.


.option server_debug_print authenticators string&!! unset
If this option is set and authentication debugging is enabled (see the &%-d%&
command line option), the string is expanded and included in the debugging
output when the authenticator is run as a server. This can help with checking
out the values of variables.
If expansion of the string fails, the error message is written to the debugging
output, and Exim carries on processing.


.option server_set_id authenticators string&!! unset
.vindex "&$authenticated_id$&"
.vindex "&$authenticated_fail_id$&"
When an Exim server successfully authenticates a client, this string is
expanded using data from the authentication, and preserved for any incoming
messages in the variable &$authenticated_id$&. It is also included in the log
lines for incoming messages. For example, a user/password authenticator
configuration might preserve the user name that was used to authenticate, and
refer to it subsequently during delivery of the message.
On a failing authentication the expansion result is instead saved in
the &$authenticated_fail_id$& variable.
If expansion fails, the option is ignored.


.option server_mail_auth_condition authenticators string&!! unset
This option allows a server to discard authenticated sender addresses supplied
as part of MAIL commands in SMTP connections that are authenticated by the
driver on which &%server_mail_auth_condition%& is set. The option is not used
as part of the authentication process; instead its (unexpanded) value is
remembered for later use.
How it is used is described in the following section.





.section "The AUTH parameter on MAIL commands" "SECTauthparamail"
.cindex "authentication" "sender; authenticated"
.cindex "AUTH" "on MAIL command"
When a client supplied an AUTH= item on a MAIL command, Exim applies
the following checks before accepting it as the authenticated sender of the
message:

.ilist
If the connection is not using extended SMTP (that is, HELO was used rather
than EHLO), the use of AUTH= is a syntax error.
.next
If the value of the AUTH= parameter is &"<>"&, it is ignored.
.next
.vindex "&$authenticated_sender$&"
If &%acl_smtp_mailauth%& is defined, the ACL it specifies is run. While it is
running, the value of &$authenticated_sender$& is set to the value obtained
from the AUTH= parameter. If the ACL does not yield &"accept"&, the value of
&$authenticated_sender$& is deleted. The &%acl_smtp_mailauth%& ACL may not
return &"drop"& or &"discard"&. If it defers, a temporary error code (451) is
given for the MAIL command.
.next
If &%acl_smtp_mailauth%& is not defined, the value of the AUTH= parameter
is accepted and placed in &$authenticated_sender$& only if the client has
authenticated.
.next
If the AUTH= value was accepted by either of the two previous rules, and
the client has authenticated, and the authenticator has a setting for the
&%server_mail_auth_condition%&, the condition is checked at this point. The
valued that was saved from the authenticator is expanded. If the expansion
fails, or yields an empty string, &"0"&, &"no"&, or &"false"&, the value of
&$authenticated_sender$& is deleted. If the expansion yields any other value,
the value of &$authenticated_sender$& is retained and passed on with the
message.
.endlist


When &$authenticated_sender$& is set for a message, it is passed on to other
hosts to which Exim authenticates as a client. Do not confuse this value with
&$authenticated_id$&, which is a string obtained from the authentication
process, and which is not usually a complete email address.

.vindex "&$sender_address$&"
Whenever an AUTH= value is ignored, the incident is logged. The ACL for
MAIL, if defined, is run after AUTH= is accepted or ignored. It can
therefore make use of &$authenticated_sender$&. The converse is not true: the
value of &$sender_address$& is not yet set up when the &%acl_smtp_mailauth%&
ACL is run.



.section "Authentication on an Exim server" "SECTauthexiser"
.cindex "authentication" "on an Exim server"
When Exim receives an EHLO command, it advertises the public names of those
authenticators that are configured as servers, subject to the following
conditions:

.ilist
The client host must match &%auth_advertise_hosts%& (default *).
.next
If the &%server_advertise_condition%& option is set, its expansion must not
yield the empty string, &"0"&, &"no"&, or &"false"&.
.endlist

The order in which the authenticators are defined controls the order in which
the mechanisms are advertised.

Some mail clients (for example, some versions of Netscape) require the user to
provide a name and password for authentication whenever AUTH is advertised,
even though authentication may not in fact be needed (for example, Exim may be
set up to allow unconditional relaying from the client by an IP address check).
You can make such clients more friendly by not advertising AUTH to them.
For example, if clients on the 10.9.8.0/24 network are permitted (by the ACL
that runs for RCPT) to relay without authentication, you should set
.code
auth_advertise_hosts = ! 10.9.8.0/24
.endd
so that no authentication mechanisms are advertised to them.

The &%server_advertise_condition%& controls the advertisement of individual
authentication mechanisms. For example, it can be used to restrict the
advertisement of a particular mechanism to encrypted connections, by a setting
such as:
.code
server_advertise_condition = ${if eq{$tls_in_cipher}{}{no}{yes}}
.endd
.vindex "&$tls_in_cipher$&"
If the session is encrypted, &$tls_in_cipher$& is not empty, and so the expansion
yields &"yes"&, which allows the advertisement to happen.

When an Exim server receives an AUTH command from a client, it rejects it
immediately if AUTH was not advertised in response to an earlier EHLO
command. This is the case if

.ilist
The client host does not match &%auth_advertise_hosts%&; or
.next
No authenticators are configured with server options; or
.next
Expansion of &%server_advertise_condition%& blocked the advertising of all the
server authenticators.
.endlist


Otherwise, Exim runs the ACL specified by &%acl_smtp_auth%& in order
to decide whether to accept the command. If &%acl_smtp_auth%& is not set,
AUTH is accepted from any client host.

If AUTH is not rejected by the ACL, Exim searches its configuration for a
server authentication mechanism that was advertised in response to EHLO and
that matches the one named in the AUTH command. If it finds one, it runs
the appropriate authentication protocol, and authentication either succeeds or
fails. If there is no matching advertised mechanism, the AUTH command is
rejected with a 504 error.

.vindex "&$received_protocol$&"
.vindex "&$sender_host_authenticated$&"
When a message is received from an authenticated host, the value of
&$received_protocol$& is set to &"esmtpa"& or &"esmtpsa"& instead of &"esmtp"&
or &"esmtps"&, and &$sender_host_authenticated$& contains the name (not the
public name) of the authenticator driver that successfully authenticated the
client from which the message was received. This variable is empty if there was
no successful authentication.

.cindex authentication "expansion item"
Successful authentication sets up information used by the
&%authresults%& expansion item.

.cindex authentication "failure event, server"
If an authenticator is run and does not succeed,
an event of type "auth:fail" is raised.
While the event is being processed the variables
&$sender_host_authenticated$& (with the authenticator name)
and &$authenticated_fail_id$& (as set by the authenticator &%server_set_id%& option)
will be valid.
If the event is serviced and a string is returned then the string will be logged
instead of the default log line.
See &<<CHAPevents>>& for details on events.


.subsection "Testing server authentication" "SECID169"
.cindex "authentication" "testing a server"
.cindex "AUTH" "testing a server"
.cindex "base64 encoding" "creating authentication test data"
Exim's &%-bh%& option can be useful for testing server authentication
configurations. The data for the AUTH command has to be sent using base64
encoding. A quick way to produce such data for testing is the following Perl
script:
.code
use MIME::Base64;
printf ("%s", encode_base64(eval "\"$ARGV[0]\""));
.endd
.cindex "binary zero" "in authentication data"
This interprets its argument as a Perl string, and then encodes it. The
interpretation as a Perl string allows binary zeros, which are required for
some kinds of authentication, to be included in the data. For example, a
command line to run this script on such data might be
.code
encode '\0user\0password'
.endd
Note the use of single quotes to prevent the shell interpreting the
backslashes, so that they can be interpreted by Perl to specify characters
whose code value is zero.

&*Warning 1*&: If either of the user or password strings starts with an octal
digit, you must use three zeros instead of one after the leading backslash. If
you do not, the octal digit that starts your string will be incorrectly
interpreted as part of the code for the first character.

&*Warning 2*&: If there are characters in the strings that Perl interprets
specially, you must use a Perl escape to prevent them being misinterpreted. For
example, a command such as
.code
encode '\0user@domain.com\0pas$$word'
.endd
gives an incorrect answer because of the unescaped &"@"& and &"$"& characters.

If you have the &%mimencode%& command installed, another way to produce
base64-encoded strings is to run the command
.code
echo -e -n `\0user\0password' | mimencode
.endd
The &%-e%& option of &%echo%& enables the interpretation of backslash escapes
in the argument, and the &%-n%& option specifies no newline at the end of its
output. However, not all versions of &%echo%& recognize these options, so you
should check your version before relying on this suggestion.



.section "Authentication by an Exim client" "SECID170"
.cindex "authentication" "on an Exim client"
The &(smtp)& transport has two options called &%hosts_require_auth%& and
&%hosts_try_auth%&. When the &(smtp)& transport connects to a server that
announces support for authentication, and the host matches an entry in either
of these options, Exim (as a client) tries to authenticate as follows:

.ilist
For each authenticator that is configured as a client, in the order in which
they are defined in the configuration, it searches the authentication
mechanisms announced by the server for one whose name matches the public name
of the authenticator.
.next
.vindex "&$host$&"
.vindex "&$host_address$&"
When it finds one that matches, it runs the authenticator's client code. The
variables &$host$& and &$host_address$& are available for any string expansions
that the client might do. They are set to the server's name and IP address. If
any expansion is forced to fail, the authentication attempt is abandoned, and
Exim moves on to the next authenticator. Otherwise an expansion failure causes
delivery to be deferred.
.next
If the result of the authentication attempt is a temporary error or a timeout,
Exim abandons trying to send the message to the host for the moment. It will
try again later. If there are any backup hosts available, they are tried in the
usual way.

.next
.cindex authentication "failure event, client"
If the response to authentication is a permanent error (5&'xx'& code),
an event of type "auth:fail" is raised.
While the event is being processed the variable
&$sender_host_authenticated$& (with the authenticator name)
will be valid.
If the event is serviced and a string is returned then the string will be logged.
See &<<CHAPevents>>& for details on events.

.next
If the response to authentication is a permanent error (5&'xx'& code), Exim
carries on searching the list of authenticators and tries another one if
possible. If all authentication attempts give permanent errors, or if there are
no attempts because no mechanisms match (or option expansions force failure),
what happens depends on whether the host matches &%hosts_require_auth%& or
&%hosts_try_auth%&. In the first case, a temporary error is generated, and
delivery is deferred. The error can be detected in the retry rules, and thereby
turned into a permanent error if you wish. In the second case, Exim tries to
deliver the message unauthenticated.
.endlist

Note that the hostlist test for whether to do authentication can be
confused if name-IP lookups change between the time the peer is decided
upon and the time that the transport runs.  For example, with a manualroute
router given a host name, and with DNS "round-robin" used by that name: if
the local resolver cache times out between the router and the transport
running, the transport may get an IP for the name for its authentication
check which does not match the connection peer IP.
No authentication will then be done, despite the names being identical.

For such cases use a separate transport which always authenticates.

.cindex "AUTH" "on MAIL command"
When Exim has authenticated itself to a remote server, it adds the AUTH
parameter to the MAIL commands it sends, if it has an authenticated sender for
the message. If the message came from a remote host, the authenticated sender
is the one that was receiving on an incoming MAIL command, provided that the
incoming connection was authenticated and the &%server_mail_auth%& condition
allowed the authenticated sender to be retained. If a local process calls Exim
to send a message, the sender address that is built from the login name and
&%qualify_domain%& is treated as authenticated. However, if the
&%authenticated_sender%& option is set on the &(smtp)& transport, it overrides
the authenticated sender that was received with the message.
.ecindex IIDauthconf1
.ecindex IIDauthconf2






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The plaintext authenticator" "CHAPplaintext"
.scindex IIDplaiauth1 "&(plaintext)& authenticator"
.scindex IIDplaiauth2 "authenticators" "&(plaintext)&"
The &(plaintext)& authenticator can be configured to support the PLAIN and
LOGIN authentication mechanisms, both of which transfer authentication data as
plain (unencrypted) text (though base64 encoded). The use of plain text is a
security risk; you are strongly advised to insist on the use of SMTP encryption
(see chapter &<<CHAPTLS>>&) if you use the PLAIN or LOGIN mechanisms. If you do
use unencrypted plain text, you should not use the same passwords for SMTP
connections as you do for login accounts.

.section "Avoiding cleartext use" "SECTplain_TLS"
The following generic option settings will disable &(plaintext)& authenticators when
TLS is not being used:
.code
  server_advertise_condition = ${if def:tls_in_cipher }
  client_condition =           ${if def:tls_out_cipher}
.endd

&*Note*&: a plaintext SMTP AUTH done inside TLS is not vulnerable to casual snooping,
but is still vulnerable to a Man In The Middle attack unless certificates
(including their names) have been properly verified.

.section "Plaintext server options" "SECID171"
.cindex "options" "&(plaintext)& authenticator (server)"
When configured as a server, &(plaintext)& uses the following options:

.option server_condition authenticators string&!! unset
This is actually a global authentication option, but it must be set in order to
configure the &(plaintext)& driver as a server. Its use is described below.

.option server_prompts plaintext "string list&!!" unset
The contents of this option, after expansion, must be a colon-separated list of
prompt strings. If expansion fails, a temporary authentication rejection is
given.

.section "Using plaintext in a server" "SECTplainserver"
.cindex "AUTH" "in &(plaintext)& authenticator"
.cindex "binary zero" "in &(plaintext)& authenticator"
.cindex "numerical variables (&$1$& &$2$& etc)" &&&
        "in &(plaintext)& authenticator"
.vindex "&$auth1$&, &$auth2$&, etc"
.cindex "base64 encoding" "in &(plaintext)& authenticator"

When running as a server, &(plaintext)& performs the authentication test by
expanding a string. The data sent by the client with the AUTH command, or in
response to subsequent prompts, is base64 encoded, and so may contain any byte
values when decoded. If any data is supplied with the command, it is treated as
a list of strings, separated by NULs (binary zeros), the first three of which
are placed in the expansion variables &$auth1$&, &$auth2$&, and &$auth3$&
(neither LOGIN nor PLAIN uses more than three strings).

For compatibility with previous releases of Exim, the values are also placed in
the expansion variables &$1$&, &$2$&, and &$3$&. However, the use of these
variables for this purpose is now deprecated, as it can lead to confusion in
string expansions that also use them for other things.

If there are more strings in &%server_prompts%& than the number of strings
supplied with the AUTH command, the remaining prompts are used to obtain more
data. Each response from the client may be a list of NUL-separated strings.

.vindex "&$authenticated_id$&"
Once a sufficient number of data strings have been received,
&%server_condition%& is expanded. If the expansion is forced to fail,
authentication fails. Any other expansion failure causes a temporary error code
to be returned. If the result of a successful expansion is an empty string,
&"0"&, &"no"&, or &"false"&, authentication fails. If the result of the
expansion is &"1"&, &"yes"&, or &"true"&, authentication succeeds and the
generic &%server_set_id%& option is expanded and saved in &$authenticated_id$&.
For any other result, a temporary error code is returned, with the expanded
string as the error text.

&*Warning*&: If you use a lookup in the expansion to find the user's
password, be sure to make the authentication fail if the user is unknown.
There are good and bad examples at the end of the next section.



.subsection "The PLAIN authentication mechanism" "SECID172"
.cindex "PLAIN authentication mechanism"
.cindex authentication PLAIN
.cindex "binary zero" "in &(plaintext)& authenticator"
The PLAIN authentication mechanism
(&url(https://www.rfc-editor.org/rfc/rfc2595,RFC 2595))
specifies that three strings be
sent as one item of data (that is, one combined string containing two NUL
separators). The data is sent either as part of the AUTH command, or
subsequently in response to an empty prompt from the server.

The second and third strings are a user name and a corresponding password.
Using a single fixed user name and password as an example, this could be
configured as follows:
.code
fixed_plain:
  driver = plaintext
  public_name = PLAIN
  server_prompts = :
  server_condition = \
    ${if and {{eq{$auth2}{username}}{eq{$auth3}{mysecret}}}}
  server_set_id = $auth2
.endd
Note that the default result strings from &%if%& (&"true"& or an empty string)
are exactly what we want here, so they need not be specified. Obviously, if the
password contains expansion-significant characters such as dollar, backslash,
or closing brace, they have to be escaped.

The &%server_prompts%& setting specifies a single, empty prompt (empty items at
the end of a string list are ignored). If all the data comes as part of the
AUTH command, as is commonly the case, the prompt is not used. This
authenticator is advertised in the response to EHLO as
.code
250-AUTH PLAIN
.endd
and a client host can authenticate itself by sending the command
.code
AUTH PLAIN AHVzZXJuYW1lAG15c2VjcmV0
.endd
As this contains three strings (more than the number of prompts), no further
data is required from the client. Alternatively, the client may just send
.code
AUTH PLAIN
.endd
to initiate authentication, in which case the server replies with an empty
prompt. The client must respond with the combined data string.

The data string is base64 encoded, as required by the RFC. This example,
when decoded, is <&'NUL'&>&`username`&<&'NUL'&>&`mysecret`&, where <&'NUL'&>
represents a zero byte. This is split up into three strings, the first of which
is empty. The &%server_condition%& option in the authenticator checks that the
second two are &`username`& and &`mysecret`& respectively.

Having just one fixed user name and password, as in this example, is not very
realistic, though for a small organization with only a handful of
authenticating clients it could make sense.

A more sophisticated instance of this authenticator could use the user name in
&$auth2$& to look up a password in a file or database, and maybe do an encrypted
comparison (see &%crypteq%& in chapter &<<CHAPexpand>>&). Here is a example of
this approach, where the passwords are looked up in a DBM file. &*Warning*&:
This is an incorrect example:
.code
server_condition = \
  ${if eq{$auth3}{${lookup{$auth2}dbm{/etc/authpwd}}}}
.endd
The expansion uses the user name (&$auth2$&) as the key to look up a password,
which it then compares to the supplied password (&$auth3$&). Why is this example
incorrect? It works fine for existing users, but consider what happens if a
non-existent user name is given. The lookup fails, but as no success/failure
strings are given for the lookup, it yields an empty string. Thus, to defeat
the authentication, all a client has to do is to supply a non-existent user
name and an empty password. The correct way of writing this test is:
.code
server_condition = ${lookup{$auth2}dbm{/etc/authpwd}\
  {${if eq{$value}{$auth3}}} {false}}
.endd
In this case, if the lookup succeeds, the result is checked; if the lookup
fails, &"false"& is returned and authentication fails. If &%crypteq%& is being
used instead of &%eq%&, the first example is in fact safe, because &%crypteq%&
always fails if its second argument is empty. However, the second way of
writing the test makes the logic clearer.


.subsection "The LOGIN authentication mechanism" "SECID173"
.cindex "LOGIN authentication mechanism"
.cindex authentication LOGIN
The LOGIN authentication mechanism is not documented in any RFC, but is in use
in a number of programs. No data is sent with the AUTH command. Instead, a
user name and password are supplied separately, in response to prompts. The
plaintext authenticator can be configured to support this as in this example:
.code
fixed_login:
  driver = plaintext
  public_name = LOGIN
  server_prompts = User Name : Password
  server_condition = \
    ${if and {{eq{$auth1}{username}}{eq{$auth2}{mysecret}}}}
  server_set_id = $auth1
.endd
Because of the way plaintext operates, this authenticator accepts data supplied
with the AUTH command (in contravention of the specification of LOGIN), but
if the client does not supply it (as is the case for LOGIN clients), the prompt
strings are used to obtain two data items.

Some clients are very particular about the precise text of the prompts. For
example, Outlook Express is reported to recognize only &"Username:"& and
&"Password:"&. Here is an example of a LOGIN authenticator that uses those
strings. It uses the &%ldapauth%& expansion condition to check the user
name and password by binding to an LDAP server:
.code
login:
  driver = plaintext
  public_name = LOGIN
  server_prompts = Username:: : Password::
  server_condition = ${if and{{ \
    !eq{}{$auth1} }{ \
    ldapauth{\
      user="uid=${quote_ldap_dn:$auth1},ou=people,o=example.org" \
      pass=${quote:$auth2} \
      ldap://ldap.example.org/} }} }
  server_set_id = uid=$auth1,ou=people,o=example.org
.endd
We have to check that the username is not empty before using it, because LDAP
does not permit empty DN components. We must also use the &%quote_ldap_dn%&
operator to correctly quote the DN for authentication. However, the basic
&%quote%& operator, rather than any of the LDAP quoting operators, is the
correct one to use for the password, because quoting is needed only to make
the password conform to the Exim syntax. At the LDAP level, the password is an
uninterpreted string.


.subsection "Support for different kinds of authentication" "SECID174"
A number of string expansion features are provided for the purpose of
interfacing to different ways of user authentication. These include checking
traditionally encrypted passwords from &_/etc/passwd_& (or equivalent), PAM,
Radius, &%ldapauth%&, &'pwcheck'&, and &'saslauthd'&. For details see section
&<<SECTexpcond>>&.




.section "Using plaintext in a client" "SECID175"
.cindex "options" "&(plaintext)& authenticator (client)"
The &(plaintext)& authenticator has two client options:

.option client_ignore_invalid_base64 plaintext boolean false
If the client receives a server prompt that is not a valid base64 string,
authentication is abandoned by default. However, if this option is set true,
the error in the challenge is ignored and the client sends the response as
usual.

.option client_send plaintext string&!! unset
The string is a colon-separated list of authentication data strings. Each
string is independently expanded before being sent to the server. The first
string is sent with the AUTH command; any more strings are sent in response
to prompts from the server. Before each string is expanded, the value of the
most recent prompt is placed in the next &$auth$&<&'n'&> variable, starting
with &$auth1$& for the first prompt. Up to three prompts are stored in this
way. Thus, the prompt that is received in response to sending the first string
(with the AUTH command) can be used in the expansion of the second string, and
so on. If an invalid base64 string is received when
&%client_ignore_invalid_base64%& is set, an empty string is put in the
&$auth$&<&'n'&> variable.

&*Note*&: You cannot use expansion to create multiple strings, because
splitting takes priority and happens first.

Because the PLAIN authentication mechanism requires NUL (binary zero) bytes in
the data, further processing is applied to each string before it is sent. If
there are any single circumflex characters in the string, they are converted to
NULs. Should an actual circumflex be required as data, it must be doubled in
the string.

This is an example of a client configuration that implements the PLAIN
authentication mechanism with a fixed user name and password:
.code
fixed_plain:
  driver = plaintext
  public_name = PLAIN
  client_send = ^username^mysecret
.endd
The lack of colons means that the entire text is sent with the AUTH
command, with the circumflex characters converted to NULs.
Note that due to the ambiguity of parsing three consectutive circumflex characters
there is no way to provide a password having a leading circumflex.


A similar example
that uses the LOGIN mechanism is:
.code
fixed_login:
  driver = plaintext
  public_name = LOGIN
  client_send = : username : mysecret
.endd
The initial colon means that the first string is empty, so no data is sent with
the AUTH command itself. The remaining strings are sent in response to
prompts.
.ecindex IIDplaiauth1
.ecindex IIDplaiauth2




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The cram_md5 authenticator" "CHID9"
.scindex IIDcramauth1 "&(cram_md5)& authenticator"
.scindex IIDcramauth2 "authenticators" "&(cram_md5)&"
.cindex "CRAM-MD5 authentication mechanism"
.cindex authentication CRAM-MD5
The CRAM-MD5 authentication mechanism is described in
&url(https://www.rfc-editor.org/rfc/rfc2195,RFC 2195). The server
sends a challenge string to the client, and the response consists of a user
name and the CRAM-MD5 digest of the challenge string combined with a secret
string (password) which is known to both server and client. Thus, the secret
is not sent over the network as plain text, which makes this authenticator more
secure than &(plaintext)&. However, the downside is that the secret has to be
available in plain text at either end.


.section "Using cram_md5 as a server" "SECID176"
.cindex "options" "&(cram_md5)& authenticator (server)"
This authenticator has one server option, which must be set to configure the
authenticator as a server:

.option server_secret cram_md5 string&!! unset
.cindex "numerical variables (&$1$& &$2$& etc)" "in &(cram_md5)& authenticator"
When the server receives the client's response, the user name is placed in
the expansion variable &$auth1$&, and &%server_secret%& is expanded to
obtain the password for that user. The server then computes the CRAM-MD5 digest
that the client should have sent, and checks that it received the correct
string. If the expansion of &%server_secret%& is forced to fail, authentication
fails. If the expansion fails for some other reason, a temporary error code is
returned to the client.

For compatibility with previous releases of Exim, the user name is also placed
in &$1$&. However, the use of this variables for this purpose is now
deprecated, as it can lead to confusion in string expansions that also use
numeric variables for other things.

For example, the following authenticator checks that the user name given by the
client is &"ph10"&, and if so, uses &"secret"& as the password. For any other
user name, authentication fails.
.code
fixed_cram:
  driver = cram_md5
  public_name = CRAM-MD5
  server_secret = ${if eq{$auth1}{ph10}{secret}fail}
  server_set_id = $auth1
.endd
.vindex "&$authenticated_id$&"
If authentication succeeds, the setting of &%server_set_id%& preserves the user
name in &$authenticated_id$&. A more typical configuration might look up the
secret string in a file, using the user name as the key. For example:
.code
lookup_cram:
  driver = cram_md5
  public_name = CRAM-MD5
  server_secret = ${lookup{$auth1}lsearch{/etc/authpwd}\
                  {$value}fail}
  server_set_id = $auth1
.endd
Note that this expansion explicitly forces failure if the lookup fails
because &$auth1$& contains an unknown user name.

As another example, if you wish to re-use a Cyrus SASL sasldb2 file without
using the relevant libraries, you need to know the realm to specify in the
lookup and then ask for the &"userPassword"& attribute for that user in that
realm, with:
.code
cyrusless_crammd5:
  driver = cram_md5
  public_name = CRAM-MD5
  server_secret = ${lookup{$auth1:mail.example.org:userPassword}\
                  dbmjz{/etc/sasldb2}{$value}fail}
  server_set_id = $auth1
.endd

.section "Using cram_md5 as a client" "SECID177"
.cindex "options" "&(cram_md5)& authenticator (client)"
When used as a client, the &(cram_md5)& authenticator has two options:



.option client_name cram_md5 string&!! "the primary host name"
This string is expanded, and the result used as the user name data when
computing the response to the server's challenge.


.option client_secret cram_md5 string&!! unset
This option must be set for the authenticator to work as a client. Its value is
expanded and the result used as the secret string when computing the response.


.vindex "&$host$&"
.vindex "&$host_address$&"
Different user names and secrets can be used for different servers by referring
to &$host$& or &$host_address$& in the options. Forced failure of either
expansion string is treated as an indication that this authenticator is not
prepared to handle this case. Exim moves on to the next configured client
authenticator. Any other expansion failure causes Exim to give up trying to
send the message to the current server.

A simple example configuration of a &(cram_md5)& authenticator, using fixed
strings, is:
.code
fixed_cram:
  driver = cram_md5
  public_name = CRAM-MD5
  client_name = ph10
  client_secret = secret
.endd
.ecindex IIDcramauth1
.ecindex IIDcramauth2



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The cyrus_sasl authenticator" "CHID10"
.scindex IIDcyrauth1 "&(cyrus_sasl)& authenticator"
.scindex IIDcyrauth2 "authenticators" "&(cyrus_sasl)&"
.cindex "Cyrus" "SASL library"
.cindex "Kerberos"
The code for this authenticator was provided by Matthew Byng-Maddick while
at A L Digital Ltd.

The &(cyrus_sasl)& authenticator provides server support for the Cyrus SASL
library implementation of the
&url(https://www.rfc-editor.org/rfc/rfc2222,RFC 2222)
(&"Simple Authentication and Security
Layer"&). This library supports a number of authentication mechanisms,
including PLAIN and LOGIN, but also several others that Exim does not support
directly. In particular, there is support for Kerberos authentication.

The &(cyrus_sasl)& authenticator provides a gatewaying mechanism directly to
the Cyrus interface, so if your Cyrus library can do, for example, CRAM-MD5,
then so can the &(cyrus_sasl)& authenticator. By default it uses the public
name of the driver to determine which mechanism to support.

Where access to some kind of secret file is required, for example, in GSSAPI
or CRAM-MD5, it is worth noting that the authenticator runs as the Exim
user, and that the Cyrus SASL library has no way of escalating privileges
by default. You may also find you need to set environment variables,
depending on the driver you are using.

The application name provided by Exim is &"exim"&, so various SASL options may
be set in &_exim.conf_& in your SASL directory. If you are using GSSAPI for
Kerberos, note that because of limitations in the GSSAPI interface,
changing the server keytab might need to be communicated down to the Kerberos
layer independently. The mechanism for doing so is dependent upon the Kerberos
implementation.

For example, for older releases of Heimdal, the environment variable KRB5_KTNAME
may be set to point to an alternative keytab file. Exim will pass this
variable through from its own inherited environment when started as root or the
Exim user. The keytab file needs to be readable by the Exim user.
With newer releases of Heimdal, a setuid Exim may cause Heimdal to discard the
environment variable.  In practice, for those releases, the Cyrus authenticator
is not a suitable interface for GSSAPI (Kerberos) support.  Instead, consider
the &(heimdal_gssapi)& authenticator, described in chapter &<<CHAPheimdalgss>>&


.section "Using cyrus_sasl as a server" "SECID178"
The &(cyrus_sasl)& authenticator has four private options. It puts the username
(on a successful authentication) into &$auth1$&. For compatibility with
previous releases of Exim, the username is also placed in &$1$&. However, the
use of this variable for this purpose is now deprecated, as it can lead to
confusion in string expansions that also use numeric variables for other
things.


.option server_hostname cyrus_sasl string&!! "see below"
This option selects the hostname that is used when communicating with the
library. The default value is &`$primary_hostname`&. It is up to the underlying
SASL plug-in what it does with this data.


.option server_mech cyrus_sasl string "see below"
This option selects the authentication mechanism this driver should use. The
default is the value of the generic &%public_name%& option. This option allows
you to use a different underlying mechanism from the advertised name. For
example:
.code
sasl:
  driver = cyrus_sasl
  public_name = X-ANYTHING
  server_mech = CRAM-MD5
  server_set_id = $auth1
.endd

.option server_realm cyrus_sasl string&!! unset
This specifies the SASL realm that the server claims to be in.


.option server_service cyrus_sasl string &`smtp`&
This is the SASL service that the server claims to implement.


For straightforward cases, you do not need to set any of the authenticator's
private options. All you need to do is to specify an appropriate mechanism as
the public name. Thus, if you have a SASL library that supports CRAM-MD5 and
PLAIN, you could have two authenticators as follows:
.code
sasl_cram_md5:
  driver = cyrus_sasl
  public_name = CRAM-MD5
  server_set_id = $auth1

sasl_plain:
  driver = cyrus_sasl
  public_name = PLAIN
  server_set_id = $auth2
.endd
Cyrus SASL does implement the LOGIN authentication method, even though it is
not a standard method. It is disabled by default in the source distribution,
but it is present in many binary distributions.
.ecindex IIDcyrauth1
.ecindex IIDcyrauth2




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////
.chapter "The dovecot authenticator" "CHAPdovecot"
.scindex IIDdcotauth1 "&(dovecot)& authenticator"
.scindex IIDdcotauth2 "authenticators" "&(dovecot)&"
This authenticator is an interface to the authentication facility of the
Dovecot 2 POP/IMAP server, which can support a number of authentication methods.
Note that Dovecot must be configured to use auth-client not auth-userdb.
If you are using Dovecot to authenticate POP/IMAP clients, it might be helpful
to use the same mechanisms for SMTP authentication. This is a server
authenticator only. There is only one non-generic option:

.option server_socket dovecot string unset

This option must specify the UNIX socket that is the interface to Dovecot
authentication. The &%public_name%& option must specify an authentication
mechanism that Dovecot is configured to support. You can have several
authenticators for different mechanisms. For example:
.code
dovecot_plain:
  driver = dovecot
  public_name = PLAIN
  server_advertise_condition = ${if def:tls_in_cipher}
  server_socket = /var/run/dovecot/auth-client
  server_set_id = $auth1

dovecot_ntlm:
  driver = dovecot
  public_name = NTLM
  server_socket = /var/run/dovecot/auth-client
  server_set_id = $auth1
.endd

&*Note*&: plaintext authentication methods such as PLAIN and LOGIN
should not be advertised on cleartext SMTP connections.
See the discussion in section &<<SECTplain_TLS>>&.

If the SMTP connection is encrypted, or if &$sender_host_address$& is equal to
&$received_ip_address$& (that is, the connection is local), the &"secured"&
option is passed in the Dovecot authentication command. If, for a TLS
connection, a client certificate has been verified, the &"valid-client-cert"&
option is passed. When authentication succeeds, the identity of the user
who authenticated is placed in &$auth1$&.

The Dovecot configuration to match the above will look
something like:
.code
conf.d/10-master.conf :-

service auth {
...
#SASL
  unix_listener auth-client {
    mode = 0660
    user = mail
  }
...
}

conf.d/10-auth.conf :-

auth_mechanisms = plain login ntlm
.endd

.ecindex IIDdcotauth1
.ecindex IIDdcotauth2


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////
.chapter "The gsasl authenticator" "CHAPgsasl"
.scindex IIDgsaslauth1 "&(gsasl)& authenticator"
.scindex IIDgsaslauth2 "authenticators" "&(gsasl)&"
.cindex "authentication" "GNU SASL"
.cindex "authentication" "SASL"
.cindex "authentication" "EXTERNAL"
.cindex "authentication" "ANONYMOUS"
.cindex "authentication" "PLAIN"
.cindex "authentication" "LOGIN"
.cindex "authentication" "DIGEST-MD5"
.cindex "authentication" "CRAM-MD5"
.cindex "authentication" "SCRAM family"
The &(gsasl)& authenticator provides integration for the GNU SASL
library and the mechanisms it provides.  This is new as of the 4.80 release
and there are a few areas where the library does not let Exim smoothly
scale to handle future authentication mechanisms, so no guarantee can be
made that any particular new authentication mechanism will be supported
without code changes in Exim.

The library is expected to add support in an upcoming
realease for the SCRAM-SHA-256 method.
The macro _HAVE_AUTH_GSASL_SCRAM_SHA_256 will be defined
when this happens.

To see the list of mechanisms supported by the library run Exim with "auth" debug
enabled and look for a line containing "GNU SASL supports".
Note however that some may not have been tested from Exim.


.option client_authz gsasl string&!! unset
This option can be used to supply an &'authorization id'&
which is different to the &'authentication_id'& provided
by &%client_username%& option.
If unset or (after expansion) empty it is not used,
which is the common case.

.option client_channelbinding gsasl boolean false
See &%server_channelbinding%& below.

.option client_password gsasl string&!! unset
This option is exapanded before use, and should result in
the password to be used, in clear.

.option client_username gsasl string&!! unset
This option is exapanded before use, and should result in
the account name to be used.


.option client_spassword gsasl string&!! unset
This option is only supported for library versions 1.9.1 and greater.
The macro _HAVE_AUTH_GSASL_SCRAM_S_KEY will be defined when this is so.

If a SCRAM mechanism is being used and this option is set
and correctly sized
it is used in preference to &%client_password%&.
The value after expansion should be
a 40 (for SHA-1) or 64 (for SHA-256) character string
with the PBKDF2-prepared password, hex-encoded.

Note that this value will depend on the salt and iteration-count
supplied by the server.
The option is expanded before use.
During the expansion &$auth1$& is set with the client username,
&$auth2$& with the iteration count, and
&$auth3$& with the salt.

The intent of this option
is to support clients that can cache thes salted password
to save on recalculation costs.
The cache lookup should return an unusable value
(eg. an empty string)
if the salt or iteration count has changed

If the authentication succeeds then the above variables are set,
.vindex "&$auth4$&"
plus the calculated salted password value value in &$auth4$&,
during the expansion of the &%client_set_id%& option.
A side-effect of this expansion can be used to prime the cache.


.option server_channelbinding gsasl boolean false
Some authentication mechanisms are able to use external context at both ends
of the session to bind the authentication to that context, and fail the
authentication process if that context differs.  Specifically, some TLS
ciphersuites can provide identifying information about the cryptographic
context.

This should have meant that certificate identity and verification becomes a
non-issue, as a man-in-the-middle attack will cause the correct client and
server to see different identifiers and authentication will fail.

This is
only usable by mechanisms which support "channel binding"; at time of
writing, that's the SCRAM family.
When using this feature the "-PLUS" variants of the method names need to be used.

This defaults off to ensure smooth upgrade across Exim releases, in case
this option causes some clients to start failing.  Some future release
of Exim might have switched the default to be true.

. However, Channel Binding in TLS has proven to be vulnerable in current versions.
. Do not plan to rely upon this feature for security, ever, without consulting
. with a subject matter expert (a cryptographic engineer).

This option was deprecated in previous releases due to doubts over
the "Triple Handshake" vulnerability.
Exim takes suitable precausions (requiring Extended Master Secret if TLS
Session Resumption was used) for safety.


.option server_hostname gsasl string&!! "see below"
This option selects the hostname that is used when communicating with the
library. The default value is &`$primary_hostname`&.
Some mechanisms will use this data.


.option server_mech gsasl string "see below"
This option selects the authentication mechanism this driver should use. The
default is the value of the generic &%public_name%& option. This option allows
you to use a different underlying mechanism from the advertised name. For
example:
.code
sasl:
  driver = gsasl
  public_name = X-ANYTHING
  server_mech = CRAM-MD5
  server_set_id = $auth1
.endd


.option server_password gsasl string&!! unset
Various mechanisms need access to the cleartext password on the server, so
that proof-of-possession can be demonstrated on the wire, without sending
the password itself.

The data available for lookup varies per mechanism.
In all cases, &$auth1$& is set to the &'authentication id'&.
The &$auth2$& variable will always be the &'authorization id'& (&'authz'&)
if available, else the empty string.
The &$auth3$& variable will always be the &'realm'& if available,
else the empty string.

A forced failure will cause authentication to defer.

If using this option, it may make sense to set the &%server_condition%&
option to be simply "true".


.option server_realm gsasl string&!! unset
This specifies the SASL realm that the server claims to be in.
Some mechanisms will use this data.


.option server_scram_iter gsasl string&!! 4096
This option provides data for the SCRAM family of mechanisms.
The &$auth1$&, &$auth2$& and &$auth3$& variables are available
when this option is expanded.

The result of expansion should be a decimal number,
and represents both a lower-bound on the security, and
a compute cost factor imposed on the client
(if it does not cache results, or the server changes
either the iteration count or the salt).
A minimum value of 4096 is required by the standards
for all current SCRAM mechanism variants.

.option server_scram_salt gsasl string&!! unset
This option provides data for the SCRAM family of mechanisms.
The &$auth1$&, &$auth2$& and &$auth3$& variables are available
when this option is expanded.
The value should be a base64-encoded string,
of random data typically 4-to-16 bytes long.
If unset or empty after expansion the library will provides a value for the
protocol conversation.


.option server_key gsasl string&!! unset
.option server_skey gsasl string&!! unset
These options can be used for the SCRAM family of mechanisms
to provide stored information related to a password,
the storage of which is preferable to plaintext.

&%server_key%& is the value defined in the SCRAM standards as ServerKey;
&%server_skey%& is StoredKey.

They are only available for version 1.9.0 (or later) of the gsasl library.
When this is so, the macros
_OPT_AUTHENTICATOR_GSASL_SERVER_KEY
and _HAVE_AUTH_GSASL_SCRAM_S_KEY
will be defined.

The &$authN$& variables are available when these options are expanded.

If set, the results of expansion should for each
should be a 28 (for SHA-1) or 44 (for SHA-256) character string
of base64-coded data, and will be used in preference to the
&%server_password%& option.
If unset or not of the right length, &%server_password%& will be used.

The libgsasl library release includes a utility &'gsasl'& which can be used
to generate these values.


.option server_service gsasl string &`smtp`&
This is the SASL service that the server claims to implement.
Some mechanisms will use this data.


.section "&(gsasl)& auth variables" "SECTgsaslauthvar"
.vindex "&$auth1$&, &$auth2$&, etc"
These may be set when evaluating specific options, as detailed above.
They will also be set when evaluating &%server_condition%&.

Unless otherwise stated below, the &(gsasl)& integration will use the following
meanings for these variables:

.ilist
.vindex "&$auth1$&"
&$auth1$&: the &'authentication id'&
.next
.vindex "&$auth2$&"
&$auth2$&: the &'authorization id'&
.next
.vindex "&$auth3$&"
&$auth3$&: the &'realm'&
.endlist

On a per-mechanism basis:

.ilist
.cindex "authentication" "EXTERNAL"
EXTERNAL: only &$auth1$& is set, to the possibly empty &'authorization id'&;
the &%server_condition%& option must be present.
.next
.cindex "authentication" "ANONYMOUS"
ANONYMOUS: only &$auth1$& is set, to the possibly empty &'anonymous token'&;
the &%server_condition%& option must be present.
.next
.cindex "authentication" "GSSAPI"
GSSAPI: &$auth1$& will be set to the &'GSSAPI Display Name'&;
&$auth2$& will be set to the &'authorization id'&,
the &%server_condition%& option must be present.
.endlist

An &'anonymous token'& is something passed along as an unauthenticated
identifier; this is analogous to FTP anonymous authentication passing an
email address, or software-identifier@, as the "password".


An example showing the password having the realm specified in the callback
and demonstrating a Cyrus SASL to GSASL migration approach is:
.code
gsasl_cyrusless_crammd5:
  driver = gsasl
  public_name = CRAM-MD5
  server_realm = imap.example.org
  server_password = ${lookup{$auth1:$auth3:userPassword}\
                    dbmjz{/etc/sasldb2}{$value}fail}
  server_set_id = ${quote:$auth1}
  server_condition = yes
.endd


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The heimdal_gssapi authenticator" "CHAPheimdalgss"
.scindex IIDheimdalgssauth1 "&(heimdal_gssapi)& authenticator"
.scindex IIDheimdalgssauth2 "authenticators" "&(heimdal_gssapi)&"
.cindex "authentication" "GSSAPI"
.cindex "authentication" "Kerberos"
The &(heimdal_gssapi)& authenticator provides server integration for the
Heimdal GSSAPI/Kerberos library, permitting Exim to set a keytab pathname
reliably.

.option server_hostname heimdal_gssapi string&!! "see below"
This option selects the hostname that is used, with &%server_service%&,
for constructing the GSS server name, as a &'GSS_C_NT_HOSTBASED_SERVICE'&
identifier.  The default value is &`$primary_hostname`&.

.option server_keytab heimdal_gssapi string&!! unset
If set, then Heimdal will not use the system default keytab (typically
&_/etc/krb5.keytab_&) but instead the pathname given in this option.
The value should be a pathname, with no &"file:"& prefix.

.option server_service heimdal_gssapi string&!! "smtp"
This option specifies the service identifier used, in conjunction with
&%server_hostname%&, for building the identifier for finding credentials
from the keytab.


.section "&(heimdal_gssapi)& auth variables" "SECTheimdalgssauthvar"
Beware that these variables will typically include a realm, thus will appear
to be roughly like an email address already.  The &'authzid'& in &$auth2$& is
not verified, so a malicious client can set it to anything.

The &$auth1$& field should be safely trustable as a value from the Key
Distribution Center.  Note that these are not quite email addresses.
Each identifier is for a role, and so the left-hand-side may include a
role suffix.  For instance, &"joe/admin@EXAMPLE.ORG"&.

.vindex "&$auth1$&, &$auth2$&, etc"
.ilist
.vindex "&$auth1$&"
&$auth1$&: the &'authentication id'&, set to the GSS Display Name.
.next
.vindex "&$auth2$&"
&$auth2$&: the &'authorization id'&, sent within SASL encapsulation after
authentication.  If that was empty, this will also be set to the
GSS Display Name.
.endlist


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The spa authenticator" "CHAPspa"
.scindex IIDspaauth1 "&(spa)& authenticator"
.scindex IIDspaauth2 "authenticators" "&(spa)&"
.cindex "authentication" "Microsoft Secure Password"
.cindex "authentication" "NTLM"
.cindex "Microsoft Secure Password Authentication"
.cindex "NTLM authentication"
The &(spa)& authenticator provides client support for Microsoft's &'Secure
Password Authentication'& mechanism,
which is also sometimes known as NTLM (NT LanMan). The code for client side of
this authenticator was contributed by Marc Prud'hommeaux, and much of it is
taken from the &url(https://www.samba.org/,Samba project). The code for the
server side was subsequently contributed by Tom Kistner. The mechanism works as
follows:

.ilist
After the AUTH command has been accepted, the client sends an SPA
authentication request based on the user name and optional domain.
.next
The server sends back a challenge.
.next
The client builds a challenge response which makes use of the user's password
and sends it to the server, which then accepts or rejects it.
.endlist

Encryption is used to protect the password in transit.



.section "Using spa as a server" "SECID179"
.cindex "options" "&(spa)& authenticator (server)"
The &(spa)& authenticator has just one server option:

.option server_password spa string&!! unset
.cindex "numerical variables (&$1$& &$2$& etc)" "in &(spa)& authenticator"
This option is expanded, and the result must be the cleartext password for the
authenticating user, whose name is at this point in &$auth1$&. For
compatibility with previous releases of Exim, the user name is also placed in
&$1$&. However, the use of this variable for this purpose is now deprecated, as
it can lead to confusion in string expansions that also use numeric variables
for other things. For example:
.code
spa:
  driver = spa
  public_name = NTLM
  server_password = \
    ${lookup{$auth1}lsearch{/etc/exim/spa_clearpass}{$value}fail}
.endd
If the expansion is forced to fail, authentication fails. Any other expansion
failure causes a temporary error code to be returned.





.section "Using spa as a client" "SECID180"
.cindex "options" "&(spa)& authenticator (client)"
The &(spa)& authenticator has the following client options:



.option client_domain spa string&!! unset
This option specifies an optional domain for the authentication.


.option client_password spa string&!! unset
This option specifies the user's password, and must be set.


.option client_username spa string&!! unset
This option specifies the user name, and must be set. Here is an example of a
configuration of this authenticator for use with the mail servers at
&'msn.com'&:
.code
msn:
  driver = spa
  public_name = MSN
  client_username = msn/msn_username
  client_password = msn_plaintext_password
  client_domain = DOMAIN_OR_UNSET
.endd
.ecindex IIDspaauth1
.ecindex IIDspaauth2





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The external authenticator" "CHAPexternauth"
.scindex IIDexternauth1 "&(external)& authenticator"
.scindex IIDexternauth2 "authenticators" "&(external)&"
.cindex "authentication" "Client Certificate"
.cindex "authentication" "X509"
.cindex "Certificate-based authentication"
The &(external)& authenticator provides support for
authentication based on non-SMTP information.
The specification is in
&url(https://tools.ietf.org/html/rfc4422,RFC 4422) Appendix A.
It is only a transport and negotiation mechanism;
the process of authentication is entirely controlled
by the server configuration.

The client presents an identity in-clear.
It is probably wise for a server to only advertise,
and for clients to only attempt,
this authentication method on a secure (eg. under TLS) connection.

One possible use, compatible with the
&url(https://k9mail.github.io/,K-9 Mail Android client)
is for using X509 client certificates.

It thus overlaps in function with the TLS authenticator
(see &<<CHAPtlsauth>>&)
but is a full SMTP SASL authenticator
rather than being implicit for TLS-connection carried
client certificates only.

The examples and discussion in this chapter assume that
client-certificate authentication is being done.

The client must present a certificate,
for which it must have been requested via the
&%tls_verify_hosts%& or &%tls_try_verify_hosts%& main options
(see &<<CHAPTLS>>&).
For authentication to be effective the certificate should be
verifiable against a trust-anchor certificate known to the server.

.section "External options" "SECTexternsoptions"
.cindex "options" "&(external)& authenticator (server)"
The &(external)& authenticator has two server options:

.option server_param2 external string&!! unset
.option server_param3 external string&!! unset
.cindex "variables (&$auth1$& &$auth2$& etc)" "in &(external)& authenticator"
These options are expanded before the &%server_condition%& option
and the result are placed in &$auth2$& and &$auth3$& resectively.
If the expansion is forced to fail, authentication fails. Any other expansion
failure causes a temporary error code to be returned.

They can be used to clarify the coding of a complex &%server_condition%&.

.section "Using external in a server" "SECTexternserver"
.cindex "AUTH" "in &(external)& authenticator"
.cindex "numerical variables (&$1$& &$2$& etc)" &&&
        "in &(external)& authenticator"
.vindex "&$auth1$&, &$auth2$&, etc"
.cindex "base64 encoding" "in &(external)& authenticator"

When running as a server, &(external)& performs the authentication test by
expanding a string. The data sent by the client with the AUTH command, or in
response to subsequent prompts, is base64 encoded, and so may contain any byte
values when decoded. The decoded value is treated as
an identity for authentication and
placed in the expansion variable &$auth1$&.

For compatibility with previous releases of Exim, the value is also placed in
the expansion variable &$1$&. However, the use of this
variable for this purpose is now deprecated, as it can lead to confusion in
string expansions that also use them for other things.

.vindex "&$authenticated_id$&"
Once an identity has been received,
&%server_condition%& is expanded. If the expansion is forced to fail,
authentication fails. Any other expansion failure causes a temporary error code
to be returned. If the result of a successful expansion is an empty string,
&"0"&, &"no"&, or &"false"&, authentication fails. If the result of the
expansion is &"1"&, &"yes"&, or &"true"&, authentication succeeds and the
generic &%server_set_id%& option is expanded and saved in &$authenticated_id$&.
For any other result, a temporary error code is returned, with the expanded
string as the error text.

Example:
.code
ext_ccert_san_mail:
  driver =            external
  public_name =       EXTERNAL

  server_advertise_condition = $tls_in_certificate_verified
  server_param2 =     ${certextract {subj_altname,mail,>:} \
                                    {$tls_in_peercert}}
  server_condition =  ${if forany {$auth2} \
                            {eq {$item}{$auth1}}}
  server_set_id =     $auth1
.endd
This accepts a client certificate that is verifiable against any
of your configured trust-anchors
(which usually means the full set of public CAs)
and which has a mail-SAN matching the claimed identity sent by the client.

&*Note*&: up to TLS1.2, the client cert is on the wire in-clear, including the SAN.
The account name is therefore guessable by an opponent.
TLS 1.3 protects both server and client certificates, and is not vulnerable
in this way.


.section "Using external in a client" "SECTexternclient"
.cindex "options" "&(external)& authenticator (client)"
The &(external)& authenticator has one client option:

.option client_send external string&!! unset
This option is expanded and sent with the AUTH command as the
identity being asserted.

Example:
.code
ext_ccert:
  driver =      external
  public_name = EXTERNAL

  client_condition = ${if !eq{$tls_out_cipher}{}}
  client_send = myaccount@smarthost.example.net
.endd


.ecindex IIDexternauth1
.ecindex IIDexternauth2





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The tls authenticator" "CHAPtlsauth"
.scindex IIDtlsauth1 "&(tls)& authenticator"
.scindex IIDtlsauth2 "authenticators" "&(tls)&"
.cindex "authentication" "Client Certificate"
.cindex "authentication" "X509"
.cindex "Certificate-based authentication"
The &(tls)& authenticator provides server support for
authentication based on client certificates.

It is not an SMTP authentication mechanism and is not
advertised by the server as part of the SMTP EHLO response.
It is an Exim authenticator in the sense that it affects
the protocol element of the log line, can be tested for
by the &%authenticated%& ACL condition, and can set
the &$authenticated_id$& variable.

The client must present a verifiable certificate,
for which it must have been requested via the
&%tls_verify_hosts%& or &%tls_try_verify_hosts%& main options
(see &<<CHAPTLS>>&).

If an authenticator of this type is configured it is
run immediately after a TLS connection being negotiated
(due to either STARTTLS or TLS-on-connect)
and can authenticate the connection.
If it does, SMTP authentication is not subsequently offered.

A maximum of one authenticator of this type may be present.


.cindex "options" "&(tls)& authenticator (server)"
The &(tls)& authenticator has three server options:

.option server_param1 tls string&!! unset
.cindex "variables (&$auth1$& &$auth2$& etc)" "in &(tls)& authenticator"
This option is expanded after the TLS negotiation and
the result is placed in &$auth1$&.
If the expansion is forced to fail, authentication fails. Any other expansion
failure causes a temporary error code to be returned.

.option server_param2 tls string&!! unset
.option server_param3 tls string&!! unset
As above, for &$auth2$& and &$auth3$&.

&%server_param1%& may also be spelled &%server_param%&.


Example:
.code
tls:
  driver = tls
  server_param1 =     ${certextract {subj_altname,mail,>:} \
                                    {$tls_in_peercert}}
  server_condition =  ${if and { {eq{$tls_in_certificate_verified}{1}} \
                                 {forany {$auth1} \
                            {!= {0} \
                                {${lookup ldap{ldap:///\
                         mailname=${quote_ldap_dn:${lc:$item}},\
                         ou=users,LDAP_DC?mailid} {$value}{0} \
                       }    }  } }}}
  server_set_id =     ${if = {1}{${listcount:$auth1}} {$auth1}{}}
.endd
This accepts a client certificate that is verifiable against any
of your configured trust-anchors
(which usually means the full set of public CAs)
and which has a SAN with a good account name.

Note that, up to TLS1.2, the client cert is on the wire in-clear, including the SAN,
The account name is therefore guessable by an opponent.
TLS 1.3 protects both server and client certificates, and is not vulnerable
in this way.
Likewise, a traditional plaintext SMTP AUTH done inside TLS is not.

. An alternative might use
. .code
.   server_param1 = ${sha256:$tls_in_peercert}
. .endd
. to require one of a set of specific certs that define a given account
. (the verification is still required, but mostly irrelevant).
. This would help for per-device use.
.
. However, for the future we really need support for checking a
. user cert in LDAP - which probably wants a base-64 DER.

.ecindex IIDtlsauth1
.ecindex IIDtlsauth2


Note that because authentication is traditionally an SMTP operation,
the &%authenticated%& ACL condition cannot be used in
a connect- or helo-ACL.



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Encrypted SMTP connections using TLS/SSL" "CHAPTLS" &&&
         "Encrypted SMTP connections"
.scindex IIDencsmtp1 "encryption" "on SMTP connection"
.scindex IIDencsmtp2 "SMTP" "encryption"
.cindex "TLS" "on SMTP connection"
.cindex "OpenSSL"
.cindex "GnuTLS"
Support for TLS (Transport Layer Security), formerly known as SSL (Secure
Sockets Layer), is implemented by making use of the OpenSSL library or the
GnuTLS library (Exim requires GnuTLS release 1.0 or later). There is no
cryptographic code in the Exim distribution itself for implementing TLS. In
order to use this feature you must install OpenSSL or GnuTLS, and then build a
version of Exim that includes TLS support (see section &<<SECTinctlsssl>>&).
You also need to understand the basic concepts of encryption at a managerial
level, and in particular, the way that public keys, private keys, and
certificates are used.

&url(https://www.rfc-editor.org/rfc/rfc3207,RFC 3207)
defines how SMTP connections can make use of encryption. Once a
connection is established, the client issues a STARTTLS command. If the
server accepts this, the client and the server negotiate an encryption
mechanism. If the negotiation succeeds, the data that subsequently passes
between them is encrypted.

Exim's ACLs can detect whether the current SMTP session is encrypted or not,
and if so, what cipher suite is in use, whether the client supplied a
certificate, and whether or not that certificate was verified. This makes it
possible for an Exim server to deny or accept certain commands based on the
encryption state.

&*Warning*&: Certain types of firewall and certain anti-virus products can
disrupt TLS connections. You need to turn off SMTP scanning for these products
in order to get TLS to work.



.section "Support for the &""submissions""& (aka &""ssmtp""& and &""smtps""&) protocol" &&&
         "SECID284"
.cindex "submissions protocol"
.cindex "ssmtp protocol"
.cindex "smtps protocol"
.cindex "SMTP" "submissions protocol"
.cindex "SMTP" "ssmtp protocol"
.cindex "SMTP" "smtps protocol"
The history of port numbers for TLS in SMTP is a little messy and has been
contentious.  As of
&url(https://www.rfc-editor.org/rfc/rfc8314,RFC 8314),
the common practice of using the historically
allocated port 465 for "email submission but with TLS immediately upon connect
instead of using STARTTLS" is officially blessed by the IETF, and recommended
by them in preference to STARTTLS.

The name originally assigned to the port was &"ssmtp"& or &"smtps"&, but as
clarity emerged over the dual roles of SMTP, for MX delivery and Email
Submission, nomenclature has shifted.  The modern name is now &"submissions"&.

This approach was, for a while, officially abandoned when encrypted SMTP was
standardized, but many clients kept using it, even as the TCP port number was
reassigned for other use.
Thus you may encounter guidance claiming that you shouldn't enable use of
this port.
In practice, a number of mail-clients have only ever supported submissions,
not submission with STARTTLS upgrade.
Ideally, offer both submission (587) and submissions (465) service.

Exim supports TLS-on-connect by means of the &%tls_on_connect_ports%&
global option. Its value must be a list of port numbers;
the most common use is expected to be:
.code
tls_on_connect_ports = 465
.endd
The port numbers specified by this option apply to all SMTP connections, both
via the daemon and via &'inetd'&. You still need to specify all the ports that
the daemon uses (by setting &%daemon_smtp_ports%& or &%local_interfaces%& or
the &%-oX%& command line option) because &%tls_on_connect_ports%& does not add
an extra port &-- rather, it specifies different behaviour on a port that is
defined elsewhere.

There is also a &%-tls-on-connect%& command line option. This overrides
&%tls_on_connect_ports%&; it forces the TLS-only behaviour for all ports.






.section "OpenSSL vs GnuTLS" "SECTopenvsgnu"
.cindex "TLS" "OpenSSL &'vs'& GnuTLS"
TLS is supported in Exim using either the OpenSSL or GnuTLS library.
To build Exim to use OpenSSL you need to set
.code
USE_OPENSSL=yes
.endd
in Local/Makefile.

To build Exim to use GnuTLS, you need to set
.code
USE_GNUTLS=yes
.endd
in Local/Makefile.

You must also set TLS_LIBS and TLS_INCLUDE appropriately, so that the
include files and libraries for GnuTLS can be found.

There are some differences in usage when using GnuTLS instead of OpenSSL:

.ilist
The &%tls_verify_certificates%& option
cannot be the path of a directory
for GnuTLS versions before 3.3.6
(for later versions, or OpenSSL, it can be either).
.next
The default value for &%tls_dhparam%& differs for historical reasons.
.next
.vindex "&$tls_in_peerdn$&"
.vindex "&$tls_out_peerdn$&"
Distinguished Name (DN) strings reported by the OpenSSL library use a slash for
separating fields; GnuTLS uses commas, in accordance with
&url(https://www.rfc-editor.org/rfc/rfc2253,RFC 2253). This
affects the value of the &$tls_in_peerdn$& and &$tls_out_peerdn$& variables.
.next
OpenSSL identifies cipher suites using hyphens as separators, for example:
DES-CBC3-SHA. GnuTLS historically used underscores, for example:
RSA_ARCFOUR_SHA. What is more, OpenSSL complains if underscores are present
in a cipher list. To make life simpler, Exim changes underscores to hyphens
for OpenSSL and passes the string unchanged to GnuTLS (expecting the library
to handle its own older variants) when processing lists of cipher suites in the
&%tls_require_ciphers%& options (the global option and the &(smtp)& transport
option).
.next
The &%tls_require_ciphers%& options operate differently, as described in the
sections &<<SECTreqciphssl>>& and &<<SECTreqciphgnu>>&.
.next
The &%tls_dh_min_bits%& SMTP transport option is only honoured by GnuTLS.
When using OpenSSL, this option is ignored.
(If an API is found to let OpenSSL be configured in this way,
let the Exim Maintainers know and we'll likely use it).
.next
With GnuTLS, if an explicit list is used for the &%tls_privatekey%& main option,
it must be ordered to match the &%tls_certificate%& list.
.next
Some other recently added features may only be available in one or the other.
This should be documented with the feature.  If the documentation does not
explicitly state that the feature is infeasible in the other TLS
implementation, then patches are welcome.
.next
The output from "exim -bV" will show which (if any) support was included
in the build.
Also, the macro "_HAVE_OPENSSL" or "_HAVE_GNUTLS" will be defined.
.endlist


.section "GnuTLS parameter computation" "SECTgnutlsparam"
This section only applies if &%tls_dhparam%& is set to &`historic`& or to
an explicit path; if the latter, then the text about generation still applies,
but not the chosen filename.
By default, as of Exim 4.80 a hard-coded D-H prime is used.
See the documentation of &%tls_dhparam%& for more information.

GnuTLS uses D-H parameters that may take a substantial amount of time
to compute. It is unreasonable to re-compute them for every TLS session.
Therefore, Exim keeps this data in a file in its spool directory, called
&_gnutls-params-NNNN_& for some value of NNNN, corresponding to the number
of bits requested.
The file is owned by the Exim user and is readable only by
its owner. Every Exim process that start up GnuTLS reads the D-H
parameters from this file. If the file does not exist, the first Exim process
that needs it computes the data and writes it to a temporary file which is
renamed once it is complete. It does not matter if several Exim processes do
this simultaneously (apart from wasting a few resources). Once a file is in
place, new Exim processes immediately start using it.

For maximum security, the parameters that are stored in this file should be
recalculated periodically, the frequency depending on your paranoia level.
If you are avoiding using the fixed D-H primes published in RFCs, then you
are concerned about some advanced attacks and will wish to do this; if you do
not regenerate then you might as well stick to the standard primes.

Arranging this is easy in principle; just delete the file when you want new
values to be computed. However, there may be a problem. The calculation of new
parameters needs random numbers, and these are obtained from &_/dev/random_&.
If the system is not very active, &_/dev/random_& may delay returning data
until enough randomness (entropy) is available. This may cause Exim to hang for
a substantial amount of time, causing timeouts on incoming connections.

The solution is to generate the parameters externally to Exim. They are stored
in &_gnutls-params-N_& in PEM format, which means that they can be
generated externally using the &(certtool)& command that is part of GnuTLS.

To replace the parameters with new ones, instead of deleting the file
and letting Exim re-create it, you can generate new parameters using
&(certtool)& and, when this has been done, replace Exim's cache file by
renaming. The relevant commands are something like this:
.code
# ls
[ look for file; assume gnutls-params-2236 is the most recent ]
# rm -f new-params
# touch new-params
# chown exim:exim new-params
# chmod 0600 new-params
# certtool --generate-dh-params --bits 2236 >>new-params
# openssl dhparam -noout -text -in new-params | head
[ check the first line, make sure it's not more than 2236;
  if it is, then go back to the start ("rm") and repeat
  until the size generated is at most the size requested ]
# chmod 0400 new-params
# mv new-params gnutls-params-2236
.endd
If Exim never has to generate the parameters itself, the possibility of
stalling is removed.

The filename changed in Exim 4.80, to gain the -bits suffix.  The value which
Exim will choose depends upon the version of GnuTLS in use.  For older GnuTLS,
the value remains hard-coded in Exim as 1024.  As of GnuTLS 2.12.x, there is
a way for Exim to ask for the "normal" number of bits for D-H public-key usage,
and Exim does so.  This attempt to remove Exim from TLS policy decisions
failed, as GnuTLS 2.12 returns a value higher than the current hard-coded limit
of the NSS library.  Thus Exim gains the &%tls_dh_max_bits%& global option,
which applies to all D-H usage, client or server.  If the value returned by
GnuTLS is greater than &%tls_dh_max_bits%& then the value will be clamped down
to &%tls_dh_max_bits%&.  The default value has been set at the current NSS
limit, which is still much higher than Exim historically used.

The filename and bits used will change as the GnuTLS maintainers change the
value for their parameter &`GNUTLS_SEC_PARAM_NORMAL`&, as clamped by
&%tls_dh_max_bits%&.  At the time of writing (mid 2012), GnuTLS 2.12 recommends
2432 bits, while NSS is limited to 2236 bits.

In fact, the requested value will be *lower* than &%tls_dh_max_bits%&, to
increase the chance of the generated prime actually being within acceptable
bounds, as GnuTLS has been observed to overshoot.  Note the check step in the
procedure above.  There is no sane procedure available to Exim to double-check
the size of the generated prime, so it might still be too large.


.section "Requiring specific ciphers in OpenSSL" "SECTreqciphssl"
.cindex "TLS" "requiring specific ciphers (OpenSSL)"
.oindex "&%tls_require_ciphers%&" "OpenSSL"
There is a function in the OpenSSL library that can be passed a list of cipher
suites before the cipher negotiation takes place. This specifies which ciphers
are acceptable for TLS versions prior to 1.3.
The list is colon separated and may contain names like
DES-CBC3-SHA. Exim passes the expanded value of &%tls_require_ciphers%&
directly to this function call.
Many systems will install the OpenSSL manual-pages, so you may have
&'ciphers(1)'& available to you.
The following quotation from the OpenSSL
documentation specifies what forms of item are allowed in the cipher string:

.ilist
It can consist of a single cipher suite such as RC4-SHA.
.next
It can represent a list of cipher suites containing a certain algorithm,
or cipher suites of a certain type. For example SHA1 represents all
ciphers suites using the digest algorithm SHA1 and SSLv3 represents all
SSL v3 algorithms.
.next
Lists of cipher suites can be combined in a single cipher string using
the + character. This is used as a logical and operation. For example
SHA1+DES represents all cipher suites containing the SHA1 and the DES
algorithms.
.endlist

Each cipher string can be optionally preceded by one of the characters &`!`&,
&`-`& or &`+`&.
.ilist
If &`!`& is used, the ciphers are permanently deleted from the list. The
ciphers deleted can never reappear in the list even if they are explicitly
stated.
.next
If &`-`& is used, the ciphers are deleted from the list, but some or all
of the ciphers can be added again by later options.
.next
If &`+`& is used, the ciphers are moved to the end of the list. This
option does not add any new ciphers; it just moves matching existing ones.
.endlist

If none of these characters is present, the string is interpreted as
a list of ciphers to be appended to the current preference list. If the list
includes any ciphers already present they will be ignored: that is, they will
not be moved to the end of the list.
.endlist

The OpenSSL &'ciphers(1)'& command may be used to test the results of a given
string:
.code
# note single-quotes to get ! past any shell history expansion
$ openssl ciphers 'HIGH:!MD5:!SHA1'
.endd

This example will let the library defaults be permitted on the MX port, where
there's probably no identity verification anyway, but ups the ante on the
submission ports where the administrator might have some influence on the
choice of clients used:
.code
# OpenSSL variant; see man ciphers(1)
tls_require_ciphers = ${if =={$received_port}{25}\
                           {DEFAULT}\
                           {HIGH:!MD5:!SHA1}}
.endd

This example will prefer ECDSA-authenticated ciphers over RSA ones:
.code
tls_require_ciphers = ECDSA:RSA:!COMPLEMENTOFDEFAULT
.endd

For TLS version 1.3 the control available is less fine-grained
and Exim does not provide access to it at present.
The value of the &%tls_require_ciphers%& option is ignored when
TLS version 1.3 is negotiated.

As of writing the library default cipher suite list for TLSv1.3 is
.code
TLS_AES_256_GCM_SHA384:TLS_CHACHA20_POLY1305_SHA256:TLS_AES_128_GCM_SHA256
.endd


.section "Requiring specific ciphers or other parameters in GnuTLS" &&&
         "SECTreqciphgnu"
.cindex "GnuTLS" "specifying parameters for"
.cindex "TLS" "specifying ciphers (GnuTLS)"
.cindex "TLS" "specifying key exchange methods (GnuTLS)"
.cindex "TLS" "specifying MAC algorithms (GnuTLS)"
.cindex "TLS" "specifying protocols (GnuTLS)"
.cindex "TLS" "specifying priority string (GnuTLS)"
.oindex "&%tls_require_ciphers%&" "GnuTLS"
The GnuTLS library allows the caller to provide a "priority string", documented
as part of the &[gnutls_priority_init]& function.  This is very similar to the
ciphersuite specification in OpenSSL.

The &%tls_require_ciphers%& option is treated as the GnuTLS priority string
and controls both protocols and ciphers.

The &%tls_require_ciphers%& option is available both as an global option,
controlling how Exim behaves as a server, and also as an option of the
&(smtp)& transport, controlling how Exim behaves as a client.  In both cases
the value is string expanded.  The resulting string is not an Exim list and
the string is given to the GnuTLS library, so that Exim does not need to be
aware of future feature enhancements of GnuTLS.

Documentation of the strings accepted may be found in the GnuTLS manual, under
"Priority strings".  This is online as
&url(https://www.gnutls.org/manual/html_node/Priority-Strings.html),
but beware that this relates to GnuTLS 3, which may be newer than the version
installed on your system.  If you are using GnuTLS 3,
then the example code
&url(https://www.gnutls.org/manual/gnutls.html#Listing-the-ciphersuites-in-a-priority-string)
on that site can be used to test a given string.

For example:
.code
# Disable older versions of protocols
tls_require_ciphers = NORMAL:%LATEST_RECORD_VERSION:-VERS-SSL3.0
.endd

Prior to Exim 4.80, an older API of GnuTLS was used, and Exim supported three
additional options, "&%gnutls_require_kx%&", "&%gnutls_require_mac%&" and
"&%gnutls_require_protocols%&".  &%tls_require_ciphers%& was an Exim list.

This example will let the library defaults be permitted on the MX port, where
there's probably no identity verification anyway, and lowers security further
by increasing compatibility; but this ups the ante on the submission ports
where the administrator might have some influence on the choice of clients
used:
.code
# GnuTLS variant
tls_require_ciphers = ${if =={$received_port}{25}\
                           {NORMAL:%COMPAT}\
                           {SECURE128}}
.endd


.section "Configuring an Exim server to use TLS" "SECID182"
.cindex "TLS" "configuring an Exim server"
.cindex "ESMTP extensions" STARTTLS
When Exim has been built with TLS support, it advertises the availability of
the STARTTLS command to client hosts that match &%tls_advertise_hosts%&,
but not to any others. The default value of this option is *, which means
that STARTTLS is always advertised.  Set it to blank to never advertise;
this is reasonable for systems that want to use TLS only as a client.

If STARTTLS is to be used you
need to set some other options in order to make TLS available.

If a client issues a STARTTLS command and there is some configuration
problem in the server, the command is rejected with a 454 error. If the client
persists in trying to issue SMTP commands, all except QUIT are rejected
with the error
.code
554 Security failure
.endd
If a STARTTLS command is issued within an existing TLS session, it is
rejected with a 554 error code.

To enable TLS operations on a server, the &%tls_advertise_hosts%& option
must be set to match some hosts. The default is * which matches all hosts.

If this is all you do, TLS encryption will be enabled but not authentication -
meaning that the peer has no assurance it is actually you he is talking to.
You gain protection from a passive sniffer listening on the wire but not
from someone able to intercept the communication.

Further protection requires some further configuration at the server end.

To make TLS work you need to set, in the server,
.code
tls_certificate = /some/file/name
tls_privatekey = /some/file/name
.endd
These options are, in fact, expanded strings, so you can make them depend on
the identity of the client that is connected if you wish. The first file
contains the server's X509 certificate, and the second contains the private key
that goes with it. These files need to be
PEM format and readable by the Exim user, and must
always be given as full path names.
The key must not be password-protected.
They can be the same file if both the
certificate and the key are contained within it. If &%tls_privatekey%& is not
set, or if its expansion is forced to fail or results in an empty string, this
is assumed to be the case. The certificate file may also contain intermediate
certificates that need to be sent to the client to enable it to authenticate
the server's certificate.

For dual-stack (eg. RSA and ECDSA) configurations, these options can be
colon-separated lists of file paths.  Ciphers using given authentication
algorithms require the presence of a suitable certificate to supply the
public-key.  The server selects among the certificates to present to the
client depending on the selected cipher, hence the priority ordering for
ciphers will affect which certificate is used.

If you do not understand about certificates and keys, please try to find a
source of this background information, which is not Exim-specific. (There are a
few comments below in section &<<SECTcerandall>>&.)

&*Note*&: These options do not apply when Exim is operating as a client &--
they apply only in the case of a server. If you need to use a certificate in an
Exim client, you must set the options of the same names in an &(smtp)&
transport.

With just these options, an Exim server will be able to use TLS. It does not
require the client to have a certificate (but see below for how to insist on
this). There is one other option that may be needed in other situations. If
.code
tls_dhparam = /some/file/name
.endd
is set, the SSL library is initialized for the use of Diffie-Hellman ciphers
with the parameters contained in the file.
Set this to &`none`& to disable use of DH entirely, by making no prime
available:
.code
tls_dhparam = none
.endd
This may also be set to a string identifying a standard prime to be used for
DH; if it is set to &`default`& or, for OpenSSL, is unset, then the prime
used is &`ike23`&.  There are a few standard primes available, see the
documentation for &%tls_dhparam%& for the complete list.

See the command
.code
openssl dhparam
.endd
for a way of generating file data.

The strings supplied for these three options are expanded every time a client
host connects. It is therefore possible to use different certificates and keys
for different hosts, if you so wish, by making use of the client's IP address
in &$sender_host_address$& to control the expansion. If a string expansion is
forced to fail, Exim behaves as if the option is not set.

.cindex "cipher" "logging"
.cindex "log" "TLS cipher"
.vindex "&$tls_in_cipher$&"
The variable &$tls_in_cipher$& is set to the cipher suite that was negotiated for
an incoming TLS connection. It is included in the &'Received:'& header of an
incoming message (by default &-- you can, of course, change this), and it is
also included in the log line that records a message's arrival, keyed by
&"X="&, unless the &%tls_cipher%& log selector is turned off. The &%encrypted%&
condition can be used to test for specific cipher suites in ACLs.

Once TLS has been established, the ACLs that run for subsequent SMTP commands
can check the name of the cipher suite and vary their actions accordingly. The
cipher suite names vary, depending on which TLS library is being used. For
example, OpenSSL uses the name DES-CBC3-SHA for the cipher suite which in other
contexts is known as TLS_RSA_WITH_3DES_EDE_CBC_SHA. Check the OpenSSL or GnuTLS
documentation for more details.

For outgoing SMTP deliveries, &$tls_out_cipher$& is used and logged
(again depending on the &%tls_cipher%& log selector).


.subsection "Requesting and verifying client certificates"
.cindex "certificate" "verification of client"
.cindex "TLS" "client certificate verification"
If you want an Exim server to request a certificate when negotiating a TLS
session with a client, you must set either &%tls_verify_hosts%& or
&%tls_try_verify_hosts%&. You can, of course, set either of them to * to
apply to all TLS connections. For any host that matches one of these options,
Exim requests a certificate as part of the setup of the TLS session. The
contents of the certificate are verified by comparing it with a list of
expected trust-anchors or certificates.
These may be the system default set (depending on library version),
an explicit file or,
depending on library version, a directory, identified by
&%tls_verify_certificates%&.

A file can contain multiple certificates, concatenated end to end. If a
directory is used
(OpenSSL only),
each certificate must be in a separate file, with a name (or a symbolic link)
of the form <&'hash'&>.0, where <&'hash'&> is a hash value constructed from the
certificate. You can compute the relevant hash by running the command
.code
openssl x509 -hash -noout -in /cert/file
.endd
where &_/cert/file_& contains a single certificate.

There is no checking of names of the client against the certificate
Subject Name or Subject Alternate Names.

The difference between &%tls_verify_hosts%& and &%tls_try_verify_hosts%& is
what happens if the client does not supply a certificate, or if the certificate
does not match any of the certificates in the collection named by
&%tls_verify_certificates%&. If the client matches &%tls_verify_hosts%&, the
attempt to set up a TLS session is aborted, and the incoming connection is
dropped. If the client matches &%tls_try_verify_hosts%&, the (encrypted) SMTP
session continues. ACLs that run for subsequent SMTP commands can detect the
fact that no certificate was verified, and vary their actions accordingly. For
example, you can insist on a certificate before accepting a message for
relaying, but not when the message is destined for local delivery.

.vindex "&$tls_in_peerdn$&"
When a client supplies a certificate (whether it verifies or not), the value of
the Distinguished Name of the certificate is made available in the variable
&$tls_in_peerdn$& during subsequent processing of the message.

.cindex "log" "distinguished name"
Because it is often a long text string, it is not included in the log line or
&'Received:'& header by default. You can arrange for it to be logged, keyed by
&"DN="&, by setting the &%tls_peerdn%& log selector, and you can use
&%received_header_text%& to change the &'Received:'& header. When no
certificate is supplied, &$tls_in_peerdn$& is empty.


.subsection "Caching of static server configuration items" "SSECTserverTLScache"
.cindex certificate caching
.cindex privatekey caching
.cindex crl caching
.cindex ocsp caching
.cindex ciphers caching
.cindex "CA bundle" caching
.cindex "certificate authorities" caching
.cindex tls_certificate caching
.cindex tls_privatekey caching
.cindex tls_crl caching
.cindex tls_ocsp_file caching
.cindex tls_require_ciphers caching
.cindex tls_verify_certificate caching
.cindex caching certificate
.cindex caching privatekey
.cindex caching crl
.cindex caching ocsp
.cindex caching ciphers
.cindex caching "certificate authorities
If any of the main configuration options &%tls_certificate%&, &%tls_privatekey%&,
&%tls_crl%& and &%tls_ocsp_file%& have values with no
expandable elements,
then the associated information is loaded at daemon startup.
It is made available
to child processes forked for handling received SMTP connections.

This caching is currently only supported under Linux and FreeBSD.

If caching is not possible, for example if an item has to be dependent
on the peer host so contains a &$sender_host_name$& expansion, the load
of the associated information is done at the startup of the TLS connection.

The cache is invalidated and reloaded after any changes to the directories
containing files specified by these options.

The information specified by the main option &%tls_verify_certificates%&
is similarly cached so long as it specifies files explicitly
or (under GnuTLS) is the string &"system,cache"&.
The latter case is not automatically invalidated;
it is the operator's responsibility to arrange for a daemon restart
any time the system certificate authority bundle is updated.
A HUP signal is sufficient for this.
The value &"system"& results in no caching under GnuTLS.

The macro _HAVE_TLS_CA_CACHE will be defined if the suffix for "system"
is acceptable in configurations for the Exim executable.

Caching of the system Certificate Authorities bundle can
save significant time and processing on every TLS connection
accepted by Exim.




.section "Configuring an Exim client to use TLS" "SECTclientTLS"
.cindex "cipher" "logging"
.cindex "log" "TLS cipher"
.cindex "log" "distinguished name"
.cindex "TLS" "configuring an Exim client"
The &%tls_cipher%& and &%tls_peerdn%& log selectors apply to outgoing SMTP
deliveries as well as to incoming, the latter one causing logging of the
server certificate's DN. The remaining client configuration for TLS is all
within the &(smtp)& transport.

.cindex "ESMTP extensions" STARTTLS
It is not necessary to set any options to have TLS work in the &(smtp)&
transport. If Exim is built with TLS support, and TLS is advertised by a
server, the &(smtp)& transport always tries to start a TLS session. However,
this can be prevented by setting &%hosts_avoid_tls%& (an option of the
transport) to a list of server hosts for which TLS should not be used.

If you do not want Exim to attempt to send messages unencrypted when an attempt
to set up an encrypted connection fails in any way, you can set
&%hosts_require_tls%& to a list of hosts for which encryption is mandatory. For
those hosts, delivery is always deferred if an encrypted connection cannot be
set up. If there are any other hosts for the address, they are tried in the
usual way.

When the server host is not in &%hosts_require_tls%&, Exim may try to deliver
the message unencrypted. It always does this if the response to STARTTLS is
a 5&'xx'& code. For a temporary error code, or for a failure to negotiate a TLS
session after a success response code, what happens is controlled by the
&%tls_tempfail_tryclear%& option of the &(smtp)& transport. If it is false,
delivery to this host is deferred, and other hosts (if available) are tried. If
it is true, Exim attempts to deliver unencrypted after a 4&'xx'& response to
STARTTLS, and if STARTTLS is accepted, but the subsequent TLS
negotiation fails, Exim closes the current connection (because it is in an
unknown state), opens a new one to the same host, and then tries the delivery
unencrypted.

The &%tls_certificate%& and &%tls_privatekey%& options of the &(smtp)&
transport provide the client with a certificate, which is passed to the server
if it requests it.
This is an optional thing for TLS connections, although either end
may insist on it.
If the server is Exim, it will request a certificate only if
&%tls_verify_hosts%& or &%tls_try_verify_hosts%& matches the client.

&*Note*&: Do not use a certificate which has the OCSP-must-staple extension,
for client use (they are usable for server use).
As the TLS protocol has no means for the client to staple before TLS 1.3 it will result
in failed connections.

If the &%tls_verify_certificates%& option is set on the &(smtp)& transport, it
specifies a collection of expected server certificates.
These may be
the system default set (depending on library version),
a file,
or (depending on library version) a directory.
The client verifies the server's certificate
against this collection, taking into account any revoked certificates that are
in the list defined by &%tls_crl%&.
Failure to verify fails the TLS connection unless either of the
&%tls_verify_hosts%& or &%tls_try_verify_hosts%& options are set.

The &%tls_verify_hosts%& and &%tls_try_verify_hosts%& options restrict
certificate verification to the listed servers.  Verification either must
or need not succeed respectively.

The &%tls_verify_cert_hostnames%& option lists hosts for which additional
name checks are made on the server certificate.
The match against this list is, as per other Exim usage, the
IP for the host.  That is most closely associated with the
name on the DNS A (or AAAA) record for the host.
However, the name that needs to be in the certificate
is the one at the head of any CNAME chain leading to the A record.
The option defaults to always checking.

The &(smtp)& transport has two OCSP-related options:
&%hosts_require_ocsp%&; a host-list for which a Certificate Status
is requested and required for the connection to proceed.  The default
value is empty.
&%hosts_request_ocsp%&; a host-list for which (additionally)
a Certificate Status is requested (but not necessarily verified).  The default
value is "*" meaning that requests are made unless configured
otherwise.

The host(s) should also be in &%hosts_require_tls%&, and
&%tls_verify_certificates%& configured for the transport,
for OCSP to be relevant.

If
&%tls_require_ciphers%& is set on the &(smtp)& transport, it must contain a
list of permitted cipher suites. If either of these checks fails, delivery to
the current host is abandoned, and the &(smtp)& transport tries to deliver to
alternative hosts, if any.

 &*Note*&:
These options must be set in the &(smtp)& transport for Exim to use TLS when it
is operating as a client. Exim does not assume that a server certificate (set
by the global options of the same name) should also be used when operating as a
client.

.vindex "&$host$&"
.vindex "&$host_address$&"
All the TLS options in the &(smtp)& transport are expanded before use, with
&$host$& and &$host_address$& containing the name and address of the server to
which the client is connected. Forced failure of an expansion causes Exim to
behave as if the relevant option were unset.

.vindex &$tls_out_bits$&
.vindex &$tls_out_cipher$&
.vindex &$tls_out_peerdn$&
.vindex &$tls_out_sni$&
Before an SMTP connection is established, the
&$tls_out_bits$&, &$tls_out_cipher$&, &$tls_out_peerdn$& and &$tls_out_sni$&
variables are emptied. (Until the first connection, they contain the values
that were set when the message was received.) If STARTTLS is subsequently
successfully obeyed, these variables are set to the relevant values for the
outgoing connection.



.subsection "Caching of static client configuration items" SECTclientTLScache
.cindex certificate caching
.cindex privatekey caching
.cindex crl caching
.cindex ciphers caching
.cindex "CA bundle" caching
.cindex "certificate authorities" caching
.cindex tls_certificate caching
.cindex tls_privatekey caching
.cindex tls_crl caching
.cindex tls_require_ciphers caching
.cindex tls_verify_certificate caching
.cindex caching certificate
.cindex caching privatekey
.cindex caching crl
.cindex caching ciphers
.cindex caching "certificate authorities
If any of the transport configuration options &%tls_certificate%&, &%tls_privatekey%&
and &%tls_crl%& have values with no
expandable elements,
then the associated information is loaded per smtp transport
at daemon startup, at the start of a queue run, or on a
command-line specified message delivery.
It is made available
to child processes forked for handling making SMTP connections.

This caching is currently only supported under Linux.

If caching is not possible, the load
of the associated information is done at the startup of the TLS connection.

The cache is invalidated in the daemon
and reloaded after any changes to the directories
containing files specified by these options.

The information specified by the main option &%tls_verify_certificates%&
is similarly cached so long as it specifies files explicitly
or (under GnuTLS) is the string &"system,cache"&.
The latter case is not automatically invaludated;
it is the operator's responsibility to arrange for a daemon restart
any time the system certificate authority bundle is updated.
A HUP signal is sufficient for this.
The value &"system"& results in no caching under GnuTLS.

The macro _HAVE_TLS_CA_CACHE will be defined if the suffix for "system"
is acceptable in configurations for the Exim executable.

Caching of the system Certificate Authorities bundle can
save significant time and processing on every TLS connection
initiated by Exim.




.section "Use of TLS Server Name Indication" "SECTtlssni"
.cindex "TLS" "Server Name Indication"
.cindex "TLS" SNI
.cindex SNI
.vindex "&$tls_in_sni$&"
.oindex "&%tls_in_sni%&"
With TLS1.0 or above, there is an extension mechanism by which extra
information can be included at various points in the protocol.  One of these
extensions, documented in
&url(https://www.rfc-editor.org/rfc/rfc6066,RFC 6066)
(and before that &url(https://www.rfc-editor.org/rfc/rfc4366,RFC 4366)) is
&"Server Name Indication"&, commonly &"SNI"&.  This extension is sent by the
client in the initial handshake, so that the server can examine the servername
within and possibly choose to use different certificates and keys (and more)
for this session.

This is analogous to HTTP's &"Host:"& header, and is the main mechanism by
which HTTPS-enabled web-sites can be virtual-hosted, many sites to one IP
address.

With SMTP to MX, there are the same problems here as in choosing the identity
against which to validate a certificate: you can't rely on insecure DNS to
provide the identity which you then cryptographically verify.  So this will
be of limited use in that environment.

With SMTP to Submission, there is a well-defined hostname which clients are
connecting to and can validate certificates against.  Thus clients &*can*&
choose to include this information in the TLS negotiation.  If this becomes
wide-spread, then hosters can choose to present different certificates to
different clients.  Or even negotiate different cipher suites.

The &%tls_sni%& option on an SMTP transport is an expanded string; the result,
if not empty, will be sent on a TLS session as part of the handshake.  There's
nothing more to it.  Choosing a sensible value not derived insecurely is the
only point of caution.  The &$tls_out_sni$& variable will be set to this string
for the lifetime of the client connection (including during authentication).

If DANE validated the connection attempt then the value of the &%tls_sni%& option
is forced to the name of the destination host, after any MX- or CNAME-following.

Except during SMTP client sessions, if &$tls_in_sni$& is set then it is a string
received from a client.
It can be logged with the &%log_selector%& item &`+tls_sni`&.

If the string &`tls_in_sni`& appears in the main section's &%tls_certificate%&
option (prior to expansion) then the following options will be re-expanded
during TLS session handshake, to permit alternative values to be chosen:

.ilist
&%tls_certificate%&
.next
&%tls_crl%&
.next
&%tls_privatekey%&
.next
&%tls_verify_certificates%&
.next
&%tls_ocsp_file%&
.endlist

Great care should be taken to deal with matters of case, various injection
attacks in the string (&`../`& or SQL), and ensuring that a valid filename
can always be referenced; it is important to remember that &$tls_in_sni$& is
arbitrary unverified data provided prior to authentication.
Further, the initial certificate is loaded before SNI has arrived, so
an expansion for &%tls_certificate%& must have a default which is used
when &$tls_in_sni$& is empty.

The Exim developers are proceeding cautiously and so far no other TLS options
are re-expanded.

When Exim is built against OpenSSL, OpenSSL must have been built with support
for TLS Extensions.  This holds true for OpenSSL 1.0.0+ and 0.9.8+ with
enable-tlsext in EXTRACONFIGURE.  If you invoke &(openssl s_client -h)& and
see &`-servername`& in the output, then OpenSSL has support.

When Exim is built against GnuTLS, SNI support is available as of GnuTLS
0.5.10.  (Its presence predates the current API which Exim uses, so if Exim
built, then you have SNI support).

.subsection ALPN
.cindex TLS ALPN
.cindex ALPN "general information"
.cindex TLS "Application Layer Protocol Names"
There is a TLS feature related to SNI
called Application Layer Protocol Name (ALPN).
This is intended to declare, or select, what protocol layer will be using a TLS
connection.
The client for the connection proposes a set of protocol names, and
the server responds with a selected one.
It is not, as of 2021, commonly used for SMTP connections.
However, to guard against misdirected or malicious use of web clients
(which often do use ALPN) against MTA ports, Exim by default check that
there is no incompatible ALPN specified by a client for a TLS connection.
If there is, the connection is rejected.

As a client Exim does not supply ALPN by default.
The behaviour of both client and server can be configured using the options
&%tls_alpn%& and &%hosts_require_alpn%&.
There are no variables providing observability.
Some feature-specific logging may appear on denied connections, but this
depends on the behaviour of the peer
(not all peers can send a feature-specific TLS Alert).

This feature is available when Exim is built with
OpenSSL 1.1.0 or later or GnuTLS 3.2.0 or later;
the macro _HAVE_TLS_ALPN will be defined when this is so.



.section "Multiple messages on the same encrypted TCP/IP connection" &&&
         "SECTmulmessam"
.cindex "multiple SMTP deliveries with TLS"
.cindex "TLS" "multiple message deliveries"
Exim sends multiple messages down the same TCP/IP connection by starting up
an entirely new delivery process for each message, passing the socket from
one process to the next. This implementation does not fit well with the use
of TLS, because there is quite a lot of state information associated with a TLS
connection, not just a socket identification. Passing all the state information
to a new process is not feasible. Consequently, for sending using TLS Exim
starts an additional proxy process for handling the encryption, piping the
unencrypted data stream from and to the delivery processes.

An older mode of operation can be enabled on a per-host basis by the
&%hosts_noproxy_tls%& option on the &(smtp)& transport.  If the host matches
this list the proxy process described above is not used; instead Exim
shuts down an existing TLS session being run by the delivery process
before passing the socket to a new process. The new process may then
try to start a new TLS session, and if successful, may try to re-authenticate
if AUTH is in use, before sending the next message.

The RFC is not clear as to whether or not an SMTP session continues in clear
after TLS has been shut down, or whether TLS may be restarted again later, as
just described. However, if the server is Exim, this shutdown and
reinitialization works. It is not known which (if any) other servers operate
successfully if the client closes a TLS session and continues with unencrypted
SMTP, but there are certainly some that do not work. For such servers, Exim
should not pass the socket to another process, because the failure of the
subsequent attempt to use it would cause Exim to record a temporary host error,
and delay other deliveries to that host.

To test for this case, Exim sends an EHLO command to the server after
closing down the TLS session. If this fails in any way, the connection is
closed instead of being passed to a new delivery process, but no retry
information is recorded.

There is also a manual override; you can set &%hosts_nopass_tls%& on the
&(smtp)& transport to match those hosts for which Exim should not pass
connections to new processes if TLS has been used.




.section "Certificates and all that" "SECTcerandall"
.cindex "certificate" "references to discussion"
In order to understand fully how TLS works, you need to know about
certificates, certificate signing, and certificate authorities.
This is a large topic and an introductory guide is unsuitable for the Exim
reference manual, so instead we provide pointers to existing documentation.

The Apache web-server was for a long time the canonical guide, so their
documentation is a good place to start; their SSL module's Introduction
document is currently at
.display
&url(https://httpd.apache.org/docs/current/ssl/ssl_intro.html)
.endd
and their FAQ is at
.display
&url(https://httpd.apache.org/docs/current/ssl/ssl_faq.html)
.endd

Eric Rescorla's book, &'SSL and TLS'&, published by Addison-Wesley (ISBN
0-201-61598-3) in 2001, contains both introductory and more in-depth
descriptions.
More recently Ivan Ristić's book &'Bulletproof SSL and TLS'&,
published by Feisty Duck (ISBN 978-1907117046) in 2013 is good.
Ivan is the author of the popular TLS testing tools at
&url(https://www.ssllabs.com/).


.subsection "Certificate chains" SECID186
A file named by &%tls_certificate%& may contain more than one
certificate. This is useful in the case where the certificate that is being
sent is validated by an intermediate certificate which the other end does
not have. Multiple certificates must be in the correct order in the file.
First the host's certificate itself, then the first intermediate
certificate to validate the issuer of the host certificate, then the next
intermediate certificate to validate the issuer of the first intermediate
certificate, and so on, until finally (optionally) the root certificate.
The root certificate must already be trusted by the recipient for
validation to succeed, of course, but if it's not preinstalled, sending the
root certificate along with the rest makes it available for the user to
install if the receiving end is a client MUA that can interact with a user.

Note that certificates using MD5 are unlikely to work on today's Internet;
even if your libraries allow loading them for use in Exim when acting as a
server, increasingly clients will not accept such certificates.  The error
diagnostics in such a case can be frustratingly vague.



.subsection "Self-signed certificates" SECID187
.cindex "certificate" "self-signed"
You can create a self-signed certificate using the &'req'& command provided
with OpenSSL, like this:
. ==== Do not shorten the duration here without reading and considering
. ==== the text below.  Please leave it at 9999 days.
.code
openssl req -x509 -newkey rsa:1024 -keyout file1 -out file2 \
            -days 9999 -nodes
.endd
&_file1_& and &_file2_& can be the same file; the key and the certificate are
delimited and so can be identified independently. The &%-days%& option
specifies a period for which the certificate is valid. The &%-nodes%& option is
important: if you do not set it, the key is encrypted with a passphrase
that you are prompted for, and any use that is made of the key causes more
prompting for the passphrase. This is not helpful if you are going to use
this certificate and key in an MTA, where prompting is not possible.

. ==== I expect to still be working 26 years from now.  The less technical
. ==== debt I create, in terms of storing up trouble for my later years, the
. ==== happier I will be then.  We really have reached the point where we
. ==== should start, at the very least, provoking thought and making folks
. ==== pause before proceeding, instead of leaving all the fixes until two
. ==== years before 2^31 seconds after the 1970 Unix epoch.
. ==== -pdp, 2012
NB: we are now past the point where 9999 days takes us past the 32-bit Unix
epoch.  If your system uses unsigned time_t (most do) and is 32-bit, then
the above command might produce a date in the past.  Think carefully about
the lifetime of the systems you're deploying, and either reduce the duration
of the certificate or reconsider your platform deployment.  (At time of
writing, reducing the duration is the most likely choice, but the inexorable
progression of time takes us steadily towards an era where this will not
be a sensible resolution).

A self-signed certificate made in this way is sufficient for testing, and
may be adequate for all your requirements if you are mainly interested in
encrypting transfers, and not in secure identification.

However, many clients require that the certificate presented by the server be a
user (also called &"leaf"& or &"site"&) certificate, and not a self-signed
certificate. In this situation, the self-signed certificate described above
must be installed on the client host as a trusted root &'certification
authority'& (CA), and the certificate used by Exim must be a user certificate
signed with that self-signed certificate.

For information on creating self-signed CA certificates and using them to sign
user certificates, see the &'General implementation overview'& chapter of the
Open-source PKI book, available online at
&url(https://sourceforge.net/projects/ospkibook/).


.subsection "Revoked certificates"
.cindex "TLS" "revoked certificates"
.cindex "revocation list"
.cindex "certificate" "revocation list"
.cindex "OCSP" "stapling"
There are three ways for a certificate to be made unusable
before its expiry.

.ilist
Certificate issuing authorities issue Certificate Revocation Lists (CRLs) when
certificates are revoked. If you have such a list, you can pass it to an Exim
server using the global option called &%tls_crl%& and to an Exim client using
an identically named option for the &(smtp)& transport. In each case, the value
of the option is expanded and must then be the name of a file that contains a
CRL in PEM format.
The downside is that clients have to periodically re-download a potentially huge
file from every certificate authority they know of.

.next
The way with most moving parts at query time is Online Certificate
Status Protocol (OCSP), where the client verifies the certificate
against an OCSP server run by the CA.  This lets the CA track all
usage of the certs.  It requires running software with access to the
private key of the CA, to sign the responses to the OCSP queries.  OCSP
is based on HTTP and can be proxied accordingly.

The only widespread OCSP server implementation (known to this writer)
comes as part of OpenSSL and aborts on an invalid request, such as
connecting to the port and then disconnecting.  This requires
re-entering the passphrase each time some random client does this.

.next
The third way is OCSP Stapling; in this, the server using a certificate
issued by the CA periodically requests an OCSP proof of validity from
the OCSP server, then serves it up inline as part of the TLS
negotiation.   This approach adds no extra round trips, does not let the
CA track users, scales well with number of certs issued by the CA and is
resilient to temporary OCSP server failures, as long as the server
starts retrying to fetch an OCSP proof some time before its current
proof expires.  The downside is that it requires server support.

Unless Exim is built with the support disabled,
or with GnuTLS earlier than version 3.3.16 / 3.4.8
support for OCSP stapling is included.

There is a global option called &%tls_ocsp_file%&.
The file specified therein is expected to be in DER format, and contain
an OCSP proof.  Exim will serve it as part of the TLS handshake.  This
option will be re-expanded for SNI, if the &%tls_certificate%& option
contains &`tls_in_sni`&, as per other TLS options.

Exim does not at this time implement any support for fetching a new OCSP
proof.  The burden is on the administrator to handle this, outside of
Exim.  The file specified should be replaced atomically, so that the
contents are always valid.  Exim will expand the &%tls_ocsp_file%& option
on each connection, so a new file will be handled transparently on the
next connection.

When built with OpenSSL Exim will check for a valid next update timestamp
in the OCSP proof; if not present, or if the proof has expired, it will be
ignored.

For the client to be able to verify the stapled OCSP the server must
also supply, in its stapled information, any intermediate
certificates for the chain leading to the OCSP proof from the signer
of the server certificate.  There may be zero or one such. These
intermediate certificates should be added to the server OCSP stapling
file named by &%tls_ocsp_file%&.

Note that the proof only covers the terminal server certificate,
not any of the chain from CA to it.

There is no current way to staple a proof for a client certificate.

.code
  A helper script "ocsp_fetch.pl" for fetching a proof from a CA
  OCSP server is supplied.  The server URL may be included in the
  server certificate, if the CA is helpful.

  One failure mode seen was the OCSP Signer cert expiring before the end
  of validity of the OCSP proof. The checking done by Exim/OpenSSL
  noted this as invalid overall, but the re-fetch script did not.
.endd
.endlist


.ecindex IIDencsmtp1
.ecindex IIDencsmtp2


.section "TLS Resumption" "SECTresumption"
.cindex TLS resumption
TLS Session Resumption for TLS 1.2 and TLS 1.3 connections can be used (defined
in &url(https://www.rfc-editor.org/rfc/rfc5077,RFC 5077) for 1.2).
The support for this requires GnuTLS 3.6.3 or OpenSSL 1.1.1 (or later).

Session resumption (this is the "stateless" variant) involves the server sending
a "session ticket" to the client on one connection, which can be stored by the
client and used for a later session.  The ticket contains sufficient state for
the server to reconstruct the TLS session, avoiding some expensive crypto
calculation and (on TLS1.2) one full packet roundtrip time.

.ilist
Operational cost/benefit:

 The extra data being transmitted costs a minor amount, and the client has
 extra costs in storing and retrieving the data.

 In the Exim/Gnutls implementation the extra cost on an initial connection
 which is TLS1.2 over a loopback path is about 6ms on 2017-laptop class hardware.
 The saved cost on a subsequent connection is about 4ms; three or more
 connections become a net win.  On longer network paths, two or more
 connections will have an average lower startup time thanks to the one
 saved packet roundtrip.  TLS1.3 will save the crypto cpu costs but not any
 packet roundtrips.

.cindex "hints database" tls
 Since a new hints DB is used on the TLS client,
 the hints DB maintenance should be updated to additionally handle "tls".

.next
Security aspects:

 The session ticket is encrypted, but is obviously an additional security
 vulnarability surface.  An attacker able to decrypt it would have access
 all connections using the resumed session.
 The session ticket encryption key is not committed to storage by the server
 and is rotated regularly (OpenSSL: 1hr, and one previous key is used for
 overlap; GnuTLS 6hr but does not specify any overlap).
 Tickets have limited lifetime (2hr, and new ones issued after 1hr under
 OpenSSL.  GnuTLS 2hr, appears to not do overlap).

 There is a question-mark over the security of the Diffie-Helman parameters
 used for session negotiation.

.next
Observability:

 The &%log_selector%& "tls_resumption" appends an asterisk to the tls_cipher "X="
 element.

 The variables &$tls_in_resumption$& and &$tls_out_resumption$&
 have bits 0-4 indicating respectively
 support built, client requested ticket, client offered session,
 server issued ticket, resume used.  A suitable decode list is provided
 in the builtin macro _RESUME_DECODE for in &%listextract%& expansions.

.next
Control:

The &%tls_resumption_hosts%& main option specifies a hostlist for which
exim, operating as a server, will offer resumption to clients.
Current best practice is to not offer the feature to MUA connection.
Commonly this can be done like this:
.code
tls_resumption_hosts = ${if inlist {$received_port}{587:465} {:}{*}}
.endd
If the peer host matches the list after expansion then resumption
is offered and/or accepted.

The &%tls_resumption_hosts%& smtp transport option performs the
equivalent function for operation as a client.
If the peer host matches the list after expansion then resumption
is attempted (if a stored session is available) or the information
stored (if supplied by the peer).


.next
Issues:

 In a resumed session:
.ilist
  The variables &$tls_{in,out}_cipher$& will have values different
 to the original (under GnuTLS).
.next
  The variables &$tls_{in,out}_ocsp$& will be "not requested" or "no response",
 and the &%hosts_require_ocsp%& smtp trasnport option will fail.
. XXX need to do something with that hosts_require_ocsp
.endlist

.endlist


.section DANE "SECDANE"
.cindex DANE
DNS-based Authentication of Named Entities, as applied to SMTP over TLS, provides assurance to a client that
it is actually talking to the server it wants to rather than some attacker operating a Man In The Middle (MITM)
operation.  The latter can terminate the TLS connection you make, and make another one to the server (so both
you and the server still think you have an encrypted connection) and, if one of the "well known" set of
Certificate Authorities has been suborned - something which *has* been seen already (2014), a verifiable
certificate (if you're using normal root CAs, eg. the Mozilla set, as your trust anchors).

What DANE does is replace the CAs with the DNS as the trust anchor.  The assurance is limited to a) the possibility
that the DNS has been suborned, b) mistakes made by the admins of the target server.   The attack surface presented
by (a) is thought to be smaller than that of the set of root CAs.

It also allows the server to declare (implicitly) that connections to it should use TLS.  An MITM could simply
fail to pass on a server's STARTTLS.

DANE scales better than having to maintain (and communicate via side-channel) copies of server certificates
for every possible target server.  It also scales (slightly) better than having to maintain on an SMTP
client a copy of the standard CAs bundle.  It also means not having to pay a CA for certificates.

DANE requires a server operator to do three things:
.olist
Run DNSSEC.  This provides assurance to clients
that DNS lookups they do for the server have not been tampered with.  The domain MX record applying
to this server, its A record, its TLSA record and any associated CNAME records must all be covered by
DNSSEC.
.next
Add TLSA DNS records.  These say what the server certificate for a TLS connection should be.
.next
Offer a server certificate, or certificate chain, in TLS connections which is anchored by one of the TLSA records.
.endlist

There are no changes to Exim specific to server-side operation of DANE.
Support for client-side operation of DANE can be included at compile time by defining SUPPORT_DANE=yes
in &_Local/Makefile_&.
If it has been included, the macro "_HAVE_DANE" will be defined.

.subsection "DNS records"
A TLSA record consist of 4 fields, the "Certificate Usage", the
"Selector", the "Matching type", and the "Certificate Association Data".
For a detailed description of the TLSA record see
&url(https://tools.ietf.org/html/rfc7671#page-5,RFC 7671).

The TLSA record for the server may have "Certificate Usage" (1st) field of DANE-TA(2) or DANE-EE(3).
These are the "Trust Anchor" and "End Entity" variants.
The latter specifies the End Entity directly, i.e. the certificate involved is that of the server
(and if only DANE-EE is used then it should be the sole one transmitted during the TLS handshake);
this is appropriate for a single system, using a self-signed certificate.
DANE-TA usage is effectively declaring a specific CA to be used; this might be a private CA or a public,
well-known one.
A private CA at simplest is just a self-signed certificate (with certain
attributes) which is used to sign server certificates, but running one securely
does require careful arrangement.
With DANE-TA, as implemented in Exim and commonly in other MTAs,
the server TLS handshake must transmit the entire certificate chain from CA to server-certificate.
DANE-TA is commonly used for several services and/or servers, each having a TLSA query-domain CNAME record,
all of which point to a single TLSA record.
DANE-TA and DANE-EE can both be used together.

Our recommendation is to use DANE with a certificate from a public CA,
because this enables a variety of strategies for remote clients to verify
your certificate.
You can then publish information both via DANE and another technology,
"MTA-STS", described below.

When you use DANE-TA to publish trust anchor information, you ask entities
outside your administrative control to trust the Certificate Authority for
connections to you.
If using a private CA then you should expect others to still apply the
technical criteria they'd use for a public CA to your certificates.
In particular, you should probably try to follow current best practices for CA
operation around hash algorithms and key sizes.
Do not expect other organizations to lower their security expectations just
because a particular profile might be reasonable for your own internal use.

When this text was last updated, this in practice means to avoid use of SHA-1
and MD5; if using RSA to use key sizes of at least 2048 bits (and no larger
than 4096, for interoperability); to use keyUsage fields correctly; to use
random serial numbers.
The list of requirements is subject to change as best practices evolve.
If you're not already using a private CA, or it doesn't meet these
requirements, then we encourage you to avoid all these issues and use a public
CA such as &url(https://letsencrypt.org/,Let's Encrypt) instead.

The TLSA record should have a "Selector" (2nd) field of SPKI(1) and
a "Matching Type" (3rd) field of SHA2-512(2).

For the "Certificate Authority Data" (4th) field, commands like

.code
  openssl x509 -pubkey -noout <certificate.pem \
  | openssl rsa -outform der -pubin 2>/dev/null \
  | openssl sha512 \
  | awk '{print $2}'
.endd

are workable to create a hash of the certificate's public key.

An example TLSA record for DANE-EE(3), SPKI(1), and SHA-512 (2) looks like

.code
  _25._tcp.mail.example.com. TLSA 3 1 2 8BA8A336E...
.endd

At the time of writing, &url(https://www.huque.com/bin/gen_tlsa)
is useful for quickly generating TLSA records.


For use with the DANE-TA model, server certificates must have a correct name (SubjectName or SubjectAltName).

The Certificate issued by the CA published in the DANE-TA model should be
issued using a strong hash algorithm.
Exim, and importantly various other MTAs sending to you, will not
re-enable hash algorithms which have been disabled by default in TLS
libraries.
This means no MD5 and no SHA-1.  SHA2-256 is the minimum for reliable
interoperability (and probably the maximum too, in 2018).

.subsection "Interaction with OCSP"
The use of OCSP-stapling should be considered, allowing for fast revocation of certificates (which would otherwise
be limited by the DNS TTL on the TLSA records).  However, this is likely to only be usable with DANE-TA.  NOTE: the
default of requesting OCSP for all hosts is modified iff DANE is in use, to:

.code
  hosts_request_ocsp = ${if or { {= {0}{$tls_out_tlsa_usage}} \
                                 {= {4}{$tls_out_tlsa_usage}} } \
                         {*}{}}
.endd

The (new) variable &$tls_out_tlsa_usage$& is a bitfield with numbered bits set for TLSA record usage codes.
The zero above means DANE was not in use, the four means that only DANE-TA usage TLSA records were
found. If the definition of &%hosts_request_ocsp%& includes the
string "tls_out_tlsa_usage", they are re-expanded in time to
control the OCSP request.

This modification of hosts_request_ocsp is only done if it has the default value of "*".  Admins who change it, and
those who use &%hosts_require_ocsp%&, should consider the interaction with DANE in their OCSP settings.


.subsection "Client configuration"
For client-side DANE there are three new smtp transport options, &%hosts_try_dane%&, &%hosts_require_dane%&
and &%dane_require_tls_ciphers%&.
The &"require"& variant will result in failure if the target host is not
DNSSEC-secured. To get DNSSEC-secured hostname resolution, use
the &%dnssec_request_domains%& router or transport option.

DANE will only be usable if the target host has DNSSEC-secured MX, A and TLSA records.

A TLSA lookup will be done if either of the above options match and the host-lookup succeeded using DNSSEC.
If a TLSA lookup is done and succeeds, a DANE-verified TLS connection
will be required for the host.  If it does not, the host will not
be used; there is no fallback to non-DANE or non-TLS.

If DANE is requested and usable, then the TLS cipher list configuration
prefers to use the option &%dane_require_tls_ciphers%& and falls
back to &%tls_require_ciphers%& only if that is unset.
This lets you configure "decent crypto" for DANE and "better than nothing
crypto" as the default.  Note though that while GnuTLS lets the string control
which versions of TLS/SSL will be negotiated, OpenSSL does not and you're
limited to ciphersuite constraints.

If DANE is requested and useable (see above) the following transport options are ignored:
.code
  hosts_require_tls
  tls_verify_hosts
  tls_try_verify_hosts
  tls_verify_certificates
  tls_crl
  tls_verify_cert_hostnames
  tls_sni
.endd

If DANE is not usable, whether requested or not, and CA-anchored
verification evaluation is wanted, the above variables should be set appropriately.

The router and transport option &%dnssec_request_domains%& must not be
set to &"never"&, and &%dnssec_require_domains%& is ignored.

.subsection Observability
If verification was successful using DANE then the "CV" item in the delivery log line will show as "CV=dane".

There is a new variable &$tls_out_dane$& which will have "yes" if
verification succeeded using DANE and "no" otherwise (only useful
in combination with events; see &<<CHAPevents>>&),
and a new variable &$tls_out_tlsa_usage$& (detailed above).

.cindex DANE reporting
An event (see &<<CHAPevents>>&) of type "dane:fail" will be raised on failures
to achieve DANE-verified connection, if one was either requested and offered, or
required.  This is intended to support TLS-reporting as defined in
&url(https://tools.ietf.org/html/draft-ietf-uta-smtp-tlsrpt-17).
The &$event_data$& will be one of the Result Types defined in
Section 4.3 of that document.

.subsection General
Under GnuTLS, DANE is only supported from version 3.0.0 onwards.

DANE is specified in &url(https://www.rfc-editor.org/rfc/rfc6698,RFC 6698).
It decouples certificate authority trust
selection from a "race to the bottom" of "you must trust everything for mail
to get through".
It does retain the need to trust the assurances provided by the DNSSEC tree.

There is an alternative technology called MTA-STS
(&url(https://www.rfc-editor.org/rfc/rfc8461,RFC 8461)), which
instead publishes MX trust anchor information on an HTTPS website.
The discovery of the address for that website does not (per standard)
require DNSSEC, and could be regarded as being less secure than DANE
as a result.

Exim has no support for MTA-STS as a client, but Exim mail server operators
can choose to publish information describing their TLS configuration using
MTA-STS to let those clients who do use that protocol derive trust
information.

The MTA-STS design requires a certificate from a public Certificate Authority
which is recognized by clients sending to you.
That selection of which CAs are trusted by others is outside your control.

The most interoperable course of action is probably to use
&url(https://letsencrypt.org/,Let's Encrypt), with automated certificate
renewal; to publish the anchor information in DNSSEC-secured DNS via TLSA
records for DANE clients (such as Exim and Postfix) and to publish anchor
information for MTA-STS as well.  This is what is done for the &'exim.org'&
domain itself (with caveats around occasionally broken MTA-STS because of
incompatible specification changes prior to reaching RFC status).



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Access control lists" "CHAPACL"
.scindex IIDacl "&ACL;" "description"
.cindex "control of incoming mail"
.cindex "message" "controlling incoming"
.cindex "policy control" "access control lists"
Access Control Lists (ACLs) are defined in a separate section of the runtime
configuration file, headed by &"begin acl"&. Each ACL definition starts with a
name, terminated by a colon. Here is a complete ACL section that contains just
one very small ACL:
.code
begin acl
small_acl:
  accept   hosts = one.host.only
.endd
You can have as many lists as you like in the ACL section, and the order in
which they appear does not matter. The lists are self-terminating.

The majority of ACLs are used to control Exim's behaviour when it receives
certain SMTP commands. This applies both to incoming TCP/IP connections, and
when a local process submits a message using SMTP by specifying the &%-bs%&
option. The most common use is for controlling which recipients are accepted
in incoming messages. In addition, you can define an ACL that is used to check
local non-SMTP messages. The default configuration file contains an example of
a realistic ACL for checking RCPT commands. This is discussed in chapter
&<<CHAPdefconfil>>&.


.section "Testing ACLs" "SECID188"
The &%-bh%& command line option provides a way of testing your ACL
configuration locally by running a fake SMTP session with which you interact.


.section "Specifying when ACLs are used" "SECID189"
.cindex "&ACL;" "options for specifying"
In order to cause an ACL to be used, you have to name it in one of the relevant
options in the main part of the configuration. These options are:
.cindex "ATRN" "ACL for"
.cindex "AUTH" "ACL for"
.cindex "DATA" "ACLs for"
.cindex "DKIM" "ACL for"
.cindex "ETRN" "ACL for"
.cindex "EXPN" "ACL for"
.cindex "HELO" "ACL for"
.cindex "EHLO" "ACL for"
.cindex "MAIL" "ACL for"
.cindex "QUIT" "ACL for"
.cindex "PRDR" "ACL for"
.cindex "RCPT" "ACL for"
.cindex "STARTTLS, ACL for"
.cindex "VRFY" "ACL for"
.cindex "WELLKNOWN" "ACL for"
.cindex "SMTP" "connection, ACL for"
.cindex "non-SMTP messages" "ACLs for"
.cindex "MIME content scanning" "ACL for"

.table2 140pt
.irow &%acl_not_smtp%&         "ACL for non-SMTP messages"
.irow &%acl_not_smtp_mime%&    "ACL for non-SMTP MIME parts"
.irow &%acl_not_smtp_start%&   "ACL at start of non-SMTP message"
.irow &%acl_smtp_atrn%&        "ACL for ATRN"
.irow &%acl_smtp_auth%&        "ACL for AUTH"
.irow &%acl_smtp_connect%&     "ACL for start of SMTP connection"
.irow &%acl_smtp_data%&        "ACL after DATA is complete"
.irow &%acl_smtp_data_prdr%&   "ACL for each recipient, after DATA is complete"
.irow &%acl_smtp_dkim%&        "ACL for each DKIM signer"
.irow &%acl_smtp_etrn%&        "ACL for ETRN"
.irow &%acl_smtp_expn%&        "ACL for EXPN"
.irow &%acl_smtp_helo%&        "ACL for HELO or EHLO"
.irow &%acl_smtp_mail%&        "ACL for MAIL"
.irow &%acl_smtp_mailauth%&    "ACL for the AUTH parameter of MAIL"
.irow &%acl_smtp_mime%&        "ACL for content-scanning MIME parts"
.irow &%acl_smtp_notquit%&     "ACL for non-QUIT terminations"
.irow &%acl_smtp_predata%&     "ACL at start of DATA command"
.irow &%acl_smtp_quit%&        "ACL for QUIT"
.irow &%acl_smtp_rcpt%&        "ACL for RCPT"
.irow &%acl_smtp_starttls%&    "ACL for STARTTLS"
.irow &%acl_smtp_vrfy%&        "ACL for VRFY"
.irow &%acl_smtp_wellknown%&   "ACL for WELLKNOWN"
.endtable

For example, if you set
.code
acl_smtp_rcpt = small_acl
.endd
the little ACL defined above is used whenever Exim receives a RCPT command
in an SMTP dialogue. The majority of policy tests on incoming messages can be
done when RCPT commands arrive. A rejection of RCPT should cause the
sending MTA to give up on the recipient address contained in the RCPT
command, whereas rejection at other times may cause the client MTA to keep on
trying to deliver the message. It is therefore recommended that you do as much
testing as possible at RCPT time.


.subsection "The non-SMTP ACLs" SECnonSMTP
.cindex "non-SMTP messages" "ACLs for"
The non-SMTP ACLs apply to all non-interactive incoming messages, that is, they
apply to batched SMTP as well as to non-SMTP messages. (Batched SMTP is not
really SMTP.) Many of the ACL conditions (for example, host tests, and tests on
the state of the SMTP connection such as encryption and authentication) are not
relevant and are forbidden in these ACLs. However, the sender and recipients
are known, so the &%senders%& and &%sender_domains%& conditions and the
&$sender_address$& and &$recipients$& variables can be used. Variables such as
&$authenticated_sender$& are also available. You can specify added header lines
in any of these ACLs.

The &%acl_not_smtp_start%& ACL is run right at the start of receiving a
non-SMTP message, before any of the message has been read. (This is the
analogue of the &%acl_smtp_predata%& ACL for SMTP input.) In the case of
batched SMTP input, it runs after the DATA command has been reached. The
result of this ACL is ignored; it cannot be used to reject a message. If you
really need to, you could set a value in an ACL variable here and reject based
on that in the &%acl_not_smtp%& ACL. However, this ACL can be used to set
controls, and in particular, it can be used to set
.code
control = suppress_local_fixups
.endd
This cannot be used in the other non-SMTP ACLs because by the time they are
run, it is too late.

The &%acl_not_smtp_mime%& ACL is available only when Exim is compiled with the
content-scanning extension. For details, see chapter &<<CHAPexiscan>>&.

The &%acl_not_smtp%& ACL is run just before the &[local_scan()]& function. Any
kind of rejection is treated as permanent, because there is no way of sending a
temporary error for these kinds of message.


.subsection "The SMTP connect ACL" SECconnectACL
.cindex "SMTP" "connection, ACL for"
.oindex &%smtp_banner%&
The ACL test specified by &%acl_smtp_connect%& happens at the start of an SMTP
session, after the test specified by &%host_reject_connection%& (which is now
an anomaly) and any TCP Wrappers testing (if configured). If the connection is
accepted by an &%accept%& verb that has a &%message%& modifier, the contents of
the message override the banner message that is otherwise specified by the
&%smtp_banner%& option.

For tls-on-connect connections, the ACL is run before the TLS connection
is accepted; if the ACL does not accept then the TCP connection is dropped without
any TLS startup attempt and without any SMTP response being transmitted.


.subsection "The EHLO/HELO ACL" SECheloACL
.cindex "EHLO" "ACL for"
.cindex "HELO" "ACL for"
The ACL test specified by &%acl_smtp_helo%& happens when the client issues an
EHLO or HELO command, after the tests specified by &%helo_accept_junk_hosts%&,
&%helo_allow_chars%&, &%helo_verify_hosts%&, and &%helo_try_verify_hosts%&.
Note that a client may issue more than one EHLO or HELO command in an SMTP
session, and indeed is required to issue a new EHLO or HELO after successfully
setting up encryption following a STARTTLS command.

Note also that a deny neither forces the client to go away nor means that
mail will be refused on the connection.  Consider checking for
&$sender_helo_name$& being defined in a MAIL or RCPT ACL to do that.

If the command is accepted by an &%accept%& verb that has a &%message%&
modifier, the message may not contain more than one line (it will be truncated
at the first newline and a panic logged if it does). Such a message cannot
affect the EHLO options that are listed on the second and subsequent lines of
an EHLO response.


.subsection "The DATA ACLs" SECdataACLS
.cindex "DATA" "ACLs for"
Two ACLs are associated with the DATA command, because it is two-stage
command, with two responses being sent to the client.
When the DATA command is received, the ACL defined by &%acl_smtp_predata%&
is obeyed. This gives you control after all the RCPT commands, but before
the message itself is received. It offers the opportunity to give a negative
response to the DATA command before the data is transmitted. Header lines
added by MAIL or RCPT ACLs are not visible at this time, but any that
are defined here are visible when the &%acl_smtp_data%& ACL is run.

You cannot test the contents of the message, for example, to verify addresses
in the headers, at RCPT time or when the DATA command is received. Such
tests have to appear in the ACL that is run after the message itself has been
received, before the final response to the DATA command is sent. This is
the ACL specified by &%acl_smtp_data%&, which is the second ACL that is
associated with the DATA command.

.cindex CHUNKING "BDAT command"
.cindex BDAT "SMTP command"
.cindex "RFC 3030" CHUNKING
If CHUNKING was advertised and a BDAT command sequence is received,
the &%acl_smtp_predata%& ACL is not run.
. XXX why not?  It should be possible, for the first BDAT.
The &%acl_smtp_data%& is run after the last BDAT command and all of
the data specified is received.

For both of these ACLs, it is not possible to reject individual recipients. An
error response rejects the entire message. Unfortunately, it is known that some
MTAs do not treat hard (5&'xx'&) responses to the DATA command (either
before or after the data) correctly &-- they keep the message on their queues
and try again later, but that is their problem, though it does waste some of
your resources.

The &%acl_smtp_data%& ACL is run after
the &%acl_smtp_data_prdr%&,
the &%acl_smtp_dkim%&
and the &%acl_smtp_mime%& ACLs.

.subsection "The SMTP DKIM ACL" SECTDKIMACL
The &%acl_smtp_dkim%& ACL is available only when Exim is compiled with DKIM support
enabled (which is the default).

If, for a specific message, an ACL control
&*dkim_disable_verify*&
has been set, this &%acl_smtp_dkim%& ACL is not called.

The ACL test specified by &%acl_smtp_dkim%& happens after a message has been
received, and is executed for each DKIM signature found in a message.  If not
otherwise specified, the default action is to accept.

This ACL is evaluated before &%acl_smtp_mime%& and &%acl_smtp_data%&.

For details on the operation of DKIM, see section &<<SECDKIM>>&.


.subsection "The SMTP MIME ACL" SECID194
The &%acl_smtp_mime%& option is available only when Exim is compiled with the
content-scanning extension. For details, see chapter &<<CHAPexiscan>>&.

This ACL is evaluated after &%acl_smtp_dkim%& but before &%acl_smtp_data%&.


.subsection "The SMTP PRDR ACL" SECTPRDRACL
.cindex "PRDR" "ACL for"
.oindex "&%prdr_enable%&"
The &%acl_smtp_data_prdr%& ACL is available only when Exim is compiled
with PRDR support enabled (which is the default).
It becomes active only when the PRDR feature is negotiated between
client and server for a message, and more than one recipient
has been accepted.

The ACL test specified by &%acl_smtp_data_prdr%& happens after a message
has been received, and is executed once for each recipient of the message
with &$local_part$& and &$domain$& valid.
The test may accept, defer or deny for individual recipients.
The &%acl_smtp_data%& will still be called after this ACL and
can reject the message overall, even if this ACL has accepted it
for some or all recipients.

PRDR may be used to support per-user content filtering.  Without it
one must defer any recipient after the first that has a different
content-filter configuration.  With PRDR, the RCPT-time check
.cindex "PRDR" "variable for"
for this can be disabled when the variable &$prdr_requested$&
is &"yes"&.
Any required difference in behaviour of the main DATA-time
ACL should however depend on the PRDR-time ACL having run, as Exim
will avoid doing so in some situations (e.g. single-recipient mails).

See also the &%prdr_enable%& global option
and the &%hosts_try_prdr%& smtp transport option.

This ACL is evaluated after &%acl_smtp_dkim%& but before &%acl_smtp_data%&.
If the ACL is not defined, processing completes as if
the feature was not requested by the client.

.subsection "The SMTP WELLKNOWN ACL" SECTWELLKNOWNACL
.cindex "WELLKNOWN" "ACL for"
.oindex "&%acl_smtp_wellknown%&"
The &%acl_smtp_wellknown%& ACL is available only when Exim is compiled
with WELLKNOWN support enabled.

The ACL determines the response to an SMTP WELLKNOWN command, using the normal
accept/defer/deny verbs for the response code,
and a new &"control=wellknown"& modifier.
This modifier takes a single option, separated by a '/'
character, which must be the name of a file containing the response
cleartext.  The modifier is expanded before use in the usual way before
it is used.  The configuration is responsible for picking a suitable file
to return and, most importantly, not returning any unexpected file.
The argument for the SMTP verb will be available in the &$smtp_command_argument$&
variable and can be used for building the file path.
If the file path given in the modifier is empty or inacessible, the control will
fail.

For example:
.code
 check_wellknown:
  accept control = wellknown/\
                    ${lookup {${xtextd:$smtp_command_argument}} \
                        dsearch,key=path,filter=file,ret=full \
                        {$spooldir/wellknown.d}}
.endd
File content will be encoded in &"xtext"& form, and line-wrapping
for line-length limitation will be done before transmission.
A response summary line will be prepended, with the (pre-encoding) file size.

The above example uses the expansion operator ${xtextd:<coded-string>}
which is needed to decode the xtext-encoded key from the SMTP verb.

Under the util directory there is a "mailtest" utility which can be used
to test/retrieve WELLKNOWN items. Syntax is
.code
  mailtest -h host.example.com -w security.txt
.endd

WELLKNOWN is a ESMTP extension providing access to extended
information about the server.  It is modelled on the webserver
facilities documented in
&url(https://www.rfc-editor.org/rfc/rfc8615,RFC 8615)
and can be used for a security.txt file
and could be used for ACME handshaking
(&url(https://www.rfc-editor.org/rfc/rfc8555,RFC 8555)).

Exim will advertise WELLKNOWN support in the EHLO response
.oindex &%wellknown_advertise_hosts%&
(conditional on a new option &%wellknown_advertise_hosts%&)
and service WELLKNOWN smtp verbs having a single parameter
giving a key for an item of "site-wide metadata".
The verb and key are separated by whitespace,
and the key is xtext-encoded
(per &url(https://www.rfc-editor.org/rfc/rfc3461,RFC 3461) section 4).


.subsection "The QUIT ACL" SECTQUITACL
.cindex "QUIT, ACL for"
The ACL for the SMTP QUIT command is anomalous, in that the outcome of the ACL
does not affect the response code to QUIT, which is always 221. Thus, the ACL
does not in fact control any access.
For this reason, it may only accept
or warn as its final result.

This ACL can be used for tasks such as custom logging at the end of an SMTP
session. For example, you can use ACL variables in other ACLs to count
messages, recipients, etc., and log the totals at QUIT time using one or
more &%logwrite%& modifiers on a &%warn%& verb.

&*Warning*&: Only the &$acl_c$&&'x'& variables can be used for this, because
the &$acl_m$&&'x'& variables are reset at the end of each incoming message.

You do not need to have a final &%accept%&, but if you do, you can use a
&%message%& modifier to specify custom text that is sent as part of the 221
response to QUIT.

This ACL is run only for a &"normal"& QUIT. For certain kinds of disastrous
failure (for example, failure to open a log file, or when Exim is bombing out
because it has detected an unrecoverable error), all SMTP commands from the
client are given temporary error responses until QUIT is received or the
connection is closed. In these special cases, the QUIT ACL does not run.


.subsection "The not-QUIT ACL" SECTNOTQUITACL
.vindex &$acl_smtp_notquit$&
The not-QUIT ACL, specified by &%acl_smtp_notquit%&, is run in most cases when
an SMTP session ends without sending QUIT. However, when Exim itself is in bad
trouble, such as being unable to write to its log files, this ACL is not run,
because it might try to do things (such as write to log files) that make the
situation even worse.

Like the QUIT ACL, this ACL is provided to make it possible to do customized
logging or to gather statistics, and its outcome is ignored. The &%delay%&
modifier is forbidden in this ACL, and the only permitted verbs are &%accept%&
and &%warn%&.

.vindex &$smtp_notquit_reason$&
When the not-QUIT ACL is running, the variable &$smtp_notquit_reason$& is set
to a string that indicates the reason for the termination of the SMTP
connection. The possible values are:
.table2
.irow &`acl-drop`&                 "Another ACL issued a &%drop%& command"
.irow &`bad-commands`&             "Too many unknown or non-mail commands"
.irow &`command-timeout`&          "Timeout while reading SMTP commands"
.irow &`connection-lost`&          "The SMTP connection has been lost"
.irow &`data-timeout`&             "Timeout while reading message data"
.irow &`local-scan-error`&         "The &[local_scan()]& function crashed"
.irow &`local-scan-timeout`&       "The &[local_scan()]& function timed out"
.irow &`signal-exit`&              "SIGTERM or SIGINT"
.irow &`synchronization-error`&    "SMTP synchronization error"
.irow &`tls-failed`&               "TLS failed to start"
.endtable
In most cases when an SMTP connection is closed without having received QUIT,
Exim sends an SMTP response message before actually closing the connection.
With the exception of the &`acl-drop`& case, the default message can be
overridden by the &%message%& modifier in the not-QUIT ACL. In the case of a
&%drop%& verb in another ACL, it is the message from the other ACL that is
used.


.section "Finding an ACL to use" "SECID195"
.cindex "&ACL;" "finding which to use"
The value of an &%acl_smtp_%&&'xxx'& option is expanded before use, so
you can use different ACLs in different circumstances. For example,
.code
acl_smtp_rcpt = ${if ={25}{$interface_port} \
                     {acl_check_rcpt} {acl_check_rcpt_submit} }
.endd
In the default configuration file there are some example settings for
providing an
&url(https://www.rfc-editor.org/rfc/rfc4409,RFC 4409)
message &"submission"& service on port 587 and an
&url(https://www.rfc-editor.org/rfc/rfc8314,RFC 8314)
&"submissions"& service on port 465. You can use a string
expansion like this to choose an ACL for MUAs on these ports which is
more appropriate for this purpose than the default ACL on port 25.

The expanded string does not have to be the name of an ACL in the
configuration file; there are other possibilities. Having expanded the
string, Exim searches for an ACL as follows:

.ilist
If the string begins with a slash, Exim uses it as a filename, and reads its
contents as an ACL. The lines are processed in the same way as lines in the
Exim configuration file. In particular, continuation lines are supported, blank
lines are ignored, as are lines whose first non-whitespace character is &"#"&.
If the file does not exist or cannot be read, an error occurs (typically
causing a temporary failure of whatever caused the ACL to be run). For example:
.code
acl_smtp_data = /etc/acls/\
  ${lookup{$sender_host_address}lsearch\
  {/etc/acllist}{$value}{default}}
.endd
This looks up an ACL file to use on the basis of the host's IP address, falling
back to a default if the lookup fails. If an ACL is successfully read from a
file, it is retained in memory for the duration of the Exim process, so that it
can be re-used without having to re-read the file.
.next
If the string does not start with a slash, and does not contain any spaces,
Exim searches the ACL section of the configuration for an ACL whose name
matches the string.
.next
If no named ACL is found, or if the string contains spaces, Exim parses
the string as an inline ACL. This can save typing in cases where you just
want to have something like
.code
acl_smtp_vrfy = accept
.endd
in order to allow free use of the VRFY command. Such a string may contain
newlines; it is processed in the same way as an ACL that is read from a file.
.endlist




.section "ACL return codes" "SECID196"
.cindex "&ACL;" "return codes"
Except for the QUIT ACL, which does not affect the SMTP return code (see
section &<<SECTQUITACL>>& above), the result of running an ACL is either
&"accept"& or &"deny"&, or, if some test cannot be completed (for example, if a
database is down), &"defer"&. These results cause 2&'xx'&, 5&'xx'&, and 4&'xx'&
return codes, respectively, to be used in the SMTP dialogue. A fourth return,
&"error"&, occurs when there is an error such as invalid syntax in the ACL.
This also causes a 4&'xx'& return code.

For the non-SMTP ACL, &"defer"& and &"error"& are treated in the same way as
&"deny"&, because there is no mechanism for passing temporary errors to the
submitters of non-SMTP messages.


ACLs that are relevant to message reception may also return &"discard"&. This
has the effect of &"accept"&, but causes either the entire message or an
individual recipient address to be discarded. In other words, it is a
blackholing facility. Use it with care.

If the ACL for MAIL returns &"discard"&, all recipients are discarded, and no
ACL is run for subsequent RCPT commands. The effect of &"discard"& in a
RCPT ACL is to discard just the one recipient address. If there are no
recipients left when the message's data is received, the DATA ACL is not
run. A &"discard"& return from the DATA or the non-SMTP ACL discards all the
remaining recipients. The &"discard"& return is not permitted for the
&%acl_smtp_predata%& ACL.

If the ACL for VRFY returns &"accept"&, a recipient verify (without callout)
is done on the address and the result determines the SMTP response.


.cindex "&[local_scan()]& function" "when all recipients discarded"
The &[local_scan()]& function is always run, even if there are no remaining
recipients; it may create new recipients.



.section "Unset ACL options" "SECID197"
.cindex "&ACL;" "unset options"
The default actions when any of the &%acl_%&&'xxx'& options are unset are not
all the same. &*Note*&: These defaults apply only when the relevant ACL is
not defined at all. For any defined ACL, the default action when control
reaches the end of the ACL statements is &"deny"&.

For &%acl_smtp_quit%& and &%acl_not_smtp_start%& there is no default because
these two are ACLs that are used only for their side effects. They cannot be
used to accept or reject anything.

For &%acl_not_smtp%&, &%acl_smtp_auth%&, &%acl_smtp_connect%&,
&%acl_smtp_data%&, &%acl_smtp_helo%&, &%acl_smtp_mail%&, &%acl_smtp_mailauth%&,
&%acl_smtp_mime%&, &%acl_smtp_predata%&, and &%acl_smtp_starttls%&, the action
when the ACL is not defined is &"accept"&.

.new
For the others (&%acl_smtp_atrn%&,
.wen
&%acl_smtp_etrn%&, &%acl_smtp_expn%&, &%acl_smtp_rcpt%&,
&%acl_smtp_vrfy%& and &%acl_smtp_wellknown%&),
the action when the ACL
is not defined is &"deny"&.  This means that &%acl_smtp_rcpt%& must be
defined in order to receive any messages over an SMTP connection.
For an example, see the ACL in the default configuration file.



.section "Data for message ACLs" "SECID198"
.cindex "&ACL;" "data for message ACL"
.vindex &$domain$&
.vindex &$local_part$&
.vindex &$sender_address$&
.vindex &$sender_host_address$&
.vindex &$smtp_command$&
When a MAIL or RCPT ACL, or either of the DATA ACLs, is running, the variables
that contain information about the host and the message's sender (for example,
&$sender_host_address$& and &$sender_address$&) are set, and can be used in ACL
statements. In the case of RCPT (but not MAIL or DATA), &$domain$& and
&$local_part$& are set from the argument address. The entire SMTP command
is available in &$smtp_command$&.

When an ACL for the AUTH parameter of MAIL is running, the variables that
contain information about the host are set, but &$sender_address$& is not yet
set. Section &<<SECTauthparamail>>& contains a discussion of this parameter and
how it is used.

.vindex "&$message_size$&"
The &$message_size$& variable is set to the value of the SIZE parameter on
the MAIL command at MAIL, RCPT and pre-data time, or to -1 if
that parameter is not given. The value is updated to the true message size by
the time the final DATA ACL is run (after the message data has been
received).

.vindex "&$rcpt_count$&"
.vindex "&$recipients_count$&"
The &$rcpt_count$& variable increases by one for each RCPT command received.
The &$recipients_count$& variable increases by one each time a RCPT command is
accepted, so while an ACL for RCPT is being processed, it contains the number
of previously accepted recipients. At DATA time (for both the DATA ACLs),
&$rcpt_count$& contains the total number of RCPT commands, and
&$recipients_count$& contains the total number of accepted recipients.





.section "Data for non-message ACLs" "SECTdatfornon"
.cindex "&ACL;" "data for non-message ACL"
.vindex &$smtp_command_argument$&
.vindex &$smtp_command$&
When an ACL is being run for ATRN, AUTH, EHLO, ETRN, EXPN, HELO, STARTTLS, or VRFY,
the remainder of the SMTP command line is placed in &$smtp_command_argument$&,
and the entire SMTP command is available in &$smtp_command$&.
These variables can be tested using a &%condition%& condition. For example,
here is an ACL for use with AUTH, which insists that either the session is
encrypted, or the CRAM-MD5 authentication method is used. In other words, it
does not permit authentication methods that use cleartext passwords on
unencrypted connections.
.code
acl_check_auth:
  accept encrypted = *
  accept condition = ${if eq{${uc:$smtp_command_argument}}\
                     {CRAM-MD5}}
  deny   message   = TLS encryption or CRAM-MD5 required
.endd
(Another way of applying this restriction is to arrange for the authenticators
that use cleartext passwords not to be advertised when the connection is not
encrypted. You can use the generic &%server_advertise_condition%& authenticator
option to do this.)



.section "Format of an ACL" "SECID199"
.cindex "&ACL;" "format of"
.cindex "&ACL;" "verbs, definition of"
An individual ACL definition consists of a number of statements.
Each statement starts
with a verb, optionally followed by a number of conditions and &"modifiers"&.
Modifiers can change the way the verb operates, define error and log messages,
set variables, insert delays, and vary the processing of accepted messages.

If all the conditions are met, the verb is obeyed. The same condition may be
used (with different arguments) more than once in the same statement. This
provides a means of specifying an &"and"& conjunction between conditions. For
example:
.code
deny  dnslists = list1.example
      dnslists = list2.example
.endd
If there are no conditions, the verb is always obeyed. Exim stops evaluating
the conditions and modifiers when it reaches a condition that fails. What
happens then depends on the verb (and in one case, on a special modifier). Not
all the conditions make sense at every testing point. For example, you cannot
test a sender address in the ACL that is run for a VRFY command.

The definition of an ACL ends where another starts,
or a different configuration section starts.


.section "ACL verbs" "SECID200"
The ACL verbs are as follows:

.ilist
.cindex "&%accept%& ACL verb"
&%accept%&: If all the conditions are met, the ACL returns &"accept"&. If any
of the conditions are not met, what happens depends on whether &%endpass%&
appears among the conditions (for syntax see below). If the failing condition
is before &%endpass%&, control is passed to the next ACL statement; if it is
after &%endpass%&, the ACL returns &"deny"&. Consider this statement, used to
check a RCPT command:
.code
accept domains = +local_domains
       endpass
       verify = recipient
.endd
If the recipient domain does not match the &%domains%& condition, control
passes to the next statement. If it does match, the recipient is verified, and
the command is accepted if verification succeeds. However, if verification
fails, the ACL yields &"deny"&, because the failing condition is after
&%endpass%&.

The &%endpass%& feature has turned out to be confusing to many people, so its
use is not recommended nowadays. It is always possible to rewrite an ACL so
that &%endpass%& is not needed, and it is no longer used in the default
configuration.

.cindex "&%message%& ACL modifier" "with &%accept%&"
If a &%message%& modifier appears on an &%accept%& statement, its action
depends on whether or not &%endpass%& is present. In the absence of &%endpass%&
(when an &%accept%& verb either accepts or passes control to the next
statement), &%message%& can be used to vary the message that is sent when an
SMTP command is accepted. For example, in a RCPT ACL you could have:
.display
&`accept  `&<&'some conditions'&>
&`        message = OK, I will allow you through today`&
.endd
You can specify an SMTP response code, optionally followed by an &"extended
response code"& at the start of the message, but the first digit must be the
same as would be sent by default, which is 2 for an &%accept%& verb.

If &%endpass%& is present in an &%accept%& statement, &%message%& specifies
an error message that is used when access is denied. This behaviour is retained
for backward compatibility, but current &"best practice"& is to avoid the use
of &%endpass%&.


.next
.cindex "&%defer%& ACL verb"
&%defer%&: If all the conditions are true, the ACL returns &"defer"& which, in
an SMTP session, causes a 4&'xx'& response to be given. For a non-SMTP ACL,
&%defer%& is the same as &%deny%&, because there is no way of sending a
temporary error. For a RCPT command, &%defer%& is much the same as using a
&(redirect)& router and &`:defer:`& while verifying, but the &%defer%& verb can
be used in any ACL, and even for a recipient it might be a simpler approach.


.next
.cindex "&%deny%& ACL verb"
&%deny%&: If all the conditions are met, the ACL returns &"deny"&. If any of
the conditions are not met, control is passed to the next ACL statement. For
example,
.code
deny dnslists = blackholes.mail-abuse.org
.endd
rejects commands from hosts that are on a DNS black list.


.next
.cindex "&%discard%& ACL verb"
&%discard%&: This verb behaves like &%accept%&, except that it returns
&"discard"& from the ACL instead of &"accept"&. It is permitted only on ACLs
that are concerned with receiving messages. When all the conditions are true,
the sending entity receives a &"success"& response. However, &%discard%& causes
recipients to be discarded. If it is used in an ACL for RCPT, just the one
recipient is discarded; if used for MAIL, DATA or in the non-SMTP ACL, all the
message's recipients are discarded. Recipients that are discarded before DATA
do not appear in the log line when the &%received_recipients%& log selector is set.

If the &%log_message%& modifier is set when &%discard%& operates,
its contents are added to the line that is automatically written to the log.
The &%message%& modifier operates exactly as it does for &%accept%&.


.next
.cindex "&%drop%& ACL verb"
&%drop%&: This verb behaves like &%deny%&, except that an SMTP connection is
forcibly closed after the 5&'xx'& error message has been sent. For example:
.code
drop   condition = ${if > {$rcpt_count}{20}}
       message   = I don't take more than 20 RCPTs
.endd
There is no difference between &%deny%& and &%drop%& for the connect-time ACL.
The connection is always dropped after sending a 550 response.

.next
.cindex "&%require%& ACL verb"
&%require%&: If all the conditions are met, control is passed to the next ACL
statement. If any of the conditions are not met, the ACL returns &"deny"&. For
example, when checking a RCPT command,
.code
require message = Sender did not verify
        verify  = sender
.endd
passes control to subsequent statements only if the message's sender can be
verified. Otherwise, it rejects the command. Note the positioning of the
&%message%& modifier, before the &%verify%& condition. The reason for this is
discussed in section &<<SECTcondmodproc>>&.

.next
.cindex "&%warn%& ACL verb"
&%warn%&: If all the conditions are true, a line specified by the
&%log_message%& modifier is written to Exim's main log. Control always passes
to the next ACL statement. If any condition is false, the log line is not
written. If an identical log line is requested several times in the same
message, only one copy is actually written to the log. If you want to force
duplicates to be written, use the &%logwrite%& modifier instead.

If &%log_message%& is not present, a &%warn%& verb just checks its conditions
and obeys any &"immediate"& modifiers (such as &%control%&, &%set%&,
&%logwrite%&, &%add_header%&, and &%remove_header%&) that appear before the
first failing condition. There is more about adding header lines in section
&<<SECTaddheadacl>>&.

If any condition on a &%warn%& statement cannot be completed (that is, there is
some sort of defer), the log line specified by &%log_message%& is not written.
This does not include the case of a forced failure from a lookup, which
is considered to be a successful completion. After a defer, no further
conditions or modifiers in the &%warn%& statement are processed. The incident
is logged, and the ACL continues to be processed, from the next statement
onwards.


.vindex "&$acl_verify_message$&"
When one of the &%warn%& conditions is an address verification that fails, the
text of the verification failure message is in &$acl_verify_message$&. If you
want this logged, you must set it up explicitly. For example:
.code
warn   !verify = sender
       log_message = sender verify failed: $acl_verify_message
.endd
.endlist

At the end of each ACL there is an implicit unconditional &%deny%&.

As you can see from the examples above, the conditions and modifiers are
written one to a line, with the first one on the same line as the verb, and
subsequent ones on following lines. If you have a very long condition, you can
continue it onto several physical lines by the usual backslash continuation
mechanism. It is conventional to align the conditions vertically.



.section "ACL variables" "SECTaclvariables"
.cindex "&ACL;" "variables"
There are some special variables that can be set during ACL processing. They
can be used to pass information between different ACLs, different invocations
of the same ACL in the same SMTP connection, and between ACLs and the routers,
transports, and filters that are used to deliver a message. The names of these
variables must begin with &$acl_c$& or &$acl_m$&, followed either by a digit or
an underscore, but the remainder of the name can be any sequence of
alphanumeric characters and underscores that you choose. There is no limit on
the number of ACL variables. The two sets act as follows:
.ilist
The values of those variables whose names begin with &$acl_c$& persist
throughout an SMTP connection. They are never reset. Thus, a value that is set
while receiving one message is still available when receiving the next message
on the same SMTP connection.
.next
The values of those variables whose names begin with &$acl_m$& persist only
while a message is being received. They are reset afterwards. They are also
reset by MAIL, RSET, EHLO, HELO, and after starting up a TLS session.
.endlist

When a message is accepted, the current values of all the ACL variables are
preserved with the message and are subsequently made available at delivery
time. The ACL variables are set by a modifier called &%set%&. For example:
.code
accept hosts = whatever
       set acl_m4 = some value
accept authenticated = *
       set acl_c_auth = yes
.endd
&*Note*&: A leading dollar sign is not used when naming a variable that is to
be set. If you want to set a variable without taking any action, you can use a
&%warn%& verb without any other modifiers or conditions.

.oindex &%strict_acl_vars%&
What happens if a syntactically valid but undefined ACL variable is
referenced depends on the setting of the &%strict_acl_vars%& option. If it is
false (the default), an empty string is substituted; if it is true, an
error is generated.

Versions of Exim before 4.64 have a limited set of numbered variables, but
their names are compatible, so there is no problem with upgrading.


.section "Condition and modifier processing" "SECTcondmodproc"
.cindex "&ACL;" "conditions; processing"
.cindex "&ACL;" "modifiers; processing"
An exclamation mark preceding a condition negates its result. For example:
.code
deny   domains = *.dom.example
      !verify  = recipient
.endd
causes the ACL to return &"deny"& if the recipient domain ends in
&'dom.example'& and the recipient address cannot be verified. Sometimes
negation can be used on the right-hand side of a condition. For example, these
two statements are equivalent:
.code
deny  hosts = !192.168.3.4
deny !hosts =  192.168.3.4
.endd
However, for many conditions (&%verify%& being a good example), only left-hand
side negation of the whole condition is possible.

The arguments of conditions and modifiers are expanded. A forced failure
of an expansion causes a condition to be ignored, that is, it behaves as if the
condition is true. Consider these two statements:
.code
accept  senders = ${lookup{$host_name}lsearch\
                  {/some/file}{$value}fail}
accept  senders = ${lookup{$host_name}lsearch\
                  {/some/file}{$value}{}}
.endd
Each attempts to look up a list of acceptable senders. If the lookup succeeds,
the returned list is searched, but if the lookup fails the behaviour is
different in the two cases. The &%fail%& in the first statement causes the
condition to be ignored, leaving no further conditions. The &%accept%& verb
therefore succeeds. The second statement, however, generates an empty list when
the lookup fails. No sender can match an empty list, so the condition fails,
and therefore the &%accept%& also fails.

ACL modifiers appear mixed in with conditions in ACL statements. Some of them
specify actions that are taken as the conditions for a statement are checked;
others specify text for messages that are used when access is denied or a
warning is generated. The &%control%& modifier affects the way an incoming
message is handled.

The positioning of the modifiers in an ACL statement is important, because the
processing of a verb ceases as soon as its outcome is known. Only those
modifiers that have already been encountered will take effect. For example,
consider this use of the &%message%& modifier:
.code
require message = Can't verify sender
        verify  = sender
        message = Can't verify recipient
        verify  = recipient
        message = This message cannot be used
.endd
If sender verification fails, Exim knows that the result of the statement is
&"deny"&, so it goes no further. The first &%message%& modifier has been seen,
so its text is used as the error message. If sender verification succeeds, but
recipient verification fails, the second message is used. If recipient
verification succeeds, the third message becomes &"current"&, but is never used
because there are no more conditions to cause failure.

For the &%deny%& verb, on the other hand, it is always the last &%message%&
modifier that is used, because all the conditions must be true for rejection to
happen. Specifying more than one &%message%& modifier does not make sense, and
the message can even be specified after all the conditions. For example:
.code
deny   hosts = ...
      !senders = *@my.domain.example
       message = Invalid sender from client host
.endd
The &"deny"& result does not happen until the end of the statement is reached,
by which time Exim has set up the message.



.section "ACL modifiers" "SECTACLmodi"
.cindex "&ACL;" "modifiers; list of"
The ACL modifiers are as follows:

.vlist
.vitem &*add_header*&&~=&~<&'text'&>
This modifier specifies one or more header lines that are to be added to an
incoming message, assuming, of course, that the message is ultimately
accepted. For details, see section &<<SECTaddheadacl>>&.

.vitem &*continue*&&~=&~<&'text'&>
.cindex "&%continue%& ACL modifier"
.cindex "database" "updating in ACL"
This modifier does nothing of itself, and processing of the ACL always
continues with the next condition or modifier. The value of &%continue%& is in
the side effects of expanding its argument. Typically this could be used to
update a database. It is really just a syntactic tidiness, to avoid having to
write rather ugly lines like this:
.display
&`condition = ${if eq{0}{`&<&'some expansion'&>&`}{true}{true}}`&
.endd
Instead, all you need is
.display
&`continue = `&<&'some expansion'&>
.endd

.vitem &*control*&&~=&~<&'text'&>
.cindex "&%control%& ACL modifier"
This modifier affects the subsequent processing of the SMTP connection or of an
incoming message that is accepted. The effect of the first type of control
lasts for the duration of the connection, whereas the effect of the second type
lasts only until the current message has been received. The message-specific
controls always apply to the whole message, not to individual recipients,
even if the &%control%& modifier appears in a RCPT ACL.

As there are now quite a few controls that can be applied, they are described
separately in section &<<SECTcontrols>>&. The &%control%& modifier can be used
in several different ways. For example:

. ==== As this is a nested list, any displays it contains must be indented
. ==== as otherwise they are too far to the left. That comment applies only
. ==== when xmlto and fop are used; formatting with sdop gets it right either
. ==== way.

.ilist
It can be at the end of an &%accept%& statement:
.code
    accept  ...some conditions
            control = queue
.endd
In this case, the control is applied when this statement yields &"accept"&, in
other words, when the conditions are all true.

.next
It can be in the middle of an &%accept%& statement:
.code
    accept  ...some conditions...
            control = queue
            ...some more conditions...
.endd
If the first set of conditions are true, the control is applied, even if the
statement does not accept because one of the second set of conditions is false.
In this case, some subsequent statement must yield &"accept"& for the control
to be relevant.

.next
It can be used with &%warn%& to apply the control, leaving the
decision about accepting or denying to a subsequent verb. For
example:
.code
    warn    ...some conditions...
            control = freeze
    accept  ...
.endd
This example of &%warn%& does not contain &%message%&, &%log_message%&, or
&%logwrite%&, so it does not add anything to the message and does not write a
log entry.

.next
If you want to apply a control unconditionally, you can use it with a
&%require%& verb. For example:
.code
    require  control = no_multiline_responses
.endd
.endlist

.vitem &*delay*&&~=&~<&'time'&>
.cindex "&%delay%& ACL modifier"
.oindex "&%-bh%&"
This modifier may appear in any ACL except notquit. It causes Exim to wait for
the time interval before proceeding. However, when testing Exim using the
&%-bh%& option, the delay is not actually imposed (an appropriate message is
output instead). The time is given in the usual Exim notation, and the delay
happens as soon as the modifier is processed. In an SMTP session, pending
output is flushed before the delay is imposed.

Like &%control%&, &%delay%& can be used with &%accept%& or &%deny%&, for
example:
.code
deny    ...some conditions...
        delay = 30s
.endd
The delay happens if all the conditions are true, before the statement returns
&"deny"&. Compare this with:
.code
deny    delay = 30s
        ...some conditions...
.endd
which waits for 30s before processing the conditions. The &%delay%& modifier
can also be used with &%warn%& and together with &%control%&:
.code
warn    ...some conditions...
        delay = 2m
        control = freeze
accept  ...
.endd

If &%delay%& is encountered when the SMTP PIPELINING extension is in use,
responses to several commands are no longer buffered and sent in one packet (as
they would normally be) because all output is flushed before imposing the
delay. This optimization is disabled so that a number of small delays do not
appear to the client as one large aggregated delay that might provoke an
unwanted timeout. You can, however, disable output flushing for &%delay%& by
using a &%control%& modifier to set &%no_delay_flush%&.


.vitem &*endpass*&
.cindex "&%endpass%& ACL modifier"
This modifier, which has no argument, is recognized only in &%accept%& and
&%discard%& statements. It marks the boundary between the conditions whose
failure causes control to pass to the next statement, and the conditions whose
failure causes the ACL to return &"deny"&. This concept has proved to be
confusing to some people, so the use of &%endpass%& is no longer recommended as
&"best practice"&. See the description of &%accept%& above for more details.


.vitem &*log_message*&&~=&~<&'text'&>
.cindex "&%log_message%& ACL modifier"
This modifier sets up a message that is used as part of the log message if the
ACL denies access or a &%warn%& statement's conditions are true. For example:
.code
require log_message = wrong cipher suite $tls_in_cipher
        encrypted   = DES-CBC3-SHA
.endd
&%log_message%& is also used when recipients are discarded by &%discard%&. For
example:
.display
&`discard `&<&'some conditions'&>
&`        log_message = Discarded $local_part@$domain because...`&
.endd
When access is denied, &%log_message%& adds to any underlying error message
that may exist because of a condition failure. For example, while verifying a
recipient address, a &':fail:'& redirection might have already set up a
message.

The message may be defined before the conditions to which it applies, because
the string expansion does not happen until Exim decides that access is to be
denied. This means that any variables that are set by the condition are
available for inclusion in the message. For example, the &$dnslist_$&<&'xxx'&>
variables are set after a DNS black list lookup succeeds. If the expansion of
&%log_message%& fails, or if the result is an empty string, the modifier is
ignored.

.vindex "&$acl_verify_message$&"
If you want to use a &%warn%& statement to log the result of an address
verification, you can use &$acl_verify_message$& to include the verification
error message.

If &%log_message%& is used with a &%warn%& statement, &"Warning:"& is added to
the start of the logged message. If the same warning log message is requested
more than once while receiving  a single email message, only one copy is
actually logged. If you want to log multiple copies, use &%logwrite%& instead
of &%log_message%&. In the absence of &%log_message%& and &%logwrite%&, nothing
is logged for a successful &%warn%& statement.

If &%log_message%& is not present and there is no underlying error message (for
example, from the failure of address verification), but &%message%& is present,
the &%message%& text is used for logging rejections. However, if any text for
logging contains newlines, only the first line is logged. In the absence of
both &%log_message%& and &%message%&, a default built-in message is used for
logging rejections.


.vitem "&*log_reject_target*&&~=&~<&'log name list'&>"
.cindex "&%log_reject_target%& ACL modifier"
.cindex "logging in ACL" "specifying which log"
This modifier makes it possible to specify which logs are used for messages
about ACL rejections. Its argument is a colon-separated list of words that can
be &"main"&, &"reject"&, or &"panic"&. The default is &`main:reject`&. The list
may be empty, in which case a rejection is not logged at all. For example, this
ACL fragment writes no logging information when access is denied:
.display
&`deny `&<&'some conditions'&>
&`     log_reject_target =`&
.endd
This modifier can be used in SMTP and non-SMTP ACLs. It applies to both
permanent and temporary rejections. Its effect lasts for the rest of the
current ACL.


.vitem &*logwrite*&&~=&~<&'text'&>
.cindex "&%logwrite%& ACL modifier"
.cindex "logging in ACL" "immediate"
This modifier writes a message to a log file as soon as it is encountered when
processing an ACL. (Compare &%log_message%&, which, except in the case of
&%warn%& and &%discard%&, is used only if the ACL statement denies
access.) The &%logwrite%& modifier can be used to log special incidents in
ACLs. For example:
.display
&`accept `&<&'some special conditions'&>
&`       control  = freeze`&
&`       logwrite = froze message because ...`&
.endd
By default, the message is written to the main log. However, it may begin
with a colon, followed by a comma-separated list of log names, and then
another colon, to specify exactly which logs are to be written. For
example:
.code
logwrite = :main,reject: text for main and reject logs
logwrite = :panic: text for panic log only
.endd


.vitem &*message*&&~=&~<&'text'&>
.cindex "&%message%& ACL modifier"
This modifier sets up a text string that is expanded and used as a response
message when an ACL statement terminates the ACL with an &"accept"&, &"deny"&,
or &"defer"& response. (In the case of the &%accept%& and &%discard%& verbs,
there is some complication if &%endpass%& is involved; see the description of
&%accept%& for details.)

The expansion of the message happens at the time Exim decides that the ACL is
to end, not at the time it processes &%message%&. If the expansion fails, or
generates an empty string, the modifier is ignored. Here is an example where
&%message%& must be specified first, because the ACL ends with a rejection if
the &%hosts%& condition fails:
.code
require  message = Host not recognized
         hosts = 10.0.0.0/8
.endd
(Once a condition has failed, no further conditions or modifiers are
processed.)

.cindex "SMTP" "error codes"
.oindex "&%smtp_banner%&
For ACLs that are triggered by SMTP commands, the message is returned as part
of the SMTP response. The use of &%message%& with &%accept%& (or &%discard%&)
is meaningful only for SMTP, as no message is returned when a non-SMTP message
is accepted. In the case of the connect ACL, accepting with a message modifier
overrides the value of &%smtp_banner%&. For the EHLO/HELO ACL, a customized
accept message may not contain more than one line (otherwise it will be
truncated at the first newline and a panic logged), and it cannot affect the
EHLO options.

When SMTP is involved, the message may begin with an overriding response code,
consisting of three digits optionally followed by an &"extended response code"&
of the form &'n.n.n'&, each code being followed by a space. For example:
.code
deny  message = 599 1.2.3 Host not welcome
      hosts = 192.168.34.0/24
.endd
The first digit of the supplied response code must be the same as would be sent
by default. A panic occurs if it is not. Exim uses a 550 code when it denies
access, but for the predata ACL, note that the default success code is 354, not
2&'xx'&.

Notwithstanding the previous paragraph, for the QUIT ACL, unlike the others,
the message modifier cannot override the 221 response code.

The text in a &%message%& modifier is literal; any quotes are taken as
literals, but because the string is expanded, backslash escapes are processed
anyway.
If the message contains newlines, this gives rise to a multi-line SMTP
response.
A long message line will also be split into multi-line SMTP responses,
on word boundaries if possible.

.vindex "&$acl_verify_message$&"
While the text is being expanded, the &$acl_verify_message$& variable
contains any message previously set.
Afterwards, &$acl_verify_message$& is cleared.

If &%message%& is used on a statement that verifies an address, the message
specified overrides any message that is generated by the verification process.
However, the original message is available in the variable
&$acl_verify_message$&, so you can incorporate it into your message if you
wish. In particular, if you want the text from &%:fail:%& items in &(redirect)&
routers to be passed back as part of the SMTP response, you should either not
use a &%message%& modifier, or make use of &$acl_verify_message$&.

For compatibility with previous releases of Exim, a &%message%& modifier that
is used with a &%warn%& verb behaves in a similar way to the &%add_header%&
modifier, but this usage is now deprecated. However, &%message%& acts only when
all the conditions are true, wherever it appears in an ACL command, whereas
&%add_header%& acts as soon as it is encountered. If &%message%& is used with
&%warn%& in an ACL that is not concerned with receiving a message, it has no
effect.


.vitem &*queue*&&~=&~<&'text'&>
.cindex "&%queue%& ACL modifier"
.cindex "named queues" "selecting in ACL"
This modifier specifies the use of a named queue for spool files
for the message.
It can only be used before the message is received (i.e. not in
the DATA ACL).
This could be used, for example, for known high-volume burst sources
of traffic, or for quarantine of messages.
Separate queue-runner processes will be needed for named queues.
If the text after expansion is empty, the default queue is used.


.vitem &*remove_header*&&~=&~<&'text'&>
This modifier specifies one or more header names in a colon-separated list
 that are to be removed from an incoming message, assuming, of course, that
the message is ultimately accepted. For details, see section &<<SECTremoveheadacl>>&.


.vitem &*set*&&~<&'acl_name'&>&~=&~<&'value'&>
.cindex "&%set%& ACL modifier"
This modifier puts a value into one of the ACL variables (see section
&<<SECTaclvariables>>&).


.vitem &*udpsend*&&~=&~<&'parameters'&>
.cindex "UDP communications"
This modifier sends a UDP packet, for purposes such as statistics
collection or behaviour monitoring. The parameters are expanded, and
the result of the expansion must be a colon-separated list consisting
of a destination server, port number, and the packet contents. The
server can be specified as a host name or IPv4 or IPv6 address. The
separator can be changed with the usual angle bracket syntax. For
example, you might want to collect information on which hosts connect
when:
.code
udpsend = <; 2001:dB8::dead:beef ; 1234 ;\
             $tod_zulu $sender_host_address
.endd
.endlist




.section "Use of the control modifier" "SECTcontrols"
.cindex "&%control%& ACL modifier"
The &%control%& modifier supports the following settings:

.vlist
.vitem &*control&~=&~allow_auth_unadvertised*&
This modifier allows a client host to use the SMTP AUTH command even when it
has not been advertised in response to EHLO. Furthermore, because there are
apparently some really broken clients that do this, Exim will accept AUTH after
HELO (rather than EHLO) when this control is set. It should be used only if you
really need it, and you should limit its use to those broken clients that do
not work without it. For example:
.code
warn hosts   = 192.168.34.25
     control = allow_auth_unadvertised
.endd
Normally, when an Exim server receives an AUTH command, it checks the name of
the authentication mechanism that is given in the command to ensure that it
matches an advertised mechanism. When this control is set, the check that a
mechanism has been advertised is bypassed. Any configured mechanism can be used
by the client. This control is permitted only in the connection and HELO ACLs.


.vitem &*control&~=&~caseful_local_part*& &&&
       &*control&~=&~caselower_local_part*&
.cindex "&ACL;" "case of local part in"
.cindex "case of local parts"
.vindex "&$local_part$&"
These two controls are permitted only in the ACL specified by &%acl_smtp_rcpt%&
(that is, during RCPT processing). By default, the contents of &$local_part$&
are lower cased before ACL processing. If &"caseful_local_part"& is specified,
any uppercase letters in the original local part are restored in &$local_part$&
for the rest of the ACL, or until a control that sets &"caselower_local_part"&
is encountered.

These controls affect only the current recipient. Moreover, they apply only to
local part handling that takes place directly in the ACL (for example, as a key
in lookups). If a test to verify the recipient is obeyed, the case-related
handling of the local part during the verification is controlled by the router
configuration (see the &%caseful_local_part%& generic router option).

This facility could be used, for example, to add a spam score to local parts
containing upper case letters. For example, using &$acl_m4$& to accumulate the
spam score:
.code
warn  control = caseful_local_part
      set acl_m4 = ${eval:\
                     $acl_m4 + \
                     ${if match{$local_part}{[A-Z]}{1}{0}}\
                    }
      control = caselower_local_part
.endd
Notice that we put back the lower cased version afterwards, assuming that
is what is wanted for subsequent tests.


.vitem &*control&~=&~cutthrough_delivery/*&<&'options'&>
.cindex "&ACL;" "cutthrough routing"
.cindex "cutthrough" "requesting"
This option requests delivery be attempted while the item is being received.

The option is usable in the RCPT ACL.
If enabled for a message received via smtp and routed to an smtp transport,
and only one transport, interface, destination host and port combination
is used for all recipients of the message,
then the delivery connection is made while the receiving connection is open
and data is copied from one to the other.

An attempt to set this option for any recipient but the first
for a mail will be quietly ignored.
If a recipient-verify callout
(with use_sender)
connection is subsequently
requested in the same ACL it is held open and used for
any subsequent recipients and the data,
otherwise one is made after the initial RCPT ACL completes.

Note that routers are used in verify mode,
and cannot depend on content of received headers.
Note also that headers cannot be
modified by any of the post-data ACLs (DATA, MIME and DKIM).
Headers may be modified by routers (subject to the above) and transports.
The &'Received-By:'& header is generated as soon as the body reception starts,
rather than the traditional time after the full message is received;
this will affect the timestamp.

All the usual ACLs are called; if one results in the message being
rejected, all effort spent in delivery (including the costs on
the ultimate destination) will be wasted.
Note that in the case of data-time ACLs this includes the entire
message body.

Cutthrough delivery is not supported via transport-filters or when DKIM signing
of outgoing messages is done, because it sends data to the ultimate destination
before the entire message has been received from the source.
It is not supported for messages received with the SMTP PRDR
or CHUNKING
options in use.

Should the ultimate destination system positively accept or reject the mail,
a corresponding indication is given to the source system and nothing is queued.
If the item is successfully delivered in cutthrough mode
the delivery log lines are tagged with ">>" rather than "=>" and appear
before the acceptance "<=" line.

If there is a temporary error the item is queued for later delivery in the
usual fashion.
This behaviour can be adjusted by appending the option &*defer=*&<&'value'&>
to the control; the default value is &"spool"& and the alternate value
&"pass"& copies an SMTP defer response from the target back to the initiator
and does not queue the message.
Note that this is independent of any recipient verify conditions in the ACL.

Delivery in this mode avoids the generation of a bounce mail to a
(possibly faked)
sender when the destination system is doing content-scan based rejection.


.vitem &*control&~=&~debug/*&<&'options'&>
.cindex "&ACL;" "enabling debug logging"
.cindex "debugging" "enabling from an ACL"
This control turns on debug logging, almost as though Exim had been invoked
with &`-d`&, with the output going to a new logfile in the usual logs directory,
by default called &'debuglog'&.

Logging set up by the control will be maintained across spool residency.

Options are a slash-separated list.
If an option takes an argument, the option name and argument are separated by
an equals character.
Several options are supported:
.display
tag=<&'suffix'&>         The filename can be adjusted with thise option.
                     The argument, which may access any variables already defined,
                     is appended to the default name.

opts=<&'debug&~options'&> The argument specififes what is to be logged,
                     using the same values as the &`-d`& command-line option.

stop                 Logging started with this control may be
                     stopped by using this option.

kill                 Logging started with this control may be
                     stopped by using this option.
                     Additionally the debug file will be removed,
                     providing one means for speculative debug tracing.

pretrigger=<&'size'&>    This option specifies a memory buffuer to be used
                     for pre-trigger debug capture.
                     Debug lines are recorded in the buffer until
                     and if) a trigger occurs; at which time they are
                     dumped to the debug file.  Newer lines displace the
                     oldest if the buffer is full.  After a trigger,
                     immediate writes to file are done as normal.

trigger=<&'reason'&>     This option selects cause for the pretrigger buffer
                     see above) to be copied to file.  A reason of &*now*&
                     take effect immediately; one of &*paniclog*& triggers
                     on a write to the panic log.
.endd

Some examples (which depend on variables that don't exist in all
contexts):
.code
      control = debug
      control = debug/tag=.$sender_host_address
      control = debug/opts=+expand+acl
      control = debug/tag=.$message_exim_id/opts=+expand
      control = debug/kill
      control = debug/opts=+all/pretrigger=1024/trigger=paniclog
      control = debug/trigger=now
.endd


.vitem &*control&~=&~dkim_disable_verify*&
.cindex "disable DKIM verify"
.cindex "DKIM" "disable verify"
This control turns off DKIM verification processing entirely.  For details on
the operation and configuration of DKIM, see section &<<SECDKIM>>&.


.vitem &*control&~=&~dmarc_disable_verify*& &&&
       &*control&~=&~dmarc_enable_forensic*&
.cindex "disable DMARC verify"
.cindex DMARC "disable verify"
.cindex DMARC controls
.cindex DMARC "forensic mails"
These control affect DMARC processing.  For details on
the operation and configuration of DMARC, see section &<<SECDMARC>>&.

The &"disable"& turns off DMARC verification processing entirely.


.vitem &*control&~=&~dscp/*&<&'value'&>
.cindex "&ACL;" "setting DSCP value"
.cindex "DSCP" "inbound"
This option causes the DSCP value associated with the socket for the inbound
connection to be adjusted to a given value, given as one of a number of fixed
strings or to numeric value.
The &%-bI:dscp%& option may be used to ask Exim which names it knows of.
Common values include &`throughput`&, &`mincost`&, and on newer systems
&`ef`&, &`af41`&, etc.  Numeric values may be in the range 0 to 0x3F.

The outbound packets from Exim will be marked with this value in the header
(for IPv4, the TOS field; for IPv6, the TCLASS field); there is no guarantee
that these values will have any effect, not be stripped by networking
equipment, or do much of anything without cooperation with your Network
Engineer and those of all network operators between the source and destination.


.vitem &*control&~=&~enforce_sync*& &&&
       &*control&~=&~no_enforce_sync*&
.cindex "SMTP" "synchronization checking"
.cindex "synchronization checking in SMTP"
These controls make it possible to be selective about when SMTP synchronization
is enforced. The global option &%smtp_enforce_sync%& specifies the initial
state of the switch (it is true by default). See the description of this option
in chapter &<<CHAPmainconfig>>& for details of SMTP synchronization checking.

The effect of these two controls lasts for the remainder of the SMTP
connection. They can appear in any ACL except the one for the non-SMTP
messages. The most straightforward place to put them is in the ACL defined by
&%acl_smtp_connect%&, which is run at the start of an incoming SMTP connection,
before the first synchronization check. The expected use is to turn off the
synchronization checks for badly-behaved hosts that you nevertheless need to
work with.


.vitem &*control&~=&~fakedefer/*&<&'message'&>
.cindex "fake defer"
.cindex "defer, fake"
.cindex fakedefer
This control works in exactly the same way as &%fakereject%& (described below)
except that it causes an SMTP 450 response after the message data instead of a
550 response. You must take care when using &%fakedefer%& because it causes the
messages to be duplicated when the sender retries. Therefore, you should not
use &%fakedefer%& if the message is to be delivered normally.

.vitem &*control&~=&~fakereject/*&<&'message'&>
.cindex "fake rejection"
.cindex "rejection, fake"
.cindex fakereject
This control is permitted only for the MAIL, RCPT, and DATA ACLs, in other
words, only when an SMTP message is being received. If Exim accepts the
message, instead the final 250 response, a 550 rejection message is sent.
However, Exim proceeds to deliver the message as normal. The control applies
only to the current message, not to any subsequent ones that may be received in
the same SMTP connection.

The text for the 550 response is taken from the &%control%& modifier. If no
message is supplied, the following is used:
.code
550-Your message has been rejected but is being
550-kept for evaluation.
550-If it was a legitimate message, it may still be
550 delivered to the target recipient(s).
.endd
This facility should be used with extreme caution.

.vitem &*control&~=&~freeze*&
.cindex "frozen messages" "forcing in ACL"
This control is permitted only for the MAIL, RCPT, DATA, and non-SMTP ACLs, in
other words, only when a message is being received. If the message is accepted,
it is placed on Exim's queue and frozen. The control applies only to the
current message, not to any subsequent ones that may be received in the same
SMTP connection.

This modifier can optionally be followed by &`/no_tell`&. If the global option
&%freeze_tell%& is set, it is ignored for the current message (that is, nobody
is told about the freezing), provided all the &*control=freeze*& modifiers that
are obeyed for the current message have the &`/no_tell`& option.

.vitem &*control&~=&~no_delay_flush*&
.cindex "SMTP" "output flushing, disabling for delay"
Exim normally flushes SMTP output before implementing a delay in an ACL, to
avoid unexpected timeouts in clients when the SMTP PIPELINING extension is in
use. This control, as long as it is encountered before the &%delay%& modifier,
disables such output flushing.

.vitem &*control&~=&~no_callout_flush*&
.cindex "SMTP" "output flushing, disabling for callout"
Exim normally flushes SMTP output before performing a callout in an ACL, to
avoid unexpected timeouts in clients when the SMTP PIPELINING extension is in
use. This control, as long as it is encountered before the &%verify%& condition
that causes the callout, disables such output flushing.

.vitem &*control&~=&~no_mbox_unspool*&
This control is available when Exim is compiled with the content scanning
extension. Content scanning may require a copy of the current message, or parts
of it, to be written in &"mbox format"& to a spool file, for passing to a virus
or spam scanner. Normally, such copies are deleted when they are no longer
needed. If this control is set, the copies are not deleted. The control applies
only to the current message, not to any subsequent ones that may be received in
the same SMTP connection. It is provided for debugging purposes and is unlikely
to be useful in production.

.vitem &*control&~=&~no_multiline_responses*&
.cindex "multiline responses, suppressing"
This control is permitted for any ACL except the one for non-SMTP messages.
It seems that there are broken clients in use that cannot handle multiline
SMTP responses, despite the fact that
&url(https://www.rfc-editor.org/rfc/rfc821,RFC 821)
defined them over 20 years ago.

If this control is set, multiline SMTP responses from ACL rejections are
suppressed. One way of doing this would have been to put out these responses as
one long line. However,
&url(https://www.rfc-editor.org/rfc/rfc2821,RFC 2821)
specifies a maximum of 512 bytes per response
(&"use multiline responses for more"& it says &-- ha!), and some of the
responses might get close to that. So this facility, which is after all only a
sop to broken clients, is implemented by doing two very easy things:

.ilist
Extra information that is normally output as part of a rejection caused by
sender verification failure is omitted. Only the final line (typically &"sender
verification failed"&) is sent.
.next
If a &%message%& modifier supplies a multiline response, only the first
line is output.
.endlist

The setting of the switch can, of course, be made conditional on the
calling host. Its effect lasts until the end of the SMTP connection.

.vitem &*control&~=&~no_pipelining*&
.cindex "PIPELINING" "suppressing advertising"
.cindex "ESMTP extensions" PIPELINING
This control turns off the advertising of the PIPELINING extension to SMTP in
the current session. To be useful, it must be obeyed before Exim sends its
response to an EHLO command. Therefore, it should normally appear in an ACL
controlled by &%acl_smtp_connect%& or &%acl_smtp_helo%&. See also
&%pipelining_advertise_hosts%&.

.vitem &*control&~=&~queue/*&<&'options'&>* &&&
       &*control&~=&~queue_only*&
.oindex "&%queue%&"
.oindex "&%queue_only%&"
.cindex "queueing incoming messages"
.cindex queueing "forcing in ACL"
.cindex "first pass routing"
This control is permitted only for the MAIL, RCPT, DATA, and non-SMTP ACLs, in
other words, only when a message is being received. If the message is accepted,
it is placed on Exim's queue and left there for delivery by a subsequent queue
runner.
If used with no options set,
no immediate delivery process is started. In other words, it has the
effect as the &%queue_only%& global option or &'-odq'& command-line option.

If the &'first_pass_route'& option is given then
the behaviour is like the command-line &'-oqds'& option;
a delivery process is started which stops short of making
any SMTP delivery.  The benefit is that the hints database will be updated for
the message being waiting for a specific host, and a later queue run will be
able to send all such messages on a single connection.

The control only applies to the current message, not to any subsequent ones that
 may be received in the same SMTP connection.

.vitem &*control&~=&~submission/*&<&'options'&>
.cindex "message" "submission"
.cindex "submission mode"
This control is permitted only for the MAIL, RCPT, and start of data ACLs (the
latter is the one defined by &%acl_smtp_predata%&). Setting it tells Exim that
the current message is a submission from a local MUA. In this case, Exim
operates in &"submission mode"&, and applies certain fixups to the message if
necessary. For example, it adds a &'Date:'& header line if one is not present.
This control is not permitted in the &%acl_smtp_data%& ACL, because that is too
late (the message has already been created).

Chapter &<<CHAPmsgproc>>& describes the processing that Exim applies to
messages. Section &<<SECTsubmodnon>>& covers the processing that happens in
submission mode; the available options for this control are described there.
The control applies only to the current message, not to any subsequent ones
that may be received in the same SMTP connection.

.vitem &*control&~=&~suppress_local_fixups*&
.cindex "submission fixups, suppressing"
This control applies to locally submitted (non TCP/IP) messages, and is the
complement of &`control = submission`&. It disables the fixups that are
normally applied to locally-submitted messages. Specifically:

.ilist
Any &'Sender:'& header line is left alone (in this respect, it is a
dynamic version of &%local_sender_retain%&).
.next
No &'Message-ID:'&, &'From:'&, or &'Date:'& header lines are added.
.next
There is no check that &'From:'& corresponds to the actual sender.
.endlist ilist

This control may be useful when a remotely-originated message is accepted,
passed to some scanning program, and then re-submitted for delivery. It can be
used only in the &%acl_smtp_mail%&, &%acl_smtp_rcpt%&, &%acl_smtp_predata%&,
and &%acl_not_smtp_start%& ACLs, because it has to be set before the message's
data is read.

&*Note:*& This control applies only to the current message, not to any others
that are being submitted at the same time using &%-bs%& or &%-bS%&.

.vitem &*control&~=&~utf8_downconvert*&
This control enables conversion of UTF-8 in message envelope addresses
to a-label form.
For details see section &<<SECTi18nMTA>>&.

.vitem &*control&~=&~wellknown*&
This control sets up a response data file for a WELLKNOWN SMTP command.
It may only be used in an ACL servicing that command.
For details see section &<<SECTWELLKNOWNACL>>&.
.endlist vlist


.section "Summary of message fixup control" "SECTsummesfix"
All four possibilities for message fixups can be specified:

.ilist
Locally submitted, fixups applied: the default.
.next
Locally submitted, no fixups applied: use
&`control = suppress_local_fixups`&.
.next
Remotely submitted, no fixups applied: the default.
.next
Remotely submitted, fixups applied: use &`control = submission`&.
.endlist



.section "Adding header lines in ACLs" "SECTaddheadacl"
.cindex "header lines" "adding in an ACL"
.cindex "header lines" "position of added lines"
.cindex "&%add_header%& ACL modifier"
The &%add_header%& modifier can be used to add one or more extra header lines
to an incoming message, as in this example:
.code
warn dnslists = sbl.spamhaus.org : \
                dialup.mail-abuse.org
     add_header = X-blacklisted-at: $dnslist_domain
.endd
The &%add_header%& modifier is permitted in the MAIL, RCPT, PREDATA, DATA,
MIME, DKIM, and non-SMTP ACLs (in other words, those that are concerned with
receiving a message). The message must ultimately be accepted for
&%add_header%& to have any significant effect. You can use &%add_header%& with
any ACL verb, including &%deny%& (though this is potentially useful only in a
RCPT ACL).

Headers will not be added to the message if the modifier is used in
DATA, MIME or DKIM ACLs for a message delivered by cutthrough routing.

Leading and trailing newlines are removed from
the data for the &%add_header%& modifier; if it then
contains one or more newlines that
are not followed by a space or a tab, it is assumed to contain multiple header
lines. Each one is checked for valid syntax; &`X-ACL-Warn:`& is added to the
front of any line that is not a valid header line.

Added header lines are accumulated during the MAIL, RCPT, and predata ACLs.
They are added to the message before processing the DATA and MIME ACLs.
However, if an identical header line is requested more than once, only one copy
is actually added to the message. Further header lines may be accumulated
during the DATA and MIME ACLs, after which they are added to the message, again
with duplicates suppressed. Thus, it is possible to add two identical header
lines to an SMTP message, but only if one is added before DATA and one after.
In the case of non-SMTP messages, new headers are accumulated during the
non-SMTP ACLs, and are added to the message after all the ACLs have run. If a
message is rejected after DATA or by the non-SMTP ACL, all added header lines
are included in the entry that is written to the reject log.

.cindex "header lines" "added; visibility of"
Header lines are not visible in string expansions
of message headers
until they are added to the
message. It follows that header lines defined in the MAIL, RCPT, and predata
ACLs are not visible until the DATA ACL and MIME ACLs are run. Similarly,
header lines that are added by the DATA or MIME ACLs are not visible in those
ACLs. Because of this restriction, you cannot use header lines as a way of
passing data between (for example) the MAIL and RCPT ACLs. If you want to do
this, you can use ACL variables, as described in section
&<<SECTaclvariables>>&.

The list of headers yet to be added is given by the &%$headers_added%& variable.

The &%add_header%& modifier acts immediately as it is encountered during the
processing of an ACL. Notice the difference between these two cases:
.display
&`accept add_header = ADDED: some text`&
&`       `&<&'some condition'&>

&`accept `&<&'some condition'&>
&`       add_header = ADDED: some text`&
.endd
In the first case, the header line is always added, whether or not the
condition is true. In the second case, the header line is added only if the
condition is true. Multiple occurrences of &%add_header%& may occur in the same
ACL statement. All those that are encountered before a condition fails are
honoured.

.cindex "&%warn%& ACL verb"
For compatibility with previous versions of Exim, a &%message%& modifier for a
&%warn%& verb acts in the same way as &%add_header%&, except that it takes
effect only if all the conditions are true, even if it appears before some of
them. Furthermore, only the last occurrence of &%message%& is honoured. This
usage of &%message%& is now deprecated. If both &%add_header%& and &%message%&
are present on a &%warn%& verb, both are processed according to their
specifications.

By default, new header lines are added to a message at the end of the existing
header lines. However, you can specify that any particular header line should
be added right at the start (before all the &'Received:'& lines), immediately
after the first block of &'Received:'& lines, or immediately before any line
that is not a &'Received:'& or &'Resent-something:'& header.

This is done by specifying &":at_start:"&, &":after_received:"&, or
&":at_start_rfc:"& (or, for completeness, &":at_end:"&) before the text of the
header line, respectively. (Header text cannot start with a colon, as there has
to be a header name first.) For example:
.code
warn add_header = \
       :after_received:X-My-Header: something or other...
.endd
If more than one header line is supplied in a single &%add_header%& modifier,
each one is treated independently and can therefore be placed differently. If
you add more than one line at the start, or after the Received: block, they end
up in reverse order.

&*Warning*&: This facility currently applies only to header lines that are
added in an ACL. It does NOT work for header lines that are added in a
system filter or in a router or transport.



.section "Removing header lines in ACLs" "SECTremoveheadacl"
.cindex "header lines" "removing in an ACL"
.cindex "header lines" "position of removed lines"
.cindex "&%remove_header%& ACL modifier"
The &%remove_header%& modifier can be used to remove one or more header lines
from an incoming message, as in this example:
.code
warn   message        = Remove internal headers
       remove_header  = x-route-mail1 : x-route-mail2
.endd
The &%remove_header%& modifier is permitted in the MAIL, RCPT, PREDATA, DATA,
MIME, DKIM, and non-SMTP ACLs (in other words, those that are concerned with
receiving a message). The message must ultimately be accepted for
&%remove_header%& to have any significant effect. You can use &%remove_header%&
with any ACL verb, including &%deny%&, though this is really not useful for
any verb that doesn't result in a delivered message.

Headers will not be removed from the message if the modifier is used in
DATA, MIME or DKIM ACLs for a message delivered by cutthrough routing.

More than one header can be removed at the same time by using a colon separated
list of header specifiers.
If a specifier does not start with a circumflex (^)
then it is treated as a header name.
The header name matching is case insensitive.
If it does, then it is treated as a (front-anchored)
regular expression applied to the whole header.

&*Note*&: The colon terminating a header name will need to be doubled
if used in an RE, and there can legitimately be whitepace before it.

Example:
.code
remove_header = \N^(?i)Authentication-Results\s*::\s*example.org;\N
.endd

List expansion is not performed, so you cannot use hostlists to
create a list of headers, however both connection and message variable expansion
are performed (&%$acl_c_*%& and &%$acl_m_*%&), illustrated in this example:
.code
warn   hosts           = +internal_hosts
       set acl_c_ihdrs = x-route-mail1 : x-route-mail2
warn   message         = Remove internal headers
       remove_header   = $acl_c_ihdrs
.endd
Header specifiers for removal are accumulated during the MAIL, RCPT, and predata ACLs.
Matching header lines are removed from the message before processing the DATA and MIME ACLs.
If multiple header lines match, all are removed.
There is no harm in attempting to remove the same header twice nor in removing
a non-existent header. Further header specifiers for removal may be accumulated
during the DATA and MIME ACLs, after which matching headers are removed
if present. In the case of non-SMTP messages, remove specifiers are
accumulated during the non-SMTP ACLs, and are acted on after
all the ACLs have run. If a message is rejected after DATA or by the non-SMTP
ACL, there really is no effect because there is no logging of what headers
would have been removed.

.cindex "header lines" "removed; visibility of"
Header lines are not visible in string expansions until the DATA phase when it
is received. Any header lines removed in the MAIL, RCPT, and predata ACLs are
not visible in the DATA ACL and MIME ACLs. Similarly, header lines that are
removed by the DATA or MIME ACLs are still visible in those ACLs. Because of
this restriction, you cannot use header lines as a way of controlling data
passed between (for example) the MAIL and RCPT ACLs. If you want to do this,
you should instead use ACL variables, as described in section
&<<SECTaclvariables>>&.

The &%remove_header%& modifier acts immediately as it is encountered during the
processing of an ACL. Notice the difference between these two cases:
.display
&`accept remove_header = X-Internal`&
&`       `&<&'some condition'&>

&`accept `&<&'some condition'&>
&`       remove_header = X-Internal`&
.endd
In the first case, the header line is always removed, whether or not the
condition is true. In the second case, the header line is removed only if the
condition is true. Multiple occurrences of &%remove_header%& may occur in the
same ACL statement. All those that are encountered before a condition fails
are honoured.

&*Warning*&: This facility currently applies only to header lines that are
present during ACL processing. It does NOT remove header lines that are added
in a system filter or in a router or transport.




.section "ACL conditions" "SECTaclconditions"
.cindex "&ACL;" "conditions; list of"
Some of the conditions listed in this section are available only when Exim is
compiled with the content-scanning extension. They are included here briefly
for completeness. More detailed descriptions can be found in the discussion on
content scanning in chapter &<<CHAPexiscan>>&.

Not all conditions are relevant in all circumstances. For example, testing
senders and recipients does not make sense in an ACL that is being run as the
result of the arrival of an ETRN command, and checks on message headers can be
done only in the ACLs specified by &%acl_smtp_data%& and &%acl_not_smtp%&. You
can use the same condition (with different parameters) more than once in the
same ACL statement. This provides a way of specifying an &"and"& conjunction.
The conditions are as follows:


.vlist
.vitem &*acl&~=&~*&<&'name&~of&~acl&~or&~ACL&~string&~or&~file&~name&~'&>
.cindex "&ACL;" "nested"
.cindex "&ACL;" "indirect"
.cindex "&ACL;" "arguments"
.cindex "&%acl%& ACL condition"
The possible values of the argument are the same as for the
&%acl_smtp_%&&'xxx'& options. The named or inline ACL is run. If it returns
&"accept"& the condition is true; if it returns &"deny"& the condition is
false. If it returns &"defer"&, the current ACL returns &"defer"& unless the
condition is on a &%warn%& verb. In that case, a &"defer"& return makes the
condition false. This means that further processing of the &%warn%& verb
ceases, but processing of the ACL continues.

If the argument is a named ACL, up to nine space-separated optional values
can be appended; they appear within the called ACL in $acl_arg1 to $acl_arg9,
and $acl_narg is set to the count of values.
Previous values of these variables are restored after the call returns.
The name and values are expanded separately.
Note that spaces in complex expansions which are used as arguments
will act as argument separators.

If the nested &%acl%& returns &"drop"& and the outer condition denies access,
the connection is dropped. If it returns &"discard"&, the verb must be
&%accept%& or &%discard%&, and the action is taken immediately &-- no further
conditions are tested.

ACLs may be nested up to 20 deep; the limit exists purely to catch runaway
loops. This condition allows you to use different ACLs in different
circumstances. For example, different ACLs can be used to handle RCPT commands
for different local users or different local domains.

.new
.vitem &*atrn_domains&~=&~*&<&'domain&~list'&>
.cindex ATRN "checking for queued messages"
This condition is only usable in the ATRN ACL.
It returns true if there are any messages queued for any of the domains given
in the list.
The list supplied must not be tainted
.cindex "tainted data" "de-tainting"
and should contain only domains relevant for the authenticated user
(to avoid leaking information about other users).
.wen

.vitem &*authenticated&~=&~*&<&'string&~list'&>
.cindex "&%authenticated%& ACL condition"
.cindex "authentication" "ACL checking"
.cindex "&ACL;" "testing for authentication"
If the SMTP connection is not authenticated, the condition is false. Otherwise,
the name of the authenticator is tested against the list. To test for
authentication by any authenticator, you can set
.code
authenticated = *
.endd

.vitem &*condition&~=&~*&<&'string'&>
.cindex "&%condition%& ACL condition"
.cindex "customizing" "ACL condition"
.cindex "&ACL;" "customized test"
.cindex "&ACL;" "testing, customized"
This feature allows you to make up custom conditions. If the result of
expanding the string is an empty string, the number zero, or one of the strings
&"no"& or &"false"&, the condition is false. If the result is any non-zero
number, or one of the strings &"yes"& or &"true"&, the condition is true. For
any other value, some error is assumed to have occurred, and the ACL returns
&"defer"&. However, if the expansion is forced to fail, the condition is
ignored. The effect is to treat it as true, whether it is positive or
negative.

.vitem &*decode&~=&~*&<&'location'&>
.cindex "&%decode%& ACL condition"
This condition is available only when Exim is compiled with the
content-scanning extension, and it is allowed only in the ACL defined by
&%acl_smtp_mime%&. It causes the current MIME part to be decoded into a file.
If all goes well, the condition is true. It is false only if there are
problems such as a syntax error or a memory shortage. For more details, see
chapter &<<CHAPexiscan>>&.

.vitem &*dnslists&~=&~*&<&'list&~of&~domain&~names&~and&~other&~data'&>
.cindex "&%dnslists%& ACL condition"
.cindex "DNS list" "in ACL"
.cindex "black list (DNS)"
.cindex "&ACL;" "testing a DNS list"
This condition checks for entries in DNS black lists. These are also known as
&"RBL lists"&, after the original Realtime Blackhole List, but note that the
use of the lists at &'mail-abuse.org'& now carries a charge. There are too many
different variants of this condition to describe briefly here. See sections
&<<SECTmorednslists>>&&--&<<SECTmorednslistslast>>& for details.

.vitem &*domains&~=&~*&<&'domain&~list'&>
.cindex "&%domains%& ACL condition"
.cindex "domain" "ACL checking"
.cindex "&ACL;" "testing a recipient domain"
.vindex "&$domain_data$&"
This condition is relevant only in a RCPT ACL. It checks that the domain
of the recipient address is in the domain list. If percent-hack processing is
enabled, it is done before this test is done. If the check succeeds with a
lookup, the result of the lookup is placed in &$domain_data$& until the next
&%domains%& test.

&*Note carefully*& (because many people seem to fall foul of this): you cannot
use &%domains%& in a DATA ACL.


.vitem &*encrypted&~=&~*&<&'string&~list'&>
.cindex "&%encrypted%& ACL condition"
.cindex "encryption" "checking in an ACL"
.cindex "&ACL;" "testing for encryption"
If the SMTP connection is not encrypted, the condition is false. Otherwise, the
name of the cipher suite in use is tested against the list. To test for
encryption without testing for any specific cipher suite(s), set
.code
encrypted = *
.endd


.vitem &*hosts&~=&~*&<&'host&~list'&>
.cindex "&%hosts%& ACL condition"
.cindex "host" "ACL checking"
.cindex "&ACL;" "testing the client host"
This condition tests that the calling host matches the host list. If you have
name lookups or wildcarded host names and IP addresses in the same host list,
you should normally put the IP addresses first. For example, you could have:
.code
accept hosts = 10.9.8.7 : dbm;/etc/friendly/hosts
.endd
The lookup in this example uses the host name for its key. This is implied by
the lookup type &"dbm"&. (For a host address lookup you would use &"net-dbm"&
and it wouldn't matter which way round you had these two items.)

The reason for the problem with host names lies in the left-to-right way that
Exim processes lists. It can test IP addresses without doing any DNS lookups,
but when it reaches an item that requires a host name, it fails if it cannot
find a host name to compare with the pattern. If the above list is given in the
opposite order, the &%accept%& statement fails for a host whose name cannot be
found, even if its IP address is 10.9.8.7.

If you really do want to do the name check first, and still recognize the IP
address even if the name lookup fails, you can rewrite the ACL like this:
.code
accept hosts = dbm;/etc/friendly/hosts
accept hosts = 10.9.8.7
.endd
The default action on failing to find the host name is to assume that the host
is not in the list, so the first &%accept%& statement fails. The second
statement can then check the IP address.

.vindex "&$host_data$&"
If a &%hosts%& condition is satisfied by means of a lookup, the result
of the lookup is made available in the &$host_data$& variable. This
allows you, for example, to set up a statement like this:
.code
deny  hosts = net-lsearch;/some/file
      message = $host_data
.endd
which gives a custom error message for each denied host.

.vitem &*local_parts&~=&~*&<&'local&~part&~list'&>
.cindex "&%local_parts%& ACL condition"
.cindex "local part" "ACL checking"
.cindex "&ACL;" "testing a local part"
.vindex "&$local_part_data$&"
This condition is relevant only in a RCPT ACL. It checks that the local
part of the recipient address is in the list. If percent-hack processing is
enabled, it is done before this test. If the check succeeds with a lookup, the
result of the lookup is placed in &$local_part_data$&, which remains set until
the next &%local_parts%& test.

.vitem &*malware&~=&~*&<&'option'&>
.cindex "&%malware%& ACL condition"
.cindex "&ACL;" "virus scanning"
.cindex "&ACL;" "scanning for viruses"
This condition is available only when Exim is compiled with the
content-scanning extension
and only after a DATA command.
It causes the incoming message to be scanned for
viruses. For details, see chapter &<<CHAPexiscan>>&.

.vitem &*mime_regex&~=&~*&<&'list&~of&~regular&~expressions'&>
.cindex "&%mime_regex%& ACL condition"
.cindex "&ACL;" "testing by regex matching"
This condition is available only when Exim is compiled with the
content-scanning extension, and it is allowed only in the ACL defined by
&%acl_smtp_mime%&. It causes the current MIME part to be scanned for a match
with any of the regular expressions. For details, see chapter
&<<CHAPexiscan>>&.

.vitem &*ratelimit&~=&~*&<&'parameters'&>
.cindex "rate limiting"
This condition can be used to limit the rate at which a user or host submits
messages. Details are given in section &<<SECTratelimiting>>&.

.vitem &*recipients&~=&~*&<&'address&~list'&>
.cindex "&%recipients%& ACL condition"
.cindex "recipient" "ACL checking"
.cindex "&ACL;" "testing a recipient"
This condition is relevant only in a RCPT ACL. It checks the entire
recipient address against a list of recipients.

.vitem &*regex&~=&~*&<&'list&~of&~regular&~expressions'&>
.cindex "&%regex%& ACL condition"
.cindex "&ACL;" "testing by regex matching"
This condition is available only when Exim is compiled with the
content-scanning extension, and is available only in the DATA, MIME, and
non-SMTP ACLs. It causes the incoming message to be scanned for a match with
any of the regular expressions. For details, see chapter &<<CHAPexiscan>>&.

.vitem &*seen&~=&~*&<&'parameters'&>
.cindex "&%seen%& ACL condition"
This condition can be used to test if a situation has been previously met,
for example for greylisting.
Details are given in section &<<SECTseen>>&.

.vitem &*sender_domains&~=&~*&<&'domain&~list'&>
.cindex "&%sender_domains%& ACL condition"
.cindex "sender" "ACL checking"
.cindex "&ACL;" "testing a sender domain"
.vindex "&$domain$&"
.vindex "&$sender_address_domain$&"
This condition tests the domain of the sender of the message against the given
domain list. &*Note*&: The domain of the sender address is in
&$sender_address_domain$&. It is &'not'& put in &$domain$& during the testing
of this condition. This is an exception to the general rule for testing domain
lists. It is done this way so that, if this condition is used in an ACL for a
RCPT command, the recipient's domain (which is in &$domain$&) can be used to
influence the sender checking.

&*Warning*&: It is a bad idea to use this condition on its own as a control on
relaying, because sender addresses are easily, and commonly, forged.

.vitem &*senders&~=&~*&<&'address&~list'&>
.cindex "&%senders%& ACL condition"
.cindex "sender" "ACL checking"
.cindex "&ACL;" "testing a sender"
This condition tests the sender of the message against the given list. To test
for a bounce message, which has an empty sender, set
.code
senders = :
.endd
&*Warning*&: It is a bad idea to use this condition on its own as a control on
relaying, because sender addresses are easily, and commonly, forged.

.vitem &*spam&~=&~*&<&'username'&>
.cindex "&%spam%& ACL condition"
.cindex "&ACL;" "scanning for spam"
This condition is available only when Exim is compiled with the
content-scanning extension. It causes the incoming message to be scanned by
SpamAssassin. For details, see chapter &<<CHAPexiscan>>&.

.vitem &*verify&~=&~certificate*&
.cindex "&%verify%& ACL condition"
.cindex "TLS" "client certificate verification"
.cindex "certificate" "verification of client"
.cindex "&ACL;" "certificate verification"
.cindex "&ACL;" "testing a TLS certificate"
This condition is true in an SMTP session if the session is encrypted, and a
certificate was received from the client, and the certificate was verified. The
server requests a certificate only if the client matches &%tls_verify_hosts%&
or &%tls_try_verify_hosts%& (see chapter &<<CHAPTLS>>&).

.vitem &*verify&~=&~csa*&
.cindex "CSA verification"
This condition checks whether the sending host (the client) is authorized to
send email. Details of how this works are given in section
&<<SECTverifyCSA>>&.

.vitem &*verify&~=&~header_names_ascii*&
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying header names only ASCII"
.cindex "header lines" "verifying header names only ASCII"
.cindex "verifying" "header names only ASCII"
This condition is relevant only in an ACL that is run after a message has been
received.
This usually means an ACL specified by &%acl_smtp_data%& or &%acl_not_smtp%&.
It checks all header names (not the content) to make sure
there are no non-ASCII characters, also excluding control characters.  The
allowable characters are decimal ASCII values 33 through 126.

Exim itself will handle headers with non-ASCII characters, but it can cause
problems for downstream applications, so this option will allow their
detection and rejection in the DATA ACL's.

.vitem &*verify&~=&~header_sender/*&<&'options'&>
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying sender in the header"
.cindex "header lines" "verifying the sender in"
.cindex "sender" "verifying in header"
.cindex "verifying" "sender in header"
This condition is relevant only in an ACL that is run after a message has been
received, that is, in an ACL specified by &%acl_smtp_data%& or
&%acl_not_smtp%&. It checks that there is a verifiable address in at least one
of the &'Sender:'&, &'Reply-To:'&, or &'From:'& header lines. Such an address
is loosely thought of as a &"sender"& address (hence the name of the test).
However, an address that appears in one of these headers need not be an address
that accepts bounce messages; only sender addresses in envelopes are required
to accept bounces. Therefore, if you use the callout option on this check, you
might want to arrange for a non-empty address in the MAIL command.

Details of address verification and the options are given later, starting at
section &<<SECTaddressverification>>& (callouts are described in section
&<<SECTcallver>>&). You can combine this condition with the &%senders%&
condition to restrict it to bounce messages only:
.code
deny    senders = :
       !verify  = header_sender
        message = A valid sender header is required for bounces
.endd

.vitem &*verify&~=&~header_syntax*&
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying header syntax"
.cindex "header lines" "verifying syntax"
.cindex "verifying" "header syntax"
This condition is relevant only in an ACL that is run after a message has been
received, that is, in an ACL specified by &%acl_smtp_data%& or
&%acl_not_smtp%&. It checks the syntax of all header lines that can contain
lists of addresses (&'Sender:'&, &'From:'&, &'Reply-To:'&, &'To:'&, &'Cc:'&,
and &'Bcc:'&), returning true if there are no problems.
Unqualified addresses (local parts without domains) are
permitted only in locally generated messages and from hosts that match
&%sender_unqualified_hosts%& or &%recipient_unqualified_hosts%&, as
appropriate.

Note that this condition is a syntax check only. However, a common spamming
ploy used to be to send syntactically invalid headers such as
.code
To: @
.endd
and this condition can be used to reject such messages, though they are not as
common as they used to be.

.vitem &*verify&~=&~helo*&
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying HELO/EHLO"
.cindex "HELO" "verifying"
.cindex "EHLO" "verifying"
.cindex "verifying" "EHLO"
.cindex "verifying" "HELO"
This condition is true if a HELO or EHLO command has been received from the
client host, and its contents have been verified. If there has been no previous
attempt to verify the HELO/EHLO contents, it is carried out when this
condition is encountered. See the description of the &%helo_verify_hosts%& and
&%helo_try_verify_hosts%& options for details of how to request verification
independently of this condition, and for detail of the verification.

For SMTP input that does not come over TCP/IP (the &%-bs%& command line
option), this condition is always true.


.vitem &*verify&~=&~not_blind/*&<&'options'&>
.cindex "verifying" "not blind"
.cindex "bcc recipients, verifying none"
This condition checks that there are no blind (bcc) recipients in the message.
Every envelope recipient must appear either in a &'To:'& header line or in a
&'Cc:'& header line for this condition to be true. Local parts are checked
case-sensitively; domains are checked case-insensitively. If &'Resent-To:'& or
&'Resent-Cc:'& header lines exist, they are also checked. This condition can be
used only in a DATA or non-SMTP ACL.

There is one possible option, &`case_insensitive`&.  If this is present then
local parts are checked case-insensitively.

There are, of course, many legitimate messages that make use of blind (bcc)
recipients. This check should not be used on its own for blocking messages.


.vitem &*verify&~=&~recipient/*&<&'options'&>
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying recipient"
.cindex "recipient" "verifying"
.cindex "verifying" "recipient"
.vindex "&$address_data$&"
This condition is relevant only after a RCPT command. It verifies the current
recipient. Details of address verification are given later, starting at section
&<<SECTaddressverification>>&. After a recipient has been verified, the value
of &$address_data$& is the last value that was set while routing the address.
This applies even if the verification fails. When an address that is being
verified is redirected to a single address, verification continues with the new
address, and in that case, the subsequent value of &$address_data$& is the
value for the child address.

.vitem &*verify&~=&~reverse_host_lookup/*&<&'options'&>
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying host reverse lookup"
.cindex "host" "verifying reverse lookup"
This condition ensures that a verified host name has been looked up from the IP
address of the client host. (This may have happened already if the host name
was needed for checking a host list, or if the host matched &%host_lookup%&.)
Verification ensures that the host name obtained from a reverse DNS lookup, or
one of its aliases, does, when it is itself looked up in the DNS, yield the
original IP address.

There is one possible option, &`defer_ok`&.  If this is present and a
DNS operation returns a temporary error, the verify condition succeeds.

If this condition is used for a locally generated message (that is, when there
is no client host involved), it always succeeds.

.vitem &*verify&~=&~sender/*&<&'options'&>
.cindex "&%verify%& ACL condition"
.cindex "&ACL;" "verifying sender"
.cindex "sender" "verifying"
.cindex "verifying" "sender"
This condition is relevant only after a MAIL or RCPT command, or after a
message has been received (the &%acl_smtp_data%& or &%acl_not_smtp%& ACLs). If
the message's sender is empty (that is, this is a bounce message), the
condition is true. Otherwise, the sender address is verified.

.vindex "&$address_data$&"
.vindex "&$sender_address_data$&"
If there is data in the &$address_data$& variable at the end of routing, its
value is placed in &$sender_address_data$& at the end of verification. This
value can be used in subsequent conditions and modifiers in the same ACL
statement. It does not persist after the end of the current statement. If you
want to preserve the value for longer, you can save it in an ACL variable.

Details of verification are given later, starting at section
&<<SECTaddressverification>>&. Exim caches the result of sender verification,
to avoid doing it more than once per message.

.vitem &*verify&~=&~sender=*&<&'address'&>&*/*&<&'options'&>
.cindex "&%verify%& ACL condition"
This is a variation of the previous option, in which a modified address is
verified as a sender.

Note that '/' is legal in local-parts; if the address may have such
(eg. is generated from the received message)
they must be protected from the options parsing by doubling:
.code
verify = sender=${listquote{/}{${address:$h_sender:}}}
.endd
.endlist



.section "Using DNS lists" "SECTmorednslists"
.cindex "DNS list" "in ACL"
.cindex "black list (DNS)"
.cindex "&ACL;" "testing a DNS list"
In its simplest form, the &%dnslists%& condition tests whether the calling host
is on at least one of a number of DNS lists by looking up the inverted IP
address in one or more DNS domains. (Note that DNS list domains are not mail
domains, so the &`+`& syntax for named lists doesn't work - it is used for
special options instead.) For example, if the calling host's IP
address is 192.168.62.43, and the ACL statement is
.code
deny dnslists = blackholes.mail-abuse.org : \
                dialups.mail-abuse.org
.endd
the following records are looked up:
.code
43.62.168.192.blackholes.mail-abuse.org
43.62.168.192.dialups.mail-abuse.org
.endd
As soon as Exim finds an existing DNS record, processing of the list stops.
Thus, multiple entries on the list provide an &"or"& conjunction. If you want
to test that a host is on more than one list (an &"and"& conjunction), you can
use two separate conditions:
.code
deny dnslists = blackholes.mail-abuse.org
     dnslists = dialups.mail-abuse.org
.endd
If a DNS lookup times out or otherwise fails to give a decisive answer, Exim
behaves as if the host does not match the list item, that is, as if the DNS
record does not exist. If there are further items in the DNS list, they are
processed.

This is usually the required action when &%dnslists%& is used with &%deny%&
(which is the most common usage), because it prevents a DNS failure from
blocking mail. However, you can change this behaviour by putting one of the
following special items in the list:
.itable none 0 0 2 25* left 75* left
.irow "+include_unknown"   "behave as if the item is on the list"
.irow "+exclude_unknown"   "behave as if the item is not on the list (default)"
.irow "+defer_unknown  "   "give a temporary error"
.endtable
.cindex "&`+include_unknown`&"
.cindex "&`+exclude_unknown`&"
.cindex "&`+defer_unknown`&"
Each of these applies to any subsequent items on the list. For example:
.code
deny dnslists = +defer_unknown : foo.bar.example
.endd
Testing the list of domains stops as soon as a match is found. If you want to
warn for one list and block for another, you can use two different statements:
.code
deny  dnslists = blackholes.mail-abuse.org
warn  dnslists = dialups.mail-abuse.org
      message  = X-Warn: sending host is on dialups list
.endd
.cindex caching "of dns lookup"
.cindex DNS TTL
DNS list lookups are cached by Exim for the duration of the SMTP session
(but limited by the DNS return TTL value),
so a lookup based on the IP address is done at most once for any incoming
connection (assuming long-enough TTL).
Exim does not share information between multiple incoming
connections (but your local name server cache should be active).

There are a number of DNS lists to choose from, some commercial, some free,
or free for small deployments.  An overview can be found at
&url(https://en.wikipedia.org/wiki/Comparison_of_DNS_blacklists).



.subsection "Specifying the IP address for a DNS list lookup" SECID201
.cindex "DNS list" "keyed by explicit IP address"
By default, the IP address that is used in a DNS list lookup is the IP address
of the calling host. However, you can specify another IP address by listing it
after the domain name, introduced by a slash. For example:
.code
deny dnslists = black.list.tld/192.168.1.2
.endd
This feature is not very helpful with explicit IP addresses; it is intended for
use with IP addresses that are looked up, for example, the IP addresses of the
MX hosts or nameservers of an email sender address. For an example, see section
&<<SECTmulkeyfor>>& below.




.subsection "DNS lists keyed on domain names" SECID202
.cindex "DNS list" "keyed by domain name"
There are some lists that are keyed on domain names rather than inverted IP
addresses (see, e.g., the &'domain based zones'& link at
&url(http://www.rfc-ignorant.org/)). No reversing of components is used
with these lists. You can change the name that is looked up in a DNS list by
listing it after the domain name, introduced by a slash. For example,
.code
deny  dnslists = dsn.rfc-ignorant.org/$sender_address_domain
      message  = Sender's domain is listed at $dnslist_domain
.endd
This particular example is useful only in ACLs that are obeyed after the
RCPT or DATA commands, when a sender address is available. If (for
example) the message's sender is &'user@tld.example'& the name that is looked
up by this example is
.code
tld.example.dsn.rfc-ignorant.org
.endd
A single &%dnslists%& condition can contain entries for both names and IP
addresses. For example:
.code
deny dnslists = sbl.spamhaus.org : \
                dsn.rfc-ignorant.org/$sender_address_domain
.endd
The first item checks the sending host's IP address; the second checks a domain
name. The whole condition is true if either of the DNS lookups succeeds.




.subsection "Multiple explicit keys for a DNS list" SECTmulkeyfor
.cindex "DNS list" "multiple keys for"
The syntax described above for looking up explicitly-defined values (either
names or IP addresses) in a DNS blacklist is a simplification. After the domain
name for the DNS list, what follows the slash can in fact be a list of items.
As with all lists in Exim, the default separator is a colon. However, because
this is a sublist within the list of DNS blacklist domains, it is necessary
either to double the separators like this:
.code
dnslists = black.list.tld/name.1::name.2
.endd
or to change the separator character, like this:
.code
dnslists = black.list.tld/<;name.1;name.2
.endd
If an item in the list is an IP address, it is inverted before the DNS
blacklist domain is appended. If it is not an IP address, no inversion
occurs. Consider this condition:
.code
dnslists = black.list.tld/<;192.168.1.2;a.domain
.endd
The DNS lookups that occur are:
.code
2.1.168.192.black.list.tld
a.domain.black.list.tld
.endd
Once a DNS record has been found (that matches a specific IP return
address, if specified &-- see section &<<SECTaddmatcon>>&), no further lookups
are done. If there is a temporary DNS error, the rest of the sublist of domains
or IP addresses is tried. A temporary error for the whole dnslists item occurs
only if no other DNS lookup in this sublist succeeds. In other words, a
successful lookup for any of the items in the sublist overrides a temporary
error for a previous item.

The ability to supply a list of items after the slash is in some sense just a
syntactic convenience. These two examples have the same effect:
.code
dnslists = black.list.tld/a.domain : black.list.tld/b.domain
dnslists = black.list.tld/a.domain::b.domain
.endd
However, when the data for the list is obtained from a lookup, the second form
is usually much more convenient. Consider this example:
.code
deny dnslists = sbl.spamhaus.org/<|${lookup dnsdb {>|a=<|\
                                   ${lookup dnsdb {>|mxh=\
                                   $sender_address_domain} }} }
     message  = The mail servers for the domain \
                $sender_address_domain \
                are listed at $dnslist_domain ($dnslist_value); \
                see $dnslist_text.
.endd
Note the use of &`>|`& in the dnsdb lookup to specify the separator for
multiple DNS records. The inner dnsdb lookup produces a list of MX hosts
and the outer dnsdb lookup finds the IP addresses for these hosts. The result
of expanding the condition might be something like this:
.code
dnslists = sbl.spamhaus.org/<|192.168.2.3|192.168.5.6|...
.endd
Thus, this example checks whether or not the IP addresses of the sender
domain's mail servers are on the Spamhaus black list.

The key that was used for a successful DNS list lookup is put into the variable
&$dnslist_matched$& (see section &<<SECID204>>&).




.subsection "Data returned by DNS lists" SECID203
.cindex "DNS list" "data returned from"
DNS lists are constructed using address records in the DNS. The original RBL
just used the address 127.0.0.1 on the right hand side of each record, but the
RBL+ list and some other lists use a number of values with different meanings.
The values used on the RBL+ list are:
.itable none 0 0 2 20* left 80* left
.irow 127.1.0.1  "RBL"
.irow 127.1.0.2  "DUL"
.irow 127.1.0.3  "DUL and RBL"
.irow 127.1.0.4  "RSS"
.irow 127.1.0.5  "RSS and RBL"
.irow 127.1.0.6  "RSS and DUL"
.irow 127.1.0.7  "RSS and DUL and RBL"
.endtable
Section &<<SECTaddmatcon>>& below describes how you can distinguish between
different values. Some DNS lists may return more than one address record;
see section &<<SECThanmuldnsrec>>& for details of how they are checked.

Values returned by a properly running DBSBL should be in the 127.0.0.0/8
range. If a DNSBL operator loses control of the domain, lookups on it
may start returning other addresses.  Because of this, Exim now ignores
returned values outside the 127/8 region.


.subsection "Variables set from DNS lists" SECID204
.cindex "expansion" "variables, set from DNS list"
.cindex "DNS list" "variables set from"
.vindex "&$dnslist_domain$&"
.vindex "&$dnslist_matched$&"
.vindex "&$dnslist_text$&"
.vindex "&$dnslist_value$&"
When an entry is found in a DNS list, the variable &$dnslist_domain$& contains
the name of the overall domain that matched (for example,
&`spamhaus.example`&), &$dnslist_matched$& contains the key within that domain
(for example, &`192.168.5.3`&), and &$dnslist_value$& contains the data from
the DNS record. When the key is an IP address, it is not reversed in
&$dnslist_matched$& (though it is, of course, in the actual lookup). In simple
cases, for example:
.code
deny dnslists = spamhaus.example
.endd
the key is also available in another variable (in this case,
&$sender_host_address$&). In more complicated cases, however, this is not true.
For example, using a data lookup (as described in section &<<SECTmulkeyfor>>&)
might generate a dnslists lookup like this:
.code
deny dnslists = spamhaus.example/<|192.168.1.2|192.168.6.7|...
.endd
If this condition succeeds, the value in &$dnslist_matched$& might be
&`192.168.6.7`& (for example).

If more than one address record is returned by the DNS lookup, all the IP
addresses are included in &$dnslist_value$&, separated by commas and spaces.
The variable &$dnslist_text$& contains the contents of any associated TXT
record. For lists such as RBL+ the TXT record for a merged entry is often not
very meaningful. See section &<<SECTmordetinf>>& for a way of obtaining more
information.

You can use the DNS list variables in &%message%& or &%log_message%& modifiers
&-- even if these appear before the condition in the ACL, they are not
expanded until after it has failed. For example:
.code
deny    hosts = !+local_networks
        message = $sender_host_address is listed \
                  at $dnslist_domain
        dnslists = rbl-plus.mail-abuse.example
.endd



.subsection "Additional matching conditions for DNS lists" SECTaddmatcon
.cindex "DNS list" "matching specific returned data"
You can add an equals sign and an IP address after a &%dnslists%& domain name
in order to restrict its action to DNS records with a matching right hand side.
For example,
.code
deny dnslists = rblplus.mail-abuse.org=127.0.0.2
.endd
rejects only those hosts that yield 127.0.0.2. Without this additional data,
any address record is considered to be a match. For the moment, we assume
that the DNS lookup returns just one record. Section &<<SECThanmuldnsrec>>&
describes how multiple records are handled.

More than one IP address may be given for checking, using a comma as a
separator. These are alternatives &-- if any one of them matches, the
&%dnslists%& condition is true. For example:
.code
deny  dnslists = a.b.c=127.0.0.2,127.0.0.3
.endd
If you want to specify a constraining address list and also specify names or IP
addresses to be looked up, the constraining address list must be specified
first. For example:
.code
deny dnslists = dsn.rfc-ignorant.org\
                =127.0.0.2/$sender_address_domain
.endd

If the character &`&&`& is used instead of &`=`&, the comparison for each
listed IP address is done by a bitwise &"and"& instead of by an equality test.
In other words, the listed addresses are used as bit masks. The comparison is
true if all the bits in the mask are present in the address that is being
tested. For example:
.code
dnslists = a.b.c&0.0.0.3
.endd
matches if the address is &'x.x.x.'&3, &'x.x.x.'&7, &'x.x.x.'&11, etc. If you
want to test whether one bit or another bit is present (as opposed to both
being present), you must use multiple values. For example:
.code
dnslists = a.b.c&0.0.0.1,0.0.0.2
.endd
matches if the final component of the address is an odd number or two times
an odd number.



.subsection "Negated DNS matching conditions" SECID205
You can supply a negative list of IP addresses as part of a &%dnslists%&
condition. Whereas
.code
deny  dnslists = a.b.c=127.0.0.2,127.0.0.3
.endd
means &"deny if the host is in the black list at the domain &'a.b.c'& and the
IP address yielded by the list is either 127.0.0.2 or 127.0.0.3"&,
.code
deny  dnslists = a.b.c!=127.0.0.2,127.0.0.3
.endd
means &"deny if the host is in the black list at the domain &'a.b.c'& and the
IP address yielded by the list is not 127.0.0.2 and not 127.0.0.3"&. In other
words, the result of the test is inverted if an exclamation mark appears before
the &`=`& (or the &`&&`&) sign.

&*Note*&: This kind of negation is not the same as negation in a domain,
host, or address list (which is why the syntax is different).

If you are using just one list, the negation syntax does not gain you much. The
previous example is precisely equivalent to
.code
deny  dnslists = a.b.c
     !dnslists = a.b.c=127.0.0.2,127.0.0.3
.endd
However, if you are using multiple lists, the negation syntax is clearer.
Consider this example:
.code
deny  dnslists = sbl.spamhaus.org : \
                 list.dsbl.org : \
                 dnsbl.njabl.org!=127.0.0.3 : \
                 relays.ordb.org
.endd
Using only positive lists, this would have to be:
.code
deny  dnslists = sbl.spamhaus.org : \
                 list.dsbl.org
deny  dnslists = dnsbl.njabl.org
     !dnslists = dnsbl.njabl.org=127.0.0.3
deny  dnslists = relays.ordb.org
.endd
which is less clear, and harder to maintain.

Negation can also be used with a bitwise-and restriction.
The dnslists condition with only be trus if a result is returned
by the lookup which, anded with the restriction, is all zeroes.
For example:
.code
deny dnslists = zen.spamhaus.org!&0.255.255.0
.endd




.subsection "Handling multiple DNS records from a DNS list" SECThanmuldnsrec
A DNS lookup for a &%dnslists%& condition may return more than one DNS record,
thereby providing more than one IP address. When an item in a &%dnslists%& list
is followed by &`=`& or &`&&`& and a list of IP addresses, in order to restrict
the match to specific results from the DNS lookup, there are two ways in which
the checking can be handled. For example, consider the condition:
.code
dnslists = a.b.c=127.0.0.1
.endd
What happens if the DNS lookup for the incoming IP address yields both
127.0.0.1 and 127.0.0.2 by means of two separate DNS records? Is the
condition true because at least one given value was found, or is it false
because at least one of the found values was not listed? And how does this
affect negated conditions? Both possibilities are provided for with the help of
additional separators &`==`& and &`=&&`&.

.ilist
If &`=`& or &`&&`& is used, the condition is true if any one of the looked up
IP addresses matches one of the listed addresses. For the example above, the
condition is true because 127.0.0.1 matches.
.next
If &`==`& or &`=&&`& is used, the condition is true only if every one of the
looked up IP addresses matches one of the listed addresses. If the condition is
changed to:
.code
dnslists = a.b.c==127.0.0.1
.endd
and the DNS lookup yields both 127.0.0.1 and 127.0.0.2, the condition is
false because 127.0.0.2 is not listed. You would need to have:
.code
dnslists = a.b.c==127.0.0.1,127.0.0.2
.endd
for the condition to be true.
.endlist

When &`!`& is used to negate IP address matching, it inverts the result, giving
the precise opposite of the behaviour above. Thus:
.ilist
If &`!=`& or &`!&&`& is used, the condition is true if none of the looked up IP
addresses matches one of the listed addresses. Consider:
.code
dnslists = a.b.c!&0.0.0.1
.endd
If the DNS lookup yields both 127.0.0.1 and 127.0.0.2, the condition is
false because 127.0.0.1 matches.
.next
If &`!==`& or &`!=&&`& is used, the condition is true if there is at least one
looked up IP address that does not match. Consider:
.code
dnslists = a.b.c!=&0.0.0.1
.endd
If the DNS lookup yields both 127.0.0.1 and 127.0.0.2, the condition is
true, because 127.0.0.2 does not match. You would need to have:
.code
dnslists = a.b.c!=&0.0.0.1,0.0.0.2
.endd
for the condition to be false.
.endlist
When the DNS lookup yields only a single IP address, there is no difference
between &`=`& and &`==`& and between &`&&`& and &`=&&`&.




.subsection "Detailed information from merged DNS lists" SECTmordetinf
.cindex "DNS list" "information from merged"
When the facility for restricting the matching IP values in a DNS list is used,
the text from the TXT record that is set in &$dnslist_text$& may not reflect
the true reason for rejection. This happens when lists are merged and the IP
address in the A record is used to distinguish them; unfortunately there is
only one TXT record. One way round this is not to use merged lists, but that
can be inefficient because it requires multiple DNS lookups where one would do
in the vast majority of cases when the host of interest is not on any of the
lists.

A less inefficient way of solving this problem is available. If
two domain names, comma-separated, are given, the second is used first to
do an initial check, making use of any IP value restrictions that are set.
If there is a match, the first domain is used, without any IP value
restrictions, to get the TXT record. As a byproduct of this, there is also
a check that the IP being tested is indeed on the first list. The first
domain is the one that is put in &$dnslist_domain$&. For example:
.code
deny   dnslists = \
         sbl.spamhaus.org,sbl-xbl.spamhaus.org=127.0.0.2 : \
         dul.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.10
       message  = \
         rejected because $sender_host_address is blacklisted \
         at $dnslist_domain\n$dnslist_text
.endd
For the first blacklist item, this starts by doing a lookup in
&'sbl-xbl.spamhaus.org'& and testing for a 127.0.0.2 return. If there is a
match, it then looks in &'sbl.spamhaus.org'&, without checking the return
value, and as long as something is found, it looks for the corresponding TXT
record. If there is no match in &'sbl-xbl.spamhaus.org'&, nothing more is done.
The second blacklist item is processed similarly.

If you are interested in more than one merged list, the same list must be
given several times, but because the results of the DNS lookups are cached,
the DNS calls themselves are not repeated. For example:
.code
deny dnslists = \
         http.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.2 : \
         socks.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.3 : \
         misc.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.4 : \
         dul.dnsbl.sorbs.net,dnsbl.sorbs.net=127.0.0.10
.endd
In this case there is one lookup in &'dnsbl.sorbs.net'&, and if none of the IP
values matches (or if no record is found), this is the only lookup that is
done. Only if there is a match is one of the more specific lists consulted.



.subsection "DNS lists and IPv6" SECTmorednslistslast
.cindex "IPv6" "DNS black lists"
.cindex "DNS list" "IPv6 usage"
If Exim is asked to do a dnslist lookup for an IPv6 address, it inverts it
nibble by nibble. For example, if the calling host's IP address is
3ffe:ffff:836f:0a00:000a:0800:200a:c031, Exim might look up
.code
1.3.0.c.a.0.0.2.0.0.8.0.a.0.0.0.0.0.a.0.f.6.3.8.
  f.f.f.f.e.f.f.3.blackholes.mail-abuse.org
.endd
(split over two lines here to fit on the page). Unfortunately, some of the DNS
lists contain wildcard records, intended for IPv4, that interact badly with
IPv6. For example, the DNS entry
.code
*.3.some.list.example.    A    127.0.0.1
.endd
is probably intended to put the entire 3.0.0.0/8 IPv4 network on the list.
Unfortunately, it also matches the entire 3::/4 IPv6 network.

You can exclude IPv6 addresses from DNS lookups by making use of a suitable
&%condition%& condition, as in this example:
.code
deny   condition = ${if isip4{$sender_host_address}}
       dnslists  = some.list.example
.endd

If an explicit key is being used for a DNS lookup and it may be an IPv6
address you should specify alternate list separators for both the outer
(DNS list name) list and inner (lookup keys) list:
.code
       dnslists = <; dnsbl.example.com/<|$acl_m_addrslist
.endd


.section "Previously seen user and hosts" "SECTseen"
.cindex "&%seen%& ACL condition"
.cindex greylisting
The &%seen%& ACL condition can be used to test whether a
situation has been previously met.
It uses a hints database to record a timestamp against a key.
The syntax of the condition is:
.display
&`seen =`& <&'optional flag'&><&'time interval'&> &`/`& <&'options'&>
.endd

For example,
.code
defer   seen = -5m / key=${sender_host_address}_$local_part@$domain
.endd
in a RCPT ACL will implement simple greylisting.

The parameters for the condition are
a possible minus sign,
then an interval,
then, slash-separated, a list of options.
The interval is taken as an offset before the current time,
and used for the test.
If the interval is preceded by a minus sign then the condition returns
whether a record is found which is before the test time.
Otherwise, the condition returns whether one is found which is since the
test time.

Options are read in order with later ones overriding earlier ones.

The default key is &$sender_host_address$&.
An explicit key can be set using a &%key=value%& option.

If a &%readonly%& option is given then
no record create or update is done.
If a &%write%& option is given then
a record create or update is always done.
An update is done if the test is for &"since"&.
If none of those hold and there was no existing record,
a record is created.

Creates and updates are marked with the current time.

Finally, a &"before"& test which succeeds, and for which the record
is old enough, will be refreshed with a timestamp of the test time.
This can prevent tidying of the database from removing the entry.
The interval for this is, by default, 10 days.
An explicit interval can be set using a
&%refresh=value%& option.

Note that &"seen"& should be added to the list of hints databases
for maintenance if this ACL condition is used.


.section "Rate limiting incoming messages" "SECTratelimiting"
.cindex "rate limiting" "client sending"
.cindex "limiting client sending rates"
.oindex "&%smtp_ratelimit_*%&"
The &%ratelimit%& ACL condition can be used to measure and control the rate at
which clients can send email. This is more powerful than the
&%smtp_ratelimit_*%& options, because those options control the rate of
commands in a single SMTP session only, whereas the &%ratelimit%& condition
works across all connections (concurrent and sequential) from the same client
host. The syntax of the &%ratelimit%& condition is:
.display
&`ratelimit =`& <&'m'&> &`/`& <&'p'&> &`/`& <&'options'&> &`/`& <&'key'&>
.endd
If the average client sending rate is less than &'m'& messages per time
period &'p'& then the condition is false; otherwise it is true.

As a side-effect, the &%ratelimit%& condition sets the expansion variable
&$sender_rate$& to the client's computed rate, &$sender_rate_limit$& to the
configured value of &'m'&, and &$sender_rate_period$& to the configured value
of &'p'&.

The parameter &'p'& is the smoothing time constant, in the form of an Exim
time interval, for example, &`8h`& for eight hours. A larger time constant
means that it takes Exim longer to forget a client's past behaviour. The
parameter &'m'& is the maximum number of messages that a client is permitted to
send in each time interval. It also specifies the number of messages permitted
in a fast burst. By increasing both &'m'& and &'p'& but keeping &'m/p'&
constant, you can allow a client to send more messages in a burst without
changing its long-term sending rate limit. Conversely, if &'m'& and &'p'& are
both small, messages must be sent at an even rate.

There is a script in &_util/ratelimit.pl_& which extracts sending rates from
log files, to assist with choosing appropriate settings for &'m'& and &'p'&
when deploying the &%ratelimit%& ACL condition. The script prints usage
instructions when it is run with no arguments.

The key is used to look up the data for calculating the client's average
sending rate. This data is stored in Exim's spool directory, alongside the
retry and other hints databases. The default key is &$sender_host_address$&,
which means Exim computes the sending rate of each client host IP address.
By changing the key you can change how Exim identifies clients for the purpose
of ratelimiting. For example, to limit the sending rate of each authenticated
user, independent of the computer they are sending from, set the key to
&$authenticated_id$&. You must ensure that the lookup key is meaningful; for
example, &$authenticated_id$& is only meaningful if the client has
authenticated (which you can check with the &%authenticated%& ACL condition).

The lookup key does not have to identify clients: If you want to limit the
rate at which a recipient receives messages, you can use the key
&`$local_part@$domain`& with the &%per_rcpt%& option (see below) in a RCPT
ACL.

Each &%ratelimit%& condition can have up to four options. A &%per_*%& option
specifies what Exim measures the rate of, for example, messages or recipients
or bytes. You can adjust the measurement using the &%unique=%& and/or
&%count=%& options. You can also control when Exim updates the recorded rate
using a &%strict%&, &%leaky%&, or &%readonly%& option. The options are
separated by a slash, like the other parameters. They may appear in any order.

Internally, Exim appends the smoothing constant &'p'& onto the lookup key with
any options that alter the meaning of the stored data. The limit &'m'& is not
stored, so you can alter the configured maximum rate and Exim will still
remember clients' past behaviour. If you change the &%per_*%& mode or add or
remove the &%unique=%& option, the lookup key changes so Exim will forget past
behaviour. The lookup key is not affected by changes to the update mode and
the &%count=%& option.


.subsection "Ratelimit options for what is being measured" ratoptmea
.cindex "rate limiting" "per_* options"
.vlist
.vitem per_conn
.cindex "rate limiting" per_conn
This option limits the client's connection rate. It is not
normally used in the &%acl_not_smtp%&, &%acl_not_smtp_mime%&, or
&%acl_not_smtp_start%& ACLs.

.vitem per_mail
.cindex "rate limiting" per_conn
This option limits the client's rate of sending messages. This is
the default if none of the &%per_*%& options is specified. It can be used in
&%acl_smtp_mail%&, &%acl_smtp_rcpt%&, &%acl_smtp_predata%&, &%acl_smtp_mime%&,
&%acl_smtp_data%&, or &%acl_not_smtp%&.

.vitem per_byte
.cindex "rate limiting" per_conn
This option limits the sender's email bandwidth. It can be used in
the same ACLs as the &%per_mail%& option, though it is best to use this option
in the &%acl_smtp_mime%&, &%acl_smtp_data%& or &%acl_not_smtp%& ACLs; if it is
used in an earlier ACL, Exim relies on the SIZE parameter given by the client
in its MAIL command, which may be inaccurate or completely missing. You can
follow the limit &'m'& in the configuration with K, M, or G to specify limits
in kilobytes, megabytes, or gigabytes, respectively.

.vitem per_rcpt
.cindex "rate limiting" per_rcpt
This option causes Exim to limit the rate at which recipients are
accepted. It can be used in the &%acl_smtp_rcpt%&, &%acl_smtp_predata%&,
&%acl_smtp_mime%&, or &%acl_smtp_data%& ACLs. In
&%acl_smtp_rcpt%& the rate is updated one recipient at a time; in the other
ACLs the rate is updated with the total (accepted) recipient count in one go. Note that
in either case the rate limiting engine will see a message with many
recipients as a large high-speed burst.

.vitem per_addr
.cindex "rate limiting" per_addr
This option is like the &%per_rcpt%& option, except it counts the
number of different recipients that the client has sent messages to in the
last time period. That is, if the client repeatedly sends messages to the same
recipient, its measured rate is not increased. This option can only be used in
&%acl_smtp_rcpt%&.

.vitem per_cmd
.cindex "rate limiting" per_cmd
This option causes Exim to recompute the rate every time the
condition is processed. This can be used to limit the rate of any SMTP
command. If it is used in multiple ACLs it can limit the aggregate rate of
multiple different commands.

.vitem count
.cindex "rate limiting" count
This option can be used to alter how much Exim adds to the client's
measured rate.
A value is required, after an equals sign.
For example, the &%per_byte%& option is equivalent to
&`per_mail/count=$message_size`&.
If there is no &%count=%& option, Exim
increases the measured rate by one (except for the &%per_rcpt%& option in ACLs
other than &%acl_smtp_rcpt%&).
The count does not have to be an integer.

.vitem unique
.cindex "rate limiting" unique
This option is described in section &<<ratoptuniq>>& below.
.endlist


.subsection "Ratelimit update modes" ratoptupd
.cindex "rate limiting" "reading data without updating"
You can specify one of three options with the &%ratelimit%& condition to
control when its database is updated. This section describes the &%readonly%&
mode, and the next section describes the &%strict%& and &%leaky%& modes.

If the &%ratelimit%& condition is used in &%readonly%& mode, Exim looks up a
previously-computed rate to check against the limit.

For example, you can test the client's sending rate and deny it access (when
it is too fast) in the connect ACL. If the client passes this check then it
can go on to send a message, in which case its recorded rate will be updated
in the MAIL ACL. Subsequent connections from the same client will check this
new rate.
.code
acl_check_connect:
 deny ratelimit = 100 / 5m / readonly
      log_message = RATE CHECK: $sender_rate/$sender_rate_period \
                  (max $sender_rate_limit)
# ...
acl_check_mail:
 warn ratelimit = 100 / 5m / strict
      log_message = RATE UPDATE: $sender_rate/$sender_rate_period \
                  (max $sender_rate_limit)
.endd

If Exim encounters multiple &%ratelimit%& conditions with the same key when
processing a message then it may increase the client's measured rate more than
it should. For example, this will happen if you check the &%per_rcpt%& option
in both &%acl_smtp_rcpt%& and &%acl_smtp_data%&. However it's OK to check the
same &%ratelimit%& condition multiple times in the same ACL. You can avoid any
multiple update problems by using the &%readonly%& option on later ratelimit
checks.

The &%per_*%& options described above do not make sense in some ACLs. If you
use a &%per_*%& option in an ACL where it is not normally permitted then the
update mode defaults to &%readonly%& and you cannot specify the &%strict%& or
&%leaky%& modes. In other ACLs the default update mode is &%leaky%& (see the
next section) so you must specify the &%readonly%& option explicitly.


.subsection "Ratelimit options for handling fast clients" ratoptfast
.cindex "rate limiting" "strict and leaky modes"
If a client's average rate is greater than the maximum, the rate limiting
engine can react in two possible ways, depending on the presence of the
&%strict%& or &%leaky%& update modes. This is independent of the other
counter-measures (such as rejecting the message) that may be specified by the
rest of the ACL.

The &%leaky%& (default) option means that the client's recorded rate is not
updated if it is above the limit. The effect of this is that Exim measures the
client's average rate of successfully sent email,
up to the given limit.
This is appropriate if the countermeasure when the condition is true
consists of refusing the message, and
is generally the better choice if you have clients that retry automatically.
If the action when true is anything more complex then this option is
likely not what is wanted.

The &%strict%& option means that the client's recorded rate is always
updated. The effect of this is that Exim measures the client's average rate
of attempts to send email, which can be much higher than the maximum it is
actually allowed. If the client is over the limit it may be subjected to
counter-measures by the ACL. It must slow down and allow sufficient time to
pass that its computed rate falls below the maximum before it can send email
again. The time (the number of smoothing periods) it must wait and not
attempt to send mail can be calculated with this formula:
.code
        ln(peakrate/maxrate)
.endd


.subsection "Limiting the rate of different events" ratoptuniq
.cindex "rate limiting" "counting unique events"
The &%ratelimit%& &%unique=%& option controls a mechanism for counting the
rate of different events. For example, the &%per_addr%& option uses this
mechanism to count the number of different recipients that the client has
sent messages to in the last time period; it is equivalent to
&`per_rcpt/unique=$local_part@$domain`&. You could use this feature to
measure the rate that a client uses different sender addresses with the
options &`per_mail/unique=$sender_address`&.

For each &%ratelimit%& key Exim stores the set of &%unique=%& values that it
has seen for that key. The whole set is thrown away when it is older than the
rate smoothing period &'p'&, so each different event is counted at most once
per period. In the &%leaky%& update mode, an event that causes the client to
go over the limit is not added to the set, in the same way that the client's
recorded rate is not updated in the same situation.

When you combine the &%unique=%& and &%readonly%& options, the specific
&%unique=%& value is ignored, and Exim just retrieves the client's stored
rate.

The &%unique=%& mechanism needs more space in the ratelimit database than the
other &%ratelimit%& options in order to store the event set. The number of
unique values is potentially as large as the rate limit, so the extra space
required increases with larger limits.

The uniqueification is not perfect: there is a small probability that Exim
will think a new event has happened before. If the sender's rate is less than
the limit, Exim should be more than 99.9% correct. However in &%strict%& mode
the measured rate can go above the limit, in which case Exim may under-count
events by a significant margin. Fortunately, if the rate is high enough (2.7
times the limit) that the false positive rate goes above 9%, then Exim will
throw away the over-full event set before the measured rate falls below the
limit. Therefore the only harm should be that exceptionally high sending rates
are logged incorrectly; any countermeasures you configure will be as effective
as intended.


.subsection "Using rate limiting" useratlim
Exim's other ACL facilities are used to define what counter-measures are taken
when the rate limit is exceeded. This might be anything from logging a warning
(for example, while measuring existing sending rates in order to define
policy), through time delays to slow down fast senders, up to rejecting the
message. For example:
.code
# Log all senders' rates
warn ratelimit = 0 / 1h / strict
     log_message = Sender rate $sender_rate / $sender_rate_period

# Slow down fast senders; note the need to truncate $sender_rate
# at the decimal point.
warn ratelimit = 100 / 1h / per_rcpt / strict
     delay     = ${eval: ${sg{$sender_rate}{[.].*}{}} - \
                   $sender_rate_limit }s

# Keep authenticated users under control
deny authenticated = *
     ratelimit = 100 / 1d / strict / $authenticated_id

# System-wide rate limit
defer ratelimit = 10 / 1s / $primary_hostname
      message = Sorry, too busy. Try again later.

# Restrict incoming rate from each host, with a default
# set using a macro and special cases looked up in a table.
defer ratelimit = ${lookup {$sender_host_address} \
                   cdb {DB/ratelimits.cdb} \
                   {$value} {RATELIMIT} }
      message = Sender rate exceeds $sender_rate_limit \
               messages per $sender_rate_period
.endd
&*Warning*&: If you have a busy server with a lot of &%ratelimit%& tests,
especially with the &%per_rcpt%& option, you may suffer from a performance
bottleneck caused by locking on the ratelimit hints database. Apart from
making your ACLs less complicated, you can reduce the problem by using a
RAM disk for Exim's hints directory (usually &_/var/spool/exim/db/_&). However
this means that Exim will lose its hints data after a reboot (including retry
hints, the callout cache, and ratelimit data).



.section "Address verification" "SECTaddressverification"
.cindex "verifying address" "options for"
.cindex "policy control" "address verification"
Several of the &%verify%& conditions described in section
&<<SECTaclconditions>>& cause addresses to be verified. Section
&<<SECTsenaddver>>& discusses the reporting of sender verification failures.
The verification conditions can be followed by options that modify the
verification process. The options are separated from the keyword and from each
other by slashes, and some of them contain parameters. For example:
.code
verify = sender/callout
verify = recipient/defer_ok/callout=10s,defer_ok
.endd
The first stage of address verification, which always happens, is to run the
address through the routers, in &"verify mode"&. Routers can detect the
difference between verification and routing for delivery, and their actions can
be varied by a number of generic options such as &%verify%& and &%verify_only%&
(see chapter &<<CHAProutergeneric>>&). If routing fails, verification fails.
The available options are as follows:

.ilist
If the &%callout%& option is specified, successful routing to one or more
remote hosts is followed by a &"callout"& to those hosts as an additional
check. Callouts and their sub-options are discussed in the next section.
.next
If there is a defer error while doing verification routing, the ACL
normally returns &"defer"&. However, if you include &%defer_ok%& in the
options, the condition is forced to be true instead. Note that this is a main
verification option as well as a suboption for callouts.
.next
The &%no_details%& option is covered in section &<<SECTsenaddver>>&, which
discusses the reporting of sender address verification failures.
.next
The &%success_on_redirect%& option causes verification always to succeed
immediately after a successful redirection. By default, if a redirection
generates just one address, that address is also verified. See further
discussion in section &<<SECTredirwhilveri>>&.
.next
If the &%quota%& option is specified for recipient verify,
successful routing to an appendfile transport is followed by a call into
the transport to evaluate the quota status for the recipient.
No actual delivery is done, but verification will succeed if the quota
is sufficient for the message (if the sender gave a message size) or
not already exceeded (otherwise).
.endlist

.cindex "verifying address" "differentiating failures"
.vindex "&$recipient_verify_failure$&"
.vindex "&$sender_verify_failure$&"
.vindex "&$acl_verify_message$&"
After an address verification failure, &$acl_verify_message$& contains the
error message that is associated with the failure. It can be preserved by
coding like this:
.code
warn  !verify = sender
       set acl_m0 = $acl_verify_message
.endd
If you are writing your own custom rejection message or log message when
denying access, you can use this variable to include information about the
verification failure.
This variable is cleared at the end of processing the ACL verb.

In addition, &$sender_verify_failure$& or &$recipient_verify_failure$& (as
appropriate) contains one of the following words:

.ilist
&%qualify%&: The address was unqualified (no domain), and the message
was neither local nor came from an exempted host.
.next
&%route%&: Routing failed.
.next
&%mail%&: Routing succeeded, and a callout was attempted; rejection
occurred at or before the MAIL command (that is, on initial
connection, HELO, or MAIL).
.next
&%recipient%&: The RCPT command in a callout was rejected.
.next
&%postmaster%&: The postmaster check in a callout was rejected.
.next
&%quota%&: The quota check for a local recipient did non pass.
.endlist

The main use of these variables is expected to be to distinguish between
rejections of MAIL and rejections of RCPT in callouts.

The above variables may also be set after a &*successful*&
address verification to:

.ilist
&%random%&: A random local-part callout succeeded
.endlist




.section "Callout verification" "SECTcallver"
.cindex "verifying address" "by callout"
.cindex "callout" "verification"
.cindex "SMTP" "callout verification"
For non-local addresses, routing verifies the domain, but is unable to do any
checking of the local part. There are situations where some means of verifying
the local part is desirable. One way this can be done is to make an SMTP
&'callback'& to a delivery host for the sender address or a &'callforward'& to
a subsequent host for a recipient address, to see if the host accepts the
address. We use the term &'callout'& to cover both cases. Note that for a
sender address, the callback is not to the client host that is trying to
deliver the message, but to one of the hosts that accepts incoming mail for the
sender's domain.

Exim does not do callouts by default. If you want them to happen, you must
request them by setting appropriate options on the &%verify%& condition, as
described below. This facility should be used with care, because it can add a
lot of resource usage to the cost of verifying an address. However, Exim does
cache the results of callouts, which helps to reduce the cost. Details of
caching are in section &<<SECTcallvercache>>&.

Recipient callouts are usually used only between hosts that are controlled by
the same administration. For example, a corporate gateway host could use
callouts to check for valid recipients on an internal mailserver. A successful
callout does not guarantee that a real delivery to the address would succeed;
on the other hand, a failing callout does guarantee that a delivery would fail.

If the &%callout%& option is present on a condition that verifies an address, a
second stage of verification occurs if the address is successfully routed to
one or more remote hosts. The usual case is routing by a &(dnslookup)& or a
&(manualroute)& router, where the router specifies the hosts. However, if a
router that does not set up hosts routes to an &(smtp)& transport with a
&%hosts%& setting, the transport's hosts are used. If an &(smtp)& transport has
&%hosts_override%& set, its hosts are always used, whether or not the router
supplies a host list.
Callouts are only supported on &(smtp)& transports.

The port that is used is taken from the transport, if it is specified and is a
remote transport. (For routers that do verification only, no transport need be
specified.) Otherwise, the default SMTP port is used. If a remote transport
specifies an outgoing interface, this is used; otherwise the interface is not
specified. Likewise, the text that is used for the HELO command is taken from
the transport's &%helo_data%& option; if there is no transport, the value of
&$smtp_active_hostname$& is used.

For a sender callout check, Exim makes SMTP connections to the remote hosts, to
test whether a bounce message could be delivered to the sender address. The
following SMTP commands are sent:
.display
&`HELO `&<&'local host name'&>
&`MAIL FROM:<>`&
&`RCPT TO:`&<&'the address to be tested'&>
&`QUIT`&
.endd
LHLO is used instead of HELO if the transport's &%protocol%& option is
set to &"lmtp"&.

The callout may use EHLO, AUTH and/or STARTTLS given appropriate option
settings.

A recipient callout check is similar. By default, it also uses an empty address
for the sender. This default is chosen because most hosts do not make use of
the sender address when verifying a recipient. Using the same address means
that a single cache entry can be used for each recipient. Some sites, however,
do make use of the sender address when verifying. These are catered for by the
&%use_sender%& and &%use_postmaster%& options, described in the next section.

If the response to the RCPT command is a 2&'xx'& code, the verification
succeeds. If it is 5&'xx'&, the verification fails. For any other condition,
Exim tries the next host, if any. If there is a problem with all the remote
hosts, the ACL yields &"defer"&, unless the &%defer_ok%& parameter of the
&%callout%& option is given, in which case the condition is forced to succeed.

.cindex "SMTP" "output flushing, disabling for callout"
A callout may take a little time. For this reason, Exim normally flushes SMTP
output before performing a callout in an ACL, to avoid unexpected timeouts in
clients when the SMTP PIPELINING extension is in use. The flushing can be
disabled by using a &%control%& modifier to set &%no_callout_flush%&.

.cindex "tainted data" "de-tainting"
.cindex "de-tainting" "using recipient verify"
A recipient callout which gets a 2&'xx'& code
will assign untainted values to the
&$domain_data$& and &$local_part_data$& variables,
corresponding to the domain and local parts of the recipient address.




.subsection "Additional parameters for callouts" CALLaddparcall
.cindex "callout" "additional parameters for"
The &%callout%& option can be followed by an equals sign and a number of
optional parameters, separated by commas. For example:
.code
verify = recipient/callout=10s,defer_ok
.endd
The old syntax, which had &%callout_defer_ok%& and &%check_postmaster%& as
separate verify options, is retained for backwards compatibility, but is now
deprecated. The additional parameters for &%callout%& are as follows:


.vlist
.vitem <&'a&~time&~interval'&>
.cindex "callout" "timeout, specifying"
This specifies the timeout that applies for the callout attempt to each host.
For example:
.code
verify = sender/callout=5s
.endd
The default is 30 seconds. The timeout is used for each response from the
remote host. It is also used for the initial connection, unless overridden by
the &%connect%& parameter.


.vitem &*connect&~=&~*&<&'time&~interval'&>
.cindex "callout" "connection timeout, specifying"
This parameter makes it possible to set a different (usually smaller) timeout
for making the SMTP connection. For example:
.code
verify = sender/callout=5s,connect=1s
.endd
If not specified, this timeout defaults to the general timeout value.

.vitem &*defer_ok*&
.cindex "callout" "defer, action on"
When this parameter is present, failure to contact any host, or any other kind
of temporary error, is treated as success by the ACL. However, the cache is not
updated in this circumstance.

.vitem &*fullpostmaster*&
.cindex "callout" "full postmaster check"
This operates like the &%postmaster%& option (see below), but if the check for
&'postmaster@domain'& fails, it tries just &'postmaster'&, without a domain, in
accordance with the specification in
&url(https://www.rfc-editor.org/rfc/rfc2821,RFC 2821).
The RFC states that the unqualified address &'postmaster'& should be accepted.


.vitem &*mailfrom&~=&~*&<&'email&~address'&>
.cindex "callout" "sender when verifying header"
When verifying addresses in header lines using the &%header_sender%&
verification option, Exim behaves by default as if the addresses are envelope
sender addresses from a message. Callout verification therefore tests to see
whether a bounce message could be delivered, by using an empty address in the
MAIL command. However, it is arguable that these addresses might never be used
as envelope senders, and could therefore justifiably reject bounce messages
(empty senders). The &%mailfrom%& callout parameter allows you to specify what
address to use in the MAIL command. For example:
.code
require  verify = header_sender/callout=mailfrom=abcd@x.y.z
.endd
This parameter is available only for the &%header_sender%& verification option.


.vitem &*maxwait&~=&~*&<&'time&~interval'&>
.cindex "callout" "overall timeout, specifying"
This parameter sets an overall timeout for performing a callout verification.
For example:
.code
verify = sender/callout=5s,maxwait=30s
.endd
This timeout defaults to four times the callout timeout for individual SMTP
commands. The overall timeout applies when there is more than one host that can
be tried. The timeout is checked before trying the next host. This prevents
very long delays if there are a large number of hosts and all are timing out
(for example, when network connections are timing out).


.vitem &*no_cache*&
.cindex "callout" "cache, suppressing"
.cindex "caching callout, suppressing"
When this parameter is given, the callout cache is neither read nor updated.

.vitem &*postmaster*&
.cindex "callout" "postmaster; checking"
When this parameter is set, a successful callout check is followed by a similar
check for the local part &'postmaster'& at the same domain. If this address is
rejected, the callout fails (but see &%fullpostmaster%& above). The result of
the postmaster check is recorded in a cache record; if it is a failure, this is
used to fail subsequent callouts for the domain without a connection being
made, until the cache record expires.

.vitem &*postmaster_mailfrom&~=&~*&<&'email&~address'&>
The postmaster check uses an empty sender in the MAIL command by default.
You can use this parameter to do a postmaster check using a different address.
For example:
.code
require  verify = sender/callout=postmaster_mailfrom=abc@x.y.z
.endd
If both &%postmaster%& and &%postmaster_mailfrom%& are present, the rightmost
one overrides. The &%postmaster%& parameter is equivalent to this example:
.code
require  verify = sender/callout=postmaster_mailfrom=
.endd
&*Warning*&: The caching arrangements for postmaster checking do not take
account of the sender address. It is assumed that either the empty address or
a fixed non-empty address will be used. All that Exim remembers is that the
postmaster check for the domain succeeded or failed.


.vitem &*random*&
.cindex "callout" "&""random""& check"
When this parameter is set, before doing the normal callout check, Exim does a
check for a &"random"& local part at the same domain. The local part is not
really random &-- it is defined by the expansion of the option
&%callout_random_local_part%&, which defaults to
.code
$primary_hostname-$tod_epoch-testing
.endd
The idea here is to try to determine whether the remote host accepts all local
parts without checking. If it does, there is no point in doing callouts for
specific local parts. If the &"random"& check succeeds, the result is saved in
a cache record, and used to force the current and subsequent callout checks to
succeed without a connection being made, until the cache record expires.

.vitem &*use_postmaster*&
.cindex "callout" "sender for recipient check"
This parameter applies to recipient callouts only. For example:
.code
deny  !verify = recipient/callout=use_postmaster
.endd
.vindex "&$qualify_domain$&"
It causes a non-empty postmaster address to be used in the MAIL command when
performing the callout for the recipient, and also for a &"random"& check if
that is configured. The local part of the address is &`postmaster`& and the
domain is the contents of &$qualify_domain$&.

.vitem &*use_sender*&
This option applies to recipient callouts only. For example:
.code
require  verify = recipient/callout=use_sender
.endd
It causes the message's actual sender address to be used in the MAIL
command when performing the callout, instead of an empty address. There is no
need to use this option unless you know that the called hosts make use of the
sender when checking recipients. If used indiscriminately, it reduces the
usefulness of callout caching.

.vitem &*hold*&
This option applies to recipient callouts only. For example:
.code
require  verify = recipient/callout=use_sender,hold
.endd
It causes the connection to be held open and used for any further recipients
and for eventual delivery (should that be done quickly).
Doing this saves on TCP and SMTP startup costs, and TLS costs also
when that is used for the connections.
The advantage is only gained if there are no callout cache hits
(which could be enforced by the no_cache option),
if the use_sender option is used,
if neither the random nor the use_postmaster option is used,
and if no other callouts intervene.
.endlist

If you use any of the parameters that set a non-empty sender for the MAIL
command (&%mailfrom%&, &%postmaster_mailfrom%&, &%use_postmaster%&, or
&%use_sender%&), you should think about possible loops. Recipient checking is
usually done between two hosts that are under the same management, and the host
that receives the callouts is not normally configured to do callouts itself.
Therefore, it is normally safe to use &%use_postmaster%& or &%use_sender%& in
these circumstances.

However, if you use a non-empty sender address for a callout to an arbitrary
host, there is the likelihood that the remote host will itself initiate a
callout check back to your host. As it is checking what appears to be a message
sender, it is likely to use an empty address in MAIL, thus avoiding a
callout loop. However, to be on the safe side it would be best to set up your
own ACLs so that they do not do sender verification checks when the recipient
is the address you use for header sender or postmaster callout checking.

Another issue to think about when using non-empty senders for callouts is
caching. When you set &%mailfrom%& or &%use_sender%&, the cache record is keyed
by the sender/recipient combination; thus, for any given recipient, many more
actual callouts are performed than when an empty sender or postmaster is used.




.subsection "Callout caching" SECTcallvercache
.cindex "hints database" "callout cache"
.cindex "callout" "cache, description of"
.cindex "caching" "callout"
Exim caches the results of callouts in order to reduce the amount of resources
used, unless you specify the &%no_cache%& parameter with the &%callout%&
option. A hints database called &"callout"& is used for the cache. Two
different record types are used: one records the result of a callout check for
a specific address, and the other records information that applies to the
entire domain (for example, that it accepts the local part &'postmaster'&).

When an original callout fails, a detailed SMTP error message is given about
the failure. However, for subsequent failures that use the cache data, this message
is not available.

The expiry times for negative and positive address cache records are
independent, and can be set by the global options &%callout_negative_expire%&
(default 2h) and &%callout_positive_expire%& (default 24h), respectively.

If a host gives a negative response to an SMTP connection, or rejects any
commands up to and including
.code
MAIL FROM:<>
.endd
(but not including the MAIL command with a non-empty address),
any callout attempt is bound to fail. Exim remembers such failures in a
domain cache record, which it uses to fail callouts for the domain without
making new connections, until the domain record times out. There are two
separate expiry times for domain cache records:
&%callout_domain_negative_expire%& (default 3h) and
&%callout_domain_positive_expire%& (default 7d).

Domain records expire when the negative expiry time is reached if callouts
cannot be made for the domain, or if the postmaster check failed.
Otherwise, they expire when the positive expiry time is reached. This
ensures that, for example, a host that stops accepting &"random"& local parts
will eventually be noticed.

The callout caching mechanism is based on the domain of the address that is
being tested. If the domain routes to several hosts, it is assumed that their
behaviour will be the same.



.section "Quota caching" "SECTquotacache"
.cindex "hints database" "quota cache"
.cindex "quota" "cache, description of"
.cindex "caching" "quota"
Exim caches the results of quota verification
in order to reduce the amount of resources used.
The &"callout"& hints database is used.

The default cache periods are five minutes for a positive (good) result
and one hour for a negative result.
To change the periods the &%quota%& option can be followed by an equals sign
and a number of optional paramemters, separated by commas.
For example:
.code
verify = recipient/quota=cachepos=1h,cacheneg=1d
.endd
Possible parameters are:
.vlist
.vitem &*cachepos&~=&~*&<&'time&~interval'&>
.cindex "quota cache" "positive entry expiry, specifying"
Set the lifetime for a positive cache entry.
A value of zero seconds is legitimate.

.vitem &*cacheneg&~=&~*&<&'time&~interval'&>
.cindex "quota cache" "negative entry expiry, specifying"
As above, for a negative entry.

.vitem &*no_cache*&
Set both positive and negative lifetimes to zero.

.section "Sender address verification reporting" "SECTsenaddver"
.cindex "verifying" "suppressing error details"
See section &<<SECTaddressverification>>& for a general discussion of
verification. When sender verification fails in an ACL, the details of the
failure are given as additional output lines before the 550 response to the
relevant SMTP command (RCPT or DATA). For example, if sender callout is in use,
you might see:
.code
MAIL FROM:<xyz@abc.example>
250 OK
RCPT TO:<pqr@def.example>
550-Verification failed for <xyz@abc.example>
550-Called:   192.168.34.43
550-Sent:     RCPT TO:<xyz@abc.example>
550-Response: 550 Unknown local part xyz in <xyz@abc.example>
550 Sender verification failed
.endd
If more than one RCPT command fails in the same way, the details are given
only for the first of them. However, some administrators do not want to send
out this much information. You can suppress the details by adding
&`/no_details`& to the ACL statement that requests sender verification. For
example:
.code
verify = sender/no_details
.endd

.section "Redirection while verifying" "SECTredirwhilveri"
.cindex "verifying" "redirection while"
.cindex "address redirection" "while verifying"
A dilemma arises when a local address is redirected by aliasing or forwarding
during verification: should the generated addresses themselves be verified,
or should the successful expansion of the original address be enough to verify
it? By default, Exim takes the following pragmatic approach:

.ilist
When an incoming address is redirected to just one child address, verification
continues with the child address, and if that fails to verify, the original
verification also fails.
.next
When an incoming address is redirected to more than one child address,
verification does not continue. A success result is returned.
.endlist

This seems the most reasonable behaviour for the common use of aliasing as a
way of redirecting different local parts to the same mailbox. It means, for
example, that a pair of alias entries of the form
.code
A.Wol:   aw123
aw123:   :fail: Gone away, no forwarding address
.endd
work as expected, with both local parts causing verification failure. When a
redirection generates more than one address, the behaviour is more like a
mailing list, where the existence of the alias itself is sufficient for
verification to succeed.

It is possible, however, to change the default behaviour so that all successful
redirections count as successful verifications, however many new addresses are
generated. This is specified by the &%success_on_redirect%& verification
option. For example:
.code
require verify = recipient/success_on_redirect/callout=10s
.endd
In this example, verification succeeds if a router generates a new address, and
the callout does not occur, because no address was routed to a remote host.

When verification is being tested via the &%-bv%& option, the treatment of
redirections is as just described, unless the &%-v%& or any debugging option is
also specified. In that case, full verification is done for every generated
address and a report is output for each of them.



.section "Client SMTP authorization (CSA)" "SECTverifyCSA"
.cindex "CSA" "verifying"
Client SMTP Authorization is a system that allows a site to advertise
which machines are and are not permitted to send email. This is done by placing
special SRV records in the DNS; these are looked up using the client's HELO
domain. At the time of writing, CSA is still an Internet Draft. Client SMTP
Authorization checks in Exim are performed by the ACL condition:
.code
verify = csa
.endd
This fails if the client is not authorized. If there is a DNS problem, or if no
valid CSA SRV record is found, or if the client is authorized, the condition
succeeds. These three cases can be distinguished using the expansion variable
&$csa_status$&, which can take one of the values &"fail"&, &"defer"&,
&"unknown"&, or &"ok"&. The condition does not itself defer because that would
be likely to cause problems for legitimate email.

The error messages produced by the CSA code include slightly more
detail. If &$csa_status$& is &"defer"&, this may be because of problems
looking up the CSA SRV record, or problems looking up the CSA target
address record. There are four reasons for &$csa_status$& being &"fail"&:

.ilist
The client's host name is explicitly not authorized.
.next
The client's IP address does not match any of the CSA target IP addresses.
.next
The client's host name is authorized but it has no valid target IP addresses
(for example, the target's addresses are IPv6 and the client is using IPv4).
.next
The client's host name has no CSA SRV record but a parent domain has asserted
that all subdomains must be explicitly authorized.
.endlist

The &%csa%& verification condition can take an argument which is the domain to
use for the DNS query. The default is:
.code
verify = csa/$sender_helo_name
.endd
This implementation includes an extension to CSA. If the query domain
is an address literal such as [192.0.2.95], or if it is a bare IP
address, Exim searches for CSA SRV records in the reverse DNS as if
the HELO domain was (for example) &'95.2.0.192.in-addr.arpa'&. Therefore it is
meaningful to say:
.code
verify = csa/$sender_host_address
.endd
In fact, this is the check that Exim performs if the client does not say HELO.
This extension can be turned off by setting the main configuration option
&%dns_csa_use_reverse%& to be false.

If a CSA SRV record is not found for the domain itself, a search
is performed through its parent domains for a record which might be
making assertions about subdomains. The maximum depth of this search is limited
using the main configuration option &%dns_csa_search_limit%&, which is 5 by
default. Exim does not look for CSA SRV records in a top level domain, so the
default settings handle HELO domains as long as seven
(&'hostname.five.four.three.two.one.com'&). This encompasses the vast majority
of legitimate HELO domains.

The &'dnsdb'& lookup also has support for CSA. Although &'dnsdb'& also supports
direct SRV lookups, this is not sufficient because of the extra parent domain
search behaviour of CSA, and (as with PTR lookups) &'dnsdb'& also turns IP
addresses into lookups in the reverse DNS space. The result of a successful
lookup such as:
.code
${lookup dnsdb {csa=$sender_helo_name}}
.endd
has two space-separated fields: an authorization code and a target host name.
The authorization code can be &"Y"& for yes, &"N"& for no, &"X"& for explicit
authorization required but absent, or &"?"& for unknown.




.section "Bounce address tag validation" "SECTverifyPRVS"
.cindex "BATV, verifying"
Bounce address tag validation (BATV) is a scheme whereby the envelope senders
of outgoing messages have a cryptographic, timestamped &"tag"& added to them.
Genuine incoming bounce messages should therefore always be addressed to
recipients that have a valid tag. This scheme is a way of detecting unwanted
bounce messages caused by sender address forgeries (often called &"collateral
spam"&), because the recipients of such messages do not include valid tags.

There are two expansion items to help with the implementation of the BATV
&"prvs"& (private signature) scheme in an Exim configuration. This scheme signs
the original envelope sender address by using a simple key to add a hash of the
address and some time-based randomizing information. The &%prvs%& expansion
item creates a signed address, and the &%prvscheck%& expansion item checks one.
The syntax of these expansion items is described in section
&<<SECTexpansionitems>>&.
The validity period on signed addresses is seven days.

As an example, suppose the secret per-address keys are stored in an MySQL
database. A query to look up the key for an address could be defined as a macro
like this:
.code
PRVSCHECK_SQL = ${lookup mysql{SELECT secret FROM batv_prvs \
                WHERE sender='${quote_mysql:$prvscheck_address}'\
                }{$value}}
.endd
Suppose also that the senders who make use of BATV are defined by an address
list called &%batv_senders%&. Then, in the ACL for RCPT commands, you could
use this:
.code
# Bounces: drop unsigned addresses for BATV senders
deny senders = :
     recipients = +batv_senders
     message = This address does not send an unsigned reverse path

# Bounces: In case of prvs-signed address, check signature.
deny senders = :
     condition  = ${prvscheck {$local_part@$domain}\
                  {PRVSCHECK_SQL}{1}}
     !condition = $prvscheck_result
     message = Invalid reverse path signature.
.endd
The first statement rejects recipients for bounce messages that are addressed
to plain BATV sender addresses, because it is known that BATV senders do not
send out messages with plain sender addresses. The second statement rejects
recipients that are prvs-signed, but with invalid signatures (either because
the key is wrong, or the signature has timed out).

A non-prvs-signed address is not rejected by the second statement, because the
&%prvscheck%& expansion yields an empty string if its first argument is not a
prvs-signed address, thus causing the &%condition%& condition to be false. If
the first argument is a syntactically valid prvs-signed address, the yield is
the third string (in this case &"1"&), whether or not the cryptographic and
timeout checks succeed. The &$prvscheck_result$& variable contains the result
of the checks (empty for failure, &"1"& for success).

There is one more issue you must consider when implementing prvs-signing:
you have to ensure that the routers accept prvs-signed addresses and
deliver them correctly. The easiest way to handle this is to use a &(redirect)&
router to remove the signature with a configuration along these lines:
.code
batv_redirect:
  driver = redirect
  data = ${prvscheck {$local_part@$domain}{PRVSCHECK_SQL}}
.endd
This works because, if the third argument of &%prvscheck%& is empty, the result
of the expansion of a prvs-signed address is the decoded value of the original
address. This router should probably be the first of your routers that handles
local addresses.

To create BATV-signed addresses in the first place, a transport of this form
can be used:
.code
external_smtp_batv:
  driver = smtp
  return_path = ${prvs {$return_path} \
                       {${lookup mysql{SELECT \
                       secret FROM batv_prvs WHERE \
                       sender='${quote_mysql:$sender_address}'} \
                       {$value}fail}}}
.endd
If no key can be found for the existing return path, no signing takes place.



.section "Using an ACL to control relaying" "SECTrelaycontrol"
.cindex "&ACL;" "relay control"
.cindex "relaying" "control by ACL"
.cindex "policy control" "relay control"
An MTA is said to &'relay'& a message if it receives it from some host and
delivers it directly to another host as a result of a remote address contained
within it. Redirecting a local address via an alias or forward file and then
passing the message on to another host is not relaying,
.cindex "&""percent hack""&"
but a redirection as a result of the &"percent hack"& is.

Two kinds of relaying exist, which are termed &"incoming"& and &"outgoing"&.
A host which is acting as a gateway or an MX backup is concerned with incoming
relaying from arbitrary hosts to a specific set of domains. On the other hand,
a host which is acting as a smart host for a number of clients is concerned
with outgoing relaying from those clients to the Internet at large. Often the
same host is fulfilling both functions,
. ///
. as illustrated in the diagram below,
. ///
but in principle these two kinds of relaying are entirely independent. What is
not wanted is the transmission of mail from arbitrary remote hosts through your
system to arbitrary domains.


You can implement relay control by means of suitable statements in the ACL that
runs for each RCPT command. For convenience, it is often easiest to use
Exim's named list facility to define the domains and hosts involved. For
example, suppose you want to do the following:

.ilist
Deliver a number of domains to mailboxes on the local host (or process them
locally in some other way). Let's say these are &'my.dom1.example'& and
&'my.dom2.example'&.
.next
Relay mail for a number of other domains for which you are the secondary MX.
These might be &'friend1.example'& and &'friend2.example'&.
.next
Relay mail from the hosts on your local LAN, to whatever domains are involved.
Suppose your LAN is 192.168.45.0/24.
.endlist


In the main part of the configuration, you put the following definitions:
.code
domainlist local_domains    = my.dom1.example : my.dom2.example
domainlist relay_to_domains = friend1.example : friend2.example
hostlist   relay_from_hosts = 192.168.45.0/24
.endd
Now you can use these definitions in the ACL that is run for every RCPT
command:
.code
acl_check_rcpt:
  accept domains = +local_domains : +relay_to_domains
  accept hosts   = +relay_from_hosts
.endd
The first statement accepts any RCPT command that contains an address in
the local or relay domains. For any other domain, control passes to the second
statement, which accepts the command only if it comes from one of the relay
hosts. In practice, you will probably want to make your ACL more sophisticated
than this, for example, by including sender and recipient verification. The
default configuration includes a more comprehensive example, which is described
in chapter &<<CHAPdefconfil>>&.



.section "Checking a relay configuration" "SECTcheralcon"
.cindex "relaying" "checking control of"
You can check the relay characteristics of your configuration in the same way
that you can test any ACL behaviour for an incoming SMTP connection, by using
the &%-bh%& option to run a fake SMTP session with which you interact.
.ecindex IIDacl



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Content scanning at ACL time" "CHAPexiscan"
.scindex IIDcosca "content scanning" "at ACL time"
The extension of Exim to include content scanning at ACL time, formerly known
as &"exiscan"&, was originally implemented as a patch by Tom Kistner. The code
was integrated into the main source for Exim release 4.50, and Tom continues to
maintain it. Most of the wording of this chapter is taken from Tom's
specification.

It is also possible to scan the content of messages at other times. The
&[local_scan()]& function (see chapter &<<CHAPlocalscan>>&) allows for content
scanning after all the ACLs have run. A transport filter can be used to scan
messages at delivery time (see the &%transport_filter%& option, described in
chapter &<<CHAPtransportgeneric>>&).

If you want to include the ACL-time content-scanning features when you compile
Exim, you need to arrange for WITH_CONTENT_SCAN to be defined in your
&_Local/Makefile_&. When you do that, the Exim binary is built with:

.ilist
Two additional ACLs (&%acl_smtp_mime%& and &%acl_not_smtp_mime%&) that are run
for all MIME parts for SMTP and non-SMTP messages, respectively.
.next
Additional ACL conditions and modifiers: &%decode%&, &%malware%&,
&%mime_regex%&, &%regex%&, and &%spam%&. These can be used in the ACL that is
run at the end of message reception (the &%acl_smtp_data%& ACL).
.next
An additional control feature (&"no_mbox_unspool"&) that saves spooled copies
of messages, or parts of messages, for debugging purposes.
.next
Additional expansion variables that are set in the new ACL and by the new
conditions.
.next
Two new main configuration options: &%av_scanner%& and &%spamd_address%&.
.endlist

Content-scanning is continually evolving, and new features are still being
added. While such features are still unstable and liable to incompatible
changes, they are made available in Exim by setting options whose names begin
EXPERIMENTAL_ in &_Local/Makefile_&. Such features are not documented in
this manual. You can find out about them by reading the file called
&_doc/experimental.txt_&.

All the content-scanning facilities work on a MBOX copy of the message that is
temporarily created in a file called:
.display
<&'spool_directory'&>&`/scan/`&<&'message_id'&>/<&'message_id'&>&`.eml`&
.endd
The &_.eml_& extension is a friendly hint to virus scanners that they can
expect an MBOX-like structure inside that file. The file is created when the
first content scanning facility is called. Subsequent calls to content
scanning conditions open the same file again. The directory is recursively
removed when the &%acl_smtp_data%& ACL has finished running, unless
.code
control = no_mbox_unspool
.endd
has been encountered. When the MIME ACL decodes files, they are put into the
same directory by default.



.section "Scanning for viruses" "SECTscanvirus"
.cindex "virus scanning"
.cindex "content scanning" "for viruses"
.cindex "content scanning" "the &%malware%& condition"
The &%malware%& ACL condition lets you connect virus scanner software to Exim.
It supports a &"generic"& interface to scanners called via the shell, and
specialized interfaces for &"daemon"& type virus scanners, which are resident
in memory and thus are much faster.

Since message data needs to have arrived,
the condition may be only called in ACL defined by
&%acl_smtp_data%&,
&%acl_smtp_data_prdr%&,
&%acl_smtp_mime%& or
&%acl_smtp_dkim%&

A timeout of 2 minutes is applied to a scanner call (by default);
if it expires then a defer action is taken.

.oindex "&%av_scanner%&"
You can set the &%av_scanner%& option in the main part of the configuration
to specify which scanner to use, together with any additional options that
are needed. The basic syntax is as follows:
.display
&`av_scanner = <`&&'scanner-type'&&`>:<`&&'option1'&&`>:<`&&'option2'&&`>:[...]`&
.endd
If you do not set &%av_scanner%&, it defaults to
.code
av_scanner = sophie:/var/run/sophie
.endd
If the value of &%av_scanner%& starts with a dollar character, it is expanded
before use.
The usual list-parsing of the content (see &<<SECTlistconstruct>>&) applies.
The following scanner types are supported in this release,
though individual ones can be included or not at build time:

.vlist
.vitem &%avast%&
.cindex "virus scanners" "avast"
This is the scanner daemon of Avast. It has been tested with Avast Core
Security (currently at version 2.2.0).
You can get a trial version at &url(https://www.avast.com) or for Linux
at &url(https://www.avast.com/linux-server-antivirus).
This scanner type takes one option,
which can be either a full path to a UNIX socket,
or host and port specifiers separated by white space.
The host may be a name or an IP address; the port is either a
single number or a pair of numbers with a dash between.
A list of options may follow. These options are interpreted on the
Exim's side of the malware scanner, or are given on separate lines to
the daemon as options before the main scan command.

.cindex &`pass_unscanned`& "avast"
If &`pass_unscanned`&
is set, any files the Avast scanner can't scan (e.g.
decompression bombs, or invalid archives) are considered clean. Use with
care.

For example:
.code
av_scanner = avast:/var/run/avast/scan.sock:FLAGS -fullfiles:SENSITIVITY -pup
av_scanner = avast:/var/run/avast/scan.sock:pass_unscanned:FLAGS -fullfiles:SENSITIVITY -pup
av_scanner = avast:192.168.2.22 5036
.endd
If you omit the argument, the default path
&_/var/run/avast/scan.sock_&
is used.
If you use a remote host,
you need to make Exim's spool directory available to it,
as the scanner is passed a file path, not file contents.
For information about available commands and their options you may use
.code
$ socat UNIX:/var/run/avast/scan.sock STDIO:
    FLAGS
    SENSITIVITY
    PACK
.endd

If the scanner returns a temporary failure (e.g. license issues, or
permission problems), the message is deferred and a paniclog entry is
written.  The usual &`defer_ok`& option is available.

.vitem &%aveserver%&
.cindex "virus scanners" "Kaspersky"
This is the scanner daemon of Kaspersky Version 5. You can get a trial version
at &url(https://www.kaspersky.com/). This scanner type takes one option,
which is the path to the daemon's UNIX socket. The default is shown in this
example:
.code
av_scanner = aveserver:/var/run/aveserver
.endd


.vitem &%clamd%&
.cindex "virus scanners" "clamd"
This daemon-type scanner is GPL and free. You can get it at
&url(https://www.clamav.net/). Some older versions of clamd do not seem to
unpack MIME containers, so it used to be recommended to unpack MIME attachments
in the MIME ACL. This is no longer believed to be necessary.

The options are a list of server specifiers, which may be
a UNIX socket specification,
a TCP socket specification,
or a (global) option.

A socket specification consists of a space-separated list.
For a Unix socket the first element is a full path for the socket,
for a TCP socket the first element is the IP address
and the second a port number,
Any further elements are per-server (non-global) options.
These per-server options are supported:
.code
retry=<timespec>        Retry on connect fail
.endd

The &`retry`& option specifies a time after which a single retry for
a failed connect is made.  The default is to not retry.

If a Unix socket file is specified, only one server is supported.

Examples:
.code
av_scanner = clamd:/opt/clamd/socket
av_scanner = clamd:192.0.2.3 1234
av_scanner = clamd:192.0.2.3 1234:local
av_scanner = clamd:192.0.2.3 1234 retry=10s
av_scanner = clamd:192.0.2.3 1234 : 192.0.2.4 1234
.endd
If the value of av_scanner points to a UNIX socket file or contains the
&`local`&
option, then the ClamAV interface will pass a filename containing the data
to be scanned, which should normally result in less I/O happening and be
more efficient.  Normally in the TCP case, the data is streamed to ClamAV as
Exim does not assume that there is a common filesystem with the remote host.

The final example shows that multiple TCP targets can be specified.  Exim will
randomly use one for each incoming email (i.e. it load balances them).  Note
that only TCP targets may be used if specifying a list of scanners; a UNIX
socket cannot be mixed in with TCP targets.  If one of the servers becomes
unavailable, Exim will try the remaining one(s) until it finds one that works.
When a clamd server becomes unreachable, Exim will log a message.  Exim does
not keep track of scanner state between multiple messages, and the scanner
selection is random, so the message will get logged in the mainlog for each
email that the down scanner gets chosen first (message wrapped to be readable):
.code
2013-10-09 14:30:39 1VTumd-0000Y8-BQ malware acl condition:
   clamd: connection to localhost, port 3310 failed
   (Connection refused)
.endd

If the option is unset, the default is &_/tmp/clamd_&. Thanks to David Saez for
contributing the code for this scanner.

.vitem &%cmdline%&
.cindex "virus scanners" "command line interface"
This is the keyword for the generic command line scanner interface. It can be
used to attach virus scanners that are invoked from the shell. This scanner
type takes 3 mandatory options:

.olist
The full path and name of the scanner binary, with all command line options,
and a placeholder (&`%s`&) for the directory to scan.

.next
A regular expression to match against the STDOUT and STDERR output of the
virus scanner. If the expression matches, a virus was found. You must make
absolutely sure that this expression matches on &"virus found"&. This is called
the &"trigger"& expression.

.next
Another regular expression, containing exactly one pair of parentheses, to
match the name of the virus found in the scanners output. This is called the
&"name"& expression.
.endlist olist

For example, Sophos Sweep reports a virus on a line like this:
.code
Virus 'W32/Magistr-B' found in file ./those.bat
.endd
For the trigger expression, we can match the phrase &"found in file"&. For the
name expression, we want to extract the W32/Magistr-B string, so we can match
for the single quotes left and right of it. Altogether, this makes the
configuration setting:
.code
av_scanner = cmdline:\
             /path/to/sweep -ss -all -rec -archive %s:\
             found in file:'(.+)'
.endd
.vitem &%drweb%&
.cindex "virus scanners" "DrWeb"
The DrWeb daemon scanner (&url(https://www.sald.ru/)) interface
takes one option,
either a full path to a UNIX socket,
or host and port specifiers separated by white space.
The host may be a name or an IP address; the port is either a
single number or a pair of numbers with a dash between.
For example:
.code
av_scanner = drweb:/var/run/drwebd.sock
av_scanner = drweb:192.168.2.20 31337
.endd
If you omit the argument, the default path &_/usr/local/drweb/run/drwebd.sock_&
is used. Thanks to Alex Miller for contributing the code for this scanner.

.vitem &%f-protd%&
.cindex "virus scanners" "f-protd"
The f-protd scanner is accessed via HTTP over TCP.
One argument is taken, being a space-separated hostname and port number
(or port-range).
For example:
.code
av_scanner = f-protd:localhost 10200-10204
.endd
If you omit the argument, the default values shown above are used.

.vitem &%f-prot6d%&
.cindex "virus scanners" "f-prot6d"
The f-prot6d scanner is accessed using the FPSCAND protocol over TCP.
One argument is taken, being a space-separated hostname and port number.
For example:
.code
av_scanner = f-prot6d:localhost 10200
.endd
If you omit the argument, the default values show above are used.

.vitem &%fsecure%&
.cindex "virus scanners" "F-Secure"
The F-Secure daemon scanner (&url(https://www.f-secure.com/)) takes one
argument which is the path to a UNIX socket. For example:
.code
av_scanner = fsecure:/path/to/.fsav
.endd
If no argument is given, the default is &_/var/run/.fsav_&. Thanks to Johan
Thelmen for contributing the code for this scanner.

.vitem &%kavdaemon%&
.cindex "virus scanners" "Kaspersky"
This is the scanner daemon of Kaspersky Version 4. This version of the
Kaspersky scanner is outdated. Please upgrade (see &%aveserver%& above). This
scanner type takes one option, which is the path to the daemon's UNIX socket.
For example:
.code
av_scanner = kavdaemon:/opt/AVP/AvpCtl
.endd
The default path is &_/var/run/AvpCtl_&.

.vitem &%mksd%&
.cindex "virus scanners" "mksd"
This was a daemon type scanner that is aimed mainly at Polish users,
though some documentation was available in English.
The history can be shown at &url(https://en.wikipedia.org/wiki/Mks_vir)
and this appears to be a candidate for removal from Exim, unless
we are informed of other virus scanners which use the same protocol
to integrate.
The only option for this scanner type is
the maximum number of processes used simultaneously to scan the attachments,
provided that mksd has
been run with at least the same number of child processes. For example:
.code
av_scanner = mksd:2
.endd
You can safely omit this option (the default value is 1).

.vitem &%sock%&
.cindex "virus scanners" "simple socket-connected"
This is a general-purpose way of talking to simple scanner daemons
running on the local machine.
There are four options:
an address (which may be an IP address and port, or the path of a Unix socket),
a commandline to send (may include a single %s which will be replaced with
the path to the mail file to be scanned),
an RE to trigger on from the returned data,
and an RE to extract malware_name from the returned data.
For example:
.code
av_scanner = sock:127.0.0.1 6001:%s:(SPAM|VIRUS):(.*)$
.endd
Note that surrounding whitespace is stripped from each option, meaning
there is no way to specify a trailing newline.
The socket specifier and both regular-expressions are required.
Default for the commandline is &_%s\n_& (note this does have a trailing newline);
specify an empty element to get this.

.vitem &%sophie%&
.cindex "virus scanners" "Sophos and Sophie"
Sophie is a daemon that uses Sophos' &%libsavi%& library to scan for viruses.
You can get Sophie at &url(http://sophie.sourceforge.net/). The only option
for this scanner type is the path to the UNIX socket that Sophie uses for
client communication. For example:
.code
av_scanner = sophie:/tmp/sophie
.endd
The default path is &_/var/run/sophie_&, so if you are using this, you can omit
the option.
.endlist

When &%av_scanner%& is correctly set, you can use the &%malware%& condition in
the DATA ACL. &*Note*&: You cannot use the &%malware%& condition in the MIME
ACL.

The &%av_scanner%& option is expanded each time &%malware%& is called. This
makes it possible to use different scanners. See further below for an example.
The &%malware%& condition caches its results, so when you use it multiple times
for the same message, the actual scanning process is only carried out once.
However, using expandable items in &%av_scanner%& disables this caching, in
which case each use of the &%malware%& condition causes a new scan of the
message.

The &%malware%& condition takes a right-hand argument that is expanded before
use and taken as a list, slash-separated by default.
The first element can then be one of

.ilist
&"true"&, &"*"&, or &"1"&, in which case the message is scanned for viruses.
The condition succeeds if a virus was found, and fail otherwise. This is the
recommended usage.
.next
&"false"& or &"0"& or an empty string, in which case no scanning is done and
the condition fails immediately.
.next
A regular expression, in which case the message is scanned for viruses. The
condition succeeds if a virus is found and its name matches the regular
expression. This allows you to take special actions on certain types of virus.
Note that &"/"& characters in the RE must be doubled due to the list-processing,
unless the separator is changed (in the usual way &<<SECTlistsepchange>>&).
.endlist

You can append a &`defer_ok`& element to the &%malware%& argument list to accept
messages even if there is a problem with the virus scanner.
Otherwise, such a problem causes the ACL to defer.

You can append a &`tmo=<val>`& element to the &%malware%& argument list to
specify a non-default timeout.  The default is two minutes.
For example:
.code
malware = * / defer_ok / tmo=10s
.endd
A timeout causes the ACL to defer.

.vindex "&$callout_address$&"
When a connection is made to the scanner the expansion variable &$callout_address$&
is set to record the actual address used.

.vindex "&$malware_name$&"
When a virus is found, the condition sets up an expansion variable called
&$malware_name$& that contains the name of the virus. You can use it in a
&%message%& modifier that specifies the error returned to the sender, and/or in
logging data.

Beware the interaction of Exim's &%message_size_limit%& with any size limits
imposed by your anti-virus scanner.

Here is a very simple scanning example:
.code
deny malware = *
     message = This message contains malware ($malware_name)
.endd
The next example accepts messages when there is a problem with the scanner:
.code
deny malware = */defer_ok
     message = This message contains malware ($malware_name)
.endd
The next example shows how to use an ACL variable to scan with both sophie and
aveserver. It assumes you have set:
.code
av_scanner = $acl_m0
.endd
in the main Exim configuration.
.code
deny set acl_m0 = sophie
     malware = *
     message = This message contains malware ($malware_name)

deny set acl_m0 = aveserver
     malware = *
     message = This message contains malware ($malware_name)
.endd


.section "Scanning with SpamAssassin and Rspamd" "SECTscanspamass"
.cindex "content scanning" "for spam"
.cindex "spam scanning"
.cindex "SpamAssassin"
.cindex "Rspamd"
The &%spam%& ACL condition calls SpamAssassin's &%spamd%& daemon to get a spam
score and a report for the message.
Support is also provided for Rspamd.

For more information about installation and configuration of SpamAssassin or
Rspamd refer to their respective websites at
&url(https://spamassassin.apache.org/) and &url(https://www.rspamd.com/)

SpamAssassin can be installed with CPAN by running:
.code
perl -MCPAN -e 'install Mail::SpamAssassin'
.endd
SpamAssassin has its own set of configuration files. Please review its
documentation to see how you can tweak it. The default installation should work
nicely, however.

.oindex "&%spamd_address%&"
By default, SpamAssassin listens on 127.0.0.1, TCP port 783 and if you
intend to use an instance running on the local host you do not need to set
&%spamd_address%&. If you intend to use another host or port for SpamAssassin,
you must set the &%spamd_address%& option in the global part of the Exim
configuration as follows (example):
.code
spamd_address = 192.168.99.45 783
.endd
The SpamAssassin protocol relies on a TCP half-close from the client.
If your SpamAssassin client side is running a Linux system with an
iptables firewall, consider setting
&%net.netfilter.nf_conntrack_tcp_timeout_close_wait%& to at least the
timeout, Exim uses when waiting for a response from the SpamAssassin
server (currently defaulting to 120s).  With a lower value the Linux
connection tracking may consider your half-closed connection as dead too
soon.


To use Rspamd (which by default listens on all local addresses
on TCP port 11333)
you should add &%variant=rspamd%& after the address/port pair, for example:
.code
spamd_address = 127.0.0.1 11333 variant=rspamd
.endd

As of version 2.60, &%SpamAssassin%& also supports communication over UNIX
sockets. If you want to us these, supply &%spamd_address%& with an absolute
filename instead of an address/port pair:
.code
spamd_address = /var/run/spamd_socket
.endd
You can have multiple &%spamd%& servers to improve scalability. These can
reside on other hardware reachable over the network. To specify multiple
&%spamd%& servers, put multiple address/port pairs in the &%spamd_address%&
option, separated with colons (the separator can be changed in the usual way &<<SECTlistsepchange>>&):
.code
spamd_address = 192.168.2.10 783 : \
                192.168.2.11 783 : \
                192.168.2.12 783
.endd
Up to 32 &%spamd%& servers are supported.
When a server fails to respond to the connection attempt, all other
servers are tried until one succeeds. If no server responds, the &%spam%&
condition defers.

Unix and TCP socket specifications may be mixed in any order.
Each element of the list is a list itself, space-separated by default
and changeable in the usual way (&<<SECTlistsepchange>>&);
take care to not double the separator.

For TCP socket specifications a host name or IP (v4 or v6, but
subject to list-separator quoting rules) address can be used,
and the port can be one or a dash-separated pair.
In the latter case, the range is tried in strict order.

Elements after the first for Unix sockets, or second for TCP socket,
are options.
The supported options are:
.code
pri=<priority>      Selection priority
weight=<value>      Selection bias
time=<start>-<end>  Use only between these times of day
retry=<timespec>    Retry on connect fail
tmo=<timespec>      Connection time limit
variant=rspamd      Use Rspamd rather than SpamAssassin protocol
.endd

The &`pri`& option specifies a priority for the server within the list,
higher values being tried first.
The default priority is 1.

The &`weight`& option specifies a selection bias.
Within a priority set
servers are queried in a random fashion, weighted by this value.
The default value for selection bias is 1.

Time specifications for the &`time`& option are <hour>.<minute>.<second>
in the local time zone; each element being one or more digits.
Either the seconds or both minutes and seconds, plus the leading &`.`&
characters, may be omitted and will be taken as zero.

Timeout specifications for the &`retry`& and &`tmo`& options
are the usual Exim time interval standard, e.g. &`20s`& or &`1m`&.

The &`tmo`& option specifies an overall timeout for communication.
The default value is two minutes.

The &`retry`& option specifies a time after which a single retry for
a failed connect is made.
The default is to not retry.

The &%spamd_address%& variable is expanded before use if it starts with
a dollar sign. In this case, the expansion may return a string that is
used as the list so that multiple spamd servers can be the result of an
expansion.

.vindex "&$callout_address$&"
When a connection is made to the server the expansion variable &$callout_address$&
is set to record the actual address used.

.section "Calling SpamAssassin from an Exim ACL" "SECID206"
Here is a simple example of the use of the &%spam%& condition in a DATA ACL:
.code
deny spam = joe
     message = This message was classified as SPAM
.endd
The right-hand side of the &%spam%& condition specifies a name. This is
relevant if you have set up multiple SpamAssassin profiles. If you do not want
to scan using a specific profile, but rather use the SpamAssassin system-wide
default profile, you can scan for an unknown name, or simply use &"nobody"&.
Rspamd does not use this setting. However, you must put something on the
right-hand side.

The name allows you to use per-domain or per-user antispam profiles in
principle, but this is not straightforward in practice, because a message may
have multiple recipients, not necessarily all in the same domain. Because the
&%spam%& condition has to be called from a DATA-time ACL in order to be able to
read the contents of the message, the variables &$local_part$& and &$domain$&
are not set.
Careful enforcement of single-recipient messages
(e.g. by responding with defer in the recipient ACL for all recipients
after the first),
or the use of PRDR,
.cindex "PRDR" "use for per-user SpamAssassin profiles"
are needed to use this feature.

The right-hand side of the &%spam%& condition is expanded before being used, so
you can put lookups or conditions there. When the right-hand side evaluates to
&"0"& or &"false"&, no scanning is done and the condition fails immediately.


Scanning with SpamAssassin uses a lot of resources. If you scan every message,
large ones may cause significant performance degradation. As most spam messages
are quite small, it is recommended that you do not scan the big ones. For
example:
.code
deny condition = ${if < {$message_size}{10K}}
     spam = nobody
     message = This message was classified as SPAM
.endd

The &%spam%& condition returns true if the threshold specified in the user's
SpamAssassin profile has been matched or exceeded. If you want to use the
&%spam%& condition for its side effects (see the variables below), you can make
it always return &"true"& by appending &`:true`& to the username.

.cindex "spam scanning" "returned variables"
When the &%spam%& condition is run, it sets up a number of expansion
variables.
Except for &$spam_report$&,
these variables are saved with the received message so are
available for use at delivery time.

.vlist
.vitem &$spam_score$&
The spam score of the message, for example, &"3.4"& or &"30.5"&. This is useful
for inclusion in log or reject messages.

.vitem &$spam_score_int$&
The spam score of the message, multiplied by ten, as an integer value. For
example &"34"& or &"305"&. It may appear to disagree with &$spam_score$&
because &$spam_score$& is rounded and &$spam_score_int$& is truncated.
The integer value is useful for numeric comparisons in conditions.

.vitem &$spam_bar$&
A string consisting of a number of &"+"& or &"-"& characters, representing the
integer part of the spam score value. A spam score of 4.4 would have a
&$spam_bar$& value of &"++++"&. This is useful for inclusion in warning
headers, since MUAs can match on such strings. The maximum length of the
spam bar is 50 characters.

.vitem &$spam_report$&
A multiline text table, containing the full SpamAssassin report for the
message. Useful for inclusion in headers or reject messages.
This variable is only usable in a DATA-time ACL.
Beware that SpamAssassin may return non-ASCII characters, especially
when running in country-specific locales, which are not legal
unencoded in headers.

.vitem &$spam_action$&
For SpamAssassin either 'reject' or 'no action' depending on the
spam score versus threshold.
For Rspamd, the recommended action.

.endlist

The &%spam%& condition caches its results unless expansion in
spamd_address was used. If you call it again with the same user name, it
does not scan again, but rather returns the same values as before.

The &%spam%& condition returns DEFER if there is any error while running
the message through SpamAssassin or if the expansion of spamd_address
failed. If you want to treat DEFER as FAIL (to pass on to the next ACL
statement block), append &`/defer_ok`& to the right-hand side of the
spam condition, like this:
.code
deny spam    = joe/defer_ok
     message = This message was classified as SPAM
.endd
This causes messages to be accepted even if there is a problem with &%spamd%&.

Here is a longer, commented example of the use of the &%spam%&
condition:
.code
# put headers in all messages (no matter if spam or not)
warn  spam = nobody:true
      add_header = X-Spam-Score: $spam_score ($spam_bar)
      add_header = X-Spam-Report: $spam_report

# add second subject line with *SPAM* marker when message
# is over threshold
warn  spam = nobody
      add_header = Subject: *SPAM* $h_Subject:

# reject spam at high scores (> 12)
deny  spam = nobody:true
      condition = ${if >{$spam_score_int}{120}{1}{0}}
      message = This message scored $spam_score spam points.
.endd



.section "Scanning MIME parts" "SECTscanmimepart"
.cindex "content scanning" "MIME parts"
.cindex "MIME content scanning"
.oindex "&%acl_smtp_mime%&"
.oindex "&%acl_not_smtp_mime%&"
The &%acl_smtp_mime%& global option specifies an ACL that is called once for
each MIME part of an SMTP message, including multipart types, in the sequence
of their position in the message. Similarly, the &%acl_not_smtp_mime%& option
specifies an ACL that is used for the MIME parts of non-SMTP messages. These
options may both refer to the same ACL if you want the same processing in both
cases.

These ACLs are called (possibly many times) just before the &%acl_smtp_data%&
ACL in the case of an SMTP message, or just before the &%acl_not_smtp%& ACL in
the case of a non-SMTP message. However, a MIME ACL is called only if the
message contains a &'Content-Type:'& header line. When a call to a MIME
ACL does not yield &"accept"&, ACL processing is aborted and the appropriate
result code is sent to the client. In the case of an SMTP message, the
&%acl_smtp_data%& ACL is not called when this happens.

You cannot use the &%malware%& or &%spam%& conditions in a MIME ACL; these can
only be used in the DATA or non-SMTP ACLs. However, you can use the &%regex%&
condition to match against the raw MIME part. You can also use the
&%mime_regex%& condition to match against the decoded MIME part (see section
&<<SECTscanregex>>&).

At the start of a MIME ACL, a number of variables are set from the header
information for the relevant MIME part. These are described below. The contents
of the MIME part are not by default decoded into a disk file except for MIME
parts whose content-type is &"message/rfc822"&. If you want to decode a MIME
part into a disk file, you can use the &%decode%& condition. The general
syntax is:
.display
&`decode = [/`&<&'path'&>&`/]`&<&'filename'&>
.endd
The right hand side is expanded before use. After expansion,
the value can be:

.olist
&"0"& or &"false"&, in which case no decoding is done.
.next
The string &"default"&. In that case, the file is put in the temporary
&"default"& directory <&'spool_directory'&>&_/scan/_&<&'message_id'&>&_/_& with
a sequential filename consisting of the message id and a sequence number. The
full path and name is available in &$mime_decoded_filename$& after decoding.
.next
A full path name starting with a slash. If the full name is an existing
directory, it is used as a replacement for the default directory. The filename
is then sequentially assigned. If the path does not exist, it is used as
the full path and filename.
.next
If the string does not start with a slash, it is used as the
filename, and the default path is then used.
.endlist
The &%decode%& condition normally succeeds. It is only false for syntax
errors or unusual circumstances such as memory shortages.
The variable &$mime_filename$& will have the suggested name for the file.
Note however that this might contain anything, and is very difficult
to safely use as all or even part of the filename.
If you place files outside of the default path, they are not
automatically unlinked.

For RFC822 attachments (these are messages attached to messages, with a
content-type of &"message/rfc822"&), the ACL is called again in the same manner
as for the primary message, only that the &$mime_is_rfc822$& expansion
variable is set (see below). Attached messages are always decoded to disk
before being checked, and the files are unlinked once the check is done.

The MIME ACL supports the &%regex%& and &%mime_regex%& conditions. These can be
used to match regular expressions against raw and decoded MIME parts,
respectively. They are described in section &<<SECTscanregex>>&.

.cindex "MIME content scanning" "returned variables"
The following list describes all expansion variables that are
available in the MIME ACL:

.vlist
.vitem &$mime_anomaly_level$& &&&
       &$mime_anomaly_text$&
.vindex &$mime_anomaly_level$&
.vindex &$mime_anomaly_text$&
If there are problems decoding, these variables contain information on
the detected issue.

.vitem &$mime_boundary$&
.vindex &$mime_boundary$&
If the current part is a multipart (see &$mime_is_multipart$& below), it should
have a boundary string, which is stored in this variable. If the current part
has no boundary parameter in the &'Content-Type:'& header, this variable
contains the empty string.

.vitem &$mime_charset$&
.vindex &$mime_charset$&
This variable contains the character set identifier, if one was found in the
&'Content-Type:'& header. Examples for charset identifiers are:
.code
us-ascii
gb2312 (Chinese)
iso-8859-1
.endd
Please note that this value is not normalized, so you should do matches
case-insensitively.

.vitem &$mime_content_description$&
.vindex &$mime_content_description$&
This variable contains the normalized content of the &'Content-Description:'&
header. It can contain a human-readable description of the parts content. Some
implementations repeat the filename for attachments here, but they are usually
only used for display purposes.

.vitem &$mime_content_disposition$&
.vindex &$mime_content_disposition$&
This variable contains the normalized content of the &'Content-Disposition:'&
header. You can expect strings like &"attachment"& or &"inline"& here.

.vitem &$mime_content_id$&
.vindex &$mime_content_id$&
This variable contains the normalized content of the &'Content-ID:'& header.
This is a unique ID that can be used to reference a part from another part.

.vitem &$mime_content_size$&
.vindex &$mime_content_size$&
This variable is set only after the &%decode%& modifier (see above) has been
successfully run. It contains the size of the decoded part in kilobytes. The
size is always rounded up to full kilobytes, so only a completely empty part
has a &$mime_content_size$& of zero.

.vitem &$mime_content_transfer_encoding$&
.vindex &$mime_content_transfer_encoding$&
This variable contains the normalized content of the
&'Content-transfer-encoding:'& header. This is a symbolic name for an encoding
type. Typical values are &"base64"& and &"quoted-printable"&.

.vitem &$mime_content_type$&
.vindex &$mime_content_type$&
If the MIME part has a &'Content-Type:'& header, this variable contains its
value, lowercased, and without any options (like &"name"& or &"charset"&). Here
are some examples of popular MIME types, as they may appear in this variable:
.code
text/plain
text/html
application/octet-stream
image/jpeg
audio/midi
.endd
If the MIME part has no &'Content-Type:'& header, this variable contains the
empty string.

.vitem &$mime_decoded_filename$&
.vindex &$mime_decoded_filename$&
This variable is set only after the &%decode%& modifier (see above) has been
successfully run. It contains the full path and filename of the file
containing the decoded data.
.endlist

.cindex "RFC 2047"
.vlist
.vitem &$mime_filename$&
.vindex &$mime_filename$&
This is perhaps the most important of the MIME variables. It contains a
proposed filename for an attachment, if one was found in either the
&'Content-Type:'& or &'Content-Disposition:'& headers. The filename will be
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047) or
&url(https://www.rfc-editor.org/rfc/rfc2231,RFC 2231)
decoded, but no additional sanity checks are done.
 If no filename was
found, this variable contains the empty string.

.vitem &$mime_is_coverletter$&
.vindex &$mime_is_coverletter$&
This variable attempts to differentiate the &"cover letter"& of an e-mail from
attached data. It can be used to clamp down on flashy or unnecessarily encoded
content in the cover letter, while not restricting attachments at all.

The variable contains 1 (true) for a MIME part believed to be part of the
cover letter, and 0 (false) for an attachment. At present, the algorithm is as
follows:

.olist
The outermost MIME part of a message is always a cover letter.

.next
If a multipart/alternative or multipart/related MIME part is a cover letter,
so are all MIME subparts within that multipart.

.next
If any other multipart is a cover letter, the first subpart is a cover letter,
and the rest are attachments.

.next
All parts contained within an attachment multipart are attachments.
.endlist olist

As an example, the following will ban &"HTML mail"& (including that sent with
alternative plain text), while allowing HTML files to be attached. HTML
coverletter mail attached to non-HTML coverletter mail will also be allowed:
.code
deny !condition = $mime_is_rfc822
     condition = $mime_is_coverletter
     condition = ${if eq{$mime_content_type}{text/html}{1}{0}}
     message = HTML mail is not accepted here
.endd

.vitem &$mime_is_multipart$&
.vindex &$mime_is_multipart$&
This variable has the value 1 (true) when the current part has the main type
&"multipart"&, for example, &"multipart/alternative"& or &"multipart/mixed"&.
Since multipart entities only serve as containers for other parts, you may not
want to carry out specific actions on them.

.vitem &$mime_is_rfc822$&
.vindex &$mime_is_rfc822$&
This variable has the value 1 (true) if the current part is not a part of the
checked message itself, but part of an attached message. Attached message
decoding is fully recursive.

.vitem &$mime_part_count$&
.vindex &$mime_part_count$&
This variable is a counter that is raised for each processed MIME part. It
starts at zero for the very first part (which is usually a multipart). The
counter is per-message, so it is reset when processing RFC822 attachments (see
&$mime_is_rfc822$&). The counter stays set after &%acl_smtp_mime%& is
complete, so you can use it in the DATA ACL to determine the number of MIME
parts of a message. For non-MIME messages, this variable contains the value -1.
.endlist



.section "Scanning with regular expressions" "SECTscanregex"
.cindex "content scanning" "with regular expressions"
.cindex "regular expressions" "content scanning with"
You can specify your own custom regular expression matches on the full body of
the message, or on individual MIME parts.

The &%regex%& condition takes one or more regular expressions as arguments and
matches them against the full message (when called in the DATA ACL) or a raw
MIME part (when called in the MIME ACL). The &%regex%& condition matches
linewise, with a maximum line length of 32K characters. That means you cannot
have multiline matches with the &%regex%& condition.

The &%mime_regex%& condition can be called only in the MIME ACL. It matches up
to 32K of decoded content (the whole content at once, not linewise). If the
part has not been decoded with the &%decode%& modifier earlier in the ACL, it
is decoded automatically when &%mime_regex%& is executed (using default path
and filename values). If the decoded data is larger than  32K, only the first
32K characters are checked.

The regular expressions are passed as a colon-separated list. To include a
literal colon, you must double it. Since the whole right-hand side string is
expanded before being used, you must also escape dollar signs and backslashes
with more backslashes, or use the &`\N`& facility to disable expansion.
Here is a simple example that contains two regular expressions:
.code
deny regex = [Mm]ortgage : URGENT BUSINESS PROPOSAL
     message = contains blacklisted regex ($regex_match_string)
.endd
The conditions returns true if any one of the regular expressions matches. The
&$regex_match_string$& expansion variable is then set up and contains the
matching regular expression.
The expansion variables &$regex1$& &$regex2$& etc
are set to any substrings captured by the regular expression.

&*Warning*&: With large messages, these conditions can be fairly
CPU-intensive.

.ecindex IIDcosca




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Adding a local scan function to Exim" "CHAPlocalscan" &&&
         "Local scan function"
.scindex IIDlosca "&[local_scan()]& function" "description of"
.cindex "customizing" "input scan using C function"
.cindex "policy control" "by local scan function"
In these days of email worms, viruses, and ever-increasing spam, some sites
want to apply a lot of checking to messages before accepting them.

The content scanning extension (chapter &<<CHAPexiscan>>&) has facilities for
passing messages to external virus and spam scanning software. You can also do
a certain amount in Exim itself through string expansions and the &%condition%&
condition in the ACL that runs after the SMTP DATA command or the ACL for
non-SMTP messages (see chapter &<<CHAPACL>>&), but this has its limitations.

To allow for further customization to a site's own requirements, there is the
possibility of linking Exim with a private message scanning function, written
in C. If you want to run code that is written in something other than C, you
can of course use a little C stub to call it.

The local scan function is run once for every incoming message, at the point
when Exim is just about to accept the message.
It can therefore be used to control non-SMTP messages from local processes as
well as messages arriving via SMTP.

Exim applies a timeout to calls of the local scan function, and there is an
option called &%local_scan_timeout%& for setting it. The default is 5 minutes.
Zero means &"no timeout"&.
Exim also sets up signal handlers for SIGSEGV, SIGILL, SIGFPE, and SIGBUS
before calling the local scan function, so that the most common types of crash
are caught. If the timeout is exceeded or one of those signals is caught, the
incoming message is rejected with a temporary error if it is an SMTP message.
For a non-SMTP message, the message is dropped and Exim ends with a non-zero
code. The incident is logged on the main and reject logs.



.section "Building Exim to use a local scan function" "SECID207"
.cindex "&[local_scan()]& function" "building Exim to use"
To make use of the local scan function feature, you must tell Exim where your
function is before building Exim, by setting
both HAVE_LOCAL_SCAN and
LOCAL_SCAN_SOURCE in your
&_Local/Makefile_&. A recommended place to put it is in the &_Local_&
directory, so you might set
.code
HAVE_LOCAL_SCAN=yes
LOCAL_SCAN_SOURCE=Local/local_scan.c
.endd
for example. The function must be called &[local_scan()]&;
the source file(s) for it should first #define LOCAL_SCAN
and then #include "local_scan.h".
It is called by
Exim after it has received a message, when the success return code is about to
be sent. This is after all the ACLs have been run. The return code from your
function controls whether the message is actually accepted or not. There is a
commented template function (that just accepts the message) in the file
_src/local_scan.c_.

If you want to make use of Exim's runtime configuration file to set options
for your &[local_scan()]& function, you must also set
.code
LOCAL_SCAN_HAS_OPTIONS=yes
.endd
in &_Local/Makefile_& (see section &<<SECTconoptloc>>& below).




.section "API for local_scan()" "SECTapiforloc"
.cindex "&[local_scan()]& function" "API description"
.cindex &%dlfunc%& "API description"
You must include this line near the start of your code:
.code
#define LOCAL_SCAN
#include "local_scan.h"
.endd
This header file defines a number of variables and other values, and the
prototype for the function itself. Exim is coded to use unsigned char values
almost exclusively, and one of the things this header defines is a shorthand
for &`unsigned char`& called &`uschar`&.
It also makes available the following macro definitions, to simplify casting character
strings and pointers to character strings:
.code
#define CS   (char *)
#define CCS  (const char *)
#define CSS  (char **)
#define US   (unsigned char *)
#define CUS  (const unsigned char *)
#define USS  (unsigned char **)
.endd
The function prototype for &[local_scan()]& is:
.code
extern int local_scan(int fd, uschar **return_text);
.endd
The arguments are as follows:

.ilist
&%fd%& is a file descriptor for the file that contains the body of the message
(the -D file). The file is open for reading and writing, but updating it is not
recommended. &*Warning*&: You must &'not'& close this file descriptor.

The descriptor is positioned at character 26 of the file, which is the first
character of the body itself, because the first 26 characters (19 characters
before Exim  4.97) are the message id followed by &`-D`& and a newline.
If you rewind the file, you should use the
macro SPOOL_DATA_START_OFFSET to reset to the start of the data, just in
case this changes in some future version.
.next
&%return_text%& is an address which you can use to return a pointer to a text
string at the end of the function. The value it points to on entry is NULL.
.endlist

The function must return an &%int%& value which is one of the following macros:

.vlist
.vitem &`LOCAL_SCAN_ACCEPT`&
.vindex "&$local_scan_data$&"
The message is accepted. If you pass back a string of text, it is saved with
the message, and made available in the variable &$local_scan_data$&. No
newlines are permitted (if there are any, they are turned into spaces) and the
maximum length of text is 1000 characters.

.vitem &`LOCAL_SCAN_ACCEPT_FREEZE`&
This behaves as LOCAL_SCAN_ACCEPT, except that the accepted message is
queued without immediate delivery, and is frozen.

.vitem &`LOCAL_SCAN_ACCEPT_QUEUE`&
This behaves as LOCAL_SCAN_ACCEPT, except that the accepted message is
queued without immediate delivery.

.vitem &`LOCAL_SCAN_REJECT`&
The message is rejected; the returned text is used as an error message which is
passed back to the sender and which is also logged. Newlines are permitted &--
they cause a multiline response for SMTP rejections, but are converted to
&`\n`& in log lines. If no message is given, &"Administrative prohibition"& is
used.

.vitem &`LOCAL_SCAN_TEMPREJECT`&
The message is temporarily rejected; the returned text is used as an error
message as for LOCAL_SCAN_REJECT. If no message is given, &"Temporary local
problem"& is used.

.vitem &`LOCAL_SCAN_REJECT_NOLOGHDR`&
This behaves as LOCAL_SCAN_REJECT, except that the header of the rejected
message is not written to the reject log. It has the effect of unsetting the
&%rejected_header%& log selector for just this rejection. If
&%rejected_header%& is already unset (see the discussion of the
&%log_selection%& option in section &<<SECTlogselector>>&), this code is the
same as LOCAL_SCAN_REJECT.

.vitem &`LOCAL_SCAN_TEMPREJECT_NOLOGHDR`&
This code is a variation of LOCAL_SCAN_TEMPREJECT in the same way that
LOCAL_SCAN_REJECT_NOLOGHDR is a variation of LOCAL_SCAN_REJECT.
.endlist

If the message is not being received by interactive SMTP, rejections are
reported by writing to &%stderr%& or by sending an email, as configured by the
&%-oe%& command line options.



.section "Configuration options for local_scan()" "SECTconoptloc"
.cindex "&[local_scan()]& function" "configuration options"
It is possible to have option settings in the main configuration file
that set values in static variables in the &[local_scan()]& module. If you
want to do this, you must have the line
.code
LOCAL_SCAN_HAS_OPTIONS=yes
.endd
in your &_Local/Makefile_& when you build Exim. (This line is in
&_OS/Makefile-Default_&, commented out). Then, in the &[local_scan()]& source
file, you must define static variables to hold the option values, and a table
to define them.

The table must be a vector called &%local_scan_options%&, of type
&`optionlist`&. Each entry is a triplet, consisting of a name, an option type,
and a pointer to the variable that holds the value. The entries must appear in
alphabetical order. Following &%local_scan_options%& you must also define a
variable called &%local_scan_options_count%& that contains the number of
entries in the table. Here is a short example, showing two kinds of option:
.code
static int my_integer_option = 42;
static uschar *my_string_option = US"a default string";

optionlist local_scan_options[] = {
  { "my_integer", opt_int,       &my_integer_option },
  { "my_string",  opt_stringptr, &my_string_option }
};

int local_scan_options_count =
  sizeof(local_scan_options)/sizeof(optionlist);
.endd
The values of the variables can now be changed from Exim's runtime
configuration file by including a local scan section as in this example:
.code
begin local_scan
my_integer = 99
my_string = some string of text...
.endd
The available types of option data are as follows:

.vlist
.vitem &*opt_bool*&
This specifies a boolean (true/false) option. The address should point to a
variable of type &`BOOL`&, which will be set to TRUE or FALSE, which are macros
that are defined as &"1"& and &"0"&, respectively. If you want to detect
whether such a variable has been set at all, you can initialize it to
TRUE_UNSET. (BOOL variables are integers underneath, so can hold more than two
values.)

.vitem &*opt_fixed*&
This specifies a fixed point number, such as is used for load averages.
The address should point to a variable of type &`int`&. The value is stored
multiplied by 1000, so, for example, 1.4142 is truncated and stored as 1414.

.vitem &*opt_int*&
This specifies an integer; the address should point to a variable of type
&`int`&. The value may be specified in any of the integer formats accepted by
Exim.

.vitem &*opt_mkint*&
This is the same as &%opt_int%&, except that when such a value is output in a
&%-bP%& listing, if it is an exact number of kilobytes or megabytes, it is
printed with the suffix K or M.

.vitem &*opt_octint*&
This also specifies an integer, but the value is always interpreted as an
octal integer, whether or not it starts with the digit zero, and it is
always output in octal.

.vitem &*opt_stringptr*&
This specifies a string value; the address must be a pointer to a
variable that points to a string (for example, of type &`uschar *`&).

.vitem &*opt_time*&
This specifies a time interval value. The address must point to a variable of
type &`int`&. The value that is placed there is a number of seconds.
.endlist

If the &%-bP%& command line option is followed by &`local_scan`&, Exim prints
out the values of all the &[local_scan()]& options.



.section "Available Exim variables" "SECID208"
.cindex "&[local_scan()]& function" "available Exim variables"
The header &_local_scan.h_& gives you access to a number of C variables. These
are the only ones that are guaranteed to be maintained from release to release.
Note, however, that you can obtain the value of any Exim expansion variable,
including &$recipients$&, by calling &'expand_string()'&. The exported
C variables are as follows:

.vlist
.vitem &*int&~body_linecount*&
This variable contains the number of lines in the message's body.
It is not valid if the &%spool_wireformat%& option is used.

.vitem &*int&~body_zerocount*&
This variable contains the number of binary zero bytes in the message's body.
It is not valid if the &%spool_wireformat%& option is used.

.vitem &*unsigned&~int&~debug_selector*&
This variable is set to zero when no debugging is taking place. Otherwise, it
is a bitmap of debugging selectors. Two bits are identified for use in
&[local_scan()]&; they are defined as macros:

.ilist
The &`D_v`& bit is set when &%-v%& was present on the command line. This is a
testing option that is not privileged &-- any caller may set it. All the
other selector bits can be set only by admin users.

.next
The &`D_local_scan`& bit is provided for use by &[local_scan()]&; it is set
by the &`+local_scan`& debug selector. It is not included in the default set
of debugging bits.
.endlist ilist

Thus, to write to the debugging output only when &`+local_scan`& has been
selected, you should use code like this:
.code
if ((debug_selector & D_local_scan) != 0)
  debug_printf("xxx", ...);
.endd
.vitem &*uschar&~*expand_string_message*&
After a failing call to &'expand_string()'& (returned value NULL), the
variable &%expand_string_message%& contains the error message, zero-terminated.

.vitem &*header_line&~*header_list*&
A pointer to a chain of header lines. The &%header_line%& structure is
discussed below.

.vitem &*header_line&~*header_last*&
A pointer to the last of the header lines.

.vitem &*const&~uschar&~*headers_charset*&
The value of the &%headers_charset%& configuration option.

.vitem &*BOOL&~host_checking*&
This variable is TRUE during a host checking session that is initiated by the
&%-bh%& command line option.

.vitem &*uschar&~*interface_address*&
The IP address of the interface that received the message, as a string. This
is NULL for locally submitted messages.

.vitem &*int&~interface_port*&
The port on which this message was received. When testing with the &%-bh%&
command line option, the value of this variable is -1 unless a port has been
specified via the &%-oMi%& option.

.vitem &*uschar&~*message_id*&
This variable contains Exim's message id for the incoming message (the value of
&$message_exim_id$&) as a zero-terminated string.

.vitem &*uschar&~*received_protocol*&
The name of the protocol by which the message was received.

.vitem &*int&~recipients_count*&
The number of accepted recipients.

.vitem &*recipient_item&~*recipients_list*&
.cindex "recipient" "adding in local scan"
.cindex "recipient" "removing in local scan"
The list of accepted recipients, held in a vector of length
&%recipients_count%&. The &%recipient_item%& structure is discussed below. You
can add additional recipients by calling &'receive_add_recipient()'& (see
below). You can delete recipients by removing them from the vector and
adjusting the value in &%recipients_count%&. In particular, by setting
&%recipients_count%& to zero you remove all recipients. If you then return the
value &`LOCAL_SCAN_ACCEPT`&, the message is accepted, but immediately
blackholed. To replace the recipients, you can set &%recipients_count%& to zero
and then call &'receive_add_recipient()'& as often as needed.

.vitem &*uschar&~*sender_address*&
The envelope sender address. For bounce messages this is the empty string.

.vitem &*uschar&~*sender_host_address*&
The IP address of the sending host, as a string. This is NULL for
locally-submitted messages.

.vitem &*uschar&~*sender_host_authenticated*&
The name of the authentication mechanism that was used, or NULL if the message
was not received over an authenticated SMTP connection.

.vitem &*uschar&~*sender_host_name*&
The name of the sending host, if known.

.vitem &*int&~sender_host_port*&
The port on the sending host.

.vitem &*BOOL&~smtp_input*&
This variable is TRUE for all SMTP input, including BSMTP.

.vitem &*BOOL&~smtp_batched_input*&
This variable is TRUE for BSMTP input.

.vitem &*int&~store_pool*&
The contents of this variable control which pool of memory is used for new
requests. See section &<<SECTmemhanloc>>& for details.
.endlist


.section "Structure of header lines" "SECID209"
The &%header_line%& structure contains the members listed below.
You can add additional header lines by calling the &'header_add()'& function
(see below). You can cause header lines to be ignored (deleted) by setting
their type to *.


.vlist
.vitem &*struct&~header_line&~*next*&
A pointer to the next header line, or NULL for the last line.

.vitem &*int&~type*&
A code identifying certain headers that Exim recognizes. The codes are printing
characters, and are documented in chapter &<<CHAPspool>>& of this manual.
Notice in particular that any header line whose type is * is not transmitted
with the message. This flagging is used for header lines that have been
rewritten, or are to be removed (for example, &'Envelope-sender:'& header
lines.) Effectively, * means &"deleted"&.

.vitem &*int&~slen*&
The number of characters in the header line, including the terminating and any
internal newlines.

.vitem &*uschar&~*text*&
A pointer to the text of the header. It always ends with a newline, followed by
a zero byte. Internal newlines are preserved.
.endlist



.section "Structure of recipient items" "SECID210"
The &%recipient_item%& structure contains these members:

.vlist
.vitem &*uschar&~*address*&
This is a pointer to the recipient address as it was received.

.vitem &*int&~pno*&
This is used in later Exim processing when top level addresses are created by
the &%one_time%& option. It is not relevant at the time &[local_scan()]& is run
and must always contain -1 at this stage.

.vitem &*uschar&~*errors_to*&
If this value is not NULL, bounce messages caused by failing to deliver to the
recipient are sent to the address it contains. In other words, it overrides the
envelope sender for this one recipient. (Compare the &%errors_to%& generic
router option.) If a &[local_scan()]& function sets an &%errors_to%& field to
an unqualified address, Exim qualifies it using the domain from
&%qualify_recipient%&. When &[local_scan()]& is called, the &%errors_to%& field
is NULL for all recipients.
.endlist



.section "Available Exim functions" "SECID211"
.cindex "&[local_scan()]& function" "available Exim functions"
The header &_local_scan.h_& gives you access to a number of Exim functions.
These are the only ones that are guaranteed to be maintained from release to
release:

.vlist
.vitem "&*pid_t&~child_open(uschar&~**argv,&~uschar&~**envp,&~int&~newumask,&&&
       &~int&~*infdptr,&~int&~*outfdptr, &~&~BOOL&~make_leader)*&"

This function creates a child process that runs the command specified by
&%argv%&. The environment for the process is specified by &%envp%&, which can
be NULL if no environment variables are to be passed. A new umask is supplied
for the process in &%newumask%&.

Pipes to the standard input and output of the new process are set up
and returned to the caller via the &%infdptr%& and &%outfdptr%& arguments. The
standard error is cloned to the standard output. If there are any file
descriptors &"in the way"& in the new process, they are closed. If the final
argument is TRUE, the new process is made into a process group leader.

The function returns the pid of the new process, or -1 if things go wrong.

.vitem &*int&~child_close(pid_t&~pid,&~int&~timeout)*&
This function waits for a child process to terminate, or for a timeout (in
seconds) to expire. A timeout value of zero means wait as long as it takes. The
return value is as follows:

.ilist
>= 0

The process terminated by a normal exit and the value is the process
ending status.

.next
< 0 and > &--256

The process was terminated by a signal and the value is the negation of the
signal number.

.next
&--256

The process timed out.
.next
&--257

The was some other error in wait(); &%errno%& is still set.
.endlist

.vitem &*pid_t&~child_open_exim(int&~*fd)*&
This function provide you with a means of submitting a new message to
Exim. (Of course, you can also call &_/usr/sbin/sendmail_& yourself if you
want, but this packages it all up for you.) The function creates a pipe,
forks a subprocess that is running
.code
exim -t -oem -oi -f <>
.endd
and returns to you (via the &`int *`& argument) a file descriptor for the pipe
that is connected to the standard input. The yield of the function is the PID
of the subprocess. You can then write a message to the file descriptor, with
recipients in &'To:'&, &'Cc:'&, and/or &'Bcc:'& header lines.

When you have finished, call &'child_close()'& to wait for the process to
finish and to collect its ending status. A timeout value of zero is usually
fine in this circumstance. Unless you have made a mistake with the recipient
addresses, you should get a return code of zero.


.vitem &*pid_t&~child_open_exim2(int&~*fd,&~uschar&~*sender,&~uschar&~&&&
       *sender_authentication)*&
This function is a more sophisticated version of &'child_open()'&. The command
that it runs is:
.display
&`exim -t -oem -oi -f `&&'sender'&&` -oMas `&&'sender_authentication'&
.endd
The third argument may be NULL, in which case the &%-oMas%& option is omitted.


.vitem &*void&~debug_printf(char&~*,&~...)*&
This is Exim's debugging function, with arguments as for &'printf()'&. The
output is written to the standard error stream. If no debugging is selected,
calls to &'debug_printf()'& have no effect. Normally, you should make calls
conditional on the &`local_scan`& debug selector by coding like this:
.code
if ((debug_selector & D_local_scan) != 0)
  debug_printf("xxx", ...);
.endd

.vitem &*uschar&~*expand_string(uschar&~*string)*&
This is an interface to Exim's string expansion code. The return value is the
expanded string, or NULL if there was an expansion failure.
The C variable &%expand_string_message%& contains an error message after an
expansion failure. If expansion does not change the string, the return value is
the pointer to the input string. Otherwise, the return value points to a new
block of memory that was obtained by a call to &'store_get()'&. See section
&<<SECTmemhanloc>>& below for a discussion of memory handling.

.vitem &*void&~header_add(int&~type,&~char&~*format,&~...)*&
This function allows you to an add additional header line at the end of the
existing ones. The first argument is the type, and should normally be a space
character. The second argument is a format string and any number of
substitution arguments as for &[sprintf()]&. You may include internal newlines
if you want, and you must ensure that the string ends with a newline.

.vitem "&*void&~header_add_at_position(BOOL&~after,&~uschar&~*name,&~&&&
        BOOL&~topnot,&~int&~type,&~char&~*format, &~&~...)*&"
This function adds a new header line at a specified point in the header
chain. The header itself is specified as for &'header_add()'&.

If &%name%& is NULL, the new header is added at the end of the chain if
&%after%& is true, or at the start if &%after%& is false. If &%name%& is not
NULL, the header lines are searched for the first non-deleted header that
matches the name. If one is found, the new header is added before it if
&%after%& is false. If &%after%& is true, the new header is added after the
found header and any adjacent subsequent ones with the same name (even if
marked &"deleted"&). If no matching non-deleted header is found, the &%topnot%&
option controls where the header is added. If it is true, addition is at the
top; otherwise at the bottom. Thus, to add a header after all the &'Received:'&
headers, or at the top if there are no &'Received:'& headers, you could use
.code
header_add_at_position(TRUE, US"Received", TRUE,
  ' ', "X-xxx: ...");
.endd
Normally, there is always at least one non-deleted &'Received:'& header, but
there may not be if &%received_header_text%& expands to an empty string.


.vitem &*void&~header_remove(int&~occurrence,&~uschar&~*name)*&
This function removes header lines. If &%occurrence%& is zero or negative, all
occurrences of the header are removed. If occurrence is greater than zero, that
particular instance of the header is removed. If no header(s) can be found that
match the specification, the function does nothing.


.vitem "&*BOOL&~header_testname(header_line&~*hdr,&~uschar&~*name,&~&&&
        int&~length,&~BOOL&~notdel)*&"
This function tests whether the given header has the given name. It is not just
a string comparison, because white space is permitted between the name and the
colon. If the &%notdel%& argument is true, a false return is forced for all
&"deleted"& headers; otherwise they are not treated specially. For example:
.code
if (header_testname(h, US"X-Spam", 6, TRUE)) ...
.endd
.vitem &*uschar&~*lss_b64encode(uschar&~*cleartext,&~int&~length)*&
.cindex "base64 encoding" "functions for &[local_scan()]& use"
This function base64-encodes a string, which is passed by address and length.
The text may contain bytes of any value, including zero. The result is passed
back in dynamic memory that is obtained by calling &'store_get()'&. It is
zero-terminated.

.vitem &*int&~lss_b64decode(uschar&~*codetext,&~uschar&~**cleartext)*&
This function decodes a base64-encoded string. Its arguments are a
zero-terminated base64-encoded string and the address of a variable that is set
to point to the result, which is in dynamic memory. The length of the decoded
string is the yield of the function. If the input is invalid base64 data, the
yield is -1. A zero byte is added to the end of the output string to make it
easy to interpret as a C string (assuming it contains no zeros of its own). The
added zero byte is not included in the returned count.

.vitem &*int&~lss_match_domain(uschar&~*domain,&~uschar&~*list)*&
This function checks for a match in a domain list. Domains are always
matched caselessly. The return value is one of the following:
.itable none 0 0 2 15* left 85* left
.irow &`OK`&    "match succeeded"
.irow &`FAIL`&  "match failed"
.irow &`DEFER`& "match deferred"
.endtable
DEFER is usually caused by some kind of lookup defer, such as the
inability to contact a database.

.vitem "&*int&~lss_match_local_part(uschar&~*localpart,&~uschar&~*list,&~&&&
        BOOL&~caseless)*&"
This function checks for a match in a local part list. The third argument
controls case-sensitivity. The return values are as for
&'lss_match_domain()'&.

.vitem "&*int&~lss_match_address(uschar&~*address,&~uschar&~*list,&~&&&
        BOOL&~caseless)*&"
This function checks for a match in an address list. The third argument
controls the case-sensitivity of the local part match. The domain is always
matched caselessly. The return values are as for &'lss_match_domain()'&.

.vitem "&*int&~lss_match_host(uschar&~*host_name,&~uschar&~*host_address,&~&&&
        uschar&~*list)*&"
This function checks for a match in a host list. The most common usage is
expected to be
.code
lss_match_host(sender_host_name, sender_host_address, ...)
.endd
.vindex "&$sender_host_address$&"
An empty address field matches an empty item in the host list. If the host name
is NULL, the name corresponding to &$sender_host_address$& is automatically
looked up if a host name is required to match an item in the list. The return
values are as for &'lss_match_domain()'&, but in addition, &'lss_match_host()'&
returns ERROR in the case when it had to look up a host name, but the lookup
failed.

.vitem "&*void&~log_write(unsigned&~int&~selector,&~int&~which,&~char&~&&&
        *format,&~...)*&"
This function writes to Exim's log files. The first argument should be zero (it
is concerned with &%log_selector%&). The second argument can be &`LOG_MAIN`& or
&`LOG_REJECT`& or &`LOG_PANIC`& or the inclusive &"or"& of any combination of
them. It specifies to which log or logs the message is written. The remaining
arguments are a format and relevant insertion arguments. The string should not
contain any newlines, not even at the end.


.vitem &*void&~receive_add_recipient(uschar&~*address,&~int&~pno)*&
This function adds an additional recipient to the message. The first argument
is the recipient address. If it is unqualified (has no domain), it is qualified
with the &%qualify_recipient%& domain. The second argument must always be -1.

This function does not allow you to specify a private &%errors_to%& address (as
described with the structure of &%recipient_item%& above), because it pre-dates
the addition of that field to the structure. However, it is easy to add such a
value afterwards. For example:
.code
 receive_add_recipient(US"monitor@mydom.example", -1);
 recipients_list[recipients_count-1].errors_to =
   US"postmaster@mydom.example";
.endd

.vitem &*BOOL&~receive_remove_recipient(uschar&~*recipient)*&
This is a convenience function to remove a named recipient from the list of
recipients. It returns true if a recipient was removed, and false if no
matching recipient could be found. The argument must be a complete email
address.
.endlist


.cindex "RFC 2047"
.vlist
.vitem "&*uschar&~rfc2047_decode(uschar&~*string,&~BOOL&~lencheck,&&&
  &~uschar&~*target,&~int&~zeroval,&~int&~*lenptr, &~&~uschar&~**error)*&"
This function decodes strings that are encoded according to
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047). Typically
these are the contents of header lines. First, each &"encoded word"& is decoded
from the Q or B encoding into a byte-string. Then, if provided with the name of
a charset encoding, and if the &[iconv()]& function is available, an attempt is
made  to translate the result to the named character set. If this fails, the
binary string is returned with an error message.

The first argument is the string to be decoded. If &%lencheck%& is TRUE, the
maximum MIME word length is enforced. The third argument is the target
encoding, or NULL if no translation is wanted.

.cindex "binary zero" "in RFC 2047 decoding"
.cindex "RFC 2047" "binary zero in"
If a binary zero is encountered in the decoded string, it is replaced by the
contents of the &%zeroval%& argument. For use with Exim headers, the value must
not be 0 because header lines are handled as zero-terminated strings.

The function returns the result of processing the string, zero-terminated; if
&%lenptr%& is not NULL, the length of the result is set in the variable to
which it points. When &%zeroval%& is 0, &%lenptr%& should not be NULL.

If an error is encountered, the function returns NULL and uses the &%error%&
argument to return an error message. The variable pointed to by &%error%& is
set to NULL if there is no error; it may be set non-NULL even when the function
returns a non-NULL value if decoding was successful, but there was a problem
with translation.


.vitem &*int&~smtp_fflush(void)*&
This function is used in conjunction with &'smtp_printf()'&, as described
below.

.vitem &*void&~smtp_printf(char&~*,BOOL,&~...)*&
The arguments of this function are almost like &[printf()]&; it writes to the SMTP
output stream. You should use this function only when there is an SMTP output
stream, that is, when the incoming message is being received via interactive
SMTP. This is the case when &%smtp_input%& is TRUE and &%smtp_batched_input%&
is FALSE. If you want to test for an incoming message from another host (as
opposed to a local process that used the &%-bs%& command line option), you can
test the value of &%sender_host_address%&, which is non-NULL when a remote host
is involved.

If an SMTP TLS connection is established, &'smtp_printf()'& uses the TLS
output function, so it can be used for all forms of SMTP connection.

The second argument is used to request that the data be buffered
(when TRUE) or flushed (along with any previously buffered, when FALSE).
This is advisory only, but likely to save on system-calls and packets
sent when a sequence of calls to the function are made.

The argument was added in Exim version 4.90 - changing the API/ABI.
Nobody noticed until 4.93 was imminent, at which point the
ABI version number was incremented.

Strings that are written by &'smtp_printf()'& from within &[local_scan()]&
must start with an appropriate response code: 550 if you are going to return
LOCAL_SCAN_REJECT, 451 if you are going to return
LOCAL_SCAN_TEMPREJECT, and 250 otherwise. Because you are writing the
initial lines of a multi-line response, the code must be followed by a hyphen
to indicate that the line is not the final response line. You must also ensure
that the lines you write terminate with CRLF. For example:
.code
smtp_printf("550-this is some extra info\r\n");
return LOCAL_SCAN_REJECT;
.endd
Note that you can also create multi-line responses by including newlines in
the data returned via the &%return_text%& argument. The added value of using
&'smtp_printf()'& is that, for instance, you could introduce delays between
multiple output lines.

The &'smtp_printf()'& function does not return any error indication, because it
does not
guarantee a flush of
pending output, and therefore does not test
the state of the stream. (In the main code of Exim, flushing and error
detection is done when Exim is ready for the next SMTP input command.) If
you want to flush the output and check for an error (for example, the
dropping of a TCP/IP connection), you can call &'smtp_fflush()'&, which has no
arguments. It flushes the output stream, and returns a non-zero value if there
is an error.

.vitem &*void&~*store_get(int,BOOL)*&
This function accesses Exim's internal store (memory) manager. It gets a new
chunk of memory whose size is given by the first argument.
The second argument should be given as TRUE if the memory will be used for
data possibly coming from an attacker (eg. the message content),
FALSE if it is locally-sourced.
Exim bombs out if it ever
runs out of memory. See the next section for a discussion of memory handling.

.vitem &*void&~*store_get_perm(int,BOOL)*&
This function is like &'store_get()'&, but it always gets memory from the
permanent pool. See the next section for a discussion of memory handling.

.vitem &*uschar&~*string_copy(uschar&~*string)*&
See below.

.vitem &*uschar&~*string_copyn(uschar&~*string,&~int&~length)*&
See below.

.vitem &*uschar&~*string_sprintf(char&~*format,&~...)*&
These three functions create strings using Exim's dynamic memory facilities.
The first makes a copy of an entire string. The second copies up to a maximum
number of characters, indicated by the second argument. The third uses a format
and insertion arguments to create a new string. In each case, the result is a
pointer to a new string in the current memory pool. See the next section for
more discussion.
.endlist



.section "More about Exim's memory handling" "SECTmemhanloc"
.cindex "&[local_scan()]& function" "memory handling"
No function is provided for freeing memory, because that is never needed.
The dynamic memory that Exim uses when receiving a message is automatically
recycled if another message is received by the same process (this applies only
to incoming SMTP connections &-- other input methods can supply only one
message at a time). After receiving the last message, a reception process
terminates.

Because it is recycled, the normal dynamic memory cannot be used for holding
data that must be preserved over a number of incoming messages on the same SMTP
connection. However, Exim in fact uses two pools of dynamic memory; the second
one is not recycled, and can be used for this purpose.

If you want to allocate memory that remains available for subsequent messages
in the same SMTP connection, you should set
.code
store_pool = POOL_PERM
.endd
before calling the function that does the allocation. There is no need to
restore the value if you do not need to; however, if you do want to revert to
the normal pool, you can either restore the previous value of &%store_pool%& or
set it explicitly to POOL_MAIN.

The pool setting applies to all functions that get dynamic memory, including
&'expand_string()'&, &'store_get()'&, and the &'string_xxx()'& functions.
There is also a convenience function called &'store_get_perm()'& that gets a
block of memory from the permanent pool while preserving the value of
&%store_pool%&.
.ecindex IIDlosca




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "System-wide message filtering" "CHAPsystemfilter"
.scindex IIDsysfil1 "filter" "system filter"
.scindex IIDsysfil2 "filtering all mail"
.scindex IIDsysfil3 "system filter"
The previous chapters (on ACLs and the local scan function) describe checks
that can be applied to messages before they are accepted by a host. There is
also a mechanism for checking messages once they have been received, but before
they are delivered. This is called the &'system filter'&.

The system filter operates in a similar manner to users' filter files, but it
is run just once per message (however many recipients the message has).
It should not normally be used as a substitute for routing, because &%deliver%&
commands in a system router provide new envelope recipient addresses.
The system filter must be an Exim filter. It cannot be a Sieve filter.

The system filter is run at the start of a delivery attempt, before any routing
is done. If a message fails to be completely delivered at the first attempt,
the system filter is run again at the start of every retry.
If you want your filter to do something only once per message, you can make use
.cindex retry condition
of the &%first_delivery%& condition in an &%if%& command in the filter to
prevent it happening on retries.

.vindex "&$domain$&"
.vindex "&$local_part$&"
&*Warning*&: Because the system filter runs just once, variables that are
specific to individual recipient addresses, such as &$local_part$& and
&$domain$&, are not set, and the &"personal"& condition is not meaningful. If
you want to run a centrally-specified filter for each recipient address
independently, you can do so by setting up a suitable &(redirect)& router, as
described in section &<<SECTperaddfil>>& below.


.section "Specifying a system filter" "SECID212"
.cindex "uid (user id)" "system filter"
.cindex "gid (group id)" "system filter"
The name of the file that contains the system filter must be specified by
setting &%system_filter%&. If you want the filter to run under a uid and gid
other than root, you must also set &%system_filter_user%& and
&%system_filter_group%& as appropriate. For example:
.code
system_filter = /etc/mail/exim.filter
system_filter_user = exim
.endd
If a system filter generates any deliveries directly to files or pipes (via the
&%save%& or &%pipe%& commands), transports to handle these deliveries must be
specified by setting &%system_filter_file_transport%& and
&%system_filter_pipe_transport%&, respectively. Similarly,
&%system_filter_reply_transport%& must be set to handle any messages generated
by the &%reply%& command.


.section "Testing a system filter" "SECID213"
You can run simple tests of a system filter in the same way as for a user
filter, but you should use &%-bF%& rather than &%-bf%&, so that features that
are permitted only in system filters are recognized.

If you want to test the combined effect of a system filter and a user filter,
you can use both &%-bF%& and &%-bf%& on the same command line.



.section "Contents of a system filter" "SECID214"
The language used to specify system filters is the same as for users' filter
files. It is described in the separate end-user document &'Exim's interface to
mail filtering'&. However, there are some additional features that are
available only in system filters; these are described in subsequent sections.
If they are encountered in a user's filter file or when testing with &%-bf%&,
they cause errors.

.cindex "frozen messages" "manual thaw; testing in filter"
There are two special conditions which, though available in users' filter
files, are designed for use in system filters. The condition &%first_delivery%&
is true only for the first attempt at delivering a message, and
&%manually_thawed%& is true only if the message has been frozen, and
subsequently thawed by an admin user. An explicit forced delivery counts as a
manual thaw, but thawing as a result of the &%auto_thaw%& setting does not.

&*Warning*&: If a system filter uses the &%first_delivery%& condition to
specify an &"unseen"& (non-significant) delivery, and that delivery does not
succeed, it will not be tried again.
If you want Exim to retry an unseen delivery until it succeeds, you should
arrange to set it up every time the filter runs.

When a system filter finishes running, the values of the variables &$n0$& &--
&$n9$& are copied into &$sn0$& &-- &$sn9$& and are thereby made available to
users' filter files. Thus a system filter can, for example, set up &"scores"&
to which users' filter files can refer.



.section "Additional variable for system filters" "SECID215"
.vindex "&$recipients$&"
The expansion variable &$recipients$&, containing a list of all the recipients
of the message (separated by commas and white space), is available in system
filters. It is not available in users' filters for privacy reasons.



.section "Defer, freeze, and fail commands for system filters" "SECID216"
.cindex "freezing messages"
.cindex "message" "freezing"
.cindex "message" "forced failure"
.cindex "&%fail%&" "in system filter"
.cindex "&%freeze%& in system filter"
.cindex "&%defer%& in system filter"
There are three extra commands (&%defer%&, &%freeze%& and &%fail%&) which are
always available in system filters, but are not normally enabled in users'
filters. (See the &%allow_defer%&, &%allow_freeze%& and &%allow_fail%& options
for the &(redirect)& router.) These commands can optionally be followed by the
word &%text%& and a string containing an error message, for example:
.code
fail text "this message looks like spam to me"
.endd
The keyword &%text%& is optional if the next character is a double quote.

The &%defer%& command defers delivery of the original recipients of the
message. The &%fail%& command causes all the original recipients to be failed,
and a bounce message to be created. The &%freeze%& command suspends all
delivery attempts for the original recipients. In all cases, any new deliveries
that are specified by the filter are attempted as normal after the filter has
run.

The &%freeze%& command is ignored if the message has been manually unfrozen and
not manually frozen since. This means that automatic freezing by a system
filter can be used as a way of checking out suspicious messages. If a message
is found to be all right, manually unfreezing it allows it to be delivered.

.cindex "log" "&%fail%& command log line"
.cindex "&%fail%&" "log line; reducing"
The text given with a fail command is used as part of the bounce message as
well as being written to the log. If the message is quite long, this can fill
up a lot of log space when such failures are common. To reduce the size of the
log message, Exim interprets the text in a special way if it starts with the
two characters &`<<`& and contains &`>>`& later. The text between these two
strings is written to the log, and the rest of the text is used in the bounce
message. For example:
.code
fail "<<filter test 1>>Your message is rejected \
     because it contains attachments that we are \
     not prepared to receive."
.endd

.cindex "loop" "caused by &%fail%&"
Take great care with the &%fail%& command when basing the decision to fail on
the contents of the message, because the bounce message will of course include
the contents of the original message and will therefore trigger the &%fail%&
command again (causing a mail loop) unless steps are taken to prevent this.
Testing the &%error_message%& condition is one way to prevent this. You could
use, for example
.code
if $message_body contains "this is spam" and not error_message
then fail text "spam is not wanted here" endif
.endd
though of course that might let through unwanted bounce messages. The
alternative is clever checking of the body and/or headers to detect bounces
generated by the filter.

The interpretation of a system filter file ceases after a
&%defer%&,
&%freeze%&, or &%fail%& command is obeyed. However, any deliveries that were
set up earlier in the filter file are honoured, so you can use a sequence such
as
.code
mail ...
freeze
.endd
to send a specified message when the system filter is freezing (or deferring or
failing) a message. The normal deliveries for the message do not, of course,
take place.



.section "Adding and removing headers in a system filter" "SECTaddremheasys"
.cindex "header lines" "adding; in system filter"
.cindex "header lines" "removing; in system filter"
.cindex "filter" "header lines; adding/removing"
Two filter commands that are available only in system filters are:
.code
headers add <string>
headers remove <string>
.endd
The argument for the &%headers add%& is a string that is expanded and then
added to the end of the message's headers. It is the responsibility of the
filter maintainer to make sure it conforms to
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822) syntax. Leading white
space is ignored, and if the string is otherwise empty, or if the expansion is
forced to fail, the command has no effect.

You can use &"\n"& within the string, followed by white space, to specify
continued header lines. More than one header may be added in one command by
including &"\n"& within the string without any following white space. For
example:
.code
headers add "X-header-1: ....\n  \
             continuation of X-header-1 ...\n\
             X-header-2: ...."
.endd
Note that the header line continuation white space after the first newline must
be placed before the backslash that continues the input string, because white
space after input continuations is ignored.

The argument for &%headers remove%& is a colon-separated list of header names.
This command applies only to those headers that are stored with the message;
those that are added at delivery time (such as &'Envelope-To:'& and
&'Return-Path:'&) cannot be removed by this means. If there is more than one
header with the same name, they are all removed.

The &%headers%& command in a system filter makes an immediate change to the set
of header lines that was received with the message (with possible additions
from ACL processing). Subsequent commands in the system filter operate on the
modified set, which also forms the basis for subsequent message delivery.
Unless further modified during routing or transporting, this set of headers is
used for all recipients of the message.

During routing and transporting, the variables that refer to the contents of
header lines refer only to those lines that are in this set. Thus, header lines
that are added by a system filter are visible to users' filter files and to all
routers and transports. This contrasts with the manipulation of header lines by
routers and transports, which is not immediate, but which instead is saved up
until the message is actually being written (see section
&<<SECTheadersaddrem>>&).

If the message is not delivered at the first attempt, header lines that were
added by the system filter are stored with the message, and so are still
present at the next delivery attempt. Header lines that were removed are still
present, but marked &"deleted"& so that they are not transported with the
message. For this reason, it is usual to make the &%headers%& command
conditional on &%first_delivery%& so that the set of header lines is not
modified more than once.

Because header modification in a system filter acts immediately, you have to
use an indirect approach if you want to modify the contents of a header line.
For example:
.code
headers add "Old-Subject: $h_subject:"
headers remove "Subject"
headers add "Subject: new subject (was: $h_old-subject:)"
headers remove "Old-Subject"
.endd



.section "Setting an errors address in a system filter" "SECID217"
.cindex "envelope from"
.cindex "envelope sender"
In a system filter, if a &%deliver%& command is followed by
.code
errors_to <some address>
.endd
in order to change the envelope sender (and hence the error reporting) for that
delivery, any address may be specified. (In a user filter, only the current
user's address can be set.) For example, if some mail is being monitored, you
might use
.code
unseen deliver monitor@spying.example errors_to root@local.example
.endd
to take a copy which would not be sent back to the normal error reporting
address if its delivery failed.



.section "Per-address filtering" "SECTperaddfil"
.vindex "&$domain_data$&"
.vindex "&$local_part_data$&"
In contrast to the system filter, which is run just once per message for each
delivery attempt, it is also possible to set up a system-wide filtering
operation that runs once for each recipient address. In this case, variables
such as &$local_part_data$& and &$domain_data$& can be used,
and indeed, the choice of filter file could be made dependent on them.
This is an example of a router which implements such a filter:
.code
central_filter:
  check_local_user
  driver = redirect
  domains = +local_domains
  file = /central/filters/$local_part_data
  no_verify
  allow_filter
  allow_freeze
.endd
The filter is run in a separate process under its own uid. Therefore, either
&%check_local_user%& must be set (as above), in which case the filter is run as
the local user, or the &%user%& option must be used to specify which user to
use. If both are set, &%user%& overrides.

Care should be taken to ensure that none of the commands in the filter file
specify a significant delivery if the message is to go on to be delivered to
its intended recipient. The router will not then claim to have dealt with the
address, so it will be passed on to subsequent routers to be delivered in the
normal way.
.ecindex IIDsysfil1
.ecindex IIDsysfil2
.ecindex IIDsysfil3






. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Message processing" "CHAPmsgproc"
.scindex IIDmesproc "message" "general processing"
Exim performs various transformations on the sender and recipient addresses of
all messages that it handles, and also on the messages' header lines. Some of
these are optional and configurable, while others always take place. All of
this processing, except rewriting as a result of routing, and the addition or
removal of header lines while delivering, happens when a message is received,
before it is placed on Exim's queue.

Some of the automatic processing takes place by default only for
&"locally-originated"& messages. This adjective is used to describe messages
that are not received over TCP/IP, but instead are passed to an Exim process on
its standard input. This includes the interactive &"local SMTP"& case that is
set up by the &%-bs%& command line option.

&*Note*&: Messages received over TCP/IP on the loopback interface (127.0.0.1
or ::1) are not considered to be locally-originated. Exim does not treat the
loopback interface specially in any way.

If you want the loopback interface to be treated specially, you must ensure
that there are appropriate entries in your ACLs.




.section "Submission mode for non-local messages" "SECTsubmodnon"
.cindex "message" "submission"
.cindex "submission mode"
Processing that happens automatically for locally-originated messages (unless
&%suppress_local_fixups%& is set) can also be requested for messages that are
received over TCP/IP. The term &"submission mode"& is used to describe this
state. Submission mode is set by the modifier
.code
control = submission
.endd
in a MAIL, RCPT, or pre-data ACL for an incoming message (see sections
&<<SECTACLmodi>>& and &<<SECTcontrols>>&). This makes Exim treat the message as
a local submission, and is normally used when the source of the message is
known to be an MUA running on a client host (as opposed to an MTA). For
example, to set submission mode for messages originating on the IPv4 loopback
interface, you could include the following in the MAIL ACL:
.code
warn  hosts = 127.0.0.1
      control = submission
.endd
.cindex "&%sender_retain%& submission option"
There are some options that can be used when setting submission mode. A slash
is used to separate options. For example:
.code
control = submission/sender_retain
.endd
Specifying &%sender_retain%& has the effect of setting &%local_sender_retain%&
true and &%local_from_check%& false for the current incoming message. The first
of these allows an existing &'Sender:'& header in the message to remain, and
the second suppresses the check to ensure that &'From:'& matches the
authenticated sender. With this setting, Exim still fixes up messages by adding
&'Date:'& and &'Message-ID:'& header lines if they are missing, but makes no
attempt to check sender authenticity in header lines.

When &%sender_retain%& is not set, a submission mode setting may specify a
domain to be used when generating a &'From:'& or &'Sender:'& header line. For
example:
.code
control = submission/domain=some.domain
.endd
The domain may be empty. How this value is used is described in sections
&<<SECTthefrohea>>& and &<<SECTthesenhea>>&. There is also a &%name%& option
that allows you to specify the user's full name for inclusion in a created
&'Sender:'& or &'From:'& header line. For example:
.code
accept authenticated = *
       control = submission/domain=wonderland.example/\
                            name=${lookup {$authenticated_id} \
                                   lsearch {/etc/exim/namelist}}
.endd
Because the name may contain any characters, including slashes, the &%name%&
option must be given last. The remainder of the string is used as the name. For
the example above, if &_/etc/exim/namelist_& contains:
.code
bigegg:  Humpty Dumpty
.endd
then when the sender has authenticated as &'bigegg'&, the generated &'Sender:'&
line would be:
.code
Sender: Humpty Dumpty <bigegg@wonderland.example>
.endd
.cindex "return path" "in submission mode"
By default, submission mode forces the return path to the same address as is
used to create the &'Sender:'& header. However, if &%sender_retain%& is
specified, the return path is also left unchanged.

&*Note*&: The changes caused by submission mode take effect after the predata
ACL. This means that any sender checks performed before the fix-ups use the
untrusted sender address specified by the user, not the trusted sender address
specified by submission mode. Although this might be slightly unexpected, it
does mean that you can configure ACL checks to spot that a user is trying to
spoof another's address.

.section "Line endings" "SECTlineendings"
.cindex "line endings"
.cindex "carriage return"
.cindex "linefeed"
&url(https://www.rfc-editor.org/rfc/rfc2821,RFC 2821)
specifies that CRLF (two characters: carriage-return, followed by
linefeed) is the line ending for messages transmitted over the Internet using
SMTP over TCP/IP. However, within individual operating systems, different
conventions are used. For example, Unix-like systems use just LF, but others
use CRLF or just CR.

Exim was designed for Unix-like systems, and internally, it stores messages
using the system's convention of a single LF as a line terminator. When
receiving a message, all line endings are translated to this standard format.
Originally, it was thought that programs that passed messages directly to an
MTA within an operating system would use that system's convention. Experience
has shown that this is not the case; for example, there are Unix applications
that use CRLF in this circumstance. For this reason, and for compatibility with
other MTAs, the way Exim handles line endings for all messages is now as
follows:

.ilist
CR is treated as a line ending; if it is immediately followed by LF, the LF
is ignored.
.next
The sequence &"CR, dot, CR"& does not terminate an incoming SMTP message,
nor a local message in the state where a line containing only a dot is a
terminator.
.next
If a bare CR is encountered within a header line, an extra space is added after
the line terminator so as not to end the header line. The reasoning behind this
is that bare CRs in header lines are most likely either to be mistakes, or
people trying to play silly games.
.next
If the first header line received in a message ends with CRLF, a subsequent
bare LF in a header line is treated in the same way as a bare CR in a header
line and a bare LF in a body line is replaced with a space.
.next
If the first header line received in a message does not end with CRLF, a subsequent
LF not preceded by CR is treated as a line ending.
.endlist





.section "Unqualified addresses" "SECID218"
.cindex "unqualified addresses"
.cindex "address" "qualification"
By default, Exim expects every envelope address it receives from an external
host to be fully qualified. Unqualified addresses cause negative responses to
SMTP commands. However, because SMTP is used as a means of transporting
messages from MUAs running on personal workstations, there is sometimes a
requirement to accept unqualified addresses from specific hosts or IP networks.

Exim has two options that separately control which hosts may send unqualified
sender or recipient addresses in SMTP commands, namely
&%sender_unqualified_hosts%& and &%recipient_unqualified_hosts%&. In both
cases, if an unqualified address is accepted, it is qualified by adding the
value of &%qualify_domain%& or &%qualify_recipient%&, as appropriate.

.oindex "&%qualify_domain%&"
.oindex "&%qualify_recipient%&"
Unqualified addresses in header lines are automatically qualified for messages
that are locally originated, unless the &%-bnq%& option is given on the command
line. For messages received over SMTP, unqualified addresses in header lines
are qualified only if unqualified addresses are permitted in SMTP commands. In
other words, such qualification is also controlled by
&%sender_unqualified_hosts%& and &%recipient_unqualified_hosts%&,




.section "The UUCP From line" "SECID219"
.cindex "&""From""& line"
.cindex "UUCP" "&""From""& line"
.cindex "sender" "address"
.oindex "&%uucp_from_pattern%&"
.oindex "&%uucp_from_sender%&"
.cindex "envelope from"
.cindex "envelope sender"
.cindex "Sendmail compatibility" "&""From""& line"
Messages that have come from UUCP (and some other applications) often begin
with a line containing the envelope sender and a timestamp, following the word
&"From"&. Examples of two common formats are:
.code
From a.oakley@berlin.mus Fri Jan  5 12:35 GMT 1996
From f.butler@berlin.mus Fri, 7 Jan 97 14:00:00 GMT
.endd
This line precedes the
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
header lines. For compatibility with Sendmail,
Exim recognizes such lines at the start of messages that are submitted to it
via the command line (that is, on the standard input). It does not recognize
such lines in incoming SMTP messages, unless the sending host matches
&%ignore_fromline_hosts%& or the &%-bs%& option was used for a local message
and &%ignore_fromline_local%& is set. The recognition is controlled by a
regular expression that is defined by the &%uucp_from_pattern%& option, whose
default value matches the two common cases shown above and puts the address
that follows &"From"& into &$1$&.

.cindex "numerical variables (&$1$& &$2$& etc)" "in &""From ""& line handling"
When the caller of Exim for a non-SMTP message that contains a &"From"& line is
a trusted user, the message's sender address is constructed by expanding the
contents of &%uucp_sender_address%&, whose default value is &"$1"&. This is
then parsed as an
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
address. If there is no domain, the local part is
qualified with &%qualify_domain%& unless it is the empty string. However, if
the command line &%-f%& option is used, it overrides the &"From"& line.

If the caller of Exim is not trusted, the &"From"& line is recognized, but the
sender address is not changed. This is also the case for incoming SMTP messages
that are permitted to contain &"From"& lines.

Only one &"From"& line is recognized. If there is more than one, the second is
treated as a data line that starts the body of the message, as it is not valid
as a header line. This also happens if a &"From"& line is present in an
incoming SMTP message from a source that is not permitted to send them.



.section "Header lines"
.subsection "Resent- header lines" SECID220
.chindex Resent-
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
makes provision for sets of header lines starting with the string
&`Resent-`& to be added to a message when it is resent by the original
recipient to somebody else. These headers are &'Resent-Date:'&,
&'Resent-From:'&, &'Resent-Sender:'&, &'Resent-To:'&, &'Resent-Cc:'&,
&'Resent-Bcc:'& and &'Resent-Message-ID:'&. The RFC says:

.blockquote
&'Resent fields are strictly informational. They MUST NOT be used in the normal
processing of replies or other such automatic actions on messages.'&
.endblockquote

This leaves things a bit vague as far as other processing actions such as
address rewriting are concerned. Exim treats &%Resent-%& header lines as
follows:

.ilist
A &'Resent-From:'& line that just contains the login id of the submitting user
is automatically rewritten in the same way as &'From:'& (see below).
.next
If there's a rewriting rule for a particular header line, it is also applied to
&%Resent-%& header lines of the same type. For example, a rule that rewrites
&'From:'& also rewrites &'Resent-From:'&.
.next
For local messages, if &'Sender:'& is removed on input, &'Resent-Sender:'& is
also removed.
.next
For a locally-submitted message,
if there are any &%Resent-%& header lines but no &'Resent-Date:'&,
&'Resent-From:'&, or &'Resent-Message-Id:'&, they are added as necessary. It is
the contents of &'Resent-Message-Id:'& (rather than &'Message-Id:'&) which are
included in log lines in this case.
.next
The logic for adding &'Sender:'& is duplicated for &'Resent-Sender:'& when any
&%Resent-%& header lines are present.
.endlist




.subsection Auto-Submitted: SECID221
Whenever Exim generates an autoreply, a bounce, or a delay warning message, it
includes the header line:
.code
Auto-Submitted: auto-replied
.endd

.subsection Bcc: SECID222
.cindex "&'Bcc:'& header line"
If Exim is called with the &%-t%& option, to take recipient addresses from a
message's header, it removes any &'Bcc:'& header line that may exist (after
extracting its addresses). If &%-t%& is not present on the command line, any
existing &'Bcc:'& is not removed.


.subsection Date: SECID223
.cindex Date:
If a locally-generated or submission-mode message has no &'Date:'& header line,
Exim adds one, using the current date and time, unless the
&%suppress_local_fixups%& control has been specified.

.subsection Delivery-date: SECID224
.cindex "&'Delivery-date:'& header line"
.oindex "&%delivery_date_remove%&"
&'Delivery-date:'& header lines are not part of the standard
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
header set.
Exim can be configured to add them to the final delivery of messages. (See
the generic &%delivery_date_add%& transport option.) They should not be present
in messages in transit. If the &%delivery_date_remove%& configuration option is
set (the default), Exim removes &'Delivery-date:'& header lines from incoming
messages.


.subsection Envelope-to: SECID225
.chindex Envelope-to:
.oindex "&%envelope_to_remove%&"
&'Envelope-to:'& header lines are not part of the standard
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822) header set.
Exim can be configured to add them to the final delivery of messages. (See the
generic &%envelope_to_add%& transport option.) They should not be present in
messages in transit. If the &%envelope_to_remove%& configuration option is set
(the default), Exim removes &'Envelope-to:'& header lines from incoming
messages.


.subsection From: SECTthefrohea
.chindex From:
.cindex "Sendmail compatibility" "&""From""& line"
.cindex "message" "submission"
.cindex "submission mode"
If a submission-mode message does not contain a &'From:'& header line, Exim
adds one if either of the following conditions is true:

.ilist
The envelope sender address is not empty (that is, this is not a bounce
message). The added header line copies the envelope sender address.
.next
.vindex "&$authenticated_id$&"
The SMTP session is authenticated and &$authenticated_id$& is not empty.
.olist
.vindex "&$qualify_domain$&"
If no domain is specified by the submission control, the local part is
&$authenticated_id$& and the domain is &$qualify_domain$&.
.next
If a non-empty domain is specified by the submission control, the local
part is &$authenticated_id$&, and the domain is the specified domain.
.next
If an empty domain is specified by the submission control,
&$authenticated_id$& is assumed to be the complete address.
.endlist
.endlist

A non-empty envelope sender takes precedence.

If a locally-generated incoming message does not contain a &'From:'& header
line, and the &%suppress_local_fixups%& control is not set, Exim adds one
containing the sender's address. The calling user's login name and full name
are used to construct the address, as described in section &<<SECTconstr>>&.
They are obtained from the password data by calling &[getpwuid()]& (but see the
&%unknown_login%& configuration option). The address is qualified with
&%qualify_domain%&.

For compatibility with Sendmail, if an incoming, non-SMTP message has a
&'From:'& header line containing just the unqualified login name of the calling
user, this is replaced by an address containing the user's login name and full
name as described in section &<<SECTconstr>>&.


.subsection Message-ID: SECID226
.chindex Message-ID:
.cindex "message" "submission"
.oindex "&%message_id_header_text%&"
If a locally-generated or submission-mode incoming message does not contain a
&'Message-ID:'& or &'Resent-Message-ID:'& header line, and the
&%suppress_local_fixups%& control is not set, Exim adds a suitable header line
to the message. If there are any &'Resent-:'& headers in the message, it
creates &'Resent-Message-ID:'&. The id is constructed from Exim's internal
message id, preceded by the letter E to ensure it starts with a letter, and
followed by @ and the primary host name. Additional information can be included
in this header line by setting the &%message_id_header_text%& and/or
&%message_id_header_domain%& options.


.subsection Received: SECID227
.chindex Received:
A &'Received:'& header line is added at the start of every message. The
contents are defined by the &%received_header_text%& configuration option, and
Exim automatically adds a semicolon and a timestamp to the configured string.

The &'Received:'& header is generated as soon as the message's header lines
have been received. At this stage, the timestamp in the &'Received:'& header
line is the time that the message started to be received. This is the value
that is seen by the DATA ACL and by the &[local_scan()]& function.

Once a message is accepted, the timestamp in the &'Received:'& header line is
changed to the time of acceptance, which is (apart from a small delay while the
-H spool file is written) the earliest time at which delivery could start.


.subsection References: SECID228
.chindex References:
Messages created by the &(autoreply)& transport include a &'References:'&
header line. This is constructed according to the rules that are described in
section 3.64 of
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
(which states that replies should contain such a header line),
and section 3.14 of &url(https://datatracker.ietf.org/doc/html/rfc3834,RFC 3834)
(which states that automatic
responses are not different in this respect). However, because some mail
processing software does not cope well with very long header lines, no more
than 12 message IDs are copied from the &'References:'& header line in the
incoming message. If there are more than 12, the first one and then the final
11 are copied, before adding the message ID of the incoming message.



.subsection Return-path: SECID229
.chindex Return-path:
.oindex "&%return_path_remove%&"
&'Return-path:'& header lines are defined as something an MTA may insert when
it does the final delivery of messages. (See the generic &%return_path_add%&
transport option.) Therefore, they should not be present in messages in
transit. If the &%return_path_remove%& configuration option is set (the
default), Exim removes &'Return-path:'& header lines from incoming messages.



.subsection Sender: SECTthesenhea
.cindex "&'Sender:'& header line"
.cindex "message" "submission"
.chindex Sender:
For a locally-originated message from an untrusted user, Exim may remove an
existing &'Sender:'& header line, and it may add a new one. You can modify
these actions by setting the &%local_sender_retain%& option true, the
&%local_from_check%& option false, or by using the &%suppress_local_fixups%&
control setting.

When a local message is received from an untrusted user and
&%local_from_check%& is true (the default), and the &%suppress_local_fixups%&
control has not been set, a check is made to see if the address given in the
&'From:'& header line is the correct (local) sender of the message. The address
that is expected has the login name as the local part and the value of
&%qualify_domain%& as the domain. Prefixes and suffixes for the local part can
be permitted by setting &%local_from_prefix%& and &%local_from_suffix%&
appropriately. If &'From:'& does not contain the correct sender, a &'Sender:'&
line is added to the message.

If you set &%local_from_check%& false, this checking does not occur. However,
the removal of an existing &'Sender:'& line still happens, unless you also set
&%local_sender_retain%& to be true. It is not possible to set both of these
options true at the same time.

.cindex "submission mode"
By default, no processing of &'Sender:'& header lines is done for messages
received over TCP/IP or for messages submitted by trusted users. However, when
a message is received over TCP/IP in submission mode, and &%sender_retain%& is
not specified on the submission control, the following processing takes place:

.vindex "&$authenticated_id$&"
First, any existing &'Sender:'& lines are removed. Then, if the SMTP session is
authenticated, and &$authenticated_id$& is not empty, a sender address is
created as follows:

.ilist
.vindex "&$qualify_domain$&"
If no domain is specified by the submission control, the local part is
&$authenticated_id$& and the domain is &$qualify_domain$&.
.next
If a non-empty domain is specified by the submission control, the local part
is &$authenticated_id$&, and the domain is the specified domain.
.next
If an empty domain is specified by the submission control,
&$authenticated_id$& is assumed to be the complete address.
.endlist

This address is compared with the address in the &'From:'& header line. If they
are different, a &'Sender:'& header line containing the created address is
added. Prefixes and suffixes for the local part in &'From:'& can be permitted
by setting &%local_from_prefix%& and &%local_from_suffix%& appropriately.

.cindex "return path" "created from &'Sender:'&"
&*Note*&: Whenever a &'Sender:'& header line is created, the return path for
the message (the envelope sender address) is changed to be the same address,
except in the case of submission mode when &%sender_retain%& is specified.



.section "Adding and removing header lines in routers and transports" &&&
         "SECTheadersaddrem"
.cindex "header lines" "adding; in router or transport"
.cindex "header lines" "removing; in router or transport"
When a message is delivered, the addition and removal of header lines can be
specified in a system filter, or on any of the routers and transports that
process the message. Section &<<SECTaddremheasys>>& contains details about
modifying headers in a system filter. Header lines can also be added in an ACL
as a message is received (see section &<<SECTaddheadacl>>&).

In contrast to what happens in a system filter, header modifications that are
specified on routers and transports apply only to the particular recipient
addresses that are being processed by those routers and transports. These
changes do not actually take place until a copy of the message is being
transported. Therefore, they do not affect the basic set of header lines, and
they do not affect the values of the variables that refer to header lines.

&*Note*&: In particular, this means that any expansions in the configuration of
the transport cannot refer to the modified header lines, because such
expansions all occur before the message is actually transported.

For both routers and transports, the argument of a &%headers_add%&
option must be in the form of one or more
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
header lines, separated by newlines (coded as &"\n"&). For example:
.code
headers_add = X-added-header: added by $primary_hostname\n\
              X-added-second: another added header line
.endd
Exim does not check the syntax of these added header lines.

Multiple &%headers_add%& options for a single router or transport can be
specified; the values will append to a single list of header lines.
Each header-line is separately expanded.

The argument of a &%headers_remove%& option must consist of a colon-separated
list of header names. This is confusing, because header names themselves are
often terminated by colons. In this case, the colons are the list separators,
not part of the names. For example:
.code
headers_remove = return-receipt-to:acknowledge-to
.endd

Multiple &%headers_remove%& options for a single router or transport can be
specified; the arguments will append to a single header-names list.
Each item is separately expanded.
Note that colons in complex expansions which are used to
form all or part of a &%headers_remove%& list
will act as list separators.

When &%headers_add%& or &%headers_remove%& is specified on a router,
items are expanded at routing time,
and then associated with all addresses that are
accepted by that router, and also with any new addresses that it generates. If
an address passes through several routers as a result of aliasing or
forwarding, the changes are cumulative.

.oindex "&%unseen%&"
However, this does not apply to multiple routers that result from the use of
the &%unseen%& option. Any header modifications that were specified by the
&"unseen"& router or its predecessors apply only to the &"unseen"& delivery.

Addresses that end up with different &%headers_add%& or &%headers_remove%&
settings cannot be delivered together in a batch, so a transport is always
dealing with a set of addresses that have the same header-processing
requirements.

The transport starts by writing the original set of header lines that arrived
with the message, possibly modified by the system filter. As it writes out
these lines, it consults the list of header names that were attached to the
recipient address(es) by &%headers_remove%& options in routers, and it also
consults the transport's own &%headers_remove%& option. Header lines whose
names are on either of these lists are not written out. If there are multiple
instances of any listed header, they are all skipped.

After the remaining original header lines have been written, new header
lines that were specified by routers' &%headers_add%& options are written, in
the order in which they were attached to the address. These are followed by any
header lines specified by the transport's &%headers_add%& option.

This way of handling header line modifications in routers and transports has
the following consequences:

.ilist
The original set of header lines, possibly modified by the system filter,
remains &"visible"&, in the sense that the &$header_$&&'xxx'& variables refer
to it, at all times.
.next
Header lines that are added by a router's
&%headers_add%& option are not accessible by means of the &$header_$&&'xxx'&
expansion syntax in subsequent routers or the transport.
.next
Conversely, header lines that are specified for removal by &%headers_remove%&
in a router remain visible to subsequent routers and the transport.
.next
Headers added to an address by &%headers_add%& in a router cannot be removed by
a later router or by a transport.
.next
An added header can refer to the contents of an original header that is to be
removed, even it has the same name as the added header. For example:
.code
headers_remove = subject
headers_add = Subject: new subject (was: $h_subject:)
.endd
.endlist

&*Warning*&: The &%headers_add%& and &%headers_remove%& options cannot be used
for a &(redirect)& router that has the &%one_time%& option set.





.section "Constructed addresses" "SECTconstr"
.cindex "address" "constructed"
.cindex "constructed address"
When Exim constructs a sender address for a locally-generated message, it uses
the form
.display
<&'user name'&>&~&~<&'login'&&`@`&&'qualify_domain'&>
.endd
For example:
.code
Zaphod Beeblebrox <zaphod@end.univ.example>
.endd
The user name is obtained from the &%-F%& command line option if set, or
otherwise by looking up the calling user by &[getpwuid()]& and extracting the
&"gecos"& field from the password entry. If the &"gecos"& field contains an
ampersand character, this is replaced by the login name with the first letter
upper cased, as is conventional in a number of operating systems. See the
&%gecos_name%& option for a way to tailor the handling of the &"gecos"& field.
The &%unknown_username%& option can be used to specify user names in cases when
there is no password file entry.

.cindex "RFC 2047"
In all cases, the user name is made to conform to
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
by quoting all or
parts of it if necessary. In addition, if it contains any non-printing
characters, it is encoded as described in
&url(https://www.rfc-editor.org/rfc/rfc2047,RFC 2047), which defines a way of
including non-ASCII characters in header lines. The value of the
&%headers_charset%& option specifies the name of the encoding that is used (the
characters are assumed to be in this encoding). The setting of
&%print_topbitchars%& controls whether characters with the top bit set (that
is, with codes greater than 127) count as printing characters or not.



.section "Case of local parts" "SECID230"
.cindex "case of local parts"
.cindex "local part" "case of"
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
states that the case of letters in the local parts of addresses cannot
be assumed to be non-significant. Exim preserves the case of local parts of
addresses, but by default it uses a lower-cased form when it is routing,
because on most Unix systems, usernames are in lower case and case-insensitive
routing is required. However, any particular router can be made to use the
original case for local parts by setting the &%caseful_local_part%& generic
router option.

.cindex "mixed-case login names"
If you must have mixed-case user names on your system, the best way to proceed,
assuming you want case-independent handling of incoming email, is to set up
your first router to convert incoming local parts in your domains to the
correct case by means of a file lookup. For example:
.code
correct_case:
  driver = redirect
  domains = +local_domains
  data = ${lookup{$local_part}cdb\
              {/etc/usercased.cdb}{$value}fail}\
              @$domain
.endd
For this router, the local part is forced to lower case by the default action
(&%caseful_local_part%& is not set). The lower-cased local part is used to look
up a new local part in the correct case. If you then set &%caseful_local_part%&
on any subsequent routers which process your domains, they will operate on
local parts with the correct case in a case-sensitive manner.



.section "Dots in local parts" "SECID231"
.cindex "dot" "in local part"
.cindex "local part" "dots in"
&url(https://www.rfc-editor.org/rfc/rfc2822,RFC 2822)
forbids empty components in local parts. That is, an unquoted local
part may not begin or end with a dot, nor have two consecutive dots in the
middle. However, it seems that many MTAs do not enforce this, so Exim permits
empty components for compatibility.



.section "Rewriting addresses" "SECID232"
.cindex "rewriting" "addresses"
Rewriting of sender and recipient addresses, and addresses in headers, can
happen automatically, or as the result of configuration options, as described
in chapter &<<CHAPrewrite>>&. The headers that may be affected by this are
&'Bcc:'&, &'Cc:'&, &'From:'&, &'Reply-To:'&, &'Sender:'&, and &'To:'&.

Automatic rewriting includes qualification, as mentioned above. The other case
in which it can happen is when an incomplete non-local domain is given. The
routing process may cause this to be expanded into the full domain name. For
example, a header such as
.code
To: hare@teaparty
.endd
might get rewritten as
.code
To: hare@teaparty.wonderland.fict.example
.endd
Rewriting as a result of routing is the one kind of message processing that
does not happen at input time, as it cannot be done until the address has
been routed.

Strictly, one should not do &'any'& deliveries of a message until all its
addresses have been routed, in case any of the headers get changed as a
result of routing. However, doing this in practice would hold up many
deliveries for unreasonable amounts of time, just because one address could not
immediately be routed. Exim therefore does not delay other deliveries when
routing of one or more addresses is deferred.
.ecindex IIDmesproc



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "SMTP processing" "CHAPSMTP"
.scindex IIDsmtpproc1 "SMTP" "processing details"
.scindex IIDsmtpproc2 "LMTP" "processing details"
Exim supports a number of different ways of using the SMTP protocol, and its
LMTP variant, which is an interactive protocol for transferring messages into a
closed mail store application. This chapter contains details of how SMTP is
processed. For incoming mail, the following are available:

.ilist
SMTP over TCP/IP (Exim daemon or &'inetd'&);
.next
SMTP over the standard input and output (the &%-bs%& option);
.next
Batched SMTP on the standard input (the &%-bS%& option).
.endlist

For mail delivery, the following are available:

.ilist
SMTP over TCP/IP (the &(smtp)& transport);
.next
LMTP over TCP/IP (the &(smtp)& transport with the &%protocol%& option set to
&"lmtp"&);
.next
LMTP over a pipe to a process running in the local host (the &(lmtp)&
transport);
.next
Batched SMTP to a file or pipe (the &(appendfile)& and &(pipe)& transports with
the &%use_bsmtp%& option set).
.endlist

&'Batched SMTP'& is the name for a process in which batches of messages are
stored in or read from files (or pipes), in a format in which SMTP commands are
used to contain the envelope information.



.section "Outgoing SMTP and LMTP over TCP/IP" "SECToutSMTPTCP"
.cindex "SMTP" "outgoing over TCP/IP"
.cindex "outgoing SMTP over TCP/IP"
.cindex "LMTP" "over TCP/IP"
.cindex "outgoing LMTP over TCP/IP"
.cindex "EHLO"
.cindex "HELO"
.cindex "SIZE" "option on MAIL command"
Outgoing SMTP and LMTP over TCP/IP is implemented by the &(smtp)& transport.
The &%protocol%& option selects which protocol is to be used, but the actual
processing is the same in both cases.

.cindex "ESMTP extensions" SIZE
If, in response to its EHLO command, Exim is told that the SIZE
extension is supported, it adds SIZE=<&'n'&> to each subsequent MAIL
command. The value of <&'n'&> is the message size plus the value of the
&%size_addition%& option (default 1024) to allow for additions to the message
such as per-transport header lines, or changes made in a
.cindex "transport" "filter"
.cindex "filter" "transport filter"
transport filter. If &%size_addition%& is set negative, the use of SIZE is
suppressed.

If the remote server advertises support for PIPELINING, Exim uses the
pipelining extension to SMTP
(&url(https://datatracker.ietf.org/doc/html/rfc2197,RFC 2197))
to reduce the number of TCP/IP packets required for the transaction.

If the remote server advertises support for the STARTTLS command, and Exim
was built to support TLS encryption, it tries to start a TLS session unless the
server matches &%hosts_avoid_tls%&. See chapter &<<CHAPTLS>>& for more details.
Either a match in that or &%hosts_verify_avoid_tls%& apply when the transport
is called for verification.

If the remote server advertises support for the AUTH command, Exim scans
the authenticators configuration for any suitable client settings, as described
in chapter &<<CHAPSMTPAUTH>>&.

.cindex "carriage return"
.cindex "linefeed"
Responses from the remote host are supposed to be terminated by CR followed by
LF. However, there are known to be hosts that do not send CR characters, so in
order to be able to interwork with such hosts, Exim treats LF on its own as a
line terminator.

If a message contains a number of different addresses, all those with the same
characteristics (for example, the same envelope sender) that resolve to the
same set of hosts, in the same order, are sent in a single SMTP transaction,
even if they are for different domains, unless there are more than the setting
of the &%max_rcpt%&s option in the &(smtp)& transport allows, in which case
they are split into groups containing no more than &%max_rcpt%&s addresses
each. If &%remote_max_parallel%& is greater than one, such groups may be sent
in parallel sessions. The order of hosts with identical MX values is not
significant when checking whether addresses can be batched in this way.

When the &(smtp)& transport suffers a temporary failure that is not
message-related, Exim updates its transport-specific database, which contains
records indexed by host name that remember which messages are waiting for each
particular host. It also updates the retry database with new retry times.

.cindex "hints database" "retry keys"
Exim's retry hints are based on host name plus IP address, so if one address of
a multi-homed host is broken, it will soon be skipped most of the time.
See the next section for more detail about error handling.

.cindex "SMTP" "passed connection"
.cindex "SMTP" "batching over TCP/IP"
When a message is successfully delivered over a TCP/IP SMTP connection, Exim
looks in the hints database for the transport to see if there are any queued
messages waiting for the host to which it is connected.
If it finds one, it arranges to attempt that message on the same connection.

The &%connection_max_messages%& option of the &(smtp)& transport can be used to
limit the number of messages sent down a single TCP/IP connection.

.cindex "asterisk" "after IP address"
The second and subsequent messages delivered down an existing connection are
identified in the main log by the addition of an asterisk after the closing
square bracket of the IP address.




.subsection "Errors in outgoing SMTP" SECToutSMTPerr
.cindex "error" "in outgoing SMTP"
.cindex "SMTP" "errors in outgoing"
.cindex "host" "error"
Three different kinds of error are recognized for outgoing SMTP: host errors,
message errors, and recipient errors.

.vlist
.vitem "&*Host errors*&"
A host error is not associated with a particular message or with a
particular recipient of a message. The host errors are:

.ilist
Connection refused or timed out,
.next
Any error response code on connection,
.next
Any error response code to EHLO or HELO,
.next
Loss of connection at any time, except after &"."&,
.next
I/O errors at any time,
.next
Timeouts during the session, other than in response to MAIL, RCPT or
the &"."& at the end of the data.
.endlist ilist

For a host error, a permanent error response on connection, or in response to
EHLO, causes all addresses routed to the host to be failed. Any other host
error causes all addresses to be deferred, and retry data to be created for the
host. It is not tried again, for any message, until its retry time arrives. If
the current set of addresses are not all delivered in this run (to some
alternative host), the message is added to the list of those waiting for this
host, so if it is still undelivered when a subsequent successful delivery is
made to the host, it will be sent down the same SMTP connection.

.vitem "&*Message errors*&"
.cindex "message" "error"
A message error is associated with a particular message when sent to a
particular host, but not with a particular recipient of the message. The
message errors are:

.ilist
Any error response code to MAIL, DATA, or the &"."& that terminates
the data,
.next
Timeout after MAIL,
.next
Timeout or loss of connection after the &"."& that terminates the data. A
timeout after the DATA command itself is treated as a host error, as is loss of
connection at any other time.
.endlist ilist

For a message error, a permanent error response (5&'xx'&) causes all addresses
to be failed, and a delivery error report to be returned to the sender. A
temporary error response (4&'xx'&), or one of the timeouts, causes all
addresses to be deferred. Retry data is not created for the host, but instead,
a retry record for the combination of host plus message id is created. The
message is not added to the list of those waiting for this host. This ensures
that the failing message will not be sent to this host again until the retry
time arrives. However, other messages that are routed to the host are not
affected, so if it is some property of the message that is causing the error,
it will not stop the delivery of other mail.

If the remote host specified support for the SIZE parameter in its response
to EHLO, Exim adds SIZE=&'nnn'& to the MAIL command, so an
over-large message will cause a message error because the error arrives as a
response to MAIL.

.vitem "&*Recipient errors*&"
.cindex "recipient" "error"
A recipient error is associated with a particular recipient of a message. The
recipient errors are:

.ilist
Any error response to RCPT,
.next
Timeout after RCPT.
.endlist

For a recipient error, a permanent error response (5&'xx'&) causes the
recipient address to be failed, and a bounce message to be returned to the
sender. A temporary error response (4&'xx'&) or a timeout causes the failing
address to be deferred, and routing retry data to be created for it. This is
used to delay processing of the address in subsequent queue runs, until its
routing retry time arrives. This applies to all messages, but because it
operates only in queue runs, one attempt will be made to deliver a new message
to the failing address before the delay starts to operate. This ensures that,
if the failure is really related to the message rather than the recipient
(&"message too big for this recipient"& is a possible example), other messages
have a chance of getting delivered. If a delivery to the address does succeed,
the retry information gets cleared, so all stuck messages get tried again, and
the retry clock is reset.

The message is not added to the list of those waiting for this host. Use of the
host for other messages is unaffected, and except in the case of a timeout,
other recipients are processed independently, and may be successfully delivered
in the current SMTP session. After a timeout it is of course impossible to
proceed with the session, so all addresses get deferred. However, those other
than the one that failed do not suffer any subsequent retry delays. Therefore,
if one recipient is causing trouble, the others have a chance of getting
through when a subsequent delivery attempt occurs before the failing
recipient's retry time.
.endlist

In all cases, if there are other hosts (or IP addresses) available for the
current set of addresses (for example, from multiple MX records), they are
tried in this run for any undelivered addresses, subject of course to their
own retry data. In other words, recipient error retry data does not take effect
until the next delivery attempt.

Some hosts have been observed to give temporary error responses to every
MAIL command at certain times (&"insufficient space"& has been seen). It
would be nice if such circumstances could be recognized, and defer data for the
host itself created, but this is not possible within the current Exim design.
What actually happens is that retry data for every (host, message) combination
is created.

The reason that timeouts after MAIL and RCPT are treated specially is that
these can sometimes arise as a result of the remote host's verification
procedures. Exim makes this assumption, and treats them as if a temporary error
response had been received. A timeout after &"."& is treated specially because
it is known that some broken implementations fail to recognize the end of the
message if the last character of the last line is a binary zero. Thus, it is
helpful to treat this case as a message error.

Timeouts at other times are treated as host errors, assuming a problem with the
host, or the connection to it. If a timeout after MAIL, RCPT,
or &"."& is really a connection problem, the assumption is that at the next try
the timeout is likely to occur at some other point in the dialogue, causing it
then to be treated as a host error.

There is experimental evidence that some MTAs drop the connection after the
terminating &"."& if they do not like the contents of the message for some
reason, in contravention of the RFC, which indicates that a 5&'xx'& response
should be given. That is why Exim treats this case as a message rather than a
host error, in order not to delay other messages to the same host.




.section "Incoming SMTP messages over TCP/IP" "SECID233"
.cindex "SMTP" "incoming over TCP/IP"
.cindex "incoming SMTP over TCP/IP"
.cindex "inetd"
.cindex "daemon"
Incoming SMTP messages can be accepted in one of two ways: by running a
listening daemon, or by using &'inetd'&. In the latter case, the entry in
&_/etc/inetd.conf_& should be like this:
.code
smtp stream tcp nowait exim /opt/exim/bin/exim in.exim -bs
.endd
Exim distinguishes between this case and the case of a locally running user
agent using the &%-bs%& option by checking whether or not the standard input is
a socket. When it is, either the port must be privileged (less than 1024), or
the caller must be root or the Exim user. If any other user passes a socket
with an unprivileged port number, Exim prints a message on the standard error
stream and exits with an error code.

By default, Exim does not make a log entry when a remote host connects or
disconnects (either via the daemon or &'inetd'&), unless the disconnection is
unexpected. It can be made to write such log entries by setting the
&%smtp_connection%& log selector.

.cindex "carriage return"
.cindex "linefeed"
Commands from the remote host are supposed to be terminated by CR followed by
LF. However, there are known to be hosts that do not send CR characters. In
order to be able to interwork with such hosts, Exim treats LF on its own as a
line terminator.
Furthermore, because common code is used for receiving messages from all
sources, a CR on its own is also interpreted as a line terminator. However, the
sequence &"CR, dot, CR"& does not terminate incoming SMTP data.

.cindex "EHLO" "invalid data"
.cindex "HELO" "invalid data"
One area that sometimes gives rise to problems concerns the EHLO or
HELO commands. Some clients send syntactically invalid versions of these
commands, which Exim rejects by default. (This is nothing to do with verifying
the data that is sent, so &%helo_verify_hosts%& is not relevant.) You can tell
Exim not to apply a syntax check by setting &%helo_accept_junk_hosts%& to
match the broken hosts that send invalid commands.

.cindex "SIZE option on MAIL command"
.cindex "MAIL" "SIZE option"
The amount of disk space available is checked whenever SIZE is received on
a MAIL command, independently of whether &%message_size_limit%& or
&%check_spool_space%& is configured, unless &%smtp_check_spool_space%& is set
false. A temporary error is given if there is not enough space. If
&%check_spool_space%& is set, the check is for that amount of space plus the
value given with SIZE, that is, it checks that the addition of the incoming
message will not reduce the space below the threshold.

When a message is successfully received, Exim includes the local message id in
its response to the final &"."& that terminates the data. If the remote host
logs this text it can help with tracing what has happened to a message.

The Exim daemon can limit the number of simultaneous incoming connections it is
prepared to handle (see the &%smtp_accept_max%& option). It can also limit the
number of simultaneous incoming connections from a single remote host (see the
&%smtp_accept_max_per_host%& option). Additional connection attempts are
rejected using the SMTP temporary error code 421.

The Exim daemon does not rely on the SIGCHLD signal to detect when a
subprocess has finished, as this can get lost at busy times. Instead, it looks
for completed subprocesses every time it wakes up. Provided there are other
things happening (new incoming calls, starts of queue runs), completed
processes will be noticed and tidied away. On very quiet systems you may
sometimes see a &"defunct"& Exim process hanging about. This is not a problem;
it will be noticed when the daemon next wakes up.

When running as a daemon, Exim can reserve some SMTP slots for specific hosts,
and can also be set up to reject SMTP calls from non-reserved hosts at times of
high system load &-- for details see the &%smtp_accept_reserve%&,
&%smtp_load_reserve%&, and &%smtp_reserve_hosts%& options. The load check
applies in both the daemon and &'inetd'& cases.

Exim normally starts a delivery process for each message received, though this
can be varied by means of the &%-odq%& command line option and the
&%queue_only%&, &%queue_only_file%&, and &%queue_only_load%& options. The
number of simultaneously running delivery processes started in this way from
SMTP input can be limited by the &%smtp_accept_queue%& and
&%smtp_accept_queue_per_connection%& options. When either limit is reached,
subsequently received messages are just put on the input queue without starting
a delivery process.

The controls that involve counts of incoming SMTP calls (&%smtp_accept_max%&,
&%smtp_accept_queue%&, &%smtp_accept_reserve%&) are not available when Exim is
started up from the &'inetd'& daemon, because in that case each connection is
handled by an entirely independent Exim process. Control by load average is,
however, available with &'inetd'&.

Exim can be configured to verify addresses in incoming SMTP commands as they
are received. See chapter &<<CHAPACL>>& for details. It can also be configured
to rewrite addresses at this time &-- before any syntax checking is done. See
section &<<SSECTrewriteS>>&.

Exim can also be configured to limit the rate at which a client host submits
MAIL and RCPT commands in a single SMTP session. See the
&%smtp_ratelimit_hosts%& option.



.subsection "Unrecognized SMTP commands" SECID234
.cindex "SMTP" "unrecognized commands"
If Exim receives more than &%smtp_max_unknown_commands%& unrecognized SMTP
commands during a single SMTP connection, it drops the connection after sending
the error response to the last command. The default value for
&%smtp_max_unknown_commands%& is 3. This is a defence against some kinds of
abuse that subvert web servers into making connections to SMTP ports; in these
circumstances, a number of non-SMTP lines are sent first.


.subsection "Syntax and protocol errors in SMTP commands" SECID235
.cindex "SMTP" "syntax errors"
.cindex "SMTP" "protocol errors"
A syntax error is detected if an SMTP command is recognized, but there is
something syntactically wrong with its data, for example, a malformed email
address in a RCPT command. Protocol errors include invalid command
sequencing such as RCPT before MAIL. If Exim receives more than
&%smtp_max_synprot_errors%& such commands during a single SMTP connection, it
drops the connection after sending the error response to the last command. The
default value for &%smtp_max_synprot_errors%& is 3. This is a defence against
broken clients that loop sending bad commands (yes, it has been seen).



.subsection "Use of non-mail SMTP commands" SECID236
.cindex "SMTP" "non-mail commands"
The &"non-mail"& SMTP commands are those other than MAIL, RCPT, and
DATA. Exim counts such commands, and drops the connection if there are too
many of them in a single SMTP session. This action catches some
denial-of-service attempts and things like repeated failing AUTHs, or a mad
client looping sending EHLO. The global option &%smtp_accept_max_nonmail%&
defines what &"too many"& means. Its default value is 10.

When a new message is expected, one occurrence of RSET is not counted. This
allows a client to send one RSET between messages (this is not necessary,
but some clients do it). Exim also allows one uncounted occurrence of HELO
or EHLO, and one occurrence of STARTTLS between messages. After
starting up a TLS session, another EHLO is expected, and so it too is not
counted.

The first occurrence of AUTH in a connection, or immediately following
STARTTLS is also not counted. Otherwise, all commands other than MAIL,
RCPT, DATA, and QUIT are counted.

You can control which hosts are subject to the limit set by
&%smtp_accept_max_nonmail%& by setting
&%smtp_accept_max_nonmail_hosts%&. The default value is &`*`&, which makes
the limit apply to all hosts. This option means that you can exclude any
specific badly-behaved hosts that you have to live with.




.subsection "The VRFY and EXPN commands" SECID237
When Exim receives a VRFY or EXPN command on a TCP/IP connection, it
runs the ACL specified by &%acl_smtp_vrfy%& or &%acl_smtp_expn%& (as
appropriate) in order to decide whether the command should be accepted or not.

.cindex "VRFY" "processing"
When no ACL is defined for VRFY, or if it rejects without
setting an explicit response code, the command is accepted
(with a 252 SMTP response code)
in order to support awkward clients that do a VRFY before every RCPT.
When VRFY is accepted, it runs exactly the same code as when Exim is
called with the &%-bv%& option, and returns 250/451/550
SMTP response codes.

.cindex "EXPN" "processing"
If no ACL for EXPN is defined, the command is rejected.
When EXPN is accepted, a single-level expansion of the address is done.
EXPN is treated as an &"address test"& (similar to the &%-bt%& option) rather
than a verification (the &%-bv%& option). If an unqualified local part is given
as the argument to EXPN, it is qualified with &%qualify_domain%&. Rejections
of VRFY and EXPN commands are logged on the main and reject logs, and
VRFY verification failures are logged in the main log for consistency with
RCPT failures.



.subsection "The ETRN command" SECTETRN
.cindex "ETRN" "processing"
.cindex "ESMTP extensions" ETRN
Most modern installations never need to use this.
It is used for managing messages queued for an intermittently-connecting
destination (eg. one using a dialup connection).

.oindex "&%acl_smtp_etrn%&"
The command is only available if permitted by an ACL
specfied by the main-section &%acl_smtp_etrn%& option.

&url(https://www.rfc-editor.org/rfc/rfc1985,RFC 1985)
describes an ESMTP command called ETRN that is designed to
overcome the security problems of the TURN command (which has fallen into
disuse). When Exim receives an ETRN command on a TCP/IP connection, it runs
the ACL specified by &%acl_smtp_etrn%& in order to decide whether the command
should be accepted or not. If no ACL is defined, the command is rejected.

The ETRN command is concerned with &"releasing"& messages that are awaiting
delivery to certain hosts. As Exim does not organize its message queue by host,
the only form of ETRN that is supported by default is the one where the
text starts with the &"#"& prefix, in which case the remainder of the text is
specific to the SMTP server. A valid ETRN command causes a run of Exim with
the &%-R%& option to happen, with the remainder of the ETRN text as its
argument. For example,
.code
ETRN #brigadoon
.endd
runs the command
.code
exim -R brigadoon
.endd
which causes a delivery attempt on all messages with undelivered addresses
containing the text &"brigadoon"&. When &%smtp_etrn_serialize%& is set (the
default), Exim prevents the simultaneous execution of more than one queue run
for the same argument string as a result of an ETRN command. This stops
a misbehaving client from starting more than one queue runner at once.

.cindex "hints database" "ETRN serialization"
Exim implements the serialization by means of a hints database in which a
record is written whenever a process is started by ETRN, and deleted when
the process completes. However, Exim does not keep the SMTP session waiting for
the ETRN process to complete. Once ETRN is accepted, the client is sent
a &"success"& return code. Obviously there is scope for hints records to get
left lying around if there is a system or program crash. To guard against this,
Exim ignores any records that are more than six hours old.

.oindex "&%smtp_etrn_command%&"
For more control over what ETRN does, the &%smtp_etrn_command%& option can
used. This specifies a command that is run whenever ETRN is received,
whatever the form of its argument. For
example:
.code
smtp_etrn_command = /etc/etrn_command $domain \
                    $sender_host_address
.endd
.vindex "&$domain$&"
The string is split up into arguments which are independently expanded. The
expansion variable &$domain$& is set to the argument of the ETRN command,
and no syntax checking is done on the contents of this argument. Exim does not
wait for the command to complete, so its status code is not checked. Exim runs
under its own uid and gid when receiving incoming SMTP, so it is not possible
for it to change them before running the command.


.new
.subsection "The ATRN command" SECTATRN
.cindex ATRN processing
.cindex "ESMTP extensions" ATRN
A second method for intermittently-connecting destinations
is specified by
&url(https://www.rfc-editor.org/rfc/rfc2645.html,RFC 2645).

This describes an ESMTP command called ATRN which requests
a swap in server/client roles of the communicating endpoints, and delivery
of queued messages.
Note that this supports customers having IP addresses that
change frequently.

Exim supports the &"provider"& side of ATRN, using the terms
of that specification:
initially as an SMTP server, then transferring to an SMTP client
role if an ATRN command is accepted.

.oindex "&%acl_smtp_atrn%&"
The command is only available if permitted by an ACL
specfied by the main-section &%acl_smtp_atrn%& option.
Per the standard, this should only be for a specific
provider port number (386, named "odmr");
Exim should be configured to listen on that port
(in addition to other duties) via &%daemon_smtp_ports%&
or equivalent commandline options, and restrict the
advertising of the facility to the port:
.code
acl_smtp_atrn = ${if = {$received_port}{386} {check_atrn}{}}
.endd

A recieved ATRN command will be rejected unless
authentication has previously been done on the connection.

Any arguments supplied with an ATRN command are (per standard)
a comma-separated list of requested domains,
and will be available in the &$smtp_command_argument$&
variable.

The ACL configured may return &"deny"& for any policy reaons
(for example, the authenticated user is not permitted the facility).
Otherwise it should use the ACL &"atrn_domains"& condition,
which returns true if there are queued messages for any of
the given list of domains.
If that condition fails the ACL should return &"defer"&
with a "453 You have no mail" response;
else it should return &"accept"&.

For example (with default domain handling, and one possible de-taint method) :
.code
check_atrn:
  warn  set acl_m0 = clientdom.net
  deny  condition = ${if def:smtp_command_argument}
        set acl_m0 = ${map \
          {<, $smtp_command_argument} \
          {${if inlist{$item}{clientdom.net:cl2dom.net} {$value}}} \
          }
        condition = ${if !def:acl_m0}
  defer !atrn_domains = <, $acl_m0
        message = 453 You have no mail
  accept
.endd

Acceptance by the ACL will result in a queue-run for messages
having addresses with the given domains.
A suitable router and transport must be configured for the deliveries.

To access a named queue
.cindex queue named
the ACL should use a "queue =" modifier before the "atrn_domains"
condition.
If the ACL does not accept, re-set the queue to an empty value
so as to not disrupt any later SMTP operations on the connection.

Use of the &"atrn_domains"& condition additionally sets up
the &$atrn_host$& variable, which can be used by a manualroute
router.  Being otherwise empty, this router will decline in
other situations so can be safely placed in a general router chain.

For example:
.code
begin routers
odmr_client:
  driver =      manualroute
  route_data =  <;$atrn_host
  transport =   client_smtp

begin transports
client_smtp:
  driver =      smtp
.endd

Although not discssed in the specification document,
Exim supports use of ATRN within a STARTTLS-
or TLS-on-connect- encrypted connection
(which is wise if a plaintext authentication mechanism is used).
In such cases the TLS connection will remain open across the
role-swap, and be used for the sending of queued messages.

Note that the RFC requires that the CRAM-MD5 authentication
method be supported.
Exim does not enforce this, but leaves it up to the configuration;
see chapter &<<CHID9>>&.

.wen



.section "Incoming local SMTP" "SECID238"
.cindex "SMTP" "local incoming"
Some user agents use SMTP to pass messages to their local MTA using the
standard input and output, as opposed to passing the envelope on the command
line and writing the message to the standard input. This is supported by the
&%-bs%& option. This form of SMTP is handled in the same way as incoming
messages over TCP/IP (including the use of ACLs), except that the envelope
sender given in a MAIL command is ignored unless the caller is trusted. In
an ACL you can detect this form of SMTP input by testing for an empty host
identification. It is common to have this as the first line in the ACL that
runs for RCPT commands:
.code
accept hosts = :
.endd
This accepts SMTP messages from local processes without doing any other tests.



.section "Batched SMTP" "SECTgenbatchSMTP"
.subsection "Outgoing batched SMTP" "SECTbatchSMTP"
.cindex "SMTP" "batched outgoing"
.cindex "batched SMTP output"
Both the &(appendfile)& and &(pipe)& transports can be used for handling
batched SMTP. Each has an option called &%use_bsmtp%& which causes messages to
be output in BSMTP format. No SMTP responses are possible for this form of
delivery. All it is doing is using SMTP commands as a way of transmitting the
envelope along with the message.

The message is written to the file or pipe preceded by the SMTP commands
MAIL and RCPT, and followed by a line containing a single dot. Lines in
the message that start with a dot have an extra dot added. The SMTP command
HELO is not normally used. If it is required, the &%message_prefix%& option
can be used to specify it.

Because &(appendfile)& and &(pipe)& are both local transports, they accept only
one recipient address at a time by default. However, you can arrange for them
to handle several addresses at once by setting the &%batch_max%& option. When
this is done for BSMTP, messages may contain multiple RCPT commands. See
chapter &<<CHAPbatching>>& for more details.

.vindex "&$host$&"
When one or more addresses are routed to a BSMTP transport by a router that
sets up a host list, the name of the first host on the list is available to the
transport in the variable &$host$&. Here is an example of such a transport and
router:
.code
begin routers
route_append:
  driver = manualroute
  transport = smtp_appendfile
  route_list = domain.example  batch.host.example

begin transports
smtp_appendfile:
  driver = appendfile
  directory = /var/bsmtp/$host
  batch_max = 1000
  use_bsmtp
  user = exim
.endd
This causes messages addressed to &'domain.example'& to be written in BSMTP
format to &_/var/bsmtp/batch.host.example_&, with only a single copy of each
message (unless there are more than 1000 recipients).



.subsection "Incoming batched SMTP" "SECTincomingbatchedSMTP"
.cindex "SMTP" "batched incoming"
.cindex "batched SMTP input"
The &%-bS%& command line option causes Exim to accept one or more messages by
reading SMTP on the standard input, but to generate no responses. If the caller
is trusted, the senders in the MAIL commands are believed; otherwise the
sender is always the caller of Exim. Unqualified senders and receivers are not
rejected (there seems little point) but instead just get qualified. HELO
and EHLO act as RSET; VRFY, EXPN, ETRN and  HELP, act
as NOOP; QUIT quits.

Minimal policy checking is done for BSMTP input. Only the non-SMTP
ACL is run in the same way as for non-SMTP local input.

If an error is detected while reading a message, including a missing &"."& at
the end, Exim gives up immediately. It writes details of the error to the
standard output in a stylized way that the calling program should be able to
make some use of automatically, for example:
.code
554 Unexpected end of file
Transaction started in line 10
Error detected in line 14
.endd
It writes a more verbose version, for human consumption, to the standard error
file, for example:
.code
An error was detected while processing a file of BSMTP input.
The error message was:

501 '>' missing at end of address

The SMTP transaction started in line 10.
The error was detected in line 12.
The SMTP command at fault was:

rcpt to:<malformed@in.com.plete

1 previous message was successfully processed.
The rest of the batch was abandoned.
.endd
The return code from Exim is zero only if there were no errors. It is 1 if some
messages were accepted before an error was detected, and 2 if no messages were
accepted.
.ecindex IIDsmtpproc1
.ecindex IIDsmtpproc2



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Customizing bounce and warning messages" "CHAPemsgcust" &&&
         "Customizing messages"
When a message fails to be delivered, or remains in the queue for more than a
configured amount of time, Exim sends a message to the original sender, or
to an alternative configured address. The text of these messages is built into
the code of Exim, but it is possible to change it, either by adding a single
string, or by replacing each of the paragraphs by text supplied in a file.

The &'From:'& and &'To:'& header lines are automatically generated; you can
cause a &'Reply-To:'& line to be added by setting the &%errors_reply_to%&
option. Exim also adds the line
.code
Auto-Submitted: auto-generated
.endd
to all warning and bounce messages,


.section "Customizing bounce messages" "SECID239"
.cindex "customizing" "bounce message"
.cindex "bounce message" "customizing"
If &%bounce_message_text%& is set, its contents are included in the default
message immediately after &"This message was created automatically by mail
delivery software."& The string is not expanded. It is not used if
&%bounce_message_file%& is set.

When &%bounce_message_file%& is set, it must point to a template file for
constructing error messages. The file consists of a series of text items,
separated by lines consisting of exactly four asterisks. If the file cannot be
opened, default text is used and a message is written to the main and panic
logs. If any text item in the file is empty, default text is used for that
item.

.vindex "&$bounce_recipient$&"
.vindex "&$bounce_return_size_limit$&"
Each item of text that is read from the file is expanded, and there are two
expansion variables which can be of use here: &$bounce_recipient$& is set to
the recipient of an error message while it is being created, and
&$bounce_return_size_limit$& contains the value of the &%return_size_limit%&
option, rounded to a whole number.

The items must appear in the file in the following order:

.ilist
The first item is included in the headers, and should include at least a
&'Subject:'& header. Exim does not check the syntax of these headers.
.next
The second item forms the start of the error message. After it, Exim lists the
failing addresses with their error messages.
.next
The third item is used to introduce any text from pipe transports that is to be
returned to the sender. It is omitted if there is no such text.
.next
The fourth, fifth and sixth items will be ignored and may be empty.
The fields exist for back-compatibility
.endlist

The default state (&%bounce_message_file%& unset) is equivalent to the
following file, in which the sixth item is empty. The &'Subject:'& and some
other lines have been split in order to fit them on the page:
.code
Subject: Mail delivery failed
  ${if eq{$sender_address}{$bounce_recipient}
  {: returning message to sender}}
****
This message was created automatically by mail delivery software.

A message ${if eq{$sender_address}{$bounce_recipient}
  {that you sent }{sent by

<$sender_address>

}}could not be delivered to all of its recipients.
This is a permanent error. The following address(es) failed:
****
The following text was generated during the delivery attempt(s):
****
------ This is a copy of the message, including all the headers.
  ------
****
------ The body of the message is $message_size characters long;
  only the first
------ $bounce_return_size_limit or so are included here.
****
.endd
.section "Customizing warning messages" "SECTcustwarn"
.cindex "customizing" "warning message"
.cindex "warning of delay" "customizing the message"
The option &%warn_message_file%& can be pointed at a template file for use when
warnings about message delays are created. In this case there are only three
text sections:

.ilist
The first item is included in the headers, and should include at least a
&'Subject:'& header. Exim does not check the syntax of these headers.
.next
The second item forms the start of the warning message. After it, Exim lists
the delayed addresses.
.next
The third item then ends the message.
.endlist

The default state is equivalent to the following file, except that some lines
have been split here, in order to fit them on the page:
.code
Subject: Warning: message $message_exim_id delayed
  $warn_message_delay
****
This message was created automatically by mail delivery software.

A message ${if eq{$sender_address}{$warn_message_recipients}
{that you sent }{sent by

<$sender_address>

}}has not been delivered to all of its recipients after
more than $warn_message_delay in the queue on $primary_hostname.

The message identifier is:     $message_exim_id
The subject of the message is: $h_subject
The date of the message is:    $h_date

The following address(es) have not yet been delivered:
****
No action is required on your part. Delivery attempts will
continue for some time, and this warning may be repeated at
intervals if the message remains undelivered. Eventually the
mail delivery software will give up, and when that happens,
the message will be returned to you.
.endd
.vindex "&$warn_message_delay$&"
.vindex "&$warn_message_recipients$&"
However, in the default state the subject and date lines are omitted if no
appropriate headers exist. During the expansion of this file,
&$warn_message_delay$& is set to the delay time in one of the forms &"<&'n'&>
minutes"& or &"<&'n'&> hours"&, and &$warn_message_recipients$& contains a list
of recipients for the warning message. There may be more than one if there are
multiple addresses with different &%errors_to%& settings on the routers that
handled them.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Some common configuration settings" "CHAPcomconreq"
This chapter discusses some configuration settings that seem to be fairly
common. More examples and discussion can be found in the Exim book.



.section "Sending mail to a smart host" "SECID240"
.cindex "smart host" "example router"
If you want to send all mail for non-local domains to a &"smart host"&, you
should replace the default &(dnslookup)& router with a router which does the
routing explicitly:
.code
send_to_smart_host:
  driver = manualroute
  route_list = !+local_domains smart.host.name
  transport = remote_smtp
.endd
You can use the smart host's IP address instead of the name if you wish.
If you are using Exim only to submit messages to a smart host, and not for
receiving incoming messages, you can arrange for it to do the submission
synchronously by setting the &%mua_wrapper%& option (see chapter
&<<CHAPnonqueueing>>&).




.section "Using Exim to handle mailing lists" "SECTmailinglists"
.cindex "mailing lists"
Exim can be used to run simple mailing lists, but for large and/or complicated
requirements, the use of additional specialized mailing list software such as
Majordomo or Mailman is recommended.

The &(redirect)& router can be used to handle mailing lists where each list
is maintained in a separate file, which can therefore be managed by an
independent manager. The &%domains%& router option can be used to run these
lists in a separate domain from normal mail. For example:
.code
lists:
  driver = redirect
  domains = lists.example
  file = ${lookup {$local_part} dsearch,ret=full {/usr/lists}}
  forbid_pipe
  forbid_file
  errors_to = ${quote_local_part:$local_part-request}@lists.example
  no_more
.endd
This router is skipped for domains other than &'lists.example'&. For addresses
in that domain, it looks for a file that matches the local part. If there is no
such file, the router declines, but because &%no_more%& is set, no subsequent
routers are tried, and so the whole delivery fails.

The &%forbid_pipe%& and &%forbid_file%& options prevent a local part from being
expanded into a filename or a pipe delivery, which is usually inappropriate in
a mailing list.

.oindex "&%errors_to%&"
The &%errors_to%& option specifies that any delivery errors caused by addresses
taken from a mailing list are to be sent to the given address rather than the
original sender of the message. However, before acting on this, Exim verifies
the error address, and ignores it if verification fails.

For example, using the configuration above, mail sent to
&'dicts@lists.example'& is passed on to those addresses contained in
&_/usr/lists/dicts_&, with error reports directed to
&'dicts-request@lists.example'&, provided that this address can be verified.
There could be a file called &_/usr/lists/dicts-request_& containing
the address(es) of this particular list's manager(s), but other approaches,
such as setting up an earlier router (possibly using the &%local_part_prefix%&
or &%local_part_suffix%& options) to handle addresses of the form
&%owner-%&&'xxx'& or &%xxx-%&&'request'&, are also possible.



.section "Syntax errors in mailing lists" "SECID241"
.cindex "mailing lists" "syntax errors in"
If an entry in redirection data contains a syntax error, Exim normally defers
delivery of the original address. That means that a syntax error in a mailing
list holds up all deliveries to the list. This may not be appropriate when a
list is being maintained automatically from data supplied by users, and the
addresses are not rigorously checked.

If the &%skip_syntax_errors%& option is set, the &(redirect)& router just skips
entries that fail to parse, noting the incident in the log. If in addition
&%syntax_errors_to%& is set to a verifiable address, a message is sent to it
whenever a broken address is skipped. It is usually appropriate to set
&%syntax_errors_to%& to the same address as &%errors_to%&.



.section "Re-expansion of mailing lists" "SECID242"
.cindex "mailing lists" "re-expansion of"
Exim remembers every individual address to which a message has been delivered,
in order to avoid duplication, but it normally stores only the original
recipient addresses with a message. If all the deliveries to a mailing list
cannot be done at the first attempt, the mailing list is re-expanded when the
delivery is next tried. This means that alterations to the list are taken into
account at each delivery attempt, so addresses that have been added to
the list since the message arrived will therefore receive a copy of the
message, even though it pre-dates their subscription.

If this behaviour is felt to be undesirable, the &%one_time%& option can be set
on the &(redirect)& router. If this is done, any addresses generated by the
router that fail to deliver at the first attempt are added to the message as
&"top level"& addresses, and the parent address that generated them is marked
&"delivered"&. Thus, expansion of the mailing list does not happen again at the
subsequent delivery attempts. The disadvantage of this is that if any of the
failing addresses are incorrect, correcting them in the file has no effect on
pre-existing messages.

The original top-level address is remembered with each of the generated
addresses, and is output in any log messages. However, any intermediate parent
addresses are not recorded. This makes a difference to the log only if the
&%all_parents%& selector is set, but for mailing lists there is normally only
one level of expansion anyway.



.section "Closed mailing lists" "SECID243"
.cindex "mailing lists" "closed"
The examples so far have assumed open mailing lists, to which anybody may
send mail. It is also possible to set up closed lists, where mail is accepted
from specified senders only. This is done by making use of the generic
&%senders%& option to restrict the router that handles the list.

The following example uses the same file as a list of recipients and as a list
of permitted senders. It requires three routers:
.code
lists_request:
  driver = redirect
  domains = lists.example
  local_part_suffix = -request
  local_parts = ${lookup {$local_part} dsearch,filter=file {/usr/lists}}
  file = /usr/lists/${local_part_data}-request
  no_more

lists_post:
  driver = redirect
  domains = lists.example
  local_parts = ${lookup {$local_part} dsearch,filter=file,ret=full {/usr/lists}}
  senders = ${if exists {$local_part_data} {lsearch;$local_part_data}{*}}
  file = ${lookup {$local_part} dsearch,ret=full {/usr/lists}}
  forbid_pipe
  forbid_file
  errors_to = ${quote_local_part:$local_part-request}@lists.example
  no_more

lists_closed:
  driver = redirect
  domains = lists.example
  allow_fail
  data = :fail: $local_part@lists.example is a closed mailing list
.endd
All three routers have the same &%domains%& setting, so for any other domains,
they are all skipped. The first router runs only if the local part ends in
&%-request%&. It handles messages to the list manager(s) by means of an open
mailing list.

The second router runs only if the &%senders%& precondition is satisfied. It
checks for the existence of a list that corresponds to the local part, and then
checks that the sender is on the list by means of a linear search. It is
necessary to check for the existence of the file before trying to search it,
because otherwise Exim thinks there is a configuration error. If the file does
not exist, the expansion of &%senders%& is *, which matches all senders. This
means that the router runs, but because there is no list, declines, and
&%no_more%& ensures that no further routers are run. The address fails with an
&"unrouteable address"& error.

The third router runs only if the second router is skipped, which happens when
a mailing list exists, but the sender is not on it. This router forcibly fails
the address, giving a suitable error message.




.section "Variable Envelope Return Paths (VERP)" "SECTverp"
.cindex "VERP"
.cindex "Variable Envelope Return Paths"
.cindex "envelope from"
.cindex "envelope sender"
Variable Envelope Return Paths &-- see &url(https://cr.yp.to/proto/verp.txt) &--
are a way of helping mailing list administrators discover which subscription
address is the cause of a particular delivery failure. The idea is to encode
the original recipient address in the outgoing envelope sender address, so that
if the message is forwarded by another host and then subsequently bounces, the
original recipient can be extracted from the recipient address of the bounce.

.oindex &%errors_to%&
.oindex &%return_path%&
Envelope sender addresses can be modified by Exim using two different
facilities: the &%errors_to%& option on a router (as shown in previous mailing
list examples), or the &%return_path%& option on a transport. The second of
these is effective only if the message is successfully delivered to another
host; it is not used for errors detected on the local host (see the description
of &%return_path%& in chapter &<<CHAPtransportgeneric>>&). Here is an example
of the use of &%return_path%& to implement VERP on an &(smtp)& transport:
.code
verp_smtp:
  driver = smtp
  max_rcpt = 1
  return_path = \
    ${if match {$return_path}{^(.+?)-request@your.dom.example\$}\
      {${quote_local_part:$1-request+$local_part=$domain}@your.dom.example}fail}
.endd
This has the effect of rewriting the return path (envelope sender) on outgoing
SMTP messages, if the local part of the original return path ends in
&"-request"&, and the domain is &'your.dom.example'&. The rewriting inserts the
local part and domain of the recipient into the return path. Suppose, for
example, that a message whose return path has been set to
&'somelist-request@your.dom.example'& is sent to
&'subscriber@other.dom.example'&. In the transport, the return path is
rewritten as
.code
somelist-request+subscriber=other.dom.example@your.dom.example
.endd
.vindex "&$local_part$&"
For this to work, you must tell Exim to send multiple copies of messages that
have more than one recipient, so that each copy has just one recipient. This is
achieved by setting &%max_rcpt%& to 1. Without this, a single copy of a message
might be sent to several different recipients in the same domain, in which case
&$local_part$& is not available in the transport, because it is not unique.

Unless your host is doing nothing but mailing list deliveries, you should
probably use a separate transport for the VERP deliveries, so as not to use
extra resources in making one-per-recipient copies for other deliveries. This
can easily be done by expanding the &%transport%& option in the router:
.code
dnslookup:
  driver = dnslookup
  domains = ! +local_domains
  transport = \
    ${if match {$return_path}{^(.+?)-request@your.dom.example\$}\
      {verp_smtp}{remote_smtp}}
  no_more
.endd
If you want to change the return path using &%errors_to%& in a router instead
of using &%return_path%& in the transport, you need to set &%errors_to%& on all
routers that handle mailing list addresses. This will ensure that all delivery
errors, including those detected on the local host, are sent to the VERP
address.

On a host that does no local deliveries and has no manual routing, only the
&(dnslookup)& router needs to be changed. A special transport is not needed for
SMTP deliveries. Every mailing list recipient has its own return path value,
and so Exim must hand them to the transport one at a time. Here is an example
of a &(dnslookup)& router that implements VERP:
.code
verp_dnslookup:
  driver = dnslookup
  domains = ! +local_domains
  transport = remote_smtp
  errors_to = \
    ${if match {$return_path}{^(.+?)-request@your.dom.example\$}}
     {${quote_local_part:$1-request+$local_part=$domain}@your.dom.example}fail}
  no_more
.endd
Before you start sending out messages with VERPed return paths, you must also
configure Exim to accept the bounce messages that come back to those paths.
Typically this is done by setting a &%local_part_suffix%& option for a
router, and using this to route the messages to wherever you want to handle
them.

The overhead incurred in using VERP depends very much on the size of the
message, the number of recipient addresses that resolve to the same remote
host, and the speed of the connection over which the message is being sent. If
a lot of addresses resolve to the same host and the connection is slow, sending
a separate copy of the message for each address may take substantially longer
than sending a single copy with many recipients (for which VERP cannot be
used).






.section "Virtual domains" "SECTvirtualdomains"
.cindex "virtual domains"
.cindex "domain" "virtual"
The phrase &'virtual domain'& is unfortunately used with two rather different
meanings:

.ilist
A domain for which there are no real mailboxes; all valid local parts are
aliases for other email addresses. Common examples are organizational
top-level domains and &"vanity"& domains.
.next
One of a number of independent domains that are all handled by the same host,
with mailboxes on that host, but where the mailbox owners do not necessarily
have login accounts on that host.
.endlist

The first usage is probably more common, and does seem more &"virtual"& than
the second. This kind of domain can be handled in Exim with a straightforward
aliasing router. One approach is to create a separate alias file for each
virtual domain. Exim can test for the existence of the alias file to determine
whether the domain exists. The &(dsearch)& lookup type is useful here, leading
to a router of this form:
.code
virtual:
  driver = redirect
  domains = dsearch;/etc/mail/virtual
  data = ${lookup{$local_part}lsearch{/etc/mail/virtual/$domain_data}}
  no_more
.endd
The &%domains%& option specifies that the router is to be skipped, unless there
is a file in the &_/etc/mail/virtual_& directory whose name is the same as the
domain that is being processed.
The &(dsearch)& lookup used results in an untainted version of &$domain$&
being placed into the &$domain_data$& variable.

When the router runs, it looks up the local
part in the file to find a new address (or list of addresses). The &%no_more%&
setting ensures that if the lookup fails (leading to &%data%& being an empty
string), Exim gives up on the address without trying any subsequent routers.

This one router can handle all the virtual domains because the alias filenames
follow a fixed pattern. Permissions can be arranged so that appropriate people
can edit the different alias files. A successful aliasing operation results in
a new envelope recipient address, which is then routed from scratch.

The other kind of &"virtual"& domain can also be handled in a straightforward
way. One approach is to create a file for each domain containing a list of
valid local parts, and use it in a router like this:
.code
my_domains:
  driver = accept
  domains = dsearch;/etc/mail/domains
  local_parts = lsearch;/etc/mail/domains/$domain
  transport = my_mailboxes
.endd
The address is accepted if there is a file for the domain, and the local part
can be found in the file. The &%domains%& option is used to check for the
file's existence because &%domains%& is tested before the &%local_parts%&
option (see section &<<SECTrouprecon>>&). You cannot use &%require_files%&,
because that option is tested after &%local_parts%&. The transport is as
follows:
.code
my_mailboxes:
  driver = appendfile
  file = /var/mail/$domain_data/$local_part_data
  user = mail
.endd
This uses a directory of mailboxes for each domain. The &%user%& setting is
required, to specify which uid is to be used for writing to the mailboxes.

The configuration shown here is just one example of how you might support this
requirement. There are many other ways this kind of configuration can be set
up, for example, by using a database instead of separate files to hold all the
information about the domains.



.section "Multiple user mailboxes" "SECTmulbox"
.cindex "multiple mailboxes"
.cindex "mailbox" "multiple"
.cindex "local part" "prefix"
.cindex "local part" "suffix"
Heavy email users often want to operate with multiple mailboxes, into which
incoming mail is automatically sorted. A popular way of handling this is to
allow users to use multiple sender addresses, so that replies can easily be
identified. Users are permitted to add prefixes or suffixes to their local
parts for this purpose. The wildcard facility of the generic router options
&%local_part_prefix%& and &%local_part_suffix%& can be used for this. For
example, consider this router:
.code
userforward:
  driver = redirect
  check_local_user
  file = $home/.forward
  local_part_suffix = -*
  local_part_suffix_optional
  allow_filter
.endd
.vindex "&$local_part_suffix$&"
It runs a user's &_.forward_& file for all local parts of the form
&'username-*'&. Within the filter file the user can distinguish different
cases by testing the variable &$local_part_suffix$&. For example:
.code
if $local_part_suffix contains -special then
save /home/$local_part_data/Mail/special
endif
.endd
If the filter file does not exist, or does not deal with such addresses, they
fall through to subsequent routers, and, assuming no subsequent use of the
&%local_part_suffix%& option is made, they presumably fail. Thus, users have
control over which suffixes are valid.

Alternatively, a suffix can be used to trigger the use of a different
&_.forward_& file &-- which is the way a similar facility is implemented in
another MTA:
.code
userforward:
  driver = redirect
  check_local_user
  local_part_suffix = -*
  local_part_suffix_optional
  file = ${lookup {.forward$local_part_suffix} dsearch,ret=full {$home} {$value}fail}
  allow_filter
.endd
If there is no suffix, &_.forward_& is used; if the suffix is &'-special'&, for
example, &_.forward-special_& is used. Once again, if the appropriate file
does not exist, or does not deal with the address, it is passed on to
subsequent routers, which could, if required, look for an unqualified
&_.forward_& file to use as a default.



.section "Simplified vacation processing" "SECID244"
.cindex "vacation processing"
The traditional way of running the &'vacation'& program is for a user to set up
a pipe command in a &_.forward_& file
(see section &<<SECTspecitredli>>& for syntax details).
This is prone to error by inexperienced users. There are two features of Exim
that can be used to make this process simpler for users:

.ilist
A local part prefix such as &"vacation-"& can be specified on a router which
can cause the message to be delivered directly to the &'vacation'& program, or
alternatively can use Exim's &(autoreply)& transport. The contents of a user's
&_.forward_& file are then much simpler. For example:
.code
spqr, vacation-spqr
.endd
.next
The &%require_files%& generic router option can be used to trigger a
vacation delivery by checking for the existence of a certain file in the
user's home directory. The &%unseen%& generic option should also be used, to
ensure that the original delivery also proceeds. In this case, all the user has
to do is to create a file called, say, &_.vacation_&, containing a vacation
message.
.endlist

Another advantage of both these methods is that they both work even when the
use of arbitrary pipes by users is locked out.



.section "Taking copies of mail" "SECID245"
.cindex "message" "copying every"
Some installations have policies that require archive copies of all messages to
be made. A single copy of each message can easily be taken by an appropriate
command in a system filter, which could, for example, use a different file for
each day's messages.

There is also a shadow transport mechanism that can be used to take copies of
messages that are successfully delivered by local transports, one copy per
delivery. This could be used, &'inter alia'&, to implement automatic
notification of delivery by sites that insist on doing such things.



.section "Intermittently connected hosts" "SECID246"
.cindex "intermittently connected hosts"
It has become quite common (because it is cheaper) for hosts to connect to the
Internet periodically rather than remain connected all the time. The normal
arrangement is that mail for such hosts accumulates on a system that is
permanently connected.

Exim was designed for use on permanently connected hosts, and so it is not
particularly well-suited to use in an intermittently connected environment.
Nevertheless there are some features that can be used.


.section "Exim on the upstream server host" "SECID247"
It is tempting to arrange for incoming mail for the intermittently connected
host to remain in Exim's queue until the client connects. However, this
approach does not scale very well. Two different kinds of waiting message are
being mixed up in the same queue &-- those that cannot be delivered because of
some temporary problem, and those that are waiting for their destination host
to connect. This makes it hard to manage the queue, as well as wasting
resources, because each queue runner scans the entire queue.

A better approach is to separate off those messages that are waiting for an
intermittently connected host. This can be done by delivering these messages
into local files in batch SMTP, &"mailstore"&, or other envelope-preserving
format, from where they are transmitted by other software when their
destination connects. This makes it easy to collect all the mail for one host
in a single directory, and to apply local timeout rules on a per-message basis
if required.

On a very small scale, leaving the mail on Exim's queue can be made to work. If
you are doing this, you should configure Exim with a long retry period for the
intermittent host. For example:
.code
cheshire.wonderland.fict.example    *   F,5d,24h
.endd
This stops a lot of failed delivery attempts from occurring, but Exim remembers
which messages it has queued up for that host. Once the intermittent host comes
online, forcing delivery of one message (either by using the &%-M%& or &%-R%&
options, or by using the ETRN SMTP command (see section &<<SECTETRN>>&)
causes all the queued up messages to be delivered, often down a single SMTP
connection. While the host remains connected, any new messages get delivered
immediately.

If the connecting hosts do not have fixed IP addresses, that is, if a host is
issued with a different IP address each time it connects, Exim's retry
mechanisms on the holding host get confused, because the IP address is normally
used as part of the key string for holding retry information. This can be
avoided by unsetting &%retry_include_ip_address%& on the &(smtp)& transport.
Since this has disadvantages for permanently connected hosts, it is best to
arrange a separate transport for the intermittently connected ones.



.section "Exim on the intermittently connected client host" "SECID248"
The value of &%smtp_accept_queue_per_connection%& should probably be
increased, or even set to zero (that is, disabled) on the intermittently
connected host, so that all incoming messages down a single connection get
delivered immediately.

.cindex "SMTP" "passed connection"
.cindex "SMTP" "multiple deliveries"
.cindex "multiple SMTP deliveries"
.cindex "first pass routing"
Mail waiting to be sent from an intermittently connected host will probably
not have been routed, because without a connection DNS lookups are not
possible. This means that if a normal queue run is done at connection time,
each message is likely to be sent in a separate SMTP session. This can be
avoided by starting the queue run with a command line option beginning with
&%-qq%& instead of &%-q%&. In this case, the queue is scanned twice. In the
first pass, routing is done but no deliveries take place. The second pass is a
normal queue run; since all the messages have been previously routed, those
destined for the same host are likely to get sent as multiple deliveries in a
single SMTP connection.



. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Using Exim as a non-queueing client" "CHAPnonqueueing" &&&
         "Exim as a non-queueing client"
.cindex "client, non-queueing"
.cindex "smart host" "suppressing queueing"
On a personal computer, it is a common requirement for all
email to be sent to a &"smart host"&. There are plenty of MUAs that can be
configured to operate that way, for all the popular operating systems.
However, there are some MUAs for Unix-like systems that cannot be so
configured: they submit messages using the command line interface of
&_/usr/sbin/sendmail_&. Furthermore, utility programs such as &'cron'& submit
messages this way.

If the personal computer runs continuously, there is no problem, because it can
run a conventional MTA that handles delivery to the smart host, and deal with
any delays via its queueing mechanism. However, if the computer does not run
continuously or runs different operating systems at different times, queueing
email is not desirable.

There is therefore a requirement for something that can provide the
&_/usr/sbin/sendmail_& interface but deliver messages to a smart host without
any queueing or retrying facilities. Furthermore, the delivery to the smart
host should be synchronous, so that if it fails, the sending MUA is immediately
informed. In other words, we want something that extends an MUA that submits
to a local MTA via the command line so that it behaves like one that submits
to a remote smart host using TCP/SMTP.

There are a number of applications (for example, there is one called &'ssmtp'&)
that do this job. However, people have found them to be lacking in various
ways. For instance, you might want to allow aliasing and forwarding to be done
before sending a message to the smart host.

Exim already had the necessary infrastructure for doing this job. Just a few
tweaks were needed to make it behave as required, though it is somewhat of an
overkill to use a fully-featured MTA for this purpose.

.oindex "&%mua_wrapper%&"
There is a Boolean global option called &%mua_wrapper%&, defaulting false.
Setting &%mua_wrapper%& true causes Exim to run in a special mode where it
assumes that it is being used to &"wrap"& a command-line MUA in the manner
just described. As well as setting &%mua_wrapper%&, you also need to provide a
compatible router and transport configuration. Typically there will be just one
router and one transport, sending everything to a smart host.

When run in MUA wrapping mode, the behaviour of Exim changes in the
following ways:

.ilist
A daemon cannot be run, nor will Exim accept incoming messages from &'inetd'&.
In other words, the only way to submit messages is via the command line.
.next
Each message is synchronously delivered as soon as it is received (&%-odi%& is
assumed). All queueing options (&%queue_only%&, &%queue_smtp_domains%&,
&%control%& in an ACL, etc.) are quietly ignored. The Exim reception process
does not finish until the delivery attempt is complete. If the delivery is
successful, a zero return code is given.
.next
Address redirection is permitted, but the final routing for all addresses must
be to the same remote transport, and to the same list of hosts. Furthermore,
the return address (envelope sender) must be the same for all recipients, as
must any added or deleted header lines. In other words, it must be possible to
deliver the message in a single SMTP transaction, however many recipients there
are.
.next
If these conditions are not met, or if routing any address results in a
failure or defer status, or if Exim is unable to deliver all the recipients
successfully to one of the smart hosts, delivery of the entire message fails.
.next
Because no queueing is allowed, all failures are treated as permanent; there
is no distinction between 4&'xx'& and 5&'xx'& SMTP response codes from the
smart host. Furthermore, because only a single yes/no response can be given to
the caller, it is not possible to deliver to some recipients and not others. If
there is an error (temporary or permanent) for any recipient, all are failed.
.next
If more than one smart host is listed, Exim will try another host after a
connection failure or a timeout, in the normal way. However, if this kind of
failure happens for all the hosts, the delivery fails.
.next
When delivery fails, an error message is written to the standard error stream
(as well as to Exim's log), and Exim exits to the caller with a return code
value 1. The message is expunged from Exim's spool files. No bounce messages
are ever generated.
.next
No retry data is maintained, and any retry rules are ignored.
.next
A number of Exim options are overridden: &%deliver_drop_privilege%& is forced
true, &%max_rcpt%& in the &(smtp)& transport is forced to &"unlimited"&,
&%remote_max_parallel%& is forced to one, and fallback hosts are ignored.
.endlist

The overall effect is that Exim makes a single synchronous attempt to deliver
the message, failing if there is any kind of problem. Because no local
deliveries are done and no daemon can be run, Exim does not need root
privilege. It should be possible to run it setuid to &'exim'& instead of setuid
to &'root'&. See section &<<SECTrunexiwitpri>>& for a general discussion about
the advantages and disadvantages of running without root privilege.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Log files" "CHAPlog"
.scindex IIDloggen "log" "general description"
.cindex "log" "types of"
Exim writes three different logs, referred to as the main log, the reject log,
and the panic log:

.ilist
.cindex "main log"
The main log records the arrival of each message and each delivery in a single
line in each case. The format is as compact as possible, in an attempt to keep
down the size of log files. Two-character flag sequences make it easy to pick
out these lines. A number of other events are recorded in the main log. Some of
them are optional, in which case the &%log_selector%& option controls whether
they are included or not. A Perl script called &'eximstats'&, which does simple
analysis of main log files, is provided in the Exim distribution (see section
&<<SECTmailstat>>&).
.next
.cindex "reject log"
The reject log records information from messages that are rejected as a result
of a configuration option (that is, for policy reasons).
The first line of each rejection is a copy of the line that is also written to
the main log. Then, if the message's header has been read at the time the log
is written, its contents are written to this log. Only the original header
lines are available; header lines added by ACLs are not logged. You can use the
reject log to check that your policy controls are working correctly; on a busy
host this may be easier than scanning the main log for rejection messages. You
can suppress the writing of the reject log by setting &%write_rejectlog%&
false.
.next
.cindex "panic log"
.cindex "system log"
When certain serious errors occur, Exim writes entries to its panic log. If the
error is sufficiently disastrous, Exim bombs out afterwards. Panic log entries
are usually written to the main log as well, but can get lost amid the mass of
other entries. The panic log should be empty under normal circumstances. It is
therefore a good idea to check it (or to have a &'cron'& script check it)
regularly, in order to become aware of any problems. When Exim cannot open its
panic log, it tries as a last resort to write to the system log (syslog). This
is opened with LOG_PID+LOG_CONS and the facility code of LOG_MAIL. The
message itself is written at priority LOG_CRIT.
.endlist

Every log line starts with a timestamp, in the format shown in the following
example. Note that many of the examples shown in this chapter are line-wrapped.
In the log file, this would be all on one line:
.code
2001-09-16 16:09:47 SMTP connection from [127.0.0.1] closed
  by QUIT
.endd
By default, the timestamps are in the local timezone. There are two
ways of changing this:

.ilist
You can set the &%timezone%& option to a different time zone; in particular, if
you set
.code
timezone = UTC
.endd
the timestamps will be in UTC (aka GMT).
.next
If you set &%log_timezone%& true, the time zone is added to the timestamp, for
example:
.code
2003-04-25 11:17:07 +0100 Start queue run: pid=12762
.endd
.endlist

.cindex "log" "process ids in"
.cindex "pid (process id)" "in log lines"
Exim does not include its process id in log lines by default, but you can
request that it does so by specifying the &`pid`& log selector (see section
&<<SECTlogselector>>&). When this is set, the process id is output, in square
brackets, immediately after the time and date.




.section "Where the logs are written" "SECTwhelogwri"
.cindex "log" "destination"
.cindex "log" "to file"
.cindex "log" "to syslog"
.cindex "syslog"
The logs may be written to local files, or to syslog, or both. However, it
should be noted that many syslog implementations use UDP as a transport, and
are therefore unreliable in the sense that messages are not guaranteed to
arrive at the loghost, nor is the ordering of messages necessarily maintained.
It has also been reported that on large log files (tens of megabytes) you may
need to tweak syslog to prevent it syncing the file with each write &-- on
Linux this has been seen to make syslog take 90% plus of CPU time.

The destination for Exim's logs is configured by setting LOG_FILE_PATH in
&_Local/Makefile_& or by setting &%log_file_path%& in the runtime
configuration. This latter string is expanded, so it can contain, for example,
references to the host name:
.code
log_file_path = /var/log/$primary_hostname/exim_%slog
.endd
It is generally advisable, however, to set the string in &_Local/Makefile_&
rather than at runtime, because then the setting is available right from the
start of Exim's execution. Otherwise, if there's something it wants to log
before it has read the configuration file (for example, an error in the
configuration file) it will not use the path you want, and may not be able to
log at all.

The value of LOG_FILE_PATH or &%log_file_path%& is a colon-separated
list, currently limited to at most two items. This is one option where the
facility for changing a list separator may not be used. The list must always be
colon-separated. If an item in the list is &"syslog"& then syslog is used;
otherwise the item must either be an absolute path, containing &`%s`& at the
point where &"main"&, &"reject"&, or &"panic"& is to be inserted, or be empty,
implying the use of a default path.

When Exim encounters an empty item in the list, it searches the list defined by
LOG_FILE_PATH, and uses the first item it finds that is neither empty nor
&"syslog"&. This means that an empty item in &%log_file_path%& can be used to
mean &"use the path specified at build time"&. If no such item exists, log
files are written in the &_log_& subdirectory of the spool directory. This is
equivalent to the configuration file setting:
.code
log_file_path = $spool_directory/log/%slog
.endd
If you do not specify anything at build time or runtime,
or if you unset the option at runtime (i.e. &`log_file_path = `&),
that is where the logs are written.

A log file path may also contain &`%D`& or &`%M`& if datestamped log filenames
are in use &-- see section &<<SECTdatlogfil>>& below.

Here are some examples of possible Makefile settings:
.display
&`LOG_FILE_PATH=syslog                    `& syslog only
&`LOG_FILE_PATH=:syslog                   `& syslog and default path
&`LOG_FILE_PATH=syslog : /usr/log/exim_%s `& syslog and specified path
&`LOG_FILE_PATH=/usr/log/exim_%s          `& specified path only
.endd
If there are more than two paths in the list, the first is used and a panic
error is logged.



.section "Logging to local files that are periodically &""cycled""&" "SECID285"
.cindex "log" "cycling local files"
.cindex "cycling logs"
.cindex "&'exicyclog'&"
.cindex "log" "local files; writing to"
Some operating systems provide centralized and standardized methods for cycling
log files. For those that do not, a utility script called &'exicyclog'& is
provided (see section &<<SECTcyclogfil>>&). This renames and compresses the
main and reject logs each time it is called. The maximum number of old logs to
keep can be set. It is suggested this script is run as a daily &'cron'& job.

An Exim delivery process opens the main log when it first needs to write to it,
and it keeps the file open in case subsequent entries are required &-- for
example, if a number of different deliveries are being done for the same
message. However, remote SMTP deliveries can take a long time, and this means
that the file may be kept open long after it is renamed if &'exicyclog'& or
something similar is being used to rename log files on a regular basis. To
ensure that a switch of log files is noticed as soon as possible, Exim calls
&[stat()]& on the main log's name before reusing an open file, and if the file
does not exist, or its inode has changed, the old file is closed and Exim
tries to open the main log from scratch. Thus, an old log file may remain open
for quite some time, but no Exim processes should write to it once it has been
renamed.



.section "Datestamped log files" "SECTdatlogfil"
.cindex "log" "datestamped files"
Instead of cycling the main and reject log files by renaming them
periodically, some sites like to use files whose names contain a datestamp,
for example, &_mainlog-20031225_&. The datestamp is in the form &_yyyymmdd_& or
&_yyyymm_&. Exim has support for this way of working. It is enabled by setting
the &%log_file_path%& option to a path that includes &`%D`& or &`%M`& at the
point where the datestamp is required. For example:
.code
log_file_path = /var/spool/exim/log/%slog-%D
log_file_path = /var/log/exim-%s-%D.log
log_file_path = /var/spool/exim/log/%D-%slog
log_file_path = /var/log/exim/%s.%M
.endd
As before, &`%s`& is replaced by &"main"& or &"reject"&; the following are
examples of names generated by the above examples:
.code
/var/spool/exim/log/mainlog-20021225
/var/log/exim-reject-20021225.log
/var/spool/exim/log/20021225-mainlog
/var/log/exim/main.200212
.endd
When this form of log file is specified, Exim automatically switches to new
files at midnight. It does not make any attempt to compress old logs; you
will need to write your own script if you require this. You should not
run &'exicyclog'& with this form of logging.

The location of the panic log is also determined by &%log_file_path%&, but it
is not datestamped, because rotation of the panic log does not make sense.
When generating the name of the panic log, &`%D`& or &`%M`& are removed from
the string. In addition, if it immediately follows a slash, a following
non-alphanumeric character is removed; otherwise a preceding non-alphanumeric
character is removed. Thus, the four examples above would give these panic
log names:
.code
/var/spool/exim/log/paniclog
/var/log/exim-panic.log
/var/spool/exim/log/paniclog
/var/log/exim/panic
.endd


.section "Logging to syslog" "SECID249"
.cindex "log" "syslog; writing to"
The use of syslog does not change what Exim logs or the format of its messages,
except in one respect. If &%syslog_timestamp%& is set false, the timestamps on
Exim's log lines are omitted when these lines are sent to syslog. Apart from
that, the same strings are written to syslog as to log files. The syslog
&"facility"& is set to LOG_MAIL, and the program name to &"exim"&
by default, but you can change these by setting the &%syslog_facility%& and
&%syslog_processname%& options, respectively. If Exim was compiled with
SYSLOG_LOG_PID set in &_Local/Makefile_& (this is the default in
&_src/EDITME_&), then, on systems that permit it (all except ULTRIX), the
LOG_PID flag is set so that the &[syslog()]& call adds the pid as well as
the time and host name to each line.
The three log streams are mapped onto syslog priorities as follows:

.ilist
&'mainlog'& is mapped to LOG_INFO
.next
&'rejectlog'& is mapped to LOG_NOTICE
.next
&'paniclog'& is mapped to LOG_ALERT
.endlist

Many log lines are written to both &'mainlog'& and &'rejectlog'&, and some are
written to both &'mainlog'& and &'paniclog'&, so there will be duplicates if
these are routed by syslog to the same place. You can suppress this duplication
by setting &%syslog_duplication%& false.

Exim's log lines can sometimes be very long, and some of its &'rejectlog'&
entries contain multiple lines when headers are included. To cope with both
these cases, entries written to syslog are split into separate &[syslog()]&
calls at each internal newline, and also after a maximum of
870 data characters. (This allows for a total syslog line length of 1024, when
additions such as timestamps are added.) If you are running a syslog
replacement that can handle lines longer than the 1024 characters allowed by
&url(https://www.rfc-editor.org/rfc/rfc3164,RFC 3164), you should set
.code
SYSLOG_LONG_LINES=yes
.endd
in &_Local/Makefile_& before building Exim. That stops Exim from splitting long
lines, but it still splits at internal newlines in &'reject'& log entries.

To make it easy to re-assemble split lines later, each component of a split
entry starts with a string of the form [<&'n'&>/<&'m'&>] or [<&'n'&>\<&'m'&>]
where <&'n'&> is the component number and <&'m'&> is the total number of
components in the entry. The / delimiter is used when the line was split
because it was too long; if it was split because of an internal newline, the \
delimiter is used. For example, supposing the length limit to be 50 instead of
870, the following would be the result of a typical rejection message to
&'mainlog'& (LOG_INFO), each line in addition being preceded by the time, host
name, and pid as added by syslog:
.code
[1/5] 2002-09-16 16:09:43 16RdAL-0006pc-00 rejected from
[2/5]  [127.0.0.1] (ph10): syntax error in 'From' header
[3/5]  when scanning for sender: missing or malformed lo
[4/5] cal part in "<>" (envelope sender is <ph10@cam.exa
[5/5] mple>)
.endd
The same error might cause the following lines to be written to &"rejectlog"&
(LOG_NOTICE):
.code
[1/18] 2002-09-16 16:09:43 16RdAL-0006pc-00 rejected fro
[2/18] m [127.0.0.1] (ph10): syntax error in 'From' head
[3/18] er when scanning for sender: missing or malformed
[4/18]  local part in "<>" (envelope sender is <ph10@cam
[5\18] .example>)
[6\18] Recipients: ph10@some.domain.cam.example
[7\18] P Received: from [127.0.0.1] (ident=ph10)
[8\18]        by xxxxx.cam.example with smtp (Exim 4.00)
[9\18]        id 16RdAL-0006pc-00
[10/18]        for ph10@cam.example; Mon, 16 Sep 2002 16:
[11\18] 09:43 +0100
[12\18] F From: <>
[13\18]   Subject: this is a test header
[18\18]   X-something: this is another header
[15/18] I Message-Id: <E16RdAL-0006pc-00@xxxxx.cam.examp
[16\18] le>
[17\18] B Bcc:
[18/18]   Date: Mon, 16 Sep 2002 16:09:43 +0100
.endd
Log lines that are neither too long nor contain newlines are written to syslog
without modification.

If only syslog is being used, the Exim monitor is unable to provide a log tail
display, unless syslog is routing &'mainlog'& to a file on the local host and
the environment variable EXIMON_LOG_FILE_PATH is set to tell the monitor
where it is.



.section "Log line flags" "SECID250"
One line is written to the main log for each message received, and for each
successful, unsuccessful, and delayed delivery. These lines can readily be
picked out by the distinctive two-character flags that immediately follow the
timestamp. The flags are:
.itable none 0 0 2 10* left 90* left
.irow &%<=%&     "message arrival"
.irow &%(=%&     "message fakereject"
.irow &%=>%&     "normal message delivery"
.irow &%->%&     "additional address in same delivery"
.irow &%>>%&     "cutthrough message delivery"
.irow &%*>%&     "delivery suppressed by &%-N%&"
.irow &%**%&     "delivery failed; address bounced"
.irow &%==%&     "delivery deferred; temporary problem"
.endtable


.section "Logging message reception" "SECID251"
.cindex "log" "reception line"
The format of the single-line entry in the main log that is written for every
message received is shown in the basic example below, which is split over
several lines in order to fit it on the page:
.code
2002-10-31 08:57:53 16ZCW1-0005MB-00 <= kryten@dwarf.fict.example
  H=mailer.fict.example [192.168.123.123] U=exim
  P=smtp S=5678 id=<incoming message id>
.endd
The address immediately following &"<="& is the envelope sender address. A
bounce message is shown with the sender address &"<>"&, and if it is locally
generated, this is followed by an item of the form
.code
R=<message id>
.endd
which is a reference to the message that caused the bounce to be sent.

.cindex "HELO"
.cindex "EHLO"
For messages from other hosts, the H and U fields identify the remote host and
record the RFC 1413 identity of the user that sent the message, if one was
received. The number given in square brackets is the IP address of the sending
host. If there is a single, unparenthesized  host name in the H field, as
above, it has been verified to correspond to the IP address (see the
&%host_lookup%& option). If the name is in parentheses, it was the name quoted
by the remote host in the SMTP HELO or EHLO command, and has not been
verified. If verification yields a different name to that given for HELO or
EHLO, the verified name appears first, followed by the HELO or EHLO
name in parentheses.

Misconfigured hosts (and mail forgers) sometimes put an IP address, with or
without brackets, in the HELO or EHLO command, leading to entries in
the log containing text like these examples:
.code
H=(10.21.32.43) [192.168.8.34]
H=([10.21.32.43]) [192.168.8.34]
.endd
This can be confusing. Only the final address in square brackets can be relied
on.

For locally generated messages (that is, messages not received over TCP/IP),
the H field is omitted, and the U field contains the login name of the caller
of Exim.

.cindex "authentication" "logging"
.cindex "AUTH" "logging"
For all messages, the P field specifies the protocol used to receive the
message. This is the value that is stored in &$received_protocol$&. In the case
of incoming SMTP messages, the value indicates whether or not any SMTP
extensions (ESMTP), encryption, or authentication were used. If the SMTP
session was encrypted, there is an additional X field that records the cipher
suite that was used.

.cindex log protocol
The protocol is set to &"esmtpsa"& or &"esmtpa"& for messages received from
hosts that have authenticated themselves using the SMTP AUTH command. The first
value is used when the SMTP connection was encrypted (&"secure"&). In this case
there is an additional item A= followed by the name of the authenticator that
was used. If an authenticated identification was set up by the authenticator's
&%server_set_id%& option, this is logged too, separated by a colon from the
authenticator name.

.cindex "size" "of message"
The id field records the existing message id, if present. The size of the
received message is given by the S field. When the message is delivered,
headers may be removed or added, so that the size of delivered copies of the
message may not correspond with this value (and indeed may be different to each
other).

The &%log_selector%& option can be used to request the logging of additional
data when a message is received. See section &<<SECTlogselector>>& below.



.section "Logging deliveries" "SECID252"
.cindex "log" "delivery line"
The format of the single-line entry in the main log that is written for every
delivery is shown in one of the examples below, for local and remote
deliveries, respectively. Each example has been split into multiple lines in order
to fit it on the page:
.code
2002-10-31 08:59:13 16ZCW1-0005MB-00 => marv
  <marv@hitch.fict.example> R=localuser T=local_delivery
2002-10-31 09:00:10 16ZCW1-0005MB-00 =>
  monk@holistic.fict.example R=dnslookup T=remote_smtp
  H=holistic.fict.example [192.168.234.234]
.endd
For ordinary local deliveries, the original address is given in angle brackets
after the final delivery address, which might be a pipe or a file. If
intermediate address(es) exist between the original and the final address, the
last of these is given in parentheses after the final address. The R and T
fields record the router and transport that were used to process the address.

If SMTP AUTH was used for the delivery there is an additional item A=
followed by the name of the authenticator that was used.
If an authenticated identification was set up by the authenticator's &%client_set_id%&
option, this is logged too, as a second colon-separated list item.
Optionally (see the &%smtp_mailauth%& &%log_selector%&) there may be a third list item.

If a shadow transport was run after a successful local delivery, the log line
for the successful delivery has an item added on the end, of the form
.display
&`ST=<`&&'shadow transport name'&&`>`&
.endd
If the shadow transport did not succeed, the error message is put in
parentheses afterwards.

.cindex "asterisk" "after IP address"
When more than one address is included in a single delivery (for example, two
SMTP RCPT commands in one transaction) the second and subsequent addresses are
flagged with &`->`& instead of &`=>`&. When two or more messages are delivered
down a single SMTP connection, an asterisk follows the
remote IP address (and port if enabled)
in the log lines for the second and subsequent messages.
When two or more messages are delivered down a single TLS connection, the
DNS and some TLS-related information logged for the first message delivered
will not be present in the log lines for the second and subsequent messages.
TLS cipher information is still available.

.cindex "delivery" "cutthrough; logging"
.cindex "cutthrough" "logging"
When delivery is done in cutthrough mode it is flagged with &`>>`& and the log
line precedes the reception line, since cutthrough waits for a possible
rejection from the destination in case it can reject the sourced item.

The generation of a reply message by a filter file gets logged as a
&"delivery"& to the addressee, preceded by &">"&.

The &%log_selector%& option can be used to request the logging of additional
data when a message is delivered. See section &<<SECTlogselector>>& below.


.section "Discarded deliveries" "SECID253"
.cindex "discarded messages"
.cindex "message" "discarded"
.cindex "delivery" "discarded; logging"
When a message is discarded as a result of the command &"seen finish"& being
obeyed in a filter file which generates no deliveries, a log entry of the form
.code
2002-12-10 00:50:49 16auJc-0001UB-00 => discarded
  <low.club@bridge.example> R=userforward
.endd
is written, to record why no deliveries are logged. When a message is discarded
because it is aliased to &":blackhole:"& the log line is like this:
.code
1999-03-02 09:44:33 10HmaX-0005vi-00 => :blackhole:
  <hole@nowhere.example> R=blackhole_router
.endd


.section "Deferred deliveries" "SECID254"
When a delivery is deferred, a line of the following form is logged:
.code
2002-12-19 16:20:23 16aiQz-0002Q5-00 == marvin@endrest.example
  R=dnslookup T=smtp defer (146): Connection refused
.endd
In the case of remote deliveries, the error is the one that was given for the
last IP address that was tried. Details of individual SMTP failures are also
written to the log, so the above line would be preceded by something like
.code
2002-12-19 16:20:23 16aiQz-0002Q5-00 Failed to connect to
  mail1.endrest.example [192.168.239.239]: Connection refused
.endd
When a deferred address is skipped because its retry time has not been reached,
a message is written to the log, but this can be suppressed by setting an
appropriate value in &%log_selector%&.



.section "Delivery failures" "SECID255"
.cindex "delivery" "failure; logging"
If a delivery fails because an address cannot be routed, a line of the
following form is logged:
.code
1995-12-19 16:20:23 0tRiQz-0002Q5-00 ** jim@trek99.example
  <jim@trek99.example>: unknown mail domain
.endd
If a delivery fails at transport time, the router and transport are shown, and
the response from the remote host is included, as in this example:
.code
2002-07-11 07:14:17 17SXDU-000189-00 ** ace400@pb.example
  R=dnslookup T=remote_smtp: SMTP error from remote mailer
  after pipelined RCPT TO:<ace400@pb.example>: host
  pbmail3.py.example [192.168.63.111]: 553 5.3.0
  <ace400@pb.example>...Addressee unknown
.endd
The word &"pipelined"& indicates that the SMTP PIPELINING extension was being
used. See &%hosts_avoid_esmtp%& in the &(smtp)& transport for a way of
disabling PIPELINING. The log lines for all forms of delivery failure are
flagged with &`**`&.



.section "Fake deliveries" "SECID256"
.cindex "delivery" "fake; logging"
If a delivery does not actually take place because the &%-N%& option has been
used to suppress it, a normal delivery line is written to the log, except that
&"=>"& is replaced by &"*>"&.



.section "Completion" "SECID257"
A line of the form
.code
2002-10-31 09:00:11 16ZCW1-0005MB-00 Completed
.endd
is written to the main log when a message is about to be removed from the spool
at the end of its processing.




.section "Summary of Fields in Log Lines" "SECID258"
.cindex "log" "summary of fields"
A summary of the field identifiers that are used in log lines is shown in
the following table:
.display
&`A   `&        authenticator name (and optional id and sender)
&`C   `&        SMTP confirmation on delivery
&`Ci  `&        connection identifier
&`    `&        command list for &"no mail in SMTP session"&
&`CV  `&        certificate verification status
&`D   `&        duration of &"no mail in SMTP session"&
&`DKIM`&        domain verified in incoming message
&`DN  `&        distinguished name from peer certificate
&`DS  `&        DNSSEC secured lookups
&`DT  `&        on &`=>`&, &'=='& and &'**'& lines: time taken for, or to attempt, a delivery
&`F   `&        sender address (on delivery lines)
&`H   `&        host name and IP address
&`I   `&        local interface used
&`id  `&        message id (from header) for incoming message
&`K   `&        CHUNKING extension used
&`L   `&        on &`<=`& and &`=>`& lines: PIPELINING extension used
&`M8S `&        8BITMIME status for incoming message
&`P   `&        on &`<=`& lines: protocol used
&`    `&        on &`=>`& and &`**`& lines: return path
&`PRDR`&        PRDR extension used
&`PRX `&        on &`<=`& and &`=>`& lines: proxy address
&`Q   `&        alternate queue name
&`QT  `&        on &`=>`& lines: time spent on queue so far
&`    `&        on &"Completed"& lines: time spent on queue
&`R   `&        on &`<=`& lines: reference for local bounce
&`    `&        on &`=>`&  &`>>`& &`**`& and &`==`& lines: router name
&`RT  `&        on &`<=`& lines: time taken for reception
&`S   `&        size of message in bytes
&`SNI `&        server name indication from TLS client hello
&`ST  `&        shadow transport name
&`T   `&        on &`<=`& lines: message subject (topic)
&`TFO `&        connection took advantage of TCP Fast Open
&`    `&        on &`=>`& &`**`& and &`==`& lines: transport name
&`U   `&        local user or RFC 1413 identity
&`X   `&        TLS cipher suite
.endd


.section "Other log entries" "SECID259"
Various other types of log entry are written from time to time. Most should be
self-explanatory. Among the more common are:

.ilist
.cindex "retry" "time not reached"
&'retry time not reached'&&~&~An address previously suffered a temporary error
during routing or local delivery, and the time to retry has not yet arrived.
This message is not written to an individual message log file unless it happens
during the first delivery attempt.
.next
&'retry time not reached for any host'&&~&~An address previously suffered
temporary errors during remote delivery, and the retry time has not yet arrived
for any of the hosts to which it is routed.
.next
.cindex "spool directory" "file locked"
&'spool file locked'&&~&~An attempt to deliver a message cannot proceed because
some other Exim process is already working on the message. This can be quite
common if queue running processes are started at frequent intervals. The
&'exiwhat'& utility script can be used to find out what Exim processes are
doing.
.next
.cindex "error" "ignored"
&'error ignored'&&~&~There are several circumstances that give rise to this
message:
.olist
Exim failed to deliver a bounce message whose age was greater than
&%ignore_bounce_errors_after%&. The bounce was discarded.
.next
A filter file set up a delivery using the &"noerror"& option, and the delivery
failed. The delivery was discarded.
.next
A delivery set up by a router configured with
. ==== As this is a nested list, any displays it contains must be indented
. ==== as otherwise they are too far to the left.
.code
    errors_to = <>
.endd
failed. The delivery was discarded.
.endlist olist
.next
.cindex DKIM "log line"
&'DKIM: d='&&~&~Verbose results of a DKIM verification attempt, if enabled for
logging and the message has a DKIM signature header.
.endlist ilist





.section "Reducing or increasing what is logged" "SECTlogselector"
.cindex "log" "selectors"
By setting the &%log_selector%& global option, you can disable some of Exim's
default logging to the main log, or you can request additional logging. The value of
&%log_selector%& is made up of names preceded by plus or minus characters. For
example:
.code
log_selector = +arguments -retry_defer
.endd
The list of optional log items is in the following table, with the default
selection marked by asterisks:
.itable none 0 0 3 2.8in left 10pt center 3in left
.irow &`8bitmime`&                      &nbsp; "received 8BITMIME status"
.irow &`acl_warn_skipped`&              *       "skipped &%warn%& statement in ACL"
.irow &`address_rewrite`&               &nbsp; "address rewriting"
.irow &`all_parents`&                   &nbsp; "all parents in => lines"
.irow &`arguments`&                     &nbsp; "command line arguments"
.irow &`connection_id`&                 &nbsp;  "connection identifier"
.irow &`connection_reject`&             *       "connection rejections"
.irow &`delay_delivery`&                *       "immediate delivery delayed"
.irow &`deliver_time`&                  &nbsp; "time taken to attempt delivery"
.irow &`delivery_size`&                 &nbsp; "add &`S=`&&'nnn'& to => lines"
.irow &`dkim`&                          *       "DKIM verified domain on <= lines"
.irow &`dkim_verbose`&                  &nbsp; "separate full DKIM verification result line, per signature; DKIM signing"
.irow &`dnslist_defer`&                 *       "defers of DNS list (aka RBL) lookups"
.irow &`dnssec`&                        &nbsp; "DNSSEC secured lookups"
.irow &`etrn`&                          *       "ETRN commands"
.irow &`host_lookup_failed`&            *       "as it says"
.irow &`ident_timeout`&                 &nbsp; "timeout for ident connection"
.irow &`incoming_interface`&            &nbsp; "local interface &amp; port on <= and => lines"
.irow &`incoming_port`&                 &nbsp; "remote port on <= lines"
.irow &`lost_incoming_connection`&      *       "as it says (includes timeouts)"
.irow &`millisec`&                      &nbsp; "millisecond timestamps and RT,QT,DT,D times"
.irow &`msg_id`&                        *       "on <= lines, Message-ID: header value"
.irow &`msg_id_created`&                &nbsp; "on <= lines, Message-ID: header value when one had to be added"
.irow &`outgoing_interface`&            &nbsp; "local interface on => lines"
.irow &`outgoing_port`&                 &nbsp; "add remote port to => lines"
.irow &`queue_run`&                     *       "start and end queue runs"
.irow &`queue_time`&                    &nbsp; "time on queue for one recipient"
.irow &`queue_time_exclusive`&          &nbsp; "exclude recieve time from QT times"
.irow &`queue_time_overall`&            &nbsp; "time on queue for whole message"
.irow &`pid`&                           &nbsp; "Exim process id"
.irow &`pipelining`&                    &nbsp; "PIPELINING use, on <= and => lines"
.irow &`proxy`&                         &nbsp; "proxy address on <= and => lines"
.irow &`receive_time`&                  &nbsp; "time taken to receive message"
.irow &`received_recipients`&           &nbsp; "recipients on <= lines"
.irow &`received_sender`&               &nbsp; "sender on <= lines"
.irow &`rejected_header`&               *       "header contents on reject log"
.irow &`retry_defer`&                   *       "&<quote>&retry time not reached&</quote>&"
.irow &`return_path_on_delivery`&       &nbsp; "put return path on => and ** lines"
.irow &`sender_on_delivery`&            &nbsp; "add sender to => lines"
.irow &`sender_verify_fail`&            *       "sender verification failures"
.irow &`size_reject`&                   *       "rejection because too big"
.irow &`skip_delivery`&                 *       "delivery skipped in a queue run"
.irow &`smtp_confirmation`&             *       "SMTP confirmation on => lines"
.irow &`smtp_connection`&               &nbsp; "incoming SMTP connections"
.irow &`smtp_incomplete_transaction`&   &nbsp; "incomplete SMTP transactions"
.irow &`smtp_mailauth`&                 &nbsp; "AUTH argument to MAIL commands"
.irow &`smtp_no_mail`&                  &nbsp; "session with no MAIL commands"
.irow &`smtp_protocol_error`&           &nbsp; "SMTP protocol errors"
.irow &`smtp_syntax_error`&             &nbsp; "SMTP syntax errors"
.irow &`subject`&                       &nbsp; "contents of &'Subject:'& on <= lines"
.irow &`tls_certificate_verified`&      *       "certificate verification status"
.irow &`tls_cipher`&                    *       "TLS cipher suite on <= and => lines"
.irow &`tls_peerdn`&                    &nbsp; "TLS peer DN on <= and => lines"
.irow &`tls_resumption`&                &nbsp; "append * to cipher field"
.irow &`tls_sni`&                       &nbsp; "TLS SNI on <= lines"
.irow &`unknown_in_list`&               &nbsp; "lookup failed in list match"
.irow &`all`&                           &nbsp; "&*all of the above*&"
.endtable
See also the &%slow_lookup_log%& main configuration option,
section &<<SECID99>>&

More details on each of these items follows:

.ilist
.cindex "8BITMIME"
.cindex "log" "8BITMIME"
&%8bitmime%&: This causes Exim to log any 8BITMIME status of received messages,
which may help in tracking down interoperability issues with ancient MTAs
that are not 8bit clean.  This is added to the &"<="& line, tagged with
&`M8S=`& and a value of &`0`&, &`7`& or &`8`&, corresponding to "not given",
&`7BIT`& and &`8BITMIME`& respectively.
.next
.cindex "&%warn%& ACL verb" "log when skipping"
&%acl_warn_skipped%&: When an ACL &%warn%& statement is skipped because one of
its conditions cannot be evaluated, a log line to this effect is written if
this log selector is set.
.next
.cindex "log" "rewriting"
.cindex "rewriting" "logging"
&%address_rewrite%&: This applies both to global rewrites and per-transport
rewrites, but not to rewrites in filters run as an unprivileged user (because
such users cannot access the log).
.next
.cindex "log" "full parentage"
&%all_parents%&: Normally only the original and final addresses are logged on
delivery lines; with this selector, intermediate parents are given in
parentheses between them.
.next
.cindex "log" "Exim arguments"
.cindex "Exim arguments, logging"
&%arguments%&: This causes Exim to write the arguments with which it was called
to the main log, preceded by the current working directory. This is a debugging
feature, added to make it easier to find out how certain MUAs call
&_/usr/sbin/sendmail_&. The logging does not happen if Exim has given up root
privilege because it was called with the &%-C%& or &%-D%& options. Arguments
that are empty or that contain white space are quoted. Non-printing characters
are shown as escape sequences. This facility cannot log unrecognized arguments,
because the arguments are checked before the configuration file is read. The
only way to log such cases is to interpose a script such as &_util/logargs.sh_&
between the caller and Exim.
.next
.cindex log "connection identifier"
.cindex connection "identifier logging"
&%connection_id%&: An identifier for the accepted connection is added to
connection start and end lines and to message accept lines.
The identifier is tagged by Ci=.
The value is PID-based, so will reset on reboot and will wrap.
.next
.cindex log "connection rejections"
.cindex connection "rejection logging"
&%connection_reject%&: A log entry is written whenever an incoming SMTP
connection is rejected, for whatever reason.
.next
.cindex log "delayed delivery"
.cindex "delayed delivery, logging"
&%delay_delivery%&: A log entry is written whenever a delivery process is not
started for an incoming message because the load is too high or too many
messages were received on one connection. Logging does not occur if no delivery
process is started because &%queue_only%& is set or &%-odq%& was used.
.next
.cindex "log" "delivery duration"
&%deliver_time%&: For each delivery, the amount of real time it has taken to
perform the actual delivery is logged as DT=<&'time'&>, for example, &`DT=1s`&.
If millisecond logging is enabled, short times will be shown with greater
precision, eg. &`DT=0.304s`&.
.next
.cindex "log" "message size on delivery"
.cindex "size" "of message"
&%delivery_size%&: For each delivery, the size of message delivered is added to
the &"=>"& line, tagged with S=.
.next
.cindex log "DKIM verification"
.cindex DKIM "verification logging"
&%dkim%&: For message acceptance log lines, when an DKIM signature in the header
verifies successfully a tag of DKIM is added, with one of the verified domains.
.next
.cindex log "DKIM verification"
.cindex DKIM "verification logging"
&%dkim_verbose%&: A log entry is written for each attempted DKIM verification.
Also, on message delivery lines signing information (domain and selector)
is added, tagged with DKIM=.
.next
.cindex "log" "dnslist defer"
.cindex "DNS list" "logging defer"
.cindex "black list (DNS)"
&%dnslist_defer%&: A log entry is written if an attempt to look up a host in a
DNS black list suffers a temporary error.
.next
.cindex log dnssec
.cindex dnssec logging
&%dnssec%&: For message acceptance and (attempted) delivery log lines, when
dns lookups gave secure results a tag of DS is added.
For acceptance this covers the reverse and forward lookups for host name verification.
It does not cover helo-name verification.
For delivery this covers the SRV, MX, A and/or AAAA lookups.
.next
.cindex "log" "ETRN commands"
.cindex "ETRN" "logging"
&%etrn%&: Every valid ETRN command that is received is logged, before the ACL
is run to determine whether or not it is actually accepted. An invalid ETRN
command, or one received within a message transaction is not logged by this
selector (see &%smtp_syntax_error%& and &%smtp_protocol_error%&).
.next
.cindex "log" "host lookup failure"
&%host_lookup_failed%&: When a lookup of a host's IP addresses fails to find
any addresses, or when a lookup of an IP address fails to find a host name, a
log line is written. This logging does not apply to direct DNS lookups when
routing email addresses, but it does apply to &"byname"& lookups.
.next
.cindex "log" "ident timeout"
.cindex "RFC 1413" "logging timeout"
&%ident_timeout%&: A log line is written whenever an attempt to connect to a
client's ident port times out.
.next
.cindex "log" "incoming interface"
.cindex "log" "outgoing interface"
.cindex "log" "local interface"
.cindex "log" "local address and port"
.cindex "TCP/IP" "logging local address and port"
.cindex "interface" "logging"
&%incoming_interface%&: The interface on which a message was received is added
to the &"<="& line as an IP address in square brackets, tagged by I= and
followed by a colon and the port number. The local interface and port are also
added to other SMTP log lines, for example, &"SMTP connection from"&, to
rejection lines, and (despite the name) to outgoing
&"=>"&, &"->"&, &"=="& and &"**"& lines.
The latter can be disabled by turning off the &%outgoing_interface%& option.
.next
.cindex log "incoming proxy address"
.cindex proxy "logging proxy address"
.cindex "TCP/IP" "logging proxy address"
&%proxy%&: The internal (closest to the system running Exim) IP address
of the proxy, tagged by PRX=, on the &"<="& line for a message accepted
on a proxied connection
or the &"=>"& line for a message delivered on a proxied connection.
See &<<SECTproxyInbound>>& for more information.
.next
.cindex "log" "incoming remote port"
.cindex "port" "logging remote"
.cindex "TCP/IP" "logging incoming remote port"
.vindex "&$sender_fullhost$&"
.vindex "&$sender_rcvhost$&"
&%incoming_port%&: The remote port number from which a message was received is
added to log entries and &'Received:'& header lines, following the IP address
in square brackets, and separated from it by a colon. This is implemented by
changing the value that is put in the &$sender_fullhost$& and
&$sender_rcvhost$& variables. Recording the remote port number has become more
important with the widening use of NAT
(see &url(https://www.rfc-editor.org/rfc/rfc2505,RFC 2505)).
.next
.cindex "log" "dropped connection"
&%lost_incoming_connection%&: A log line is written when an incoming SMTP
connection is unexpectedly dropped.
.next
.cindex "log" "millisecond timestamps"
.cindex millisecond logging
.cindex timestamps "millisecond, in logs"
&%millisec%&: Timestamps have a period and three decimal places of finer granularity
appended to the seconds value.
.next
.cindex "log" "message id"
&%msg_id%&: The value of the Message-ID: header.
.next
&%msg_id_created%&: The value of the Message-ID: header, when one had to be created.
This will be either because the message is a bounce, or was submitted locally
(submission mode) without one.
The field identifier will have an asterix appended: &"id*="&.
.next
.cindex "log" "outgoing interface"
.cindex "log" "local interface"
.cindex "log" "local address and port"
.cindex "TCP/IP" "logging local address and port"
.cindex "interface" "logging"
&%outgoing_interface%&: If &%incoming_interface%& is turned on, then the
interface on which a message was sent is added to delivery lines as an I= tag
followed by IP address in square brackets. You can disable this by turning
off the &%outgoing_interface%& option.
.next
.cindex "log" "outgoing remote port"
.cindex "port" "logging outgoing remote"
.cindex "TCP/IP" "logging outgoing remote port"
&%outgoing_port%&: The remote port number is added to delivery log lines (those
containing => tags) following the IP address.
The local port is also added if &%incoming_interface%& and
&%outgoing_interface%& are both enabled.
This option is not included in the default setting, because for most ordinary
configurations, the remote port number is always 25 (the SMTP port), and the
local port is a random ephemeral port.
.next
.cindex "log" "process ids in"
.cindex "pid (process id)" "in log lines"
&%pid%&: The current process id is added to every log line, in square brackets,
immediately after the time and date.
.next
.cindex log pipelining
.cindex pipelining "logging outgoing"
&%pipelining%&: A field is added to delivery and accept
log lines when the ESMTP PIPELINING extension was used.
The field is a single "L".

On accept lines, where PIPELINING was offered but not used by the client,
the field has a minus appended.

.cindex "pipelining" "early connection"
If Exim is built without the DISABLE_PIPE_CONNECT build option
accept "L" fields have a period appended if the feature was
offered but not used, or an asterisk appended if used.
Delivery "L" fields have an asterisk appended if used.

.next
.cindex "log" "queue run"
.cindex "queue runner" "logging"
&%queue_run%&: The start and end of every queue run are logged.
.next
.cindex "log" "queue time"
&%queue_time%&: The amount of time the message has been in the queue on the
local host is logged as QT=<&'time'&> on delivery (&`=>`&) lines, for example,
&`QT=3m45s`&.
If millisecond logging is enabled, short times will be shown with greater
precision, eg. &`QT=1.578s`&.
.next
&%queue_time_overall%&: The amount of time the message has been in the queue on
the local host is logged as QT=<&'time'&> on &"Completed"& lines, for
example, &`QT=3m45s`&.
.next
.cindex "log" "receive duration"
&%receive_time%&: For each message, the amount of real time it has taken to
perform the reception is logged as RT=<&'time'&>, for example, &`RT=1s`&.
If millisecond logging is enabled, short times will be shown with greater
precision, eg. &`RT=0.204s`&.
.next
.cindex "log" "recipients"
&%received_recipients%&: The recipients of a message are listed in the main log
as soon as the message is received. The list appears at the end of the log line
that is written when a message is received, preceded by the word &"for"&. The
addresses are listed after they have been qualified, but before any rewriting
has taken place.
Recipients that were discarded by an ACL for MAIL or RCPT do not appear
in the list.
.next
.cindex "log" "sender reception"
&%received_sender%&: The unrewritten original sender of a message is added to
the end of the log line that records the message's arrival, after the word
&"from"& (before the recipients if &%received_recipients%& is also set).
.next
.cindex "log" "header lines for rejection"
&%rejected_header%&: If a message's header has been received at the time a
rejection is written to the reject log, the complete header is added to the
log. Header logging can be turned off individually for messages that are
rejected by the &[local_scan()]& function (see section &<<SECTapiforloc>>&).
.next
.cindex "log" "retry defer"
&%retry_defer%&: A log line is written if a delivery is deferred because a
retry time has not yet been reached. However, this &"retry time not reached"&
message is always omitted from individual message logs after the first delivery
attempt.
.next
.cindex "log" "return path"
&%return_path_on_delivery%&: The return path that is being transmitted with
the message is included in delivery and bounce lines, using the tag P=.
This is omitted if no delivery actually happens, for example, if routing fails,
or if delivery is to &_/dev/null_& or to &`:blackhole:`&.
.next
.cindex "log" "sender on delivery"
&%sender_on_delivery%&: The message's sender address is added to every delivery
and bounce line, tagged by F= (for &"from"&).
This is the original sender that was received with the message; it is not
necessarily the same as the outgoing return path.
.next
.cindex "log" "sender verify failure"
&%sender_verify_fail%&: If this selector is unset, the separate log line that
gives details of a sender verification failure is not written. Log lines for
the rejection of SMTP commands contain just &"sender verify failed"&, so some
detail is lost.
.next
.cindex "log" "size rejection"
&%size_reject%&: A log line is written whenever a message is rejected because
it is too big.
.next
.cindex "log" "frozen messages; skipped"
.cindex "frozen messages" "logging skipping"
&%skip_delivery%&: A log line is written whenever a message is skipped during a
queue run because it another process is already delivering it or because
it is frozen.
.cindex "&""spool file is locked""&"
.cindex "&""message is frozen""&"
The message that is written is either &"spool file is locked"& or
&"message is frozen"&.
.next
.cindex "log" "smtp confirmation"
.cindex "SMTP" "logging confirmation"
.cindex "LMTP" "logging confirmation"
&%smtp_confirmation%&: The response to the final &"."& in the SMTP or LMTP dialogue for
outgoing messages is added to delivery log lines in the form &`C=`&<&'text'&>.
A number of MTAs (including Exim) return an identifying string in this
response.
.next
.cindex "log" "SMTP connections"
.cindex "SMTP" "logging connections"
&%smtp_connection%&: A log line is written whenever an incoming SMTP connection is
established or closed, unless the connection is from a host that matches
&%hosts_connection_nolog%&. (In contrast, &%lost_incoming_connection%& applies
only when the closure is unexpected.) This applies to connections from local
processes that use &%-bs%& as well as to TCP/IP connections. If a connection is
dropped in the middle of a message, a log line is always written, whether or
not this selector is set, but otherwise nothing is written at the start and end
of connections unless this selector is enabled.

For TCP/IP connections to an Exim daemon, the current number of connections is
included in the log message for each new connection, but note that the count is
reset if the daemon is restarted.
Also, because connections are closed (and the closure is logged) in
subprocesses, the count may not include connections that have been closed but
whose termination the daemon has not yet noticed. Thus, while it is possible to
match up the opening and closing of connections in the log, the value of the
logged counts may not be entirely accurate.
.next
.cindex "log" "SMTP transaction; incomplete"
.cindex "SMTP" "logging incomplete transactions"
&%smtp_incomplete_transaction%&: When a mail transaction is aborted by
RSET, QUIT, loss of connection, or otherwise, the incident is logged,
and the message sender plus any accepted recipients are included in the log
line. This can provide evidence of dictionary attacks.
.next
.cindex "log" "non-MAIL SMTP sessions"
.cindex "MAIL" "logging session without"
&%smtp_no_mail%&: A line is written to the main log whenever an accepted SMTP
connection terminates without having issued a MAIL command. This includes both
the case when the connection is dropped, and the case when QUIT is used. It
does not include cases where the connection is rejected right at the start (by
an ACL, or because there are too many connections, or whatever). These cases
already have their own log lines.

The log line that is written contains the identity of the client in the usual
way, followed by D= and a time, which records the duration of the connection.
If the connection was authenticated, this fact is logged exactly as it is for
an incoming message, with an A= item. If the connection was encrypted, CV=,
DN=, and X= items may appear as they do for an incoming message, controlled by
the same logging options.

Finally, if any SMTP commands were issued during the connection, a C= item
is added to the line, listing the commands that were used. For example,
.code
C=EHLO,QUIT
.endd
shows that the client issued QUIT straight after EHLO. If there were fewer
than 20 commands, they are all listed. If there were more than 20 commands,
the last 20 are listed, preceded by &"..."&. However, with the default
setting of 10 for &%smtp_accept_max_nonmail%&, the connection will in any case
have been aborted before 20 non-mail commands are processed.
.next
&%smtp_mailauth%&: A third subfield with the authenticated sender,
colon-separated, is appended to the A= item for a message arrival or delivery
log line, if an AUTH argument to the SMTP MAIL command (see &<<SECTauthparamail>>&)
was accepted or used.
.next
.cindex "log" "SMTP protocol error"
.cindex "SMTP" "logging protocol error"
&%smtp_protocol_error%&: A log line is written for every SMTP protocol error
encountered. Exim does not have perfect detection of all protocol errors
because of transmission delays and the use of pipelining. If PIPELINING has
been advertised to a client, an Exim server assumes that the client will use
it, and therefore it does not count &"expected"& errors (for example, RCPT
received after rejecting MAIL) as protocol errors.
.next
.cindex "SMTP" "logging syntax errors"
.cindex "SMTP" "syntax errors; logging"
.cindex "SMTP" "unknown command; logging"
.cindex "log" "unknown SMTP command"
.cindex "log" "SMTP syntax error"
&%smtp_syntax_error%&: A log line is written for every SMTP syntax error
encountered. An unrecognized command is treated as a syntax error. For an
external connection, the host identity is given; for an internal connection
using &%-bs%& the sender identification (normally the calling user) is given.
.next
.cindex "log" "subject"
.cindex "subject, logging"
&%subject%&: The subject of the message is added to the arrival log line,
preceded by &"T="& (T for &"topic"&, since S is already used for &"size"&).
Any MIME &"words"& in the subject are decoded. The &%print_topbitchars%& option
specifies whether characters with values greater than 127 should be logged
unchanged, or whether they should be rendered as escape sequences.
.next
.cindex "log" "certificate verification"
.cindex log DANE
.cindex DANE logging
&%tls_certificate_verified%&: An extra item is added to <= and => log lines
when TLS is in use. The item is &`CV=yes`& if the peer's certificate was
verified
using a CA trust anchor,
&`CV=dane`& if using a DNS trust anchor,
and &`CV=no`& if not.
.next
.cindex "log" "TLS cipher"
.cindex "TLS" "logging cipher"
&%tls_cipher%&: When a message is sent or received over an encrypted
connection, the cipher suite used is added to the log line, preceded by X=.
.next
.cindex "log" "TLS peer DN"
.cindex "TLS" "logging peer DN"
&%tls_peerdn%&: When a message is sent or received over an encrypted
connection, and a certificate is supplied by the remote host, the peer DN is
added to the log line, preceded by DN=.
.next
.cindex "log" "TLS resumption"
.cindex "TLS" "logging session resumption"
&%tls_resumption%&: When a message is sent or received over an encrypted
connection and the TLS session resumed one used on a previous TCP connection,
an asterisk is appended to the X= cipher field in the log line.
.next
.cindex "log" "TLS SNI"
.cindex "TLS" "logging SNI"
.cindex SNI logging
&%tls_sni%&: When a message is received over an encrypted connection, and
the remote host provided the Server Name Indication extension, the SNI is
added to the log line, preceded by SNI=.
.next
.cindex "log" "DNS failure in list"
&%unknown_in_list%&: This setting causes a log entry to be written when the
result of a list match is failure because a DNS lookup failed, or because
a bad IP address was in the list.
.endlist


.section "Message log" "SECID260"
.cindex "message" "log file for"
.cindex "log" "message log; description of"
.cindex "&_msglog_& directory"
.oindex "&%preserve_message_logs%&"
In addition to the general log files, Exim writes a log file for each message
that it handles. The names of these per-message logs are the message ids, and
they are kept in the &_msglog_& sub-directory of the spool directory. Each
message log contains copies of the log lines that apply to the message. This
makes it easier to inspect the status of an individual message without having
to search the main log. A message log is deleted when processing of the message
is complete, unless &%preserve_message_logs%& is set, but this should be used
only with great care because they can fill up your disk very quickly.

On a heavily loaded system, it may be desirable to disable the use of
per-message logs, in order to reduce disk I/O. This can be done by setting the
&%message_logs%& option false.
.ecindex IIDloggen




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Exim utilities" "CHAPutils"
.scindex IIDutils "utilities"
A number of utility scripts and programs are supplied with Exim and are
described in this chapter. There is also the Exim Monitor, which is covered in
the next chapter. The utilities described here are:

.itable none 0 0 3  7* left  15* left  40* left
.irow &<<SECTfinoutwha>>&     &'exiwhat'&       &&&
  "list what Exim processes are doing"
.irow &<<SECTgreptheque>>&    &'exiqgrep'&      "grep the queue"
.irow &<<SECTsumtheque>>&     &'exiqsumm'&      "summarize the queue"
.irow &<<SECTextspeinf>>&     &'exigrep'&       "search the main log"
.irow &<<SECTexipick>>&       &'exipick'&       "select messages on &&&
                                                various criteria"
.irow &<<SECTcyclogfil>>&     &'exicyclog'&     "cycle (rotate) log files"
.irow &<<SECTmailstat>>&      &'eximstats'&     &&&
  "extract statistics from the log"
.irow &<<SECTcheckaccess>>&   &'exim_checkaccess'& &&&
  "check address acceptance from given IP"
.irow &<<SECTdbmbuild>>&      &'exim_dbmbuild'& "build a DBM file"
.irow &<<SECTfinindret>>&     &'exinext'&       "extract retry information"
.irow &<<SECTdumpdb>>&        &'exim_dumpdb'&   "dump a hints database"
.irow &<<SECTtidydb>>&        &'exim_tidydb'&   "clean up a hints database"
.irow &<<SECTfixdb>>&         &'exim_fixdb'&    "patch a hints database"
.irow &<<SECTmailboxmaint>>&  &'exim_lock'&     "lock a mailbox file"
.irow &<<SECTexim_msgdate>>&  &'exim_msgdate'&  "Message Ids for humans (exim_msgdate)"
.endtable

Another utility that might be of use to sites with many MTAs is Tom Kistner's
&'exilog'&. It provides log visualizations across multiple Exim servers. See
&url(https://duncanthrax.net/exilog/) for details.




.section "Finding out what Exim processes are doing (exiwhat)" "SECTfinoutwha"
.cindex "&'exiwhat'&"
.cindex "process, querying"
.cindex "SIGUSR1"
On operating systems that can restart a system call after receiving a signal
(most modern OS), an Exim process responds to the SIGUSR1 signal by writing
a line describing what it is doing to the file &_exim-process.info_& in the
Exim spool directory. The &'exiwhat'& script sends the signal to all Exim
processes it can find, having first emptied the file. It then waits for one
second to allow the Exim processes to react before displaying the results. In
order to run &'exiwhat'& successfully you have to have sufficient privilege to
send the signal to the Exim processes, so it is normally run as root.

&*Warning*&: This is not an efficient process. It is intended for occasional
use by system administrators. It is not sensible, for example, to set up a
script that sends SIGUSR1 signals to Exim processes at short intervals.


Unfortunately, the &'ps'& command that &'exiwhat'& uses to find Exim processes
varies in different operating systems. Not only are different options used,
but the format of the output is different. For this reason, there are some
system configuration options that configure exactly how &'exiwhat'& works. If
it doesn't seem to be working for you, check the following compile-time
options:
.itable none 0 0 2 30* left 70* left
.irow &`EXIWHAT_PS_CMD`&        "the command for running &'ps'&"
.irow &`EXIWHAT_PS_ARG`&        "the argument for &'ps'&"
.irow &`EXIWHAT_EGREP_ARG`&     "the argument for &'egrep'& to select from &'ps'& output"
.irow &`EXIWHAT_KILL_ARG`&      "the argument for the &'kill'& command"
.endtable
An example of typical output from &'exiwhat'& is
.code
164 daemon: -q1h, listening on port 25
10483 running queue: waiting for 0tAycK-0002ij-00 (10492)
10492 delivering 0tAycK-0002ij-00 to mail.ref.example
  [10.19.42.42] (editor@ref.example)
10592 handling incoming call from [192.168.243.242]
10628 accepting a local non-SMTP message
.endd
The first number in the output line is the process number. The third line has
been split here, in order to fit it on the page.



.section "Selective queue listing (exiqgrep)" "SECTgreptheque"
.cindex "&'exiqgrep'&"
.cindex "queue" "grepping"
This utility is a Perl script contributed by Matt Hubbard. It runs
.code
exim -bpu
.endd
or (in case &*-a*& switch is specified)
.code
exim -bp
.endd
to obtain a queue listing, and then greps the output to select messages
that match given criteria. The following selection options are available:

.vlist
.vitem &*-f*&&~<&'regex'&>
Match the sender address using a case-insensitive search. The field that is
tested is enclosed in angle brackets, so you can test for bounce messages with
.code
exiqgrep -f '^<>$'
.endd
.vitem &*-r*&&~<&'regex'&>
Match a recipient address using a case-insensitive search. The field that is
tested is not enclosed in angle brackets.

.vitem &*-s*&&~<&'regex'&>
Match against the size field.

.vitem &*-y*&&~<&'seconds'&>
Match messages that are younger than the given time.

.vitem &*-o*&&~<&'seconds'&>
Match messages that are older than the given time.

.vitem &*-z*&
Match only frozen messages.

.vitem &*-x*&
Match only non-frozen messages.

.vitem &*-G*&&~<&'queuename'&>
Match only messages in the given queue.  Without this, the default queue is searched.
.endlist

The following options control the format of the output:

.vlist
.vitem &*-c*&
Display only the count of matching messages.

.vitem &*-l*&
Long format &-- display the full message information as output by Exim. This is
the default.

.vitem &*-i*&
Display message ids only.

.vitem &*-b*&
Brief format &-- one line per message.

.vitem &*-R*&
Display messages in reverse order.

.vitem &*-a*&
Include delivered recipients in queue listing.
.endlist

The following options give alternates for configuration:

.vlist
.vitem &*-C*&&~<&'config&~file'&>
is used to specify an alternate &_exim.conf_& which might
contain alternate exim configuration the queue management might be using.

.vitem &*-E*&&~<&'path'&>
can be used to specify a path for the exim binary,
overriding the built-in one.
.endlist

There is one more option, &%-h%&, which outputs a list of options.
At least one selection option, or either the &*-c*& or &*-h*& option, must be given.



.section "Summarizing the queue (exiqsumm)" "SECTsumtheque"
.cindex "&'exiqsumm'&"
.cindex "queue" "summary"
The &'exiqsumm'& utility is a Perl script which reads the output of &`exim
-bp`& and produces a summary of the messages in the queue. Thus, you use it by
running a command such as
.code
exim -bp | exiqsumm
.endd
The output consists of one line for each domain that has messages waiting for
it, as in the following example:
.code
3   2322   74m   66m  msn.com.example
.endd
Each line lists the number of pending deliveries for a domain, their total
volume, and the length of time that the oldest and the newest messages have
been waiting. Note that the number of pending deliveries is greater than the
number of messages when messages have more than one recipient.

A summary line is output at the end. By default the output is sorted on the
domain name, but &'exiqsumm'& has the options &%-a%& and &%-c%&, which cause
the output to be sorted by oldest message and by count of messages,
respectively. There are also three options that split the messages for each
domain into two or more subcounts: &%-b%& separates bounce messages, &%-f%&
separates frozen messages, and &%-s%& separates messages according to their
sender.

The output of &'exim -bp'& contains the original addresses in the message, so
this also applies to the output from &'exiqsumm'&. No domains from addresses
generated by aliasing or forwarding are included (unless the &%one_time%&
option of the &(redirect)& router has been used to convert them into &"top
level"& addresses).




.section "Extracting specific information from the log (exigrep)" &&&
         "SECTextspeinf"
.cindex "&'exigrep'&"
.cindex "log" "extracts; grepping for"
The &'exigrep'& utility is a Perl script that searches one or more main log
files for entries that match a given pattern. When it finds a match, it
extracts all the log entries for the relevant message, not just those that
match the pattern. Thus, &'exigrep'& can extract complete log entries for a
given message, or all mail for a given user, or for a given host, for example.
The input files can be in Exim log format or syslog format.
If a matching log line is not associated with a specific message, it is
included in &'exigrep'&'s output without any additional lines. The usage is:
.display
&`exigrep [-t<`&&'n'&&`>] [-I] [-l] [-M] [-v] <`&&'pattern'&&`> [<`&&'log file'&&`>] ...`&
.endd
If no log filenames are given on the command line, the standard input is read.

The &%-t%& argument specifies a number of seconds. It adds an additional
condition for message selection. Messages that are complete are shown only if
they spent more than <&'n'&> seconds in the queue.

By default, &'exigrep'& does case-insensitive matching. The &%-I%& option
makes it case-sensitive. This may give a performance improvement when searching
large log files. Without &%-I%&, the Perl pattern matches use Perl's &`/i`&
option; with &%-I%& they do not. In both cases it is possible to change the
case sensitivity within the pattern by using &`(?i)`& or &`(?-i)`&.

The &%-l%& option means &"literal"&, that is, treat all characters in the
pattern as standing for themselves. Otherwise the pattern must be a Perl
regular expression.

The &%-v%& option inverts the matching condition. That is, a line is selected
if it does &'not'& match the pattern.

The &%-M%& options means &"related messages"&. &'exigrep'& will show messages
that are generated as a result/response to a message that &'exigrep'& matched
normally.

Example of &%-M%&:
user_a sends a message to user_b, which generates a bounce back to user_b. If
&'exigrep'& is used to search for &"user_a"&, only the first message will be
displayed.  But if &'exigrep'& is used to search for &"user_b"&, the first and
the second (bounce) message will be displayed. Using &%-M%& with &'exigrep'&
when searching for &"user_a"& will show both messages since the bounce is
&"related"& to or a &"result"& of the first message that was found by the
search term.

If the location of a &'zcat'& command is known from the definition of
ZCAT_COMMAND in &_Local/Makefile_&, &'exigrep'& automatically passes any file
whose name ends in COMPRESS_SUFFIX through &'zcat'& as it searches it.
If the ZCAT_COMMAND is not executable, &'exigrep'& tries to use
autodetection of some well known compression extensions.


.section "Selecting messages by various criteria (exipick)" "SECTexipick"
.cindex "&'exipick'&"
John Jetmore's &'exipick'& utility is included in the Exim distribution. It
lists messages from the queue according to a variety of criteria. For details
of &'exipick'&'s facilities, run &'exipick'& with
the &%--help%& option.


.section "Cycling log files (exicyclog)" "SECTcyclogfil"
.cindex "log" "cycling local files"
.cindex "cycling logs"
.cindex "&'exicyclog'&"
The &'exicyclog'& script can be used to cycle (rotate) &'mainlog'& and
&'rejectlog'& files. This is not necessary if only syslog is being used, or if
you are using log files with datestamps in their names (see section
&<<SECTdatlogfil>>&). Some operating systems have their own standard mechanisms
for log cycling, and these can be used instead of &'exicyclog'& if preferred.
There are two command line options for &'exicyclog'&:
.ilist
&%-k%& <&'count'&> specifies the number of log files to keep, overriding the
default that is set when Exim is built. The default default is 10.
.next
&%-l%& <&'path'&> specifies the log file path, in the same format as Exim's
&%log_file_path%& option (for example, &`/var/log/exim_%slog`&), again
overriding the script's default, which is to find the setting from Exim's
configuration.
.endlist

Each time &'exicyclog'& is run the filenames get &"shuffled down"& by one. If
the main log filename is &_mainlog_& (the default) then when &'exicyclog'& is
run &_mainlog_& becomes &_mainlog.01_&, the previous &_mainlog.01_& becomes
&_mainlog.02_& and so on, up to the limit that is set in the script or by the
&%-k%& option. Log files whose numbers exceed the limit are discarded. Reject
logs are handled similarly.

If the limit is greater than 99, the script uses 3-digit numbers such as
&_mainlog.001_&, &_mainlog.002_&, etc. If you change from a number less than 99
to one that is greater, or &'vice versa'&, you will have to fix the names of
any existing log files.

If no &_mainlog_& file exists, the script does nothing. Files that &"drop off"&
the end are deleted. All files with numbers greater than 01 are compressed,
using a compression command which is configured by the COMPRESS_COMMAND
setting in &_Local/Makefile_&. It is usual to run &'exicyclog'& daily from a
root &%crontab%& entry of the form
.code
1 0 * * * su exim -c /usr/exim/bin/exicyclog
.endd
assuming you have used the name &"exim"& for the Exim user. You can run
&'exicyclog'& as root if you wish, but there is no need.



.section "Mail statistics (eximstats)" "SECTmailstat"
.cindex "statistics"
.cindex "&'eximstats'&"
A Perl script called &'eximstats'& is provided for extracting statistical
information from log files. The output is either plain text, or HTML.
. --- 2018-09-07: LogReport's Lire appears to be dead; website is a Yahoo Japan
. --- 404 error and everything else points to that.

The &'eximstats'& script has been hacked about quite a bit over time. The
latest version is the result of some extensive revision by Steve Campbell. A
lot of information is given by default, but there are options for suppressing
various parts of it. Following any options, the arguments to the script are a
list of files, which should be main log files. For example:
.code
eximstats -nr /var/spool/exim/log/mainlog.01
.endd
By default, &'eximstats'& extracts information about the number and volume of
messages received from or delivered to various hosts. The information is sorted
both by message count and by volume, and the top fifty hosts in each category
are listed on the standard output. Similar information, based on email
addresses or domains instead of hosts can be requested by means of various
options. For messages delivered and received locally, similar statistics are
also produced per user.

The output also includes total counts and statistics about delivery errors, and
histograms showing the number of messages received and deliveries made in each
hour of the day. A delivery with more than one address in its envelope (for
example, an SMTP transaction with more than one RCPT command) is counted
as a single delivery by &'eximstats'&.

Though normally more deliveries than receipts are reported (as messages may
have multiple recipients), it is possible for &'eximstats'& to report more
messages received than delivered, even though the queue is empty at the start
and end of the period in question. If an incoming message contains no valid
recipients, no deliveries are recorded for it. A bounce message is handled as
an entirely separate message.

&'eximstats'& always outputs a grand total summary giving the volume and number
of messages received and deliveries made, and the number of hosts involved in
each case. It also outputs the number of messages that were delayed (that is,
not completely delivered at the first attempt), and the number that had at
least one address that failed.

The remainder of the output is in sections that can be independently disabled
or modified by various options. It consists of a summary of deliveries by
transport, histograms of messages received and delivered per time interval
(default per hour), information about the time messages spent in the queue,
a list of relayed messages, lists of the top fifty sending hosts, local
senders, destination hosts, and destination local users by count and by volume,
and a list of delivery errors that occurred.

The relay information lists messages that were actually relayed, that is, they
came from a remote host and were directly delivered to some other remote host,
without being processed (for example, for aliasing or forwarding) locally.

There are quite a few options for &'eximstats'& to control exactly what it
outputs. These are documented in the Perl script itself, and can be extracted
by running the command &(perldoc)& on the script. For example:
.code
perldoc /usr/exim/bin/eximstats
.endd

.section "Checking access policy (exim_checkaccess)" "SECTcheckaccess"
.cindex "&'exim_checkaccess'&"
.cindex "policy control" "checking access"
.cindex "checking access"
The &%-bh%& command line argument allows you to run a fake SMTP session with
debugging output, in order to check what Exim is doing when it is applying
policy controls to incoming SMTP mail. However, not everybody is sufficiently
familiar with the SMTP protocol to be able to make full use of &%-bh%&, and
sometimes you just want to answer the question &"Does this address have
access?"& without bothering with any further details.

The &'exim_checkaccess'& utility is a &"packaged"& version of &%-bh%&. It takes
two arguments, an IP address and an email address:
.code
exim_checkaccess 10.9.8.7 A.User@a.domain.example
.endd
The utility runs a call to Exim with the &%-bh%& option, to test whether the
given email address would be accepted in a RCPT command in a TCP/IP
connection from the host with the given IP address. The output of the utility
is either the word &"accepted"&, or the SMTP error response, for example:
.code
Rejected:
550 Relay not permitted
.endd
When running this test, the utility uses &`<>`& as the envelope sender address
for the MAIL command, but you can change this by providing additional
options. These are passed directly to the Exim command. For example, to specify
that the test is to be run with the sender address &'himself@there.example'&
you can use:
.code
exim_checkaccess 10.9.8.7 A.User@a.domain.example \
                 -f himself@there.example
.endd
Note that these additional Exim command line items must be given after the two
mandatory arguments.

Because the &%exim_checkaccess%& uses &%-bh%&, it does not perform callouts
while running its checks. You can run checks that include callouts by using
&%-bhc%&, but this is not yet available in a &"packaged"& form.



.section "Making DBM files (exim_dbmbuild)" "SECTdbmbuild"
.cindex "DBM" "building dbm files"
.cindex "building DBM files"
.cindex "&'exim_dbmbuild'&"
.cindex "lower casing"
.cindex "binary zero" "in lookup key"
The &'exim_dbmbuild'& program reads an input file containing keys and data in
the format used by the &(lsearch)& lookup (see section
&<<SECTsinglekeylookups>>&). It writes a DBM file using the lower-cased alias
names as keys and the remainder of the information as data. The lower-casing
can be prevented by calling the program with the &%-nolc%& option.

A terminating zero is included as part of the key string. This is expected by
the &(dbm)& lookup type. However, if the option &%-nozero%& is given,
&'exim_dbmbuild'& creates files without terminating zeroes in either the key
strings or the data strings. The &(dbmnz)& lookup type can be used with such
files.

The program requires two arguments: the name of the input file (which can be a
single hyphen to indicate the standard input), and the name of the output file.
It creates the output under a temporary name, and then renames it if all went
well.

.cindex "USE_DB"
If the native DB interface is in use (USE_DB is set in a compile-time
configuration file &-- this is common in free versions of Unix) the two
filenames must be different, because in this mode the Berkeley DB functions
create a single output file using exactly the name given. For example,
.code
exim_dbmbuild /etc/aliases /etc/aliases.db
.endd
reads the system alias file and creates a DBM version of it in
&_/etc/aliases.db_&.

In systems that use the &'ndbm'& routines (mostly proprietary versions of
Unix), two files are used, with the suffixes &_.dir_& and &_.pag_&. In this
environment, the suffixes are added to the second argument of
&'exim_dbmbuild'&, so it can be the same as the first. This is also the case
when the Berkeley functions are used in compatibility mode (though this is not
recommended), because in that case it adds a &_.db_& suffix to the filename.

If a duplicate key is encountered, the program outputs a warning, and when it
finishes, its return code is 1 rather than zero, unless the &%-noduperr%&
option is used. By default, only the first of a set of duplicates is used &--
this makes it compatible with &(lsearch)& lookups. There is an option
&%-lastdup%& which causes it to use the data for the last duplicate instead.
There is also an option &%-nowarn%&, which stops it listing duplicate keys to
&%stderr%&. For other errors, where it doesn't actually make a new file, the
return code is 2.




.section "Finding individual retry times (exinext)" "SECTfinindret"
.cindex "retry" "times"
.cindex "&'exinext'&"
A utility called &'exinext'& (mostly a Perl script) provides the ability to
fish specific information out of the retry database. Given a mail domain (or a
complete address), it looks up the hosts for that domain, and outputs any retry
information for the hosts or for the domain. At present, the retry information
is obtained by running &'exim_dumpdb'& (see below) and post-processing the
output. For example:
.code
$ exinext piglet@milne.fict.example
kanga.milne.example:192.168.8.1 error 146: Connection refused
  first failed: 21-Feb-1996 14:57:34
  last tried:   21-Feb-1996 14:57:34
  next try at:  21-Feb-1996 15:02:34
roo.milne.example:192.168.8.3 error 146: Connection refused
  first failed: 20-Jan-1996 13:12:08
  last tried:   21-Feb-1996 11:42:03
  next try at:  21-Feb-1996 19:42:03
  past final cutoff time
.endd
You can also give &'exinext'& a local part, without a domain, and it
will give any retry information for that local part in your default domain.
A message id can be used to obtain retry information pertaining to a specific
message. This exists only when an attempt to deliver a message to a remote host
suffers a message-specific error (see section &<<SECToutSMTPerr>>&).
&'exinext'& is not particularly efficient, but then it is not expected to be
run very often.

The &'exinext'& utility calls Exim to find out information such as the location
of the spool directory. The utility has &%-C%& and &%-D%& options, which are
passed on to the &'exim'& commands. The first specifies an alternate Exim
configuration file, and the second sets macros for use within the configuration
file. These features are mainly to help in testing, but might also be useful in
environments where more than one configuration file is in use.



.section "Hints database maintenance" "SECThindatmai"
.cindex "hints database" "maintenance"
.cindex "maintaining Exim's hints database"
Three utility programs are provided for maintaining the DBM files that Exim
uses to contain its delivery hint information. Each program requires two
arguments. The first specifies the name of Exim's spool directory, and the
second is the name of the database it is to operate on. These are as follows:

.ilist
&'retry'&: the database of retry information
.next
&'wait-'&<&'transport name'&>: databases of information about messages waiting
for remote hosts
.next
&'callout'&: the callout cache
.next
&'ratelimit'&: the data for implementing the ratelimit ACL condition
.next
&'tls'&: TLS session resumption data
.next
&'misc'&: other hints data
.endlist

The &'misc'& database is used for

.ilist
Serializing delivery to a specific host (when &%serialize_hosts%& is set in an
&(smtp)& transport)
.next
Limiting the concurrency of specific transports (when &%max_parallel%& is set
in a transport)
.next
Recording EHLO-time facilities advertised by hosts
.next
Serializing ETRN runs (when &%smtp_etrn_serialize%& is set)
.endlist



.subsection "exim_dumpdb" "SECTdumpdb"
.cindex "&'exim_dumpdb'&"
The entire contents of a database are written to the standard output by the
&'exim_dumpdb'& program,
taking as arguments the spool and database names.
An option &'-z'& may be given to request times in UTC;
otherwise times are in the local timezone.
An option &'-k'& may be given to dump only the record keys.
For example, to dump the retry database:
.code
exim_dumpdb /var/spool/exim retry
.endd
For the retry database
two lines of output are produced for each entry:
.code
T:mail.ref.example:192.168.242.242 146 77 Connection refused
31-Oct-1995 12:00:12 02-Nov-1995 12:21:39 02-Nov-1995 20:21:39 *
.endd
The first item on the first line is the key of the record. It starts with one
of the letters R, or T, depending on whether it refers to a routing or
transport retry. For a local delivery, the next part is the local address; for
a remote delivery it is the name of the remote host, followed by its failing IP
address (unless &%retry_include_ip_address%& is set false on the &(smtp)&
transport). If the remote port is not the standard one (port 25), it is added
to the IP address. Then there follows an error code, an additional error code,
and a textual description of the error.

The three times on the second line are the time of first failure, the time of
the last delivery attempt, and the computed time for the next attempt. The line
ends with an asterisk if the cutoff time for the last retry rule has been
exceeded.

Each output line from &'exim_dumpdb'& for the &'wait-xxx'& databases
consists of a host name followed by a list of ids for messages that are or were
waiting to be delivered to that host. If there are a very large number for any
one host, continuation records, with a sequence number added to the host name,
may be seen. The data in these records is often out of date, because a message
may be routed to several alternative hosts, and Exim makes no effort to keep
cross-references.



.subsection "exim_tidydb" "SECTtidydb"
.cindex "&'exim_tidydb'&"
The &'exim_tidydb'& utility program is used to tidy up the contents of a hints
database. If run with no options, it removes all records that are more than 30
days old. The age is calculated from the date and time that the record was last
updated. Note that, in the case of the retry database, it is &'not'& the time
since the first delivery failure. Information about a host that has been down
for more than 30 days will remain in the database, provided that the record is
updated sufficiently often.

The cutoff date can be altered by means of the &%-t%& option, which must be
followed by a time. For example, to remove all records older than a week from
the retry database:
.code
exim_tidydb -t 7d /var/spool/exim retry
.endd
Both the &'wait-xxx'& and &'retry'& databases contain items that involve
message ids. In the former these appear as data in records keyed by host &--
they were messages that were waiting for that host &-- and in the latter they
are the keys for retry information for messages that have suffered certain
types of error. When &'exim_tidydb'& is run, a check is made to ensure that
message ids in database records are those of messages that are still on the
queue. Message ids for messages that no longer exist are removed from
&'wait-xxx'& records, and if this leaves any records empty, they are deleted.
For the &'retry'& database, records whose keys are non-existent message ids are
removed. The &'exim_tidydb'& utility outputs comments on the standard output
whenever it removes information from the database.

Certain records are automatically removed by Exim when they are no longer
needed, but others are not. For example, if all the MX hosts for a domain are
down, a retry record is created for each one. If the primary MX host comes back
first, its record is removed when Exim successfully delivers to it, but the
records for the others remain because Exim has not tried to use those hosts.

It is important, therefore, to run &'exim_tidydb'& periodically on all the
hints databases. You should do this at a quiet time of day, because it requires
a database to be locked (and therefore inaccessible to Exim) while it does its
work. Removing records from a DBM file does not normally make the file smaller,
but all the common DBM libraries are able to re-use the space that is released.
After an initial phase of increasing in size, the databases normally reach a
point at which they no longer get any bigger, as long as they are regularly
tidied.

&*Warning*&: If you never run &'exim_tidydb'&, the space used by the hints
databases is likely to keep on increasing.




.subsection "exim_fixdb" "SECTfixdb"
.cindex "&'exim_fixdb'&"
The &'exim_fixdb'& program is a utility for interactively modifying databases.
Its main use is for testing Exim, but it might also be occasionally useful for
getting round problems in a live system. Its interface
is somewhat crude. On entry, it prompts for input with a right angle-bracket. A
key of a database record can then be entered, and the data for that record is
displayed.

If &"d"& is typed at the next prompt, the entire record is deleted. For all
except the &'retry'& database, that is the only operation that can be carried
out. For the &'retry'& database, each field is output preceded by a number, and
data for individual fields can be changed by typing the field number followed
by new data, for example:
.code
> 4 951102:1000
.endd
resets the time of the next delivery attempt. Time values are given as a
sequence of digit pairs for year, month, day, hour, and minute. Colons can be
used as optional separators.

Both displayed and input times are in the local timezone by default.
If an option &'-z'& is used on the command line, displayed times
are in UTC.




.section "Mailbox maintenance (exim_lock)" "SECTmailboxmaint"
.cindex "mailbox" "maintenance"
.cindex "&'exim_lock'&"
.cindex "locking mailboxes"
The &'exim_lock'& utility locks a mailbox file using the same algorithm as
Exim. For a discussion of locking issues, see section &<<SECTopappend>>&.
&'Exim_lock'& can be used to prevent any modification of a mailbox by Exim or
a user agent while investigating a problem. The utility requires the name of
the file as its first argument. If the locking is successful, the second
argument is run as a command (using C's &[system()]& function); if there is no
second argument, the value of the SHELL environment variable is used; if this
is unset or empty, &_/bin/sh_& is run. When the command finishes, the mailbox
is unlocked and the utility ends. The following options are available:

.vlist
.vitem &%-fcntl%&
Use &[fcntl()]& locking on the open mailbox.

.vitem &%-flock%&
Use &[flock()]& locking on the open mailbox, provided the operating system
supports it.

.vitem &%-interval%&
This must be followed by a number, which is a number of seconds; it sets the
interval to sleep between retries (default 3).

.vitem &%-lockfile%&
Create a lock file before opening the mailbox.

.vitem &%-mbx%&
Lock the mailbox using MBX rules.

.vitem &%-q%&
Suppress verification output.

.vitem &%-retries%&
This must be followed by a number; it sets the number of times to try to get
the lock (default 10).

.vitem &%-restore_time%&
This option causes &%exim_lock%& to restore the modified and read times to the
locked file before exiting. This allows you to access a locked mailbox (for
example, to take a backup copy) without disturbing the times that the user
subsequently sees.

.vitem &%-timeout%&
This must be followed by a number, which is a number of seconds; it sets a
timeout to be used with a blocking &[fcntl()]& lock. If it is not set (the
default), a non-blocking call is used.

.vitem &%-v%&
Generate verbose output.
.endlist

If none of &%-fcntl%&, &%-flock%&, &%-lockfile%& or &%-mbx%& are given, the
default is to create a lock file and also to use &[fcntl()]& locking on the
mailbox, which is the same as Exim's default. The use of &%-flock%& or
&%-fcntl%& requires that the file be writeable; the use of &%-lockfile%&
requires that the directory containing the file be writeable. Locking by lock
file does not last forever; Exim assumes that a lock file is expired if it is
more than 30 minutes old.

The &%-mbx%& option can be used with either or both of &%-fcntl%& or
&%-flock%&. It assumes &%-fcntl%& by default. MBX locking causes a shared lock
to be taken out on the open mailbox, and an exclusive lock on the file
&_/tmp/.n.m_& where &'n'& and &'m'& are the device number and inode
number of the mailbox file. When the locking is released, if an exclusive lock
can be obtained for the mailbox, the file in &_/tmp_& is deleted.

The default output contains verification of the locking that takes place. The
&%-v%& option causes some additional information to be given. The &%-q%& option
suppresses all output except error messages.

A command such as
.code
exim_lock /var/spool/mail/spqr
.endd
runs an interactive shell while the file is locked, whereas
.display
&`exim_lock -q /var/spool/mail/spqr <<End`&
<&'some commands'&>
&`End`&
.endd
runs a specific non-interactive sequence of commands while the file is locked,
suppressing all verification output. A single command can be run by a command
such as
.code
exim_lock -q /var/spool/mail/spqr \
  "cp /var/spool/mail/spqr /some/where"
.endd
Note that if a command is supplied, it must be entirely contained within the
second argument &-- hence the quotes.


.section "Message Ids for humans (exim_msgdate)" "SECTexim_msgdate"
.cindex "exim_msgdate"
The &'exim_msgdate'& utility is written by Andrew Aitchison and included in the Exim distribution.
This Perl script converts an Exim Mesage ID back into a human readable form.
For details of &'exim_msgdate'&'s options, run &'exim_msgdate'& with the &%--help%& option.

Section &<<SECTmessiden>>& (Message identification) describes Exim Mesage IDs.
.ecindex IIDutils
. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "The Exim monitor" "CHAPeximon"
.scindex IIDeximon "Exim monitor" "description"
.cindex "X-windows"
.cindex "&'eximon'&"
.cindex "Local/eximon.conf"
.cindex "&_exim_monitor/EDITME_&"
The Exim monitor is an application which displays in an X window information
about the state of Exim's queue and what Exim is doing. An admin user can
perform certain operations on messages from this GUI interface; however all
such facilities are also available from the command line, and indeed, the
monitor itself makes use of the command line to perform any actions requested.



.section "Running the monitor" "SECID264"
The monitor is started by running the script called &'eximon'&. This is a shell
script that sets up a number of environment variables, and then runs the
binary called &_eximon.bin_&. The default appearance of the monitor window can
be changed by editing the &_Local/eximon.conf_& file created by editing
&_exim_monitor/EDITME_&. Comments in that file describe what the various
parameters are for.

The parameters that get built into the &'eximon'& script can be overridden for
a particular invocation by setting up environment variables of the same names,
preceded by &`EXIMON_`&. For example, a shell command such as
.code
EXIMON_LOG_DEPTH=400 eximon
.endd
(in a Bourne-compatible shell) runs &'eximon'& with an overriding setting of
the LOG_DEPTH parameter. If EXIMON_LOG_FILE_PATH is set in the environment, it
overrides the Exim log file configuration. This makes it possible to have
&'eximon'& tailing log data that is written to syslog, provided that MAIL.INFO
syslog messages are routed to a file on the local host.

X resources can be used to change the appearance of the window in the normal
way. For example, a resource setting of the form
.code
Eximon*background: gray94
.endd
changes the colour of the background to light grey rather than white. The
stripcharts are drawn with both the data lines and the reference lines in
black. This means that the reference lines are not visible when on top of the
data. However, their colour can be changed by setting a resource called
&"highlight"& (an odd name, but that's what the Athena stripchart widget uses).
For example, if your X server is running Unix, you could set up lighter
reference lines in the stripcharts by obeying
.code
xrdb -merge <<End
Eximon*highlight: gray
End
.endd
.cindex "admin user"
In order to see the contents of messages in the queue, and to operate on them,
&'eximon'& must either be run as root or by an admin user.

The command-line parameters of &'eximon'& are passed to &_eximon.bin_& and may
contain X11 resource parameters interpreted by the X11 library.  In addition,
if the first parameter starts with the string "gdb" then it is removed and the
binary is invoked under gdb (the parameter is used as the gdb command-name, so
versioned variants of gdb can be invoked).

The monitor's window is divided into three parts. The first contains one or
more stripcharts and two action buttons, the second contains a &"tail"& of the
main log file, and the third is a display of the queue of messages awaiting
delivery, with two more action buttons. The following sections describe these
different parts of the display.




.section "The stripcharts" "SECID265"
.cindex "stripchart"
The first stripchart is always a count of messages in the queue. Its name can
be configured by setting QUEUE_STRIPCHART_NAME in the
&_Local/eximon.conf_& file. The remaining stripcharts are defined in the
configuration script by regular expression matches on log file entries, making
it possible to display, for example, counts of messages delivered to certain
hosts or using certain transports. The supplied defaults display counts of
received and delivered messages, and of local and SMTP deliveries. The default
period between stripchart updates is one minute; this can be adjusted by a
parameter in the &_Local/eximon.conf_& file.

The stripchart displays rescale themselves automatically as the value they are
displaying changes. There are always 10 horizontal lines in each chart; the
title string indicates the value of each division when it is greater than one.
For example, &"x2"& means that each division represents a value of 2.

It is also possible to have a stripchart which shows the percentage fullness of
a particular disk partition, which is useful when local deliveries are confined
to a single partition.

.cindex "&%statvfs%& function"
This relies on the availability of the &[statvfs()]& function or equivalent in
the operating system. Most, but not all versions of Unix that support Exim have
this. For this particular stripchart, the top of the chart always represents
100%, and the scale is given as &"x10%"&. This chart is configured by setting
SIZE_STRIPCHART and (optionally) SIZE_STRIPCHART_NAME in the
&_Local/eximon.conf_& file.




.section "Main action buttons" "SECID266"
.cindex "size" "of monitor window"
.cindex "Exim monitor" "window size"
.cindex "window size"
Below the stripcharts there is an action button for quitting the monitor. Next
to this is another button marked &"Size"&. They are placed here so that
shrinking the window to its default minimum size leaves just the queue count
stripchart and these two buttons visible. Pressing the &"Size"& button causes
the window to expand to its maximum size, unless it is already at the maximum,
in which case it is reduced to its minimum.

When expanding to the maximum, if the window cannot be fully seen where it
currently is, it is moved back to where it was the last time it was at full
size. When it is expanding from its minimum size, the old position is
remembered, and next time it is reduced to the minimum it is moved back there.

The idea is that you can keep a reduced window just showing one or two
stripcharts at a convenient place on your screen, easily expand it to show
the full window when required, and just as easily put it back to what it was.
The idea is copied from what the &'twm'& window manager does for its
&'f.fullzoom'& action. The minimum size of the window can be changed by setting
the MIN_HEIGHT and MIN_WIDTH values in &_Local/eximon.conf_&.

Normally, the monitor starts up with the window at its full size, but it can be
built so that it starts up with the window at its smallest size, by setting
START_SMALL=yes in &_Local/eximon.conf_&.



.section "The log display" "SECID267"
.cindex "log" "tail of; in monitor"
The second section of the window is an area in which a display of the tail of
the main log is maintained.
To save space on the screen, the timestamp on each log line is shortened by
removing the date and, if &%log_timezone%& is set, the timezone.
The log tail is not available when the only destination for logging data is
syslog, unless the syslog lines are routed to a local file whose name is passed
to &'eximon'& via the EXIMON_LOG_FILE_PATH environment variable.

The log sub-window has a scroll bar at its lefthand side which can be used to
move back to look at earlier text, and the up and down arrow keys also have a
scrolling effect. The amount of log that is kept depends on the setting of
LOG_BUFFER in &_Local/eximon.conf_&, which specifies the amount of memory
to use. When this is full, the earlier 50% of data is discarded &-- this is
much more efficient than throwing it away line by line. The sub-window also has
a horizontal scroll bar for accessing the ends of long log lines. This is the
only means of horizontal scrolling; the right and left arrow keys are not
available. Text can be cut from this part of the window using the mouse in the
normal way. The size of this subwindow is controlled by parameters in the
configuration file &_Local/eximon.conf_&.

Searches of the text in the log window can be carried out by means of the ^R
and ^S keystrokes, which default to a reverse and a forward search,
respectively. The search covers only the text that is displayed in the window.
It cannot go further back up the log.

The point from which the search starts is indicated by a caret marker. This is
normally at the end of the text in the window, but can be positioned explicitly
by pointing and clicking with the left mouse button, and is moved automatically
by a successful search. If new text arrives in the window when it is scrolled
back, the caret remains where it is, but if the window is not scrolled back,
the caret is moved to the end of the new text.

Pressing ^R or ^S pops up a window into which the search text can be typed.
There are buttons for selecting forward or reverse searching, for carrying out
the search, and for cancelling. If the &"Search"& button is pressed, the search
happens and the window remains so that further searches can be done. If the
&"Return"& key is pressed, a single search is done and the window is closed. If
^C is typed the search is cancelled.

The searching facility is implemented using the facilities of the Athena text
widget. By default this pops up a window containing both &"search"& and
&"replace"& options. In order to suppress the unwanted &"replace"& portion for
eximon, a modified version of the &%TextPop%& widget is distributed with Exim.
However, the linkers in BSDI and HP-UX seem unable to handle an externally
provided version of &%TextPop%& when the remaining parts of the text widget
come from the standard libraries. The compile-time option EXIMON_TEXTPOP can be
unset to cut out the modified &%TextPop%&, making it possible to build Eximon
on these systems, at the expense of having unwanted items in the search popup
window.



.section "The queue display" "SECID268"
.cindex "queue" "display in monitor"
The bottom section of the monitor window contains a list of all messages that
are in the queue, which includes those currently being received or delivered,
as well as those awaiting delivery. The size of this subwindow is controlled by
parameters in the configuration file &_Local/eximon.conf_&, and the frequency
at which it is updated is controlled by another parameter in the same file &--
the default is 5 minutes, since queue scans can be quite expensive. However,
there is an &"Update"& action button just above the display which can be used
to force an update of the queue display at any time.

When a host is down for some time, a lot of pending mail can build up for it,
and this can make it hard to deal with other messages in the queue. To help
with this situation there is a button next to &"Update"& called &"Hide"&. If
pressed, a dialogue box called &"Hide addresses ending with"& is put up. If you
type anything in here and press &"Return"&, the text is added to a chain of
such texts, and if every undelivered address in a message matches at least one
of the texts, the message is not displayed.

If there is an address that does not match any of the texts, all the addresses
are displayed as normal. The matching happens on the ends of addresses so, for
example, &'cam.ac.uk'& specifies all addresses in Cambridge, while
&'xxx@foo.com.example'& specifies just one specific address. When any hiding
has been set up, a button called &"Unhide"& is displayed. If pressed, it
cancels all hiding. Also, to ensure that hidden messages do not get forgotten,
a hide request is automatically cancelled after one hour.

While the dialogue box is displayed, you can't press any buttons or do anything
else to the monitor window. For this reason, if you want to cut text from the
queue display to use in the dialogue box, you have to do the cutting before
pressing the &"Hide"& button.

The queue display contains, for each unhidden queued message, the length of
time it has been in the queue, the size of the message, the message id, the
message sender, and the first undelivered recipient, all on one line. If it is
a bounce message, the sender is shown as &"<>"&. If there is more than one
recipient to which the message has not yet been delivered, subsequent ones are
listed on additional lines, up to a maximum configured number, following which
an ellipsis is displayed. Recipients that have already received the message are
not shown.

.cindex "frozen messages" "display"
If a message is frozen, an asterisk is displayed at the left-hand side.

The queue display has a vertical scroll bar, and can also be scrolled by means
of the arrow keys. Text can be cut from it using the mouse in the normal way.
The text searching facilities, as described above for the log window, are also
available, but the caret is always moved to the end of the text when the queue
display is updated.



.section "The queue menu" "SECID269"
.cindex "queue" "menu in monitor"
If the &%shift%& key is held down and the left button is clicked when the mouse
pointer is over the text for any message, an action menu pops up, and the first
line of the queue display for the message is highlighted. This does not affect
any selected text.

If you want to use some other event for popping up the menu, you can set the
MENU_EVENT parameter in &_Local/eximon.conf_& to change the default, or
set EXIMON_MENU_EVENT in the environment before starting the monitor. The
value set in this parameter is a standard X event description. For example, to
run eximon using &%ctrl%& rather than &%shift%& you could use
.code
EXIMON_MENU_EVENT='Ctrl<Btn1Down>' eximon
.endd
The title of the menu is the message id, and it contains entries which act as
follows:

.ilist
&'message log'&: The contents of the message log for the message are displayed
in a new text window.
.next
&'headers'&: Information from the spool file that contains the envelope
information and headers is displayed in a new text window. See chapter
&<<CHAPspool>>& for a description of the format of spool files.
.next
&'body'&: The contents of the spool file containing the body of the message are
displayed in a new text window. There is a default limit of 20,000 bytes to the
amount of data displayed. This can be changed by setting the BODY_MAX
option at compile time, or the EXIMON_BODY_MAX option at runtime.
.next
&'deliver message'&: A call to Exim is made using the &%-M%& option to request
delivery of the message. This causes an automatic thaw if the message is
frozen. The &%-v%& option is also set, and the output from Exim is displayed in
a new text window. The delivery is run in a separate process, to avoid holding
up the monitor while the delivery proceeds.
.next
&'freeze message'&: A call to Exim is made using the &%-Mf%& option to request
that the message be frozen.
.next
.cindex "thawing messages"
.cindex "unfreezing messages"
.cindex "frozen messages" "thawing"
&'thaw message'&: A call to Exim is made using the &%-Mt%& option to request
that the message be thawed.
.next
.cindex "delivery" "forcing failure"
&'give up on msg'&: A call to Exim is made using the &%-Mg%& option to request
that Exim gives up trying to deliver the message. A bounce message is generated
for any remaining undelivered addresses.
.next
&'remove message'&: A call to Exim is made using the &%-Mrm%& option to request
that the message be deleted from the system without generating a bounce
message.
.next
&'add recipient'&: A dialog box is displayed into which a recipient address can
be typed. If the address is not qualified and the QUALIFY_DOMAIN parameter
is set in &_Local/eximon.conf_&, the address is qualified with that domain.
Otherwise it must be entered as a fully qualified address. Pressing RETURN
causes a call to Exim to be made using the &%-Mar%& option to request that an
additional recipient be added to the message, unless the entry box is empty, in
which case no action is taken.
.next
&'mark delivered'&: A dialog box is displayed into which a recipient address
can be typed. If the address is not qualified and the QUALIFY_DOMAIN parameter
is set in &_Local/eximon.conf_&, the address is qualified with that domain.
Otherwise it must be entered as a fully qualified address. Pressing RETURN
causes a call to Exim to be made using the &%-Mmd%& option to mark the given
recipient address as already delivered, unless the entry box is empty, in which
case no action is taken.
.next
&'mark all delivered'&: A call to Exim is made using the &%-Mmad%& option to
mark all recipient addresses as already delivered.
.next
&'edit sender'&: A dialog box is displayed initialized with the current
sender's address. Pressing RETURN causes a call to Exim to be made using the
&%-Mes%& option to replace the sender address, unless the entry box is empty,
in which case no action is taken. If you want to set an empty sender (as in
bounce messages), you must specify it as &"<>"&. Otherwise, if the address is
not qualified and the QUALIFY_DOMAIN parameter is set in &_Local/eximon.conf_&,
the address is qualified with that domain.
.endlist

When a delivery is forced, a window showing the &%-v%& output is displayed. In
other cases when a call to Exim is made, if there is any output from Exim (in
particular, if the command fails) a window containing the command and the
output is displayed. Otherwise, the results of the action are normally apparent
from the log and queue displays. However, if you set ACTION_OUTPUT=yes in
&_Local/eximon.conf_&, a window showing the Exim command is always opened, even
if no output is generated.

The queue display is automatically updated for actions such as freezing and
thawing, unless ACTION_QUEUE_UPDATE=no has been set in
&_Local/eximon.conf_&. In this case the &"Update"& button has to be used to
force an update of the display after one of these actions.

In any text window that is displayed as result of a menu action, the normal
cut-and-paste facility is available, and searching can be carried out using ^R
and ^S, as described above for the log tail window.
.ecindex IIDeximon





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Security considerations" "CHAPsecurity"
.scindex IIDsecurcon "security" "discussion of"
This chapter discusses a number of issues concerned with security, some of
which are also covered in other parts of this manual.

For reasons that this author does not understand, some people have promoted
Exim as a &"particularly secure"& mailer. Perhaps it is because of the
existence of this chapter in the documentation. However, the intent of the
chapter is simply to describe the way Exim works in relation to certain
security concerns, not to make any specific claims about the effectiveness of
its security as compared with other MTAs.

What follows is a description of the way Exim is supposed to be. Best efforts
have been made to try to ensure that the code agrees with the theory, but an
absence of bugs can never be guaranteed. Any that are reported will get fixed
as soon as possible.


.section "Building a more &""hardened""& Exim" "SECID286"
.cindex "security" "build-time features"
There are a number of build-time options that can be set in &_Local/Makefile_&
to create Exim binaries that are &"harder"& to attack, in particular by a rogue
Exim administrator who does not have the root password, or by someone who has
penetrated the Exim (but not the root) account. These options are as follows:

.ilist
ALT_CONFIG_PREFIX can be set to a string that is required to match the
start of any filenames used with the &%-C%& option. When it is set, these
filenames are also not allowed to contain the sequence &"/../"&. (However, if
the value of the &%-C%& option is identical to the value of CONFIGURE_FILE in
&_Local/Makefile_&, Exim ignores &%-C%& and proceeds as usual.) There is no
default setting for &%ALT_CONFIG_PREFIX%&.

If the permitted configuration files are confined to a directory to
which only root has access, this guards against someone who has broken
into the Exim account from running a privileged Exim with an arbitrary
configuration file, and using it to break into other accounts.
.next

If a non-trusted configuration file (i.e. not the default configuration file
or one which is trusted by virtue of being listed in the TRUSTED_CONFIG_LIST
file) is specified with &%-C%&, or if macros are given with &%-D%& (but see
the next item), then root privilege is retained only if the caller of Exim is
root. This locks out the possibility of testing a configuration using &%-C%&
right through message reception and delivery, even if the caller is root. The
reception works, but by that time, Exim is running as the Exim user, so when
it re-execs to regain privilege for the delivery, the use of &%-C%& causes
privilege to be lost. However, root can test reception and delivery using two
separate commands.

.next
The WHITELIST_D_MACROS build option declares some macros to be safe to override
with &%-D%& if the real uid is one of root, the Exim run-time user or the
CONFIGURE_OWNER, if defined.  The potential impact of this option is limited by
requiring the run-time value supplied to &%-D%& to match a regex that errs on
the restrictive side.  Requiring build-time selection of safe macros is onerous
but this option is intended solely as a transition mechanism to permit
previously-working configurations to continue to work after release 4.73.
.next
If DISABLE_D_OPTION is defined, the use of the &%-D%& command line option
is disabled.
.next
FIXED_NEVER_USERS can be set to a colon-separated list of users that are
never to be used for any deliveries. This is like the &%never_users%& runtime
option, but it cannot be overridden; the runtime option adds additional users
to the list. The default setting is &"root"&; this prevents a non-root user who
is permitted to modify the runtime file from using Exim as a way to get root.
.endlist



.section "Root privilege" "SECID270"
.cindex "setuid"
.cindex "root privilege"
The Exim binary is normally setuid to root, which means that it gains root
privilege (runs as root) when it starts execution. In some special cases (for
example, when the daemon is not in use and there are no local deliveries), it
may be possible to run Exim setuid to some user other than root. This is
discussed in the next section. However, in most installations, root privilege
is required for two things:

.ilist
To set up a socket connected to the standard SMTP port (25) when initialising
the listening daemon. If Exim is run from &'inetd'&, this privileged action is
not required.
.next
To be able to change uid and gid in order to read users' &_.forward_& files and
perform local deliveries as the receiving user or as specified in the
configuration.
.endlist

It is not necessary to be root to do any of the other things Exim does, such as
receiving messages and delivering them externally over SMTP, and it is
obviously more secure if Exim does not run as root except when necessary.
For this reason, a user and group for Exim to use must be defined in
&_Local/Makefile_&. These are known as &"the Exim user"& and &"the Exim
group"&. Their values can be changed by the runtime configuration, though this
is not recommended. Often a user called &'exim'& is used, but some sites use
&'mail'& or another user name altogether.

Exim uses &[setuid()]& whenever it gives up root privilege. This is a permanent
abdication; the process cannot regain root afterwards. Prior to release 4.00,
&[seteuid()]& was used in some circumstances, but this is no longer the case.

After a new Exim process has interpreted its command line options, it changes
uid and gid in the following cases:

.ilist
.oindex "&%-C%&"
.oindex "&%-D%&"
If the &%-C%& option is used to specify an alternate configuration file, or if
the &%-D%& option is used to define macro values for the configuration, and the
calling process is not running as root, the uid and gid are changed to those of
the calling process.
However, if DISABLE_D_OPTION is defined in &_Local/Makefile_&, the &%-D%&
option may not be used at all.
If WHITELIST_D_MACROS is defined in &_Local/Makefile_&, then some macro values
can be supplied if the calling process is running as root, the Exim run-time
user or CONFIGURE_OWNER, if defined.
.next
.oindex "&%-be%&"
.oindex "&%-bf%&"
.oindex "&%-bF%&"
If the expansion test option (&%-be%&) or one of the filter testing options
(&%-bf%& or &%-bF%&) are used, the uid and gid are changed to those of the
calling process.
.next
If the process is not a daemon process or a queue runner process or a delivery
process or a process for testing address routing (started with &%-bt%&), the
uid and gid are changed to the Exim user and group. This means that Exim always
runs under its own uid and gid when receiving messages. This also applies when
testing address verification
.oindex "&%-bv%&"
.oindex "&%-bh%&"
(the &%-bv%& option) and testing incoming message policy controls (the &%-bh%&
option).
.next
For a daemon, queue runner, delivery, or address testing process, the uid
remains as root at this stage, but the gid is changed to the Exim group.
.endlist

The processes that initially retain root privilege behave as follows:

.ilist
A daemon process changes the gid to the Exim group and the uid to the Exim
user after setting up one or more listening sockets. The &[initgroups()]&
function is called, so that if the Exim user is in any additional groups, they
will be used during message reception.
.next
A queue runner process retains root privilege throughout its execution. Its
job is to fork a controlled sequence of delivery processes.

.next
A delivery process retains root privilege throughout most of its execution.,
including while the recipient addresses in a message are being routed.

.ilist
However, if a user's filter file has to be processed,
this is done in a subprocess that runs under the individual user's uid and
gid. A system filter is run as root unless &%system_filter_user%& is set.
.endlist

Any actual deliveries (that is, the transports themselves) are run in
subprocesses which always change to a non-root uid and gid.
.ilist
For local
deliveries this is typically the uid and gid of the owner of the mailbox.
.next
For remote deliveries, the Exim uid and gid are used.
.endlist

Once all the delivery
subprocesses have been run, a delivery process changes to the Exim uid and gid
while doing post-delivery tidying up such as updating the retry database and
generating bounce and warning messages.

.next
A process that is testing addresses (the &%-bt%& option) runs as root so that
the routing is done in the same environment as a message delivery.
.endlist




.section "Running Exim without privilege" "SECTrunexiwitpri"
.cindex "privilege, running without"
.cindex "unprivileged running"
.cindex "root privilege" "running without"
Some installations like to run Exim in an unprivileged state for more of its
operation, for added security. Support for this mode of operation is provided
by the global option &%deliver_drop_privilege%&. When this is set, the uid and
gid are changed to the Exim user and group at the start of a delivery process
(and also queue runner and address testing processes). This means that address
routing is no longer run as root, and the deliveries themselves cannot change
to any other uid.

.cindex SIGHUP
.cindex "daemon" "restarting"
Leaving the binary setuid to root, but setting &%deliver_drop_privilege%& means
that the daemon can still be started in the usual way, and it can respond
correctly to SIGHUP because the re-invocation regains root privilege.

An alternative approach is to make Exim setuid to the Exim user and also setgid
to the Exim group. If you do this, the daemon must be started from a root
process. (Calling Exim from a root process makes it behave in the way it does
when it is setuid root.) However, the daemon cannot restart itself after a
SIGHUP signal because it cannot regain privilege.

It is still useful to set &%deliver_drop_privilege%& in this case, because it
stops Exim from trying to re-invoke itself to do a delivery after a message has
been received. Such a re-invocation is a waste of resources because it has no
effect.

If restarting the daemon is not an issue (for example, if &%mua_wrapper%& is
set, or &'inetd'& is being used instead of a daemon), having the binary setuid
to the Exim user seems a clean approach, but there is one complication:

In this style of operation, Exim is running with the real uid and gid set to
those of the calling process, and the effective uid/gid set to Exim's values.
Ideally, any association with the calling process' uid/gid should be dropped,
that is, the real uid/gid should be reset to the effective values so as to
discard any privileges that the caller may have. While some operating systems
have a function that permits this action for a non-root effective uid, quite a
number of them do not. Because of this lack of standardization, Exim does not
address this problem at this time.

For this reason, the recommended approach for &"mostly unprivileged"& running
is to keep the Exim binary setuid to root, and to set
&%deliver_drop_privilege%&. This also has the advantage of allowing a daemon to
be used in the most straightforward way.

If you configure Exim not to run delivery processes as root, there are a
number of restrictions on what you can do:

.ilist
You can deliver only as the Exim user/group. You should  explicitly use the
&%user%& and &%group%& options to override routers or local transports that
normally deliver as the recipient. This makes sure that configurations that
work in this mode function the same way in normal mode. Any implicit or
explicit specification of another user causes an error.
.next
Use of &_.forward_& files is severely restricted, such that it is usually
not worthwhile to include them in the configuration.
.next
Users who wish to use &_.forward_& would have to make their home directory and
the file itself accessible to the Exim user. Pipe and append-to-file entries,
and their equivalents in Exim filters, cannot be used. While they could be
enabled in the Exim user's name, that would be insecure and not very useful.
.next
Unless the local user mailboxes are all owned by the Exim user (possible in
some POP3 or IMAP-only environments):

.olist
They must be owned by the Exim group and be writeable by that group. This
implies you must set &%mode%& in the appendfile configuration, as well as the
mode of the mailbox files themselves.
.next
You must set &%no_check_owner%&, since most or all of the files will not be
owned by the Exim user.
.next
You must set &%file_must_exist%&, because Exim cannot set the owner correctly
on a newly created mailbox when unprivileged. This also implies that new
mailboxes need to be created manually.
.endlist olist
.endlist ilist


These restrictions severely restrict what can be done in local deliveries.
However, there are no restrictions on remote deliveries. If you are running a
gateway host that does no local deliveries, setting &%deliver_drop_privilege%&
gives more security at essentially no cost.

If you are using the &%mua_wrapper%& facility (see chapter
&<<CHAPnonqueueing>>&), &%deliver_drop_privilege%& is forced to be true.




.section "Delivering to local files" "SECID271"
Full details of the checks applied by &(appendfile)& before it writes to a file
are given in chapter &<<CHAPappendfile>>&.



.section "Running local commands" "SECTsecconslocalcmds"
.cindex "security" "local commands"
.cindex "security" "command injection attacks"
There are a number of ways in which an administrator can configure Exim to run
commands based upon received, untrustworthy, data. Further, in some
configurations a user who can control a &_.forward_& file can also arrange to
run commands. Configuration to check includes, but is not limited to:

.ilist
Use of &%use_shell%& in the pipe transport: various forms of shell command
injection may be possible with this option present. It is dangerous and should
be used only with considerable caution. Consider constraints which whitelist
allowed characters in a variable which is to be used in a pipe transport that
has &%use_shell%& enabled.
.next
A number of options such as &%forbid_filter_run%&, &%forbid_filter_perl%&,
&%forbid_filter_dlfunc%& and so forth which restrict facilities available to
&_.forward_& files in a redirect router. If Exim is running on a central mail
hub to which ordinary users do not have shell access, but home directories are
NFS mounted (for instance) then administrators should review the list of these
forbid options available, and should bear in mind that the options that may
need forbidding can change as new features are added between releases.
.next
The &%${run...}%& expansion item does not use a shell by default, but
administrators can configure use of &_/bin/sh_& as part of the command.
Such invocations should be viewed with prejudicial suspicion.
.next
Administrators who use embedded Perl are advised to explore how Perl's
taint checking might apply to their usage.
.next
Use of &%${expand...}%& is somewhat analogous to shell's eval builtin and
administrators are well advised to view its use with suspicion, in case (for
instance) it allows a local-part to contain embedded Exim directives.
.next
Use of &%${match_local_part...}%& and friends becomes more dangerous if
Exim was built with EXPAND_LISTMATCH_RHS defined: the second string in
each can reference arbitrary lists and files, rather than just being a list
of opaque strings.
The EXPAND_LISTMATCH_RHS option was added and set false by default because of
real-world security vulnerabilities caused by its use with untrustworthy data
injected in, for SQL injection attacks.
Consider the use of the &%inlisti%& expansion condition instead.
.endlist




.section "Trust in configuration data" "SECTsecconfdata"
.cindex "security" "data sources"
.cindex "security" "regular expressions"
.cindex "regular expressions" "security"
.cindex "PCRE2" "security"
If configuration data for Exim can come from untrustworthy sources, there
are some issues to be aware of:

.ilist
Use of &%${expand...}%& may provide a path for shell injection attacks.
.next
Letting untrusted data provide a regular expression is unwise.
.next
Using &%${match...}%& to apply a fixed regular expression against untrusted
data may result in pathological behaviour within PCRE2.  Be aware of what
"backtracking" means and consider options for being more strict with a regular
expression. Avenues to explore include limiting what can match (avoiding &`.`&
when &`[a-z0-9]`& or other character class will do), use of atomic grouping and
possessive quantifiers or just not using regular expressions against untrusted
data.
.next
It can be important to correctly use &%${quote:...}%&,
&%${quote_local_part:...}%& and &%${quote_%&<&'lookup-type'&>&%:...}%& expansion
items to ensure that data is correctly constructed.
.next
Some lookups might return multiple results, even though normal usage is only
expected to yield one result.
.endlist




.section "IPv4 source routing" "SECID272"
.cindex "source routing" "in IP packets"
.cindex "IP source routing"
Many operating systems suppress IP source-routed packets in the kernel, but
some cannot be made to do this, so Exim does its own check. It logs incoming
IPv4 source-routed TCP calls, and then drops them. Things are all different in
IPv6. No special checking is currently done.



.section "The VRFY, EXPN, and ETRN commands in SMTP" "SECID273"
Support for these SMTP commands is disabled by default. If required, they can
be enabled by defining suitable ACLs.




.section "Privileged users" "SECID274"
.cindex "trusted users"
.cindex "admin user"
.cindex "privileged user"
.cindex "user" "trusted"
.cindex "user" "admin"
Exim recognizes two sets of users with special privileges. Trusted users are
able to submit new messages to Exim locally, but supply their own sender
addresses and information about a sending host. For other users submitting
local messages, Exim sets up the sender address from the uid, and doesn't
permit a remote host to be specified.

.oindex "&%-f%&"
However, an untrusted user is permitted to use the &%-f%& command line option
in the special form &%-f <>%& to indicate that a delivery failure for the
message should not cause an error report. This affects the message's envelope,
but it does not affect the &'Sender:'& header. Untrusted users may also be
permitted to use specific forms of address with the &%-f%& option by setting
the &%untrusted_set_sender%& option.

Trusted users are used to run processes that receive mail messages from some
other mail domain and pass them on to Exim for delivery either locally, or over
the Internet. Exim trusts a caller that is running as root, as the Exim user,
as any user listed in the &%trusted_users%& configuration option, or under any
group listed in the &%trusted_groups%& option.

Admin users are permitted to do things to the messages on Exim's queue. They
can freeze or thaw messages, cause them to be returned to their senders, remove
them entirely, or modify them in various ways. In addition, admin users can run
the Exim monitor and see all the information it is capable of providing, which
includes the contents of files on the spool.

.oindex "&%-M%&"
.oindex "&%-q%&"
By default, the use of the &%-M%& and &%-q%& options to cause Exim to attempt
delivery of messages on its queue is restricted to admin users. This
restriction can be relaxed by setting the &%no_prod_requires_admin%& option.
Similarly, the use of &%-bp%& (and its variants) to list the contents of the
queue is also restricted to admin users. This restriction can be relaxed by
setting &%no_queue_list_requires_admin%&.

Exim recognizes an admin user if the calling process is running as root or as
the Exim user or if any of the groups associated with the calling process is
the Exim group. It is not necessary actually to be running under the Exim
group. However, if admin users who are not root or the Exim user are to access
the contents of files on the spool via the Exim monitor (which runs
unprivileged), Exim must be built to allow group read access to its spool
files.

By default, regular users are trusted to perform basic testing and
introspection commands, as themselves.  This setting can be tightened by
setting the &%commandline_checks_require_admin%& option.
This affects most of the checking options,
such as &%-be%& and anything else &%-b*%&.


.section "Spool files" "SECID275"
.cindex "spool directory" "files"
Exim's spool directory and everything it contains is owned by the Exim user and
set to the Exim group. The mode for spool files is defined in the
&_Local/Makefile_& configuration file, and defaults to 0640. This means that
any user who is a member of the Exim group can access these files.



.section "Use of argv[0]" "SECID276"
Exim examines the last component of &%argv[0]%&, and if it matches one of a set
of specific strings, Exim assumes certain options. For example, calling Exim
with the last component of &%argv[0]%& set to &"rsmtp"& is exactly equivalent
to calling it with the option &%-bS%&. There are no security implications in
this.



.section "Use of %f formatting" "SECID277"
The only use made of &"%f"& by Exim is in formatting load average values. These
are actually stored in integer variables as 1000 times the load average.
Consequently, their range is limited and so therefore is the length of the
converted output.



.section "Embedded Exim path" "SECID278"
Exim uses its own path name, which is embedded in the code, only when it needs
to re-exec in order to regain root privilege. Therefore, it is not root when it
does so. If some bug allowed the path to get overwritten, it would lead to an
arbitrary program's being run as exim, not as root.



.section "Dynamic module directory" "SECTdynmoddir"
Any dynamically loadable modules must be installed into the directory
defined in &`LOOKUP_MODULE_DIR`& in &_Local/Makefile_& for Exim to permit
loading it.


.section "Use of sprintf()" "SECID279"
.cindex "&[sprintf()]&"
A large number of occurrences of &"sprintf"& in the code are actually calls to
&'string_sprintf()'&, a function that returns the result in malloc'd store.
The intermediate formatting is done into a large fixed buffer by a function
that runs through the format string itself, and checks the length of each
conversion before performing it, thus preventing buffer overruns.

The remaining uses of &[sprintf()]& happen in controlled circumstances where
the output buffer is known to be sufficiently long to contain the converted
string.



.section "Use of debug_printf() and log_write()" "SECID280"
Arbitrary strings are passed to both these functions, but they do their
formatting by calling the function &'string_vformat()'&, which runs through
the format string itself, and checks the length of each conversion.



.section "Use of strcat() and strcpy()" "SECID281"
These are used only in cases where the output buffer is known to be large
enough to hold the result.
.ecindex IIDsecurcon




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Format of spool files" "CHAPspool"
.scindex IIDforspo1 "format" "spool files"
.scindex IIDforspo2 "spool directory" "format of files"
.scindex IIDforspo3 "spool files" "format of"
.cindex "spool files" "editing"
A message on Exim's queue consists of two files, whose names are the message id
followed by -D and -H, respectively. The data portion of the message is kept in
the -D file on its own. The message's envelope, status, and headers are all
kept in the -H file, whose format is described in this chapter. Each of these
two files contains the final component of its own name as its first line. This
is insurance against disk crashes where the directory is lost but the files
themselves are recoverable.

The file formats may be changed, or new formats added, at any release.
Spool files are not intended as an interface to other programs
and should not be used as such.

Some people are tempted into editing -D files in order to modify messages. You
need to be extremely careful if you do this; it is not recommended and you are
on your own if you do it. Here are some of the pitfalls:

.ilist
You must ensure that Exim does not try to deliver the message while you are
fiddling with it. The safest way is to take out a write lock on the -D file,
which is what Exim itself does, using &[fcntl()]&. If you update the file in
place, the lock will be retained. If you write a new file and rename it, the
lock will be lost at the instant of rename.
.next
.vindex "&$body_linecount$&"
If you change the number of lines in the file, the value of
&$body_linecount$&, which is stored in the -H file, will be incorrect and can
cause incomplete transmission of messages or undeliverable messages.
.next
If the message is in MIME format, you must take care not to break it.
.next
If the message is cryptographically signed, any change will invalidate the
signature.
.endlist
All in all, modifying -D files is fraught with danger.

Files whose names end with -J may also be seen in the &_input_& directory (or
its subdirectories when &%split_spool_directory%& is set). These are journal
files, used to record addresses to which the message has been delivered during
the course of a delivery attempt. If there are still undelivered recipients at
the end, the -H file is updated, and the -J file is deleted. If, however, there
is some kind of crash (for example, a power outage) before this happens, the -J
file remains in existence. When Exim next processes the message, it notices the
-J file and uses it to update the -H file before starting the next delivery
attempt.

Files whose names end with -K or .eml may also be seen in the spool.
These are temporaries used for DKIM or malware processing, when that is used.
They should be tidied up by normal operations; any old ones are probably
relics of crashes and can be removed.

.section "Format of the -H file" "SECID282"
.cindex "uid (user id)" "in spool file"
.cindex "gid (group id)" "in spool file"
The second line of the -H file contains the login name for the uid of the
process that called Exim to read the message, followed by the numerical uid and
gid. For a locally generated message, this is normally the user who sent the
message. For a message received over TCP/IP via the daemon, it is
normally the Exim user.

The third line of the file contains the address of the message's sender as
transmitted in the envelope, contained in angle brackets. The sender address is
empty for bounce messages. For incoming SMTP mail, the sender address is given
in the MAIL command. For locally generated mail, the sender address is
created by Exim from the login name of the current user and the configured
&%qualify_domain%&. However, this can be overridden by the &%-f%& option or a
leading &"From&~"& line if the caller is trusted, or if the supplied address is
&"<>"& or an address that matches &%untrusted_set_senders%&.

The fourth line contains two numbers. The first is the time that the message
was received, in the conventional Unix form &-- the number of seconds since the
start of the epoch. The second number is a count of the number of messages
warning of delayed delivery that have been sent to the sender.

There follow a number of lines starting with a hyphen.
These contain variables, can appear in any
order, and are omitted when not relevant.

If there is a second hyphen after the first,
the corresponding data is tainted.
If there is a value in parentheses, the data is quoted for a lookup.

The following word specifies a variable,
and the remainder of the item depends on the variable.

.vlist
.vitem "&%-acl%&&~<&'number'&>&~<&'length'&>"
This item is obsolete, and is not generated from Exim release 4.61 onwards;
&%-aclc%& and &%-aclm%& are used instead. However, &%-acl%& is still
recognized, to provide backward compatibility. In the old format, a line of
this form is present for every ACL variable that is not empty. The number
identifies the variable; the &%acl_c%&&*x*& variables are numbered 0&--9 and
the &%acl_m%&&*x*& variables are numbered 10&--19. The length is the length of
the data string for the variable. The string itself starts at the beginning of
the next line, and is followed by a newline character. It may contain internal
newlines.

.vitem "&%-aclc%&&~<&'rest-of-name'&>&~<&'length'&>"
A line of this form is present for every ACL connection variable that is
defined. Note that there is a space between &%-aclc%& and the rest of the name.
The length is the length of the data string for the variable. The string itself
starts at the beginning of the next line, and is followed by a newline
character. It may contain internal newlines.

.vitem "&%-aclm%&&~<&'rest-of-name'&>&~<&'length'&>"
A line of this form is present for every ACL message variable that is defined.
Note that there is a space between &%-aclm%& and the rest of the name. The
length is the length of the data string for the variable. The string itself
starts at the beginning of the next line, and is followed by a newline
character. It may contain internal newlines.

.vitem "&%-active_hostname%&&~<&'hostname'&>"
This is present if, when the message was received over SMTP, the value of
&$smtp_active_hostname$& was different to the value of &$primary_hostname$&.

.vitem &%-allow_unqualified_recipient%&
This is present if unqualified recipient addresses are permitted in header
lines (to stop such addresses from being qualified if rewriting occurs at
transport time). Local messages that were input using &%-bnq%& and remote
messages from hosts that match &%recipient_unqualified_hosts%& set this flag.

.vitem &%-allow_unqualified_sender%&
This is present if unqualified sender addresses are permitted in header lines
(to stop such addresses from being qualified if rewriting occurs at transport
time). Local messages that were input using &%-bnq%& and remote messages from
hosts that match &%sender_unqualified_hosts%& set this flag.

.vitem "&%-auth_id%&&~<&'text'&>"
The id information for a message received on an authenticated SMTP connection
&-- the value of the &$authenticated_id$& variable.

.vitem "&%-auth_sender%&&~<&'address'&>"
The address of an authenticated sender &-- the value of the
&$authenticated_sender$& variable.

.vitem "&%-body_linecount%&&~<&'number'&>"
This records the number of lines in the body of the message, and is
present unless &%-spool_file_wireformat%& is.

.vitem "&%-body_zerocount%&&~<&'number'&>"
This records the number of binary zero bytes in the body of the message, and is
present if the number is greater than zero.

.vitem &%-deliver_firsttime%&
This is written when a new message is first added to the spool. When the spool
file is updated after a deferral, it is omitted.

.vitem "&%-frozen%&&~<&'time'&>"
.cindex "frozen messages" "spool data"
The message is frozen, and the freezing happened at <&'time'&>.

.vitem "&%-helo_name%&&~<&'text'&>"
This records the host name as specified by a remote host in a HELO or EHLO
command.

.vitem "&%-host_address%&&~<&'address'&>.<&'port'&>"
This records the IP address of the host from which the message was received and
the remote port number that was used. It is omitted for locally generated
messages.

.vitem "&%-host_auth%&&~<&'text'&>"
If the message was received on an authenticated SMTP connection, this records
the name of the authenticator &-- the value of the
&$sender_host_authenticated$& variable.

.vitem &%-host_lookup_failed%&
This is present if an attempt to look up the sending host's name from its IP
address failed. It corresponds to the &$host_lookup_failed$& variable.

.vitem "&%-host_name%&&~<&'text'&>"
.cindex "reverse DNS lookup"
.cindex "DNS" "reverse lookup"
This records the name of the remote host from which the message was received,
if the host name was looked up from the IP address when the message was being
received. It is not present if no reverse lookup was done.

.vitem "&%-ident%&&~<&'text'&>"
For locally submitted messages, this records the login of the originating user,
unless it was a trusted user and the &%-oMt%& option was used to specify an
ident value. For messages received over TCP/IP, this records the ident string
supplied by the remote host, if any.

.vitem "&%-interface_address%&&~<&'address'&>.<&'port'&>"
This records the IP address of the local interface and the port number through
which a message was received from a remote host. It is omitted for locally
generated messages.

.vitem &%-local%&
The message is from a local sender.

.vitem &%-localerror%&
The message is a locally-generated bounce message.

.vitem "&%-local_scan%&&~<&'string'&>"
This records the data string that was returned by the &[local_scan()]& function
when the message was received &-- the value of the &$local_scan_data$&
variable. It is omitted if no data was returned.

.vitem &%-manual_thaw%&
The message was frozen but has been thawed manually, that is, by an explicit
Exim command rather than via the auto-thaw process.

.vitem &%-N%&
A testing delivery process was started using the &%-N%& option to suppress any
actual deliveries, but delivery was deferred. At any further delivery attempts,
&%-N%& is assumed.

.vitem &%-received_protocol%&
This records the value of the &$received_protocol$& variable, which contains
the name of the protocol by which the message was received.

.vitem &%-sender_set_untrusted%&
The envelope sender of this message was set by an untrusted local caller (used
to ensure that the caller is displayed in queue listings).

.vitem "&%-spam_score_int%&&~<&'number'&>"
If a message was scanned by SpamAssassin, this is present. It records the value
of &$spam_score_int$&.

.vitem &%-spool_file_wireformat%&
The -D file for this message is in wire-format (for ESMTP CHUNKING)
rather than Unix-format.
The line-ending is CRLF rather than newline.
There is still, however, no leading-dot-stuffing.

.vitem &%-tls_certificate_verified%&
A TLS certificate was received from the client that sent this message, and the
certificate was verified by the server.

.vitem "&%-tls_cipher%&&~<&'cipher name'&>"
When the message was received over an encrypted connection, this records the
name of the cipher suite that was used.

.vitem "&%-tls_peerdn%&&~<&'peer DN'&>"
When the message was received over an encrypted connection, and a certificate
was received from the client, this records the Distinguished Name from that
certificate.
.endlist

Following the options there is a list of those addresses to which the message
is not to be delivered. This set of addresses is initialized from the command
line when the &%-t%& option is used and &%extract_addresses_remove_arguments%&
is set; otherwise it starts out empty. Whenever a successful delivery is made,
the address is added to this set. The addresses are kept internally as a
balanced binary tree, and it is a representation of that tree which is written
to the spool file. If an address is expanded via an alias or forward file, the
original address is added to the tree when deliveries to all its child
addresses are complete.

If the tree is empty, there is a single line in the spool file containing just
the text &"XX"&. Otherwise, each line consists of two letters, which are either
Y or N, followed by an address. The address is the value for the node of the
tree, and the letters indicate whether the node has a left branch and/or a
right branch attached to it, respectively. If branches exist, they immediately
follow. Here is an example of a three-node tree:
.code
YY darcy@austen.fict.example
NN alice@wonderland.fict.example
NN editor@thesaurus.ref.example
.endd
After the non-recipients tree, there is a list of the message's recipients.
This is a simple list, preceded by a count. It includes all the original
recipients of the message, including those to whom the message has already been
delivered. In the simplest case, the list contains one address per line. For
example:
.code
4
editor@thesaurus.ref.example
darcy@austen.fict.example
rdo@foundation
alice@wonderland.fict.example
.endd
However, when a child address has been added to the top-level addresses as a
result of the use of the &%one_time%& option on a &(redirect)& router, each
line is of the following form:
.display
<&'top-level address'&> <&'errors_to address'&> &&&
  <&'length'&>,<&'parent number'&>#<&'flag bits'&>
.endd
The 01 flag bit indicates the presence of the three other fields that follow
the top-level address. Other bits may be used in future to support additional
fields. The <&'parent number'&> is the offset in the recipients list of the
original parent of the &"one time"& address. The first two fields are the
envelope sender that is associated with this address and its length. If the
length is zero, there is no special envelope sender (there are then two space
characters in the line). A non-empty field can arise from a &(redirect)& router
that has an &%errors_to%& setting.


A blank line separates the envelope and status information from the headers
which follow. A header may occupy several lines of the file, and to save effort
when reading it in, each header is preceded by a number and an identifying
character. The number is the number of characters in the header, including any
embedded newlines and the terminating newline. The character is one of the
following:

.table2 50pt
.row <&'blank'&>         "header in which Exim has no special interest"
.row &`B`&               "&'Bcc:'& header"
.row &`C`&               "&'Cc:'& header"
.row &`F`&               "&'From:'& header"
.row &`I`&               "&'Message-id:'& header"
.row &`P`&               "&'Received:'& header &-- P for &""postmark""&"
.row &`R`&               "&'Reply-To:'& header"
.row &`S`&               "&'Sender:'& header"
.row &`T`&               "&'To:'& header"
.row &`*`&               "replaced or deleted header"
.endtable

Deleted or replaced (rewritten) headers remain in the spool file for debugging
purposes. They are not transmitted when the message is delivered. Here is a
typical set of headers:
.code
111P Received: by hobbit.fict.example with local (Exim 4.00)
id 14y9EI-00026G-00; Fri, 11 May 2001 10:28:59 +0100
049  Message-Id: <E14y9EI-00026G-00@hobbit.fict.example>
038* X-rewrote-sender: bb@hobbit.fict.example
042* From: Bilbo Baggins <bb@hobbit.fict.example>
049F From: Bilbo Baggins <B.Baggins@hobbit.fict.example>
099* To: alice@wonderland.fict.example, rdo@foundation,
darcy@austen.fict.example, editor@thesaurus.ref.example
104T To: alice@wonderland.fict.example, rdo@foundation.example,
darcy@austen.fict.example, editor@thesaurus.ref.example
038  Date: Fri, 11 May 2001 10:28:59 +0100
.endd
The asterisked headers indicate that the envelope sender, &'From:'& header, and
&'To:'& header have been rewritten, the last one because routing expanded the
unqualified domain &'foundation'&.
.ecindex IIDforspo1
.ecindex IIDforspo2
.ecindex IIDforspo3

.section "Format of the -D file" "SECID282a"
The data file is traditionally in Unix-standard format: lines are ended with
an ASCII newline character.
However, when the &%spool_wireformat%& main option is used some -D files
can have an alternate format.
This is flagged by a &%-spool_file_wireformat%& line in the corresponding -H file.
The -D file lines (not including the first name-component line) are
suitable for direct copying to the wire when transmitting using the
ESMTP CHUNKING option, meaning lower processing overhead.
Lines are terminated with an ASCII CRLF pair.
There is no dot-stuffing (and no dot-termination).

. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "DKIM, SPF, SRS and DMARC" "CHAPdkim" &&&
         "DKIM, SPF, SRS and DMARC Support"

.section "DKIM (DomainKeys Identified Mail)" SECDKIM
.cindex "DKIM"

DKIM is a mechanism by which messages sent by some entity can be provably
linked to a domain which that entity controls.  It permits reputation to
be tracked on a per-domain basis, rather than merely upon source IP address.
DKIM is documented in &url(https://www.rfc-editor.org/rfc/rfc6376,RFC 6376).

As DKIM relies on the message being unchanged in transit, messages handled
by a mailing-list (which traditionally adds to the message) will not match
any original DKIM signature.

DKIM support is compiled into Exim by default if TLS support is present.
It can be disabled by setting DISABLE_DKIM=yes in &_Local/Makefile_&.

Exim's DKIM implementation allows for
.olist
Signing outgoing messages: This function is implemented in the SMTP transport.
It can co-exist with all other Exim features
(including transport filters) except cutthrough delivery.
However, signing options may not depend on headers modified by
routers, the transport or a transport filter.
.next
Verifying signatures in incoming messages: This is implemented by an additional
ACL (acl_smtp_dkim), which can be called several times per message, with
different signature contexts.
.endlist

In typical Exim style, the verification implementation does not include any
default "policy". Instead it enables you to build your own policy using
Exim's standard controls.

Please note that verification of DKIM signatures in incoming mail is turned
on by default for logging (in the <= line) purposes.

Additional log detail can be enabled using the &%dkim_verbose%& log_selector.
When set, for each signature in incoming email,
exim will log a line displaying the most important signature details, and the
signature status. Here is an example (with line-breaks added for clarity):
.code
2009-09-09 10:22:28 1MlIRf-0003LU-U3 DKIM:
    d=facebookmail.com s=q1-2009b
    c=relaxed/relaxed a=rsa-sha1
    i=@facebookmail.com t=1252484542 [verification succeeded]
.endd

You might want to turn off DKIM verification processing entirely for internal
or relay mail sources. To do that, set the &%dkim_disable_verify%& ACL
control modifier. This should typically be done in the RCPT ACL, at points
where you accept mail from relay sources (internal hosts or authenticated
senders).


.subsection "Signing outgoing messages" SECDKIMSIGN
.cindex DKIM signing

For signing to be usable you must have published a DKIM record in DNS.
Note that &url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301)
(which does not cover EC keys) says:
.code
rsa-sha1 MUST NOT be used for signing or verifying.

Signers MUST use RSA keys of at least 1024 bits for all keys.
Signers SHOULD use RSA keys of at least 2048 bits.
.endd

Note also that the key content (the 'p=' field)
in the DNS record is different between RSA and EC keys;
for the former it is the base64 of the ASN.1 for the RSA public key
(equivalent to the private-key .pem with the header/trailer stripped)
but for EC keys it is the base64 of the pure key; no ASN.1 wrapping.

Signing is enabled by setting private options on the SMTP transport.
These options take (expandable) strings as arguments.

.option dkim_domain smtp "string list&!!" unset
The domain(s) you want to sign with.
After expansion, this can be a list.
Each element in turn,
lowercased,
.vindex "&$dkim_domain$&"
is put into the &%$dkim_domain%& expansion variable
while expanding the remaining signing options.
If it is empty after expansion, DKIM signing is not done,
and no error will result even if &%dkim_strict%& is set.

.option dkim_selector smtp "string list&!!" unset
This sets the key selector string.
After expansion, which can use &$dkim_domain$&, this can be a list.
Each element in turn is put in the expansion
.vindex "&$dkim_selector$&"
variable &%$dkim_selector%& which may be used in the &%dkim_private_key%&
option along with &%$dkim_domain%&.
If the option is empty after expansion, DKIM signing is not done for this domain,
and no error will result even if &%dkim_strict%& is set.

To do, for example, dual-signing with RSA and EC keys
this could be be used:
.code
dkim_selector = ec_sel : rsa_sel
dkim_private_key = KEYS_DIR/$dkim_selector
.endd

.option dkim_private_key smtp string&!! unset
This sets the private key to use.
You can use the &%$dkim_domain%& and
&%$dkim_selector%& expansion variables to determine the private key to use.
The result can either
.ilist
be a valid RSA private key in ASCII armor (.pem file), including line breaks
.next
with GnuTLS 3.6.0 or OpenSSL 1.1.1 or later,
be a valid Ed25519 private key (same format as above)
.next
start with a slash, in which case it is treated as a file that contains
the private key
.next
be "0", "false" or the empty string, in which case the message will not
be signed. This case will not result in an error, even if &%dkim_strict%&
is set.
.endlist

To generate keys under OpenSSL:
.code
openssl genrsa -out dkim_rsa.private 2048
openssl rsa -in dkim_rsa.private -out /dev/stdout -pubout -outform PEM
.endd
The result file from the first command should be retained,
permissions set so that Exim can read it,
and this option set to use it.
Take the base-64 lines from the output of the second command, concatenated,
for the DNS TXT record.
See section 3.6 of &url(https://www.rfc-editor.org/rfc/rfc6376,RFC 6376)
for the record specification.

Under GnuTLS:
.code
certtool --generate-privkey --rsa --bits=2048 --password='' -8 --outfile=dkim_rsa.private
certtool --load-privkey=dkim_rsa.private --pubkey-info
.endd

Note that &url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301) says:
.code
Signers MUST use RSA keys of at least 1024 bits for all keys.
Signers SHOULD use RSA keys of at least 2048 bits.
.endd

EC keys for DKIM are defined by
&url(https://www.rfc-editor.org/rfc/rfc8463,RFC 8463).
They are considerably smaller than RSA keys for equivalent protection.
As they are a recent development, users should consider dual-signing
(by setting a list of selectors, and an expansion for this option)
for some transition period.
The "_CRYPTO_SIGN_ED25519" macro will be defined if support is present
for EC keys.

OpenSSL 1.1.1 and GnuTLS 3.6.0 can create Ed25519 private keys:
.code
openssl genpkey -algorithm ed25519 -out dkim_ed25519.private
certtool --generate-privkey --key-type=ed25519 --outfile=dkim_ed25519.private
.endd

To produce the required public key value for a DNS record:
.code
openssl pkey -outform DER -pubout -in dkim_ed25519.private | tail -c +13 | base64
certtool --load_privkey=dkim_ed25519.private --pubkey_info --outder | tail -c +13 | base64
.endd

Exim also supports an alternate format
of Ed25519 keys in DNS which was a candidate during development
of the standard, but not adopted.
A future release will probably drop that support.

.option dkim_hash smtp string&!! sha256
Can be set to any one of the supported hash methods, which are:
.ilist
&`sha1`& &-- should not be used, is old and insecure
.next
&`sha256`& &-- the default
.next
&`sha512`& &-- possibly more secure but less well supported
.endlist

Note that &url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301) says:
.code
rsa-sha1 MUST NOT be used for signing or verifying.
.endd

.option dkim_identity smtp string&!! unset
If set after expansion, the value is used to set an "i=" tag in
the signing header.  The DKIM standards restrict the permissible
syntax of this optional tag to a mail address, with possibly-empty
local part, an @, and a domain identical to or subdomain of the "d="
tag value.  Note that Exim does not check the value.

.option dkim_canon smtp string&!! unset
This option sets the canonicalization method used when signing a message.
The DKIM RFC currently supports two methods: "simple" and "relaxed".
The option defaults to "relaxed" when unset. Note: the current implementation
only supports signing with the same canonicalization method for both headers and body.

.option dkim_strict smtp string&!! unset
This  option  defines  how  Exim  behaves  when  signing a message that
should be signed fails for some reason.  When the expansion evaluates to
either &"1"& or &"true"&, Exim will defer. Otherwise Exim will send the message
unsigned. You can use the &%$dkim_domain%& and &%$dkim_selector%& expansion
variables here.

.option dkim_sign_headers smtp string&!! "see below"
If set, this option must expand to a colon-separated
list of header names.
Headers with these names, or the absence of such a header, will be included
in the message signature.
When unspecified, the header names listed in
&url(https://www.rfc-editor.org/rfc/rfc4871,RFC 4871) will be used,
whether or not each header is present in the message.
The default list is available for the expansion in the macro
&"_DKIM_SIGN_HEADERS"&
and an oversigning variant is in &"_DKIM_OVERSIGN_HEADERS"&.

If a name is repeated, multiple headers by that name (or the absence thereof)
will be signed.  The textually later headers in the headers part of the
message are signed first, if there are multiples.

A name can be prefixed with either an &"="& or a &"+"& character.
If an &"="& prefix is used, all headers that are present with this name
will be signed.
If a &"+"& prefix if used, all headers that are present with this name
will be signed, and one signature added for a missing header with the
name will be appended.

.option dkim_timestamps smtp integer&!! unset
This option controls the inclusion of timestamp information in the signature.
If not set, no such information will be included.
Otherwise, must be an unsigned number giving an offset in seconds from the
current time for the expiry tag (e.g. 1209600 for two weeks); both creation
(t=) and expiry (x=) tags will be included unless the offset is 0 (no expiry).

&url(https://www.rfc-editor.org/rfc/rfc6376,RFC 6376)
lists these tags as RECOMMENDED.


.subsection "Verifying DKIM signatures in incoming mail" SECDKIMVFY
.cindex DKIM verification

Verification of DKIM signatures in SMTP incoming email is done for all
messages for which an ACL control &%dkim_disable_verify%& has not been set.

.cindex DKIM "selecting signature algorithms"
Individual classes of DKIM signature algorithm can be ignored by changing
the main options &%dkim_verify_hashes%& or &%dkim_verify_keytypes%&.
The &%dkim_verify_minimal%& option can be set to cease verification
processing for a message once the first passing signature is found.

.cindex authentication "expansion item"
Performing verification sets up information used by the
&%authresults%& expansion item.

For most purposes the default option settings suffice and the remainder
of this section can be ignored.

The results of verification are made available to the
&%acl_smtp_dkim%& ACL, which (for complex needs) can examine and modify them.
A missing ACL definition defaults to accept.
By default, the ACL is called once for each
syntactically(!) correct signature in the incoming message.
If any ACL call does not accept, the message is not accepted.
If a cutthrough delivery was in progress for the message, that is
summarily dropped (having wasted the transmission effort).

To evaluate the verification result in the ACL
a large number of expansion variables
containing the signature status and its details are set up during the
runtime of the ACL.

Calling the ACL only for existing signatures is not sufficient to build
more advanced policies. For that reason, the main option
&%dkim_verify_signers%&, and an expansion variable
&%$dkim_signers%& exist.

The main option &%dkim_verify_signers%& can be set to a colon-separated
list of DKIM domains or identities for which the ACL &%acl_smtp_dkim%& is
called. It is expanded when the message has been received. At this point,
the expansion variable &%$dkim_signers%& already contains a colon-separated
list of signer domains and identities for the message. When
&%dkim_verify_signers%& is not specified in the main configuration,
it defaults as:
.code
dkim_verify_signers = $dkim_signers
.endd
This leads to the default behaviour of calling &%acl_smtp_dkim%& for each
DKIM signature in the message. Current DKIM verifiers may want to explicitly
call the ACL for known domains or identities. This would be achieved as follows:
.code
dkim_verify_signers = paypal.com:ebay.com:$dkim_signers
.endd
This would result in &%acl_smtp_dkim%& always being called for "paypal.com"
and "ebay.com", plus all domains and identities that have signatures in the message.
You can also be more creative in constructing your policy. For example:
.code
dkim_verify_signers = $sender_address_domain:$dkim_signers
.endd

If a domain or identity is listed several times in the (expanded) value of
&%dkim_verify_signers%&, the ACL is only called once for that domain or identity.

Note that if the option is set using untrustworthy data
(such as the From: header)
care should be taken to force lowercase for domains
and for the domain part if identities.
The default setting can be regarded as trustworthy in this respect.

If multiple signatures match a domain (or identity), the ACL is called once
for each matching signature.


Inside the DKIM ACL, the following expansion variables are
available (from most to least important):


.vlist
.vitem &%$dkim_cur_signer%&
The signer that is being evaluated in this ACL run. This can be a domain or
an identity. This is one of the list items from the expanded main option
&%dkim_verify_signers%& (see above).

.vitem &%$dkim_verify_status%&
So long as a DKIM ACL is defined
(it need do no more than accept, which is the default),
after all the DKIM ACL runs have completed, the value becomes a
colon-separated list of the values after each run.
The value is maintained for the MIME, PRDR and DATA ACLs.

Within the DKIM ACL,
a string describing the general status of the signature. One of
.ilist
&%none%&: There is no signature in the message for the current domain or
identity (as reflected by &%$dkim_cur_signer%&).
.next
&%invalid%&: The signature could not be verified due to a processing error.
More detail is available in &%$dkim_verify_reason%&.
.next
&%fail%&: Verification of the signature failed.  More detail is
available in &%$dkim_verify_reason%&.
.next
&%pass%&: The signature passed verification. It is valid.
.endlist

This variable can be overwritten using an ACL 'set' modifier.
This might, for instance, be done to enforce a policy restriction on
hash-method or key-size:
.code
  warn condition       = ${if eq {$dkim_verify_status}{pass}}
       condition       = ${if eq {${length_3:$dkim_algo}}{rsa}}
       condition       = ${if or {{eq {$dkim_algo}{rsa-sha1}} \
                                  {< {$dkim_key_length}{1024}}}}
       logwrite        = NOTE: forcing DKIM verify fail (was pass)
       set dkim_verify_status = fail
       set dkim_verify_reason = hash too weak or key too short
.endd

.vitem &%$dkim_verify_reason%&
A string giving a little bit more detail when &%$dkim_verify_status%& is either
"fail" or "invalid". One of
.ilist
&%pubkey_unavailable%& (when &%$dkim_verify_status%&="invalid"): The public
key for the domain could not be retrieved. This may be a temporary problem.
.next
&%pubkey_syntax%& (when &%$dkim_verify_status%&="invalid"): The public key
record for the domain is syntactically invalid.
.next
&%bodyhash_mismatch%& (when &%$dkim_verify_status%&="fail"): The calculated
body hash does not match the one specified in the signature header. This
means that the message body was modified in transit.
.next
&%signature_incorrect%& (when &%$dkim_verify_status%&="fail"): The signature
could not be verified. This may mean that headers were modified,
re-written or otherwise changed in a way which is incompatible with
DKIM verification. It may of course also mean that the signature is forged.
.endlist

This variable can be overwritten, with any value, using an ACL 'set' modifier.

.vitem &%$dkim_domain%&
The signing domain. IMPORTANT: This variable is only populated if there is
an actual signature in the message for the current domain or identity (as
reflected by &%$dkim_cur_signer%&).

.vitem &%$dkim_identity%&
The signing identity, if present. IMPORTANT: This variable is only populated
if there is an actual signature in the message for the current domain or
identity (as reflected by &%$dkim_cur_signer%&).

.vitem &%$dkim_selector%&
The key record selector string.

.vitem &%$dkim_algo%&
The algorithm used. One of 'rsa-sha1' or 'rsa-sha256'.
If running under GnuTLS 3.6.0 or OpenSSL 1.1.1 or later,
may also be 'ed25519-sha256'.
The "_CRYPTO_SIGN_ED25519" macro will be defined if support is present
for EC keys.

Note that &url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301) says:
.code
rsa-sha1 MUST NOT be used for signing or verifying.

DKIM signatures identified as having been signed with historic
algorithms (currently, rsa-sha1) have permanently failed evaluation
.endd

To enforce this you must either have a DKIM ACL which checks this variable
and overwrites the &$dkim_verify_status$& variable as discussed above,
or have set the main option &%dkim_verify_hashes%& to exclude
processing of such signatures.

.vitem &%$dkim_canon_body%&
The body canonicalization method. One of 'relaxed' or 'simple'.

.vitem &%$dkim_canon_headers%&
The header canonicalization method. One of 'relaxed' or 'simple'.

.vitem &%$dkim_copiedheaders%&
A transcript of headers and their values which are included in the signature
(copied from the 'z=' tag of the signature).
Note that &url(https://www.rfc-editor.org/rfc/rfc6376,RFC 6376)
requires that verification fail if the From: header is
not included in the signature.  Exim does not enforce this; sites wishing
strict enforcement should code the check explicitly.

.vitem &%$dkim_bodylength%&
The number of signed body bytes. If zero ("0"), the body is unsigned. If no
limit was set by the signer, "9999999999999" is returned. This makes sure
that this variable always expands to an integer value.
&*Note:*& The presence of the signature tag specifying a signing body length
is one possible route to spoofing of valid DKIM signatures.
A paranoid implementation might wish to regard signature where this variable
shows less than the "no limit" return as being invalid.

.vitem &%$dkim_created%&
UNIX timestamp reflecting the date and time when the signature was created.
When this was not specified by the signer, "0" is returned.

.vitem &%$dkim_expires%&
UNIX timestamp reflecting the date and time when the signer wants the
signature to be treated as "expired". When this was not specified by the
signer, "9999999999999" is returned. This makes it possible to do useful
integer size comparisons against this value.
Note that Exim does not check this value.

.vitem &%$dkim_headernames%&
A colon-separated list of names of headers included in the signature.

.vitem &%$dkim_key_testing%&
"1" if the key record has the "testing" flag set, "0" if not.

.vitem &%$dkim_key_nosubdomains%&
"1" if the key record forbids subdomaining, "0" otherwise.

.vitem &%$dkim_key_srvtype%&
Service type (tag s=) from the key record. Defaults to "*" if not specified
in the key record.

.vitem &%$dkim_key_granularity%&
Key granularity (tag g=) from the key record. Defaults to "*" if not specified
in the key record.

.vitem &%$dkim_key_notes%&
Notes from the key record (tag n=).

.vitem &%$dkim_key_length%&
Number of bits in the key.
Valid only once the key is loaded, which is at the time the header signature
is verified, which is after the body hash is.

Note that &url(https://www.rfc-editor.org/rfc/rfc8301,RFC 8301) says:
.code
Verifiers MUST NOT consider signatures using RSA keys of
less than 1024 bits as valid signatures.
.endd

This is enforced by the default setting for the &%dkim_verify_min_keysizes%&
option.

.endlist

In addition, two ACL conditions are provided:

.vlist
.vitem &%dkim_signers%&
ACL condition that checks a colon-separated list of domains or identities
for a match against the domain or identity that the ACL is currently verifying
(reflected by &%$dkim_cur_signer%&).
This condition is only usable in a DKIM ACL.
This is typically used to restrict an ACL
verb to a group of domains or identities. For example:

.code
# Warn when Mail purportedly from GMail has no gmail signature
warn sender_domains = gmail.com
     dkim_signers = gmail.com
     dkim_status = none
     log_message = GMail sender without gmail.com DKIM signature
.endd

Note that the above does not check for a total lack of DKIM signing;
for that check for empty &$h_DKIM-Signature:$& in the data ACL.

.vitem &%dkim_status%&
ACL condition that checks a colon-separated list of possible DKIM verification
results against the actual result of verification,
given by &$dkim_verify_status$& if that is non-empty or "none" if empty.
This condition may be used in DKIM, MIME, PRDR and DATA ACLs.

A basic verification might be:
.code
deny !dkim_status = pass:none:invalid
.endd

A more complex use could be
to restrict an ACL verb to a list of verification outcomes, for example:

.code
deny sender_domains = paypal.com:paypal.de
     dkim_signers = paypal.com:paypal.de
     dkim_status = none:invalid:fail
     message = Mail from Paypal with invalid/missing signature
.endd

The possible status keywords are: 'none','invalid','fail' and 'pass'. Please
see the documentation of the &%$dkim_verify_status%& expansion variable above
for more information of what they mean.

The condition is true if the status
(or any of the list of status values)
is any one of the supplied list.
.endlist




.section "SPF (Sender Policy Framework)" SECSPF
.cindex SPF verification

SPF is a mechanism whereby a domain may assert which IP addresses may transmit
messages with its domain in the envelope from, documented by
&url(https://www.rfc-editor.org/rfc/rfc7208,RFC 7208).
For more information on SPF see &url(http://www.open-spf.org), a static copy of
the &url(http://openspf.org).
. --- 2019-10-28: still not https, open-spf.org is told to be a
. --- web-archive copy of the now dead openspf.org site
. --- See https://www.mail-archive.com/mailop@mailop.org/msg08019.html for a
. --- discussion.

Messages sent by a system not authorised will fail checking of such assertions.
This includes retransmissions done by traditional forwarders.

SPF verification support is built into Exim if SUPPORT_SPF=yes is set in
&_Local/Makefile_&.  The support uses the &_libspf2_& library
&url(https://www.libspf2.org/).
.new
.cindex "dynamic modules"
The support can be built as a dynamic-load module if desired;
see the comments in that Makefile.
.wen

There is no Exim involvement in the transmission of messages;
publishing certain DNS records is all that is required.

For verification, an ACL condition and an expansion lookup are provided.
.cindex authentication "expansion item"
Performing verification sets up information used by the
&%authresults%& expansion item.


.cindex SPF "ACL condition"
.cindex ACL "spf condition"
The ACL condition "spf" can be used at or after the MAIL ACL.
It takes as an argument a list of strings giving the outcome of the SPF check,
and will succeed for any matching outcome.
Valid strings are:
.vlist
.vitem &%pass%&
The SPF check passed, the sending host is positively verified by SPF.

.vitem &%fail%&
The SPF check failed, the sending host is NOT allowed to send mail for the
domain in the envelope-from address.

.vitem &%softfail%&
The SPF check failed, but the queried domain can't absolutely confirm that this
is a forgery.

.vitem &%none%&
The queried domain does not publish SPF records.

.vitem &%neutral%&
The SPF check returned a "neutral" state.  This means the queried domain has
published a SPF record, but wants to allow outside servers to send mail under
its domain as well.  This should be treated like "none".

.vitem &%permerror%&
This indicates a syntax error in the SPF record of the queried domain.
You may deny messages when this occurs.

.vitem &%temperror%&
This indicates a temporary error during all processing, including Exim's
SPF processing.  You may defer messages when this occurs.

.vitem &%invalid%&
There was an error during processing of the SPF lookup
.endlist

You can prefix each string with an exclamation mark to  invert
its meaning,  for example  "!fail" will  match all  results but
"fail".  The  string  list is  evaluated  left-to-right,  in a
short-circuit fashion.

Example:
.code
deny spf = fail
     message = $sender_host_address is not allowed to send mail from \
               ${if def:sender_address_domain \
                    {$sender_address_domain}{$sender_helo_name}}.  \
               Please see http://www.open-spf.org/Why;\
               identity=${if def:sender_address_domain \
                             {$sender_address}{$sender_helo_name}};\
               ip=$sender_host_address
.endd

Note: The above mentioned URL may not be as helpful as expected. You are
encouraged to replace the link with a link to a site with more
explanations.

When the spf condition has run, it sets up several expansion
variables:

.cindex SPF "verification variables"
.vlist
.vitem &$spf_header_comment$&
.vindex &$spf_header_comment$&
  This contains a human-readable string describing the outcome
  of the SPF check. You can add it to a custom header or use
  it for logging purposes.

.vitem &$spf_received$&
.vindex &$spf_received$&
  This contains a complete Received-SPF: header (name and
  content) that can be added to the message. Please note that
  according to the SPF draft, this header must be added at the
  top of the header list, i.e. with
.code
add_header = :at_start:$spf_received
.endd
  See section &<<SECTaddheadacl>>& for further details.

  Note: in case of "Best-guess" (see below), the convention is
  to put this string in a header called X-SPF-Guess: instead.

.vitem &$spf_result$&
.vindex &$spf_result$&
  This contains the outcome of the SPF check in string form,
  currently one of pass, fail, softfail, none, neutral, permerror,
  temperror, or &"(invalid)"&.

.vitem &$spf_result_guessed$&
.vindex &$spf_result_guessed$&
  This boolean is true only if a best-guess operation was used
  and required in order to obtain a result.

.vitem &$spf_smtp_comment$&
.vindex &$spf_smtp_comment$&
.vindex &%spf_smtp_comment_template%&
  This contains a string that can be used in a SMTP response
  to the calling party. Useful for "fail".
  The string is generated by the SPF library from the template configured in the main config
  option &%spf_smtp_comment_template%&.
.endlist


.cindex SPF "ACL condition"
.cindex ACL "spf_guess condition"
.cindex SPF "best guess"
In addition to SPF, you can also perform checks for so-called
"Best-guess".  Strictly speaking, "Best-guess" is not standard
SPF, but it is supported by the same framework that enables SPF
capability.
Refer to &url(http://www.open-spf.org/FAQ/Best_guess_record)
for a description of what it means.
. --- 2019-10-28: still not https:

To access this feature, simply use the spf_guess condition in place
of the spf one.  For example:

.code
deny spf_guess = fail
     message = $sender_host_address doesn't look trustworthy to me
.endd

In case you decide to reject messages based on this check, you
should note that although it uses the same framework, "Best-guess"
is not SPF, and therefore you should not mention SPF at all in your
reject message.

When the spf_guess condition has run, it sets up the same expansion
variables as when spf condition is run, described above.

Additionally, since Best-guess is not standardized, you may redefine
what "Best-guess" means to you by redefining the main configuration
&%spf_guess%& option.
For example, the following:

.code
spf_guess = v=spf1 a/16 mx/16 ptr ?all
.endd

would relax host matching rules to a broader network range.


.cindex SPF "lookup expansion"
.cindex lookup spf
A lookup expansion is also available. It takes an email
address as the key and an IP address
(v4 or v6)
as the database:

.code
  ${lookup {username@domain} spf {ip.ip.ip.ip}}
.endd

The lookup will return the same result strings as can appear in
&$spf_result$& (pass,fail,softfail,neutral,none,err_perm,err_temp).





.subsection "SRS (Sender Rewriting Scheme)" SECTSRS
.cindex SRS "sender rewriting scheme"
.cindex VERP "variable envelope return path"

SRS can be used to modify sender addresses when forwarding so that
SPF verification does not object to them.
It can also be used to identify a received bounce message as
likely (or not) having been trigged by a message from the
local system, and for identifying dead addresses in mailing lists.
It is one implementation of a VERP (Variable Envelope Return Path) method.

SRS operates by encoding the original envelope sender in a new
sender local part and using a domain run by the forwarding site
as the new domain for the sender.  Any DSN message should be returned
to this new sender at the forwarding site, which can extract the
original sender from the coded local part and forward the DSN to
the originator.

This is a way of avoiding the breakage that SPF does to forwarding.
The constructed local-part will be longer than the original,
leading to possible problems with very long addresses.
The changing of the sender address also hinders the tracing of mail
problems.

Exim can be built to include native SRS support.  To do this
SUPPORT_SRS=yes must be defined in &_Local/Makefile_&.
If this has been done, the macros _HAVE_SRS and _HAVE_NATIVE_SRS
will be defined.
The support is limited to SRS0-encoding; SRS1 is not supported.

.cindex SRS excoding
To encode an address use this expansion item:
.vlist
.vitem &*${srs_encode&~{*&<&'secret'&>&*}{*&<&'return&~path'&>&*}{*&<&'original&~domain'&>&*}}*&
.cindex "&%srs_encode%& expansion item"
.cindex SRS "expansion item"
The first argument should be a secret known and used by all systems
handling the recipient domain for the original message.
There is no need to periodically change this key; a timestamp is also
encoded.
The second argument should be given as the envelope sender address before this
encoding operation.
If this value is empty the the expansion result will be empty.
The third argument should be the recipient domain of the message when
it arrived at this system.
All arguments are expanded before use.

The result of the expansion is the replacement envelope-from (return path)
to be used.
.endlist

.cindex SRS decoding
To decode an address use this expansion condition:
.vlist
.vitem &*inbound_srs&~{*&<&'local&~part'&>&*}{*&<&'secret'&>&*}*&
The first argument should be the recipient local part as it was received.
The second argument is the site secret.
Both arguments are expanded before use.

If the messages is not for an SRS-encoded recipient the condition will
return false.
If it is, the condition will return true and the variable
&$srs_recipient$& will be set to the decoded (original) value.

If the second argument is empty then the condition returns true if
the first argument is in valid SRS formet, else false.
The variable &$srs_recipient$& is not set for this case.
.endlist

Example usage:
.code
  #macro
  SRS_SECRET = <pick something unique for your site for this. Use on all MXs.>

  #routers

  outbound:
    driver =    dnslookup
    # if outbound, and forwarding has been done, use an alternate transport
    domains =   ! +my_domains
    transport = ${if eq {$local_part@$domain} \
                        {$original_local_part@$original_domain} \
                     {remote_smtp} {remote_forwarded_smtp}}

  inbound_srs:
    driver =    redirect
    senders =   :
    domains =   +my_domains
    # detect inbound bounces which are SRS'd, and decode them
    condition = ${if inbound_srs {$local_part} {SRS_SECRET}}
    data =      $srs_recipient

  inbound_srs_failure:
    driver =    redirect
    senders =   :
    domains =   +my_domains
    # detect inbound bounces which look SRS'd but are invalid
    condition = ${if inbound_srs {$local_part} {}}
    allow_fail
    data =      :fail: Invalid SRS recipient address

  #... further routers here get inbound_srs-redirected recipients
  # and any that were not SRS'd


  # transport; should look like the non-forward outbound
  # one, plus the max_rcpt and return_path options
  remote_forwarded_smtp:
    driver =              smtp
    # single-recipient so that $original_domain is valid
    max_rcpt =            1
    # modify the envelope from, for mails that we forward
    return_path =         ${srs_encode {SRS_SECRET} {$return_path} {$original_domain}}
.endd





.section DMARC SECDMARC
.cindex DMARC verification

DMARC combines feedback from SPF, DKIM, and header From: in order
to attempt to provide better indicators of the authenticity of an
email.  This document does not explain the fundamentals; you
should read and understand how it works by visiting the
&url(http://www.dmarc.org/,DMARC website).

If Exim is built with DMARC support,
the libopendmarc library is used.

For building Exim yourself, obtain the library from
&url(http://sourceforge.net/projects/opendmarc/,sourceforge)
to obtain a copy, or find it in your favorite package
repository.  You will need to attend to the local/Makefile feature
SUPPORT_DMARC and the associated LDFLAGS addition.
This description assumes
that headers will be in /usr/local/include, and that the libraries
are in /usr/local/lib.

.subsection Configuration SSECDMARCCONFIG
.cindex DMARC configuration

There are three main-configuration options:
.cindex DMARC "configuration options"

The &%dmarc_tld_file%& option
.oindex &%dmarc_tld_file%&
defines the location of a text file of valid
top level domains the opendmarc library uses
during domain parsing. Maintained by Mozilla,
the most current version can be downloaded
from a link at &url(https://publicsuffix.org/list/public_suffix_list.dat).
See also the util/renew-opendmarc-tlds.sh script.
The default for the option is unset.
If not set, DMARC processing is disabled.


The &%dmarc_history_file%& option, if set
.oindex &%dmarc_history_file%&
defines the location of a file to log results
of dmarc verification on inbound emails. The
contents are importable by the opendmarc tools
which will manage the data, send out DMARC
reports, and expire the data. Make sure the
directory of this file is writable by the user
exim runs as.
The default is unset.

The &%dmarc_forensic_sender%& option
.oindex &%dmarc_forensic_sender%&
defines an alternate email address to use when sending a
forensic report detailing alignment failures
if a sender domain's dmarc record specifies it
and you have configured Exim to send them.
If set, this is expanded and used for the
From: header line; the address is extracted
from it and used for the envelope from.
If not set (the default), the From: header is expanded from
the dsn_from option, and <> is used for the
envelope from.

.subsection Controls SSECDMARCCONTROLS
.cindex DMARC controls

By default, the DMARC processing will run for any remote,
non-authenticated user.  It makes sense to only verify DMARC
status of messages coming from remote, untrusted sources.  You can
use standard conditions such as hosts, senders, etc, to decide that
DMARC verification should *not* be performed for them and disable
DMARC with an ACL control modifier:
.code
  control = dmarc_disable_verify
.endd
A DMARC record can also specify a "forensic address", which gives
exim an email address to submit reports about failed alignment.
Exim does not do this by default because in certain conditions it
results in unintended information leakage (what lists a user might
be subscribed to, etc).  You must configure exim to submit forensic
reports to the owner of the domain.  If the DMARC record contains a
forensic address and you specify the control statement below, then
exim will send these forensic emails.  It is also advised that you
configure a &%dmarc_forensic_sender%& because the default sender address
construction might be inadequate.
.code
  control = dmarc_enable_forensic
.endd
(AGAIN: You can choose not to send these forensic reports by simply
not putting the dmarc_enable_forensic control line at any point in
your exim config.  If you don't tell exim to send them, it will not
send them.)

There are no options to either control.  Both must appear before
the DATA acl.

.subsection ACL SSECDMARCACL
.cindex DMARC "ACL condition"

DMARC checks can be run on incoming SMTP  messages by using the
&"dmarc_status"& ACL condition in the DATA ACL.  You are required to
call the &"spf"& condition first in the ACLs, then the &"dmarc_status"&
condition.  Putting this condition in the ACLs is required in order
for a DMARC check to actually occur.  All of the variables are set
up before the DATA ACL, but there is no actual DMARC check that
occurs until a &"dmarc_status"& condition is encountered in the ACLs.

The &"dmarc_status"& condition takes a list of strings on its
right-hand side.  These strings describe recommended action based
on the DMARC check.  To understand what the policy recommendations
mean, refer to the DMARC website above.  Valid strings are:
.itable none 0 0 2 20* left 80* left
.irow &'accept'&        "The DMARC check passed and the library recommends accepting the email"
.irow &'reject'&        "The DMARC check failed and the library recommends rejecting the email"
.irow &'quarantine'&    "The DMARC check failed and the library recommends keeping it for further inspection"
.irow &'none'&          "The DMARC check passed and the library recommends no specific action, neutral"
.irow &'norecord'&      "No policy section in the DMARC record for this RFC5322.From field"
.irow &'nofrom'&        "Unable to determine the domain of the sender"
.irow &'temperror'&     "Library error or dns error"
.irow &'off'&           "The DMARC check was disabled for this email"
.endtable
You can prefix each string with an exclamation mark to invert its
meaning, for example "!accept" will match all results but
"accept".  The string list is evaluated left-to-right in a
short-circuit fashion.  When a string matches the outcome of the
DMARC check, the condition succeeds.  If none of the listed
strings matches the outcome of the DMARC check, the condition
fails.

Of course, you can also use any other lookup method that Exim
supports, including LDAP, Postgres, MySQL, etc, as long as the
result is a list of colon-separated strings.

Performing the check sets up information used by the
&%authresults%& expansion item.

Several expansion variables are set before the DATA ACL is
processed, and you can use them in this ACL.  The following
expansion variables are available:

.vlist
.vitem &$dmarc_status$&
.vindex &$dmarc_status$&
.cindex DMARC result
A one word status indicating what the DMARC library
thinks of the email.  It is a combination of the results of
DMARC record lookup and the SPF/DKIM/DMARC processing results
(if a DMARC record was found).  The actual policy declared
in the DMARC record is in a separate expansion variable.

.vitem &$dmarc_status_text$&
.vindex &$dmarc_status_text$&
Slightly longer, human readable status.

.vitem &$dmarc_used_domain$&
.vindex &$dmarc_used_domain$&
The domain which DMARC used to look up the DMARC policy record.

.vitem &$dmarc_domain_policy$&
.vindex &$dmarc_domain_policy$&
The policy declared in the DMARC record.  Valid values
are "none", "reject" and "quarantine".  It is blank when there
is any error, including no DMARC record.

.new
.vitem &$dmarc_alignment_spf$&
.vindex &$dmarc_alignment_spf$&
The result of the SPF alignment portion of the test status;
"yes" or "no".

.vitem &$dmarc_alignment_dkim$&
.vindex &$dmarc_alignment_dkim$&
The result of the DKIM alignment portion of the test status;
"yes" or "no".
.wen
.endlist

.subsection Logging SSECDMARCLOGGING
.cindex DMARC logging

By default, Exim's DMARC configuration is intended to be
non-intrusive and conservative.  To facilitate this, Exim will not
create any type of logging files without explicit configuration by
you, the admin.  Nor will Exim send out any emails/reports about
DMARC issues without explicit configuration by you, the admin (other
than typical bounce messages that may come about due to ACL
processing or failure delivery issues).

In order to log statistics suitable to be imported by the opendmarc
tools, you need to:
.ilist
Configure the global option &%dmarc_history_file%&
.next
Configure cron jobs to call the appropriate opendmarc history
import scripts and truncating the dmarc_history_file
.endlist

In order to send forensic reports, you need to:
.ilist
Configure the global option &%dmarc_forensic_sender%&
.next
Configure, somewhere before the DATA ACL, the control option to
enable sending DMARC forensic reports
.endlist

.subsection Example SSECDMARCEXAMPLE
.cindex DMARC example

Example usage:
.code
(RCPT ACL)
  warn    domains        = +local_domains
          hosts          = +local_hosts
          control        = dmarc_disable_verify

  warn    !domains       = +screwed_up_dmarc_records
          control        = dmarc_enable_forensic

  warn    condition      = (lookup if destined to mailing list)
          set acl_m_mailing_list = 1

(DATA ACL)
  warn    dmarc_status   = accept : none : off
          !authenticated = *
          log_message    = DMARC DEBUG: $dmarc_status $dmarc_used_domain

  warn    dmarc_status   = !accept
          !authenticated = *
          log_message    = DMARC DEBUG: '$dmarc_status' for $dmarc_used_domain

  warn    dmarc_status   = quarantine
          !authenticated = *
          set $acl_m_quarantine = 1
          # Do something in a transport with this flag variable

  deny    condition      = ${if eq{$dmarc_domain_policy}{reject}}
          condition      = ${if eq{$acl_m_mailing_list}{1}}
          message        = Messages from $dmarc_used_domain break mailing lists

  deny    dmarc_status   = reject
          !authenticated = *
          message        = Message from $dmarc_used_domain failed sender's DMARC policy, REJECT

  warn    add_header     = :at_start:${authresults {$primary_hostname}}
.endd





. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Proxies" "CHAPproxies" &&&
         "Proxy support"
.cindex "proxy support"
.cindex "proxy" "access via"

A proxy is an intermediate system through which communication is passed.
Proxies may provide a security, availability or load-distribution function.


.section "Inbound proxies" SECTproxyInbound
.cindex proxy inbound
.cindex proxy "server side"
.cindex proxy "Proxy protocol"
.cindex "Proxy protocol" proxy

Exim has support for receiving inbound SMTP connections via a proxy
that uses &"Proxy Protocol"& to speak to it.
To include this support, include &"SUPPORT_PROXY=yes"&
in Local/Makefile.

It was built on the HAProxy specification, found at
&url(https://www.haproxy.org/download/1.8/doc/proxy-protocol.txt).

The purpose of this facility is so that an application load balancer,
such as HAProxy, can sit in front of several Exim servers
to distribute load.
Exim uses the local protocol communication with the proxy to obtain
the remote SMTP system IP address and port information.
There is no logging if a host passes or
fails Proxy Protocol negotiation, but it can easily be determined and
recorded in an ACL (example is below).

Use of a proxy is enabled by setting the &%hosts_proxy%&
main configuration option to a hostlist; connections from these
hosts will use Proxy Protocol.
Exim supports both version 1 and version 2 of the Proxy Protocol and
automatically determines which version is in use.

The Proxy Protocol header is the first data received on a TCP connection
and is inserted before any TLS-on-connect handshake from the client; Exim
negotiates TLS between Exim-as-server and the remote client, not between
Exim and the proxy server. The Proxy Protocol header must be received
within &%proxy_protocol_timeout%&, which defaults to 3s.

The following expansion variables are usable
(&"internal"& and &"external"& here refer to the interfaces
of the proxy):
.itable none 0 0 2 30* left 70* left
.irow $proxy_external_address   "IP of host being proxied or IP of remote interface of proxy"
.irow $proxy_external_port      "Port of host being proxied or Port on remote interface of proxy"
.irow $proxy_local_address      "IP of proxy server inbound or IP of local interface of proxy"
.irow $proxy_local_port         "Port of proxy server inbound or Port on local interface of proxy"
.irow $proxy_session            "boolean: SMTP connection via proxy"
.endtable
If &$proxy_session$& is set but &$proxy_external_address$& is empty
there was a protocol error.
The variables &$sender_host_address$& and &$sender_host_port$&
will have values for the actual client system, not the proxy.

Since the real connections are all coming from the proxy, and the
per host connection tracking is done before Proxy Protocol is
evaluated, &%smtp_accept_max_per_host%& must be set high enough to
handle all of the parallel volume you expect per inbound proxy.
With the option set so high, you lose the ability
to protect your server from many connections from one IP.
In order to prevent your server from overload, you
need to add a per connection ratelimit to your connect ACL.
A possible solution is:
.display
  # Set max number of connections per host
  LIMIT   = 5
  # Or do some kind of IP lookup in a flat file or database
  # LIMIT = ${lookup{$sender_host_address}iplsearch{/etc/exim/proxy_limits}}

  defer   ratelimit      = LIMIT / 5s / per_conn / strict
          message        = Too many connections from this IP right now
.endd



.section "Outbound proxies" SECTproxySOCKS
.cindex proxy outbound
.cindex proxy "client side"
.cindex proxy SOCKS
.cindex SOCKS proxy
Exim has support for sending outbound SMTP via a proxy
using a protocol called SOCKS5 (defined by
&url(https://www.rfc-editor.org/rfc/rfc1928,RFC 1928)).
The support can be optionally included by defining SUPPORT_SOCKS=yes in
Local/Makefile.

Use of a proxy is enabled by setting the &%socks_proxy%& option
on an smtp transport.
.new
If unset (or empty after expansion) then proxying is not done.
.wen
Otherwise, expansion should result in a list
(colon-separated by default) of proxy specifiers.
Each proxy specifier is a list
(space-separated by default) where the initial element
is an IP address and any subsequent elements are options.

Each option is a string of form <name>=<value>.
The list of options is in the following table:
.itable none 0 0 2 10* left 90* left
.irow &'auth'&  "authentication method"
.irow &'name'&  "authentication username"
.irow &'pass'&  "authentication password"
.irow &'port'&  "tcp port"
.irow &'tmo'&   "connection timeout"
.irow &'pri'&   "priority"
.irow &'weight'& "selection bias"
.endtable

More details on each of these options follows:

.ilist
.cindex authentication "to proxy"
.cindex proxy authentication
&%auth%&: Either &"none"& (default) or &"name"&.
Using &"name"& selects username/password authentication per
&url(https://www.rfc-editor.org/rfc/rfc1929,RFC 1929)
for access to the proxy.
Default is &"none"&.
.next
&%name%&: sets the username for the &"name"& authentication method.
Default is empty.
.next
&%pass%&: sets the password for the &"name"& authentication method.
Default is empty.
.next
&%port%&: the TCP port number to use for the connection to the proxy.
Default is 1080.
.next
&%tmo%&: sets a connection timeout in seconds for this proxy.
Default is 5.
.next
&%pri%&: specifies a priority for the proxy within the list,
higher values being tried first.
The default priority is 1.
.next
&%weight%&: specifies a selection bias.
Within a priority set servers are queried in a random fashion,
weighted by this value.
The default value for selection bias is 1.
.endlist

Proxies from the list are tried according to their priority
and weight settings until one responds.  The timeout for the
overall connection applies to the set of proxied attempts.

.section Logging SECTproxyLog
To log the (local) IP of a proxy in the incoming or delivery logline,
add &"+proxy"& to the &%log_selector%& option.
This will add a component tagged with &"PRX="& to the line.

. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Internationalisation" "CHAPi18n" &&&
         "Internationalisation""
.cindex internationalisation "email address"
.cindex EAI
.cindex i18n
.cindex utf8 "mail name handling"

Exim has support for Internationalised mail names.
To include this it must be built with SUPPORT_I18N and the libidn library.
Standards supported are RFCs
&url(https://www.rfc-editor.org/rfc/rfc2060.html,2060),
&url(https://www.rfc-editor.org/rfc/rfc5890.html,5890),
&url(https://www.rfc-editor.org/rfc/rfc6530.html,6530) and
&url(https://www.rfc-editor.org/rfc/rfc6533.html,6533).

If Exim is built with SUPPORT_I18N_2008 (in addition to SUPPORT_I18N, not
instead of it) then IDNA2008 is supported; this adds an extra library
requirement, upon libidn2.

.section "MTA operations" SECTi18nMTA
.cindex SMTPUTF8 "ESMTP option"
.cindex "ESMTP extensions" SMTPUTF8
The main configuration option &%smtputf8_advertise_hosts%& specifies
a host list.  If this matches the sending host and
accept_8bitmime is true (the default) then the ESMTP option
SMTPUTF8 will be advertised.

If the sender specifies the SMTPUTF8 option on a MAIL command
international handling for the message is enabled and
the expansion variable &$message_smtputf8$& will have value TRUE.

The option &%allow_utf8_domains%& is set to true for this
message. All DNS lookups are converted to a-label form
whatever the setting of &%allow_utf8_domains%&
when Exim is built with SUPPORT_I18N.

Both localparts and domain are maintained as the original
UTF-8 form internally; any comparison or regular-expression use will
require appropriate care.  Filenames created, eg. by
the appendfile transport, will have UTF-8 names.

HELO names sent by the smtp transport will have any UTF-8
components expanded to a-label form,
and any certificate name checks will be done using the a-label
form of the name.

.cindex log protocol
.cindex SMTPUTF8 logging
.cindex i18n logging
Log lines and Received-by: header lines will acquire a "utf8"
prefix on the protocol element, eg. utf8esmtp.

The following expansion operators can be used:
.code
${utf8_domain_to_alabel:str}
${utf8_domain_from_alabel:str}
${utf8_localpart_to_alabel:str}
${utf8_localpart_from_alabel:str}
.endd

.cindex utf8 "address downconversion"
.cindex i18n "utf8 address downconversion"
The RCPT ACL
may use the following modifier:
.display
control = utf8_downconvert
control = utf8_downconvert/<value>
.endd
This sets a flag requiring that envelope addresses are converted to
a-label form before smtp delivery.
This is usually for use in a Message Submission Agent context,
but could be used for any message.

If a value is appended it may be:
.itable none 0 0 2 5* right 95* left
.irow &`1`&     "mandatory downconversion"
.irow &`0`&     "no downconversion"
.irow &`-1`&    "if SMTPUTF8 not supported by destination host"
.endtable
If no value is given, 1 is used.

If mua_wrapper is set, the utf8_downconvert control
is initially set to -1.

The smtp transport has an option &%utf8_downconvert%&.
If set it must expand to one of the three values described above,
or an empty string.
If non-empty it overrides value previously set
(due to mua_wrapper or by an ACL control).


There is no explicit support for VRFY and EXPN.
Configurations supporting these should inspect
&$smtp_command_argument$& for an SMTPUTF8 argument.

There is no support for LMTP on Unix sockets.
Using the "lmtp" protocol option on an smtp transport,
for LMTP over TCP, should work as expected.

There is no support for DSN unitext handling,
and no provision for converting logging from or to UTF-8.



.section "MDA operations" SECTi18nMDA
To aid in constructing names suitable for IMAP folders
the following expansion operator can be used:
.code
${imapfolder {<string>} {<sep>} {<specials>}}
.endd

The string is converted from the charset specified by
the "headers charset" command (in a filter file)
or &%headers_charset%& main configuration option (otherwise),
to the
modified UTF-7 encoding specified by
&url(https://www.rfc-editor.org/rfc/rfc2060.html,RFC 2060),
with the following exception: All occurrences of <sep>
(which has to be a single character)
are replaced with periods ("."), and all periods and slashes that are not
<sep> and are not in the <specials> string are BASE64 encoded.

The third argument can be omitted, defaulting to an empty string.
The second argument can be omitted, defaulting to "/".

This is the encoding used by Courier for Maildir names on disk, and followed
by many other IMAP servers.

Examples:
.display
&`${imapfolder {Foo/Bar}}       `& yields &`Foo.Bar`&
&`${imapfolder {Foo/Bar}{.}{/}} `& yields &`Foo&&AC8-Bar`&
&`${imapfolder {Räksmörgås}}    `& yields &`R&&AOQ-ksm&&APY-rg&&AOU-s`&
.endd

Note that the source charset setting is vital, and also that characters
must be representable in UTF-16.


. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Events" "CHAPevents" &&&
         "Events"
.cindex events

The events mechanism in Exim can be used to intercept processing at a number
of points.  It was originally invented to give a way to do customised logging
actions (for example, to a database) but can also be used to modify some
processing actions.

Most installations will never need to use Events.
The support can be left out of a build by defining DISABLE_EVENT=yes
in &_Local/Makefile_&.

There are two major classes of events: main and transport.
The main configuration option &%event_action%& controls reception events;
a transport option &%event_action%& controls delivery events.

Both options are a string which is expanded when the event fires.
An example might look like:
.cindex logging custom
.code
event_action = ${if eq {msg:delivery}{$event_name} \
{${lookup pgsql {SELECT * FROM record_Delivery( \
    '${quote_pgsql:$sender_address_domain}',\
    '${quote_pgsql:${lc:$sender_address_local_part}}', \
    '${quote_pgsql:$domain}', \
    '${quote_pgsql:${lc:$local_part}}', \
    '${quote_pgsql:$host_address}', \
    '${quote_pgsql:${lc:$host}}', \
    '${quote_pgsql:$message_exim_id}')}} \
} {}}
.endd

Events have names which correspond to the point in process at which they fire.
The name is placed in the variable &$event_name$& and the event action
expansion must check this, as it will be called for every possible event type.

.new
The current list of events is:
.itable all 0 0 4 25* left 10* center 15* center 50* left
.row auth:fail              after    both       "per driver per authentication attempt"
.row dane:fail              after    transport  "per connection"
.row dns:fail               after    both       "per lookup"
.row msg:complete           after    main       "per message"
.row msg:defer              after    transport  "per message per delivery try"
.row msg:delivery           after    transport  "per recipient"
.row msg:rcpt:host:defer    after    transport  "per recipient per host"
.row msg:rcpt:defer         after    transport  "per recipient"
.row msg:host:defer         after    transport  "per host per delivery try; host errors"
.row msg:fail:delivery      after    transport  "per recipient"
.row msg:fail:internal      after    main       "per recipient"
.row tcp:connect            before   transport  "per connection"
.row tcp:close              after    transport  "per connection"
.row tls:cert               before   both       "per certificate in verification chain"
.row tls:fail:connect       after    main       "per connection"
.row smtp:connect           after    transport  "per connection"
.row smtp:ehlo              after    transport  "per connection"
.row smtp:fail:protocol     after    main       "per connection"
.row smtp:fail:syntax       after    main       "per connection"
.endtable
.wen
New event types may be added in future.

The event name is a colon-separated list, defining the type of
event in a tree of possibilities.  It may be used as a list
or just matched on as a whole.  There will be no spaces in the name.

The second column in the table above describes whether the event fires
before or after the action is associates with.  Those which fire before
can be used to affect that action (more on this below).

The third column in the table above says what section of the configuration
should define the event action.

An additional variable, &$event_data$&, is filled with information varying
with the event type:
.new
.itable all 0 0 2 20* left 80* left
.row auth:fail            "smtp response"
.row dane:fail            "failure reason"
.row dns:fail             "failure reason, key and lookup-type"
.row msg:defer            "error string"
.row msg:delivery         "smtp confirmation message"
.row msg:fail:internal    "failure reason"
.row msg:fail:delivery    "smtp error message"
.row msg:host:defer       "error string"
.row msg:rcpt:host:defer  "error string"
.row msg:rcpt:defer       "error string"
.row tls:cert             "verification chain depth"
.row tls:fail:connect     "error string"
.row smtp:connect         "smtp banner"
.row smtp:ehlo            "smtp ehlo response"
.row smtp:fail:protocol   "error string"
.row smtp:fail:syntax     "error string"
.endtable
.wen

The :defer events populate one extra variable: &$event_defer_errno$&.

For complex operations an ACL expansion can be used in &%event_action%&,
however due to the multiple contexts that Exim operates in during
the course of its processing:
.ilist
variables set in transport events will not be visible outside that
transport call
.next
acl_m variables in a server context are lost on a new connection,
and after smtp helo/ehlo/mail/starttls/rset commands
.endlist
Using an ACL expansion with the logwrite modifier can be
a useful way of writing to the main log.

The expansion of the event_action option should normally
return an empty string.  Should it return anything else the
following will be forced:
.itable all 0 0 2 20* left 80* left
.row auth:fail        "log information to write"
.row tcp:connect      "do not connect"
.row tls:cert         "refuse verification"
.row smtp:connect     "close connection"
.endtable
All other message types ignore the result string, and
no other use is made of it.

For a tcp:connect event, if the connection is being made to a proxy
then the &$host_address$& and &$host_port$& variables
will be that of the proxy and not the target system.

For tls:cert events, if GnuTLS is in use this will trigger only per
chain element received on the connection.
For OpenSSL it will trigger for every chain element including those
loaded locally.

For dns:fail events from dnsdb lookups, a &"defer_never"& option does not
affect the reporting of DNS_AGAIN.

. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

.chapter "Adding new drivers or lookup types" "CHID13" &&&
         "Adding drivers or lookups"
.cindex "adding drivers"
.cindex "new drivers, adding"
.cindex "drivers" "adding new"
The following actions have to be taken in order to add a new router, transport,
authenticator, or lookup type to Exim:

.olist
Choose a name for the driver or lookup type that does not conflict with any
existing name; I will use &"newdriver"& in what follows.
.next
Add to &_src/EDITME_& the line:
.display
<&'type'&>&`_NEWDRIVER=yes`&
.endd
where <&'type'&> is ROUTER, TRANSPORT, AUTH, or LOOKUP. If the
code is not to be included in the binary by default, comment this line out. You
should also add any relevant comments about the driver or lookup type.
.next
Add to &_src/config.h.defaults_& the line:
.code
#define <type>_NEWDRIVER
.endd
.next
Edit &_src/drtables.c_&, adding conditional code to pull in the private header
and create a table entry as is done for all the other drivers and lookup types.
.next
Edit &_scripts/lookups-Makefile_& if this is a new lookup; there is a for-loop
near the bottom, ranging the &`name_mod`& variable over a list of all lookups.
Add your &`NEWDRIVER`& to that list.
As long as the dynamic module would be named &_newdriver.so_&, you can use the
simple form that most lookups have.
.next
Edit &_Makefile_& in the appropriate sub-directory (&_src/routers_&,
&_src/transports_&, &_src/auths_&, or &_src/lookups_&); add a line for the new
driver or lookup type and add it to the definition of OBJ.
.next
Edit &_OS/Makefile-Base_& adding a &_.o_& file for the predefined-macros, to the
definition of OBJ_MACRO.  Add a set of line to do the compile also.
.next
Create &_newdriver.h_& and &_newdriver.c_& in the appropriate sub-directory of
&_src_&.
.next
Edit &_scripts/MakeLinks_& and add commands to link the &_.h_& and &_.c_& files
as for other drivers and lookups.
.endlist

Then all you need to do is write the code! A good way to start is to make a
proforma by copying an existing module of the same type, globally changing all
occurrences of the name, and cutting out most of the code. Note that any
options you create must be listed in alphabetical order, because the tables are
searched using a binary chop procedure.

There is a &_README_& file in each of the sub-directories of &_src_& describing
the interface that is expected.




. ////////////////////////////////////////////////////////////////////////////
. ////////////////////////////////////////////////////////////////////////////

. /////////////////////////////////////////////////////////////////////////////
. These lines are processing instructions for the Simple DocBook Processor that
. Philip Hazel has developed as a less cumbersome way of making PostScript and
. PDFs than using xmlto and fop. They will be ignored by all other XML
. processors.
. /////////////////////////////////////////////////////////////////////////////

.literal xml
<?sdop
  format="newpage"
  foot_right_recto="&chaptertitle;"
  foot_right_verso="&chaptertitle;"
?>
.literal off

.makeindex "Options index"   "option"
.makeindex "Variables index" "variable"
.makeindex "Concept index"   "concept"


. /////////////////////////////////////////////////////////////////////////////
. /////////////////////////////////////////////////////////////////////////////