1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.483 2007/02/20 11:37:16 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 MH/01 Fix for bug #448, segfault in Dovecot authenticator when interface_address
10 is unset (happens when testing with -bh and -oMi isn't used). Thanks to
13 PH/01 Added a new log selector smtp_no_mail, to log SMTP sessions that do not
16 PH/02 In an ACL statement such as
18 deny dnslists = X!=127.0.0.2 : X=127.0.0.2
20 if a client was not listed at all, or was listed with a value other than
21 127.0.0.2, in the X list, but was listed with 127.0.0.2 in the Y list,
22 the condition was not true (as it should be), so access was not denied.
23 The bug was that the ! inversion was incorrectly passed on to the second
24 item. This has been fixed.
26 PH/03 Added additional dnslists conditions == and =& which are different from
27 = and & when the dns lookup returns more than one IP address.
29 PH/04 Added gnutls_require_{kx,mac,protocols} to give more control over the
30 cipher suites used by GnuTLS. These options are ignored by OpenSSL.
32 PH/05 After discussion on the list, added a compile time option ENABLE_DISABLE_
33 FSYNC, which compiles an option called disable_fsync that allows for
34 bypassing fsync(). The documentation is heavily laced with warnings.
36 SC/01 Updated eximstats to collate all SpamAssassin rejects into one bucket.
38 PH/06 Some tidies to the infrastructure of the Test Suite that is concerned
39 with the auxiliary C programs that it uses: (1) Arrange for BIND_8_COMPAT
40 to be defined when compiling on OSX (Darwin); (2) Tidies to the Makefile,
41 including adding "make clean"; (3) Added -fPIC when compiling the test
42 dynamically loaded module, to get rid of a warning.
44 MH/02 Fix for bug #451, causing paniclog entries to be written if a bounce
45 message fails, move_frozen_messages = true and ignore_bounce_errors_after
46 = 0s. The bug is otherwise harmless.
48 PH/07 There was a bug in the dovecot authenticator such that the value of
49 $auth1 could be overwritten, and so not correctly preserved, after a
50 successful authentication. This usually meant that the value preserved by
51 the server_setid option was incorrect.
53 PH/08 Added $smtp_count_at_connection_start, deliberately with a long name.
55 PH/09 Installed PCRE release 7.0.
57 PH/10 The acl_not_smtp_start ACL was, contrary to the documentation, not being
58 run for batched SMTP input. It is now run at the start of every message
59 in the batch. While fixing this I discovered that the process information
60 (output by running exiwhat) was not always getting set for -bs and -bS
61 input. This is fixed, and it now also says "batched" for BSMTP.
63 PH/11 Added control=no_pipelining.
65 PH/12 Added $sending_ip_address and $sending_port (mostly Magnus Holmgren's
66 patch, slightly modified), and move the expansion of helo_data till after
67 the connection is made in the smtp transport (so it can use these
70 PH/13 Added ${rfc2047d: to decoded RFC 2047 strings.
72 PH/14 Added log_selector = +pid.
74 PH/15 Flush SMTP output before delaying, unless control=no_delay_flush is set.
76 PH/16 Add ${if forany and ${if forall.
78 PH/17 Added dsn_from option to vary the From: line in DSNs.
80 PH/18 Flush SMTP output before performing a callout, unless control =
81 no_callout_flush is set.
83 PH/19 Change 4.64/PH/36 introduced a bug: when address_retry_include_sender
84 was true (the default) a successful delivery failed to delete the retry
85 item, thus causing premature timeout of the address. The bug is now
88 PH/20 Added hosts_avoid_pipelining to the smtp transport.
90 PH/21 Long custom messages for fakedefer and fakereject are now split up
91 into multiline reponses in the same way that messages for "deny" and
92 other ACL rejections are.
94 PH/22 Applied Jori Hamalainen's speed-up changes and typo fixes to exigrep,
95 with slight modification.
97 PH/23 Applied sieve patches from the maintainer "tracking the latest notify
98 draft, changing the syntax and factoring some duplicate code".
100 PH/24 When the log selector "outgoing_port" was set, the port was shown as -1
101 for deliveries of the second and subsequent messages over the same SMTP
104 PH/25 Applied Magnus Holmgren's patch for ${addresses, ${map, ${filter, and
105 ${reduce, with only minor "tidies".
107 SC/02 Applied Daniel Tiefnig's patch to improve the '($parent) =' pattern match.
109 PH/26 Added a "continue" ACL modifier that does nothing, for the benefit of its
110 expansion side effects.
112 PH/27 When a message times out after an over-quota error from an Exim-imposed
113 quota, the bounce message says "mailbox is full". This message was not
114 being given when it was a system quota that was exceeded. It now should
117 MH/03 Made $recipients available in local_scan(). local_scan() already has
118 better access to the recipient list through recipients_list[], but
119 $recipients can be useful in postmaster-provided expansion strings.
121 PH/28 The $smtp_command and $smtp_command_argument variables were not correct
122 in the case of a MAIL command with additional options following the
123 address, for example: MAIL FROM:<foo@bar> SIZE=1234. The option settings
124 were accidentally chopped off.
130 PH/01 Two more bugs that were introduced by 4.64/PH/07, in addition to the one
131 fixed by 4.65/MH/01 (is this a record?) are fixed:
133 (i) An empty string was always treated as zero by the numeric comparison
134 operators. This behaviour has been restored.
136 (ii) It is documented that the numeric comparison operators always treat
137 their arguments as decimal numbers. This was broken in that numbers
138 starting with 0 were being interpreted as octal.
140 While fixing these problems I realized that there was another issue that
141 hadn't been noticed. Values of message_size_limit (both the global option
142 and the transport option) were treated as octal if they started with 0.
143 The documentation was vague. These values are now always treated as
144 decimal, and I will make that clear in the documentation.
150 TK/01 Disable default definition of HAVE_LINUX_SENDFILE. Clashes with
151 Linux large file support (_FILE_OFFSET_BITS=64) on older glibc
154 MH/01 Don't check that the operands of numeric comparison operators are
155 integers when their expansion is in "skipping" mode (fixes bug
156 introduced by 4.64-PH/07).
158 PH/01 If a system filter or a router generates more than SHRT_MAX (32767)
159 child addresses, Exim now panics and dies. Previously, because the count
160 is held in a short int, deliveries were likely to be lost. As such a
161 large number of recipients for a single message is ridiculous
162 (performance will be very, very poor), I have chosen to impose a limit
163 rather than extend the field.
169 TK/01 Bugzilla #401. Fix DK spooling code so that it can overwrite a
170 leftover -K file (the existence of which was triggered by #402).
171 While we were at it, introduced process PID as part of the -K
172 filename. This should rule out race conditions when creating
175 TK/02 Bugzilla #402. Apply patch from Simon Arlott, speeding up DK signing
176 processing considerably. Previous code took too long for large mails,
177 triggering a timeout which in turn triggers #401.
179 TK/03 Introduced HAVE_LINUX_SENDFILE to os.h-Linux. Currently only used
180 in the DK code in transports.c. sendfile() is not really portable,
181 hence the _LINUX specificness.
183 TF/01 In the add_headers option to the mail command in an Exim filter,
184 there was a bug that Exim would claim a syntax error in any
185 header after the first one which had an odd number of characters
188 PH/01 If a server that rejects MAIL FROM:<> was the target of a sender
189 callout verification, Exim cached a "reject" for the entire domain. This
190 is correct for most verifications, but it is not correct for a recipient
191 verification with use_sender or use_postmaster set, because in that case
192 the callout does not use MAIL FROM:<>. Exim now distinguishes the special
193 case of MAIL FROM:<> rejection from other early rejections (e.g.
194 rejection of HELO). When verifying a recipient using a non-null MAIL
195 address, the cache is ignored if it shows MAIL FROM:<> rejection.
196 Whatever the result of the callout, the value of the domain cache is
197 left unchanged (for any other kind of callout, getting as far as trying
198 RCPT means that the domain itself is ok).
200 PH/02 Tidied a number of unused variable and signed/unsigned warnings that
203 PH/03 On Solaris, an unexpectedly close socket (dropped connection) can
204 manifest itself as EPIPE rather than ECONNECT. When tidying away a
205 session, the daemon ignores ECONNECT errors and logs others; it now
206 ignores EPIPE as well.
208 PH/04 Applied Nico Erfurth's refactoring patch to tidy up mime.c
209 (quoted-printable decoding).
211 PH/05 Applied Nico Erfurth's refactoring patch to tidy up spool_mbox.c, and
212 later the small subsequent patch to fix an introduced bug.
214 PH/06 Installed the latest Cygwin Makefile from the Cygwin maintainer.
216 PH/07 There was no check for overflow in expansions such as ${if >{1}{4096M}}.
218 PH/08 An error is now given if message_size_limit is specified negative.
220 PH/09 Applied and tidied up Jakob Hirsch's patch for allowing ACL variables
221 to be given (somewhat) arbitrary names.
223 JJ/01 exipick 20060919.0, allow for arbitrary acl_ variables introduced
226 JJ/02 exipick 20060919.0, --show-vars args can now be regular expressions,
227 miscellaneous code fixes
229 PH/10 Added the log_reject_target ACL modifier to specify where to log
232 PH/11 Callouts were setting the name used for EHLO/HELO from $smtp_active_
233 hostname. This is wrong, because it relates to the incoming message (and
234 probably the interface on which it is arriving) and not to the outgoing
235 callout (which could be using a different interface). This has been
236 changed to use the value of the helo_data option from the smtp transport
237 instead - this is what is used when a message is actually being sent. If
238 there is no remote transport (possible with a router that sets up host
239 addresses), $smtp_active_hostname is used.
241 PH/12 Installed Andrey Panin's patch to add a dovecot authenticator. Various
242 tweaks were necessary in order to get it to work (see also 21 below):
243 (a) The code assumed that strncpy() returns a negative number on buffer
244 overflow, which isn't the case. Replaced with Exim's string_format()
246 (b) There were several signed/unsigned issues. I just did the minimum
247 hacking in of casts. There is scope for a larger refactoring.
248 (c) The code used strcasecmp() which is not a standard C function.
249 Replaced with Exim's strcmpic() function.
250 (d) The code set only $1; it now sets $auth1 as well.
251 (e) A simple test gave the error "authentication client didn't specify
252 service in request". It would seem that Dovecot has changed its
253 interface. Fortunately there's a specification; I followed it and
254 changed what the client sends and it appears to be working now.
256 PH/13 Added $message_headers_raw to provide the headers without RFC 2047
259 PH/14 Corrected misleading output from -bv when -v was also used. Suppose the
260 address A is aliased to B and C, where B exists and C does not. Without
261 -v the output is "A verified" because verification stops after a
262 successful redirection if more than one address is generated. However,
263 with -v the child addresses are also verified. Exim was outputting "A
264 failed to verify" and then showing the successful verification for C,
265 with its parentage. It now outputs "B failed to verify", showing B's
266 parentage before showing the successful verification of C.
268 PH/15 Applied Michael Deutschmann's patch to allow DNS black list processing to
269 look up a TXT record in a specific list after matching in a combined
272 PH/16 It seems that the options setting for the resolver (RES_DEFNAMES and
273 RES_DNSRCH) can affect the behaviour of gethostbyname() and friends when
274 they consult the DNS. I had assumed they would set it the way they
275 wanted; and indeed my experiments on Linux seem to show that in some
276 cases they do (I could influence IPv6 lookups but not IPv4 lookups).
277 To be on the safe side, however, I have now made the interface to
278 host_find_byname() similar to host_find_bydns(), with an argument
279 containing the DNS resolver options. The host_find_byname() function now
280 sets these options at its start, just as host_find_bydns() does. The smtp
281 transport options dns_qualify_single and dns_search_parents are passed to
282 host_find_byname() when gethostbyname=TRUE in this transport. Other uses
283 of host_find_byname() use the default settings of RES_DEFNAMES
284 (qualify_single) but not RES_DNSRCH (search_parents).
286 PH/17 Applied (a modified version of) Nico Erfurth's patch to make
287 spool_read_header() do less string testing, by means of a preliminary
288 switch on the second character of optional "-foo" lines. (This is
289 overdue, caused by the large number of possibilities that now exist.
290 Originally there were few.) While I was there, I also converted the
291 str(n)cmp tests so they don't re-test the leading "-" and the first
292 character, in the hope this might squeeze out yet more improvement.
294 PH/18 Two problems with "group" syntax in header lines when verifying: (1) The
295 flag allowing group syntax was set by the header_syntax check but not
296 turned off, possible causing trouble later; (2) The flag was not being
297 set at all for the header_verify test, causing "group"-style headers to
298 be rejected. I have now set it in this case, and also caused header_
299 verify to ignore an empty address taken from a group. While doing this, I
300 came across some other cases where the code for allowing group syntax
301 while scanning a header line wasn't quite right (mostly, not resetting
302 the flag correctly in the right place). These bugs could have caused
303 trouble for malformed header lines. I hope it is now all correct.
305 PH/19 The functions {pwcheck,saslauthd}_verify_password() are always called
306 with the "reply" argument non-NULL. The code, however (which originally
307 came from elsewhere) had *some* tests for NULL when it wrote to *reply,
308 but it didn't always do it. This confused somebody who was copying the
309 code for some other use. I have removed all the tests.
311 PH/20 It was discovered that the GnuTLS code had support for RSA_EXPORT, a
312 feature that was used to support insecure browsers during the U.S. crypto
313 embargo. It requires special client support, and Exim is probably the
314 only MTA that supported it -- and would never use it because real RSA is
315 always available. This code has been removed, because it had the bad
316 effect of slowing Exim down by computing (never used) parameters for the
317 RSA_EXPORT functionality.
319 PH/21 On the advice of Timo Sirainen, added a check to the dovecot
320 authenticator to fail if there's a tab character in the incoming data
321 (there should never be unless someone is messing about, as it's supposed
322 to be base64-encoded). Also added, on Timo's advice, the "secured" option
323 if the connection is using TLS or if the remote IP is the same as the
324 local IP, and the "valid-client-cert option" if a client certificate has
327 PH/22 As suggested by Dennis Davis, added a server_condition option to *all*
328 authenticators. This can be used for authorization after authentication
329 succeeds. (In the case of plaintext, it servers for both authentication
332 PH/23 Testing for tls_required and lost_connection in a retry rule didn't work
333 if any retry times were supplied.
335 PH/24 Exim crashed if verify=helo was activated during an incoming -bs
336 connection, where there is no client IP address to check. In this
337 situation, the verify now always succeeds.
339 PH/25 Applied John Jetmore's -Mset patch.
341 PH/26 Added -bem to be like -Mset, but loading a message from a file.
343 PH/27 In a string expansion for a processed (not raw) header when multiple
344 headers of the same name were present, leading whitespace was being
345 removed from all of them, but trailing whitespace was being removed only
346 from the last one. Now trailing whitespace is removed from each header
347 before concatenation. Completely empty headers in a concatenation (as
350 PH/28 Fixed bug in backwards-compatibility feature of PH/09 (thanks to John
351 Jetmore). It would have mis-read ACL variables from pre-4.61 spool files.
353 PH/29 [Removed. This was a change that I later backed out, and forgot to
354 correct the ChangeLog entry (that I had efficiently created) before
355 committing the later change.]
357 PH/30 Exim was sometimes attempting to deliver messages that had suffered
358 address errors (4xx response to RCPT) over the same connection as other
359 messages routed to the same hosts. Such deliveries are always "forced",
360 so retry times are not inspected. This resulted in far too many retries
361 for the affected addresses. The effect occurred only when there were more
362 hosts than the hosts_max_try setting in the smtp transport when it had
363 the 4xx errors. Those hosts that it had tried were not added to the list
364 of hosts for which the message was waiting, so if all were tried, there
365 was no problem. Two fixes have been applied:
367 (i) If there are any address or message errors in an SMTP delivery, none
368 of the hosts (tried or untried) are now added to the list of hosts
369 for which the message is waiting, so the message should not be a
370 candidate for sending over the same connection that was used for a
371 successful delivery of some other message. This seems entirely
372 reasonable: after all the message is NOT "waiting for some host".
373 This is so "obvious" that I'm not sure why it wasn't done
374 previously. Hope I haven't missed anything, but it can't do any
375 harm, as the worst effect is to miss an optimization.
377 (ii) If, despite (i), such a delivery is accidentally attempted, the
378 routing retry time is respected, so at least it doesn't keep
379 hammering the server.
381 PH/31 Installed Andrew Findlay's patch to close the writing end of the socket
382 in ${readsocket because some servers need this prod.
384 PH/32 Added some extra debug output when updating a wait-xxx database.
386 PH/33 The hint "could be header name not terminated by colon", which has been
387 given for certain expansion errors for a long time, was not being given
388 for the ${if def:h_colon_omitted{... case.
390 PH/34 The spec says: "With one important exception, whenever a domain list is
391 being scanned, $domain contains the subject domain." There was at least
392 one case where this was not true.
394 PH/35 The error "getsockname() failed: connection reset by peer" was being
395 written to the panic log as well as the main log, but it isn't really
396 panic-worthy as it just means the connection died rather early on. I have
397 removed the panic log writing for the ECONNRESET error when getsockname()
400 PH/36 After a 4xx response to a RCPT error, that address was delayed (in queue
401 runs only) independently of the message's sender address. This meant
402 that, if the 4xx error was in fact related to the sender, a different
403 message to the same recipient with a different sender could confuse
404 things. In particualar, this can happen when sending to a greylisting
405 server, but other circumstances could also provoke similar problems.
406 I have changed the default so that the retry time for these errors is now
407 based a combination of the sender and recipient addresses. This change
408 can be overridden by setting address_retry_include_sender=false in the
411 PH/37 For LMTP over TCP/IP (the smtp transport), error responses from the
412 remote server are returned as part of bounce messages. This was not
413 happening for LMTP over a pipe (the lmtp transport), but now it is the
414 same for both kinds of LMTP.
416 PH/38 Despite being documented as not happening, Exim was rewriting addresses
417 in header lines that were in fact CNAMEs. This is no longer the case.
419 PH/39 If -R or -S was given with -q<time>, the effect of -R or -S was ignored,
420 and queue runs started by the daemon processed all messages. This has
421 been fixed so that -R and -S can now usefully be given with -q<time>.
423 PH/40 Import PCRE release 6.7 (fixes some bugs).
425 PH/41 Add bitwise logical operations to eval (courtesy Brad Jorsch).
427 PH/42 Give an error if -q is specified more than once.
429 PH/43 Renamed the variables $interface_address and $interface_port as
430 $received_ip_address and $received_port, to make it clear that these
431 values apply to message reception, and not to the outgoing interface when
432 a message is delivered. (The old names remain recognized, of course.)
434 PH/44 There was no timeout on the connect() call when using a Unix domain
435 socket in the ${readsocket expansion. There now is.
437 PH/45 Applied a modified version of Brad Jorsch's patch to allow "message" to
438 be meaningful with "accept".
440 SC/01 Eximstats V1.43
441 Bug fix for V1.42 with -h0 specified. Spotted by Chris Lear.
443 SC/02 Eximstats V1.44
444 Use a glob alias rather than an array ref in the generated
445 parser. This improves both readability and performance.
447 SC/03 Eximstats V1.45 (Marco Gaiarin / Steve Campbell)
448 Collect SpamAssassin and rejection statistics.
449 Don't display local sender or destination tables unless
450 there is data to show.
451 Added average volumes into the top table text output.
453 SC/04 Eximstats V1.46
454 Collect data on the number of addresses (recipients)
455 as well as the number of messages.
457 SC/05 Eximstats V1.47
458 Added 'Message too big' to the list of mail rejection
459 reasons (thanks to Marco Gaiarin).
461 SC/06 Eximstats V1.48
462 Mainlog lines which have GMT offsets and are too short to
463 have a flag are now skipped.
465 SC/07 Eximstats V1.49 (Alain Williams)
466 Added the -emptyok flag.
468 SC/08 Eximstats V1.50
469 Fixes for obtaining the IP address from reject messages.
471 JJ/03 exipick.20061117.2, made header handling as similar to exim as possible
472 (added [br]h_ prefixes, implemented RFC2047 decoding. Fixed
473 whitesspace changes from 4.64-PH/27
475 JJ/04 exipick.20061117.2, fixed format and added $message_headers_raw to
478 JJ/05 exipick.20061117.2, bug fixes (error out sooner when invalid criteria
479 are found, allow negative numbers in numeric criteria)
481 JJ/06 exipick.20061117.2, added new $message_body_missing variable
483 JJ/07 exipick.20061117.2, added $received_ip_address and $received_port
484 to match changes made in 4.64-PH/43
486 PH/46 Applied Jori Hamalainen's patch to add features to exiqsumm.
488 PH/47 Put in an explicit test for a DNS lookup of an address record where the
489 "domain" is actually an IP address, and force a failure. This locks out
490 those revolvers/nameservers that support "A-for-A" lookups, in
491 contravention of the specifications.
493 PH/48 When a host name was looked up from an IP address, and the subsequent
494 forward lookup of the name timed out, the host name was left in
495 $sender_host_name, contrary to the specification.
497 PH/49 Although default lookup types such as lsearch* or cdb*@ have always been
498 restricted to single-key lookups, Exim was not diagnosing an error if
499 * or *@ was used with a query-style lookup.
501 PH/50 Increased the value of DH_BITS in tls-gnu.c from 768 to 1024.
503 MH/01 local_scan ABI version incremented to 1.1. It should have been updated
504 long ago, but noone interested enough thought of it. Let's just say that
505 the "1.1" means that there are some new functions that weren't there at
506 some point in the past.
508 PH/51 Error processing for expansion failure of helo_data from an smtp
509 transport during callout processing was broken.
511 PH/52 Applied John Jetmore's patch to allow tls-on-connect and STARTTLS to be
512 tested/used via the -bh/-bhc/-bs options.
514 PH/53 Added missing "#include <time.h>" to pcre/pcretest.c (this was a PCRE
515 bug, fixed in subsequent PCRE releases).
517 PH/54 Applied Robert Bannocks' patch to avoid a problem with references that
518 arises when using the Solaris LDAP libraries (but not with OpenLDAP).
520 PH/55 Check for a ridiculously long file name in exim_dbmbuild.
526 SC/01 Use a glob alias rather than an array ref in eximstats generated
527 parser. This improves both readability and performance.
529 SC/02 Collect SpamAssassin and rejection statistics in eximstats.
530 Don't display local sender or destination tables in eximstats unless
531 there is data to show.
532 Added average volumes into the eximstats top table text output.
534 SC/03 Collect data on the number of addresses (recipients) as well
535 as the number of messages in eximstats.
537 TF/01 Correct an error in the documentation for the redirect router. Exim
538 does (usually) call initgroups() when daemonizing.
540 TF/02 Call initgroups() when dropping privilege in exim.c, so that Exim runs
541 with consistent privilege compared to when running as a daemon.
543 TF/03 Note in the spec that $authenticated_id is not set for local
544 submissions from trusted users.
546 TF/04 The ratelimit per_rcpt option now works correctly in acl_not_smtp.
547 Thanks to Dean Brooks <dean@iglou.com> for the patch.
549 TF/05 Make it easier to get SMTP authentication and TLS/SSL support working
550 by adding some example configuration directives to the default
551 configuration file. A little bit of work is required to uncomment the
552 directives and define how usernames and passwords are checked, but
553 there is now a framework to start from.
555 PH/01 Added #define LDAP_DEPRECATED 1 to ldap.c because some of the "old"
556 functions that Exim currently uses aren't defined in ldap.h for OpenLDAP
557 without this. I don't know how relevant this is to other LDAP libraries.
559 PH/02 Add the verb name to the "unknown ACL verb" error.
561 PH/03 Magnus Holmgren's patch for filter_prepend_home.
563 PH/03 Fixed Bugzilla #101: macro definition between ACLs doesn't work.
565 PH/04 Applied Magnus Holmgren's patch to fix Bugzilla #98: transport's home
566 directory not expanded when it should be if an expanded home directory
567 was set for the address (which is overridden by the transport).
569 PH/05 Applied Alex Kiernan's patch to fix Bugzilla #99: a problem with
572 PH/06 Added acl_not_smtp_start, based on Johannes Berg's patch, and set the
573 bit to forbid control=suppress_local_fixups in the acl_not_smtp ACL,
574 because it is too late at that time, and has no effect.
576 PH/07 Changed ${quote_pgsql to quote ' as '' instead of \' because of a
577 security issue with \' (bugzilla #107). I could not use the
578 PQescapeStringConn() function, because it needs a PGconn value as one of
581 PH/08 When testing addresses using -bt, indicate those final addresses that
582 are duplicates that would not cause an additional delivery. At least one
583 person was confused, thinking that -bt output corresponded to deliveries.
584 (Suppressing duplicates isn't a good idea as you lose the information
585 about possibly different redirections that led to the duplicates.)
587 PH/09 Applied patch from Erik to use select() instead of poll() in spam.c on
588 systems where poll() doesn't work, in particular OS X.
590 PH/10 Added more information to debugging output for retry time not reached.
592 PH/11 Applied patch from Arkadiusz Miskiewicz to apply a timeout to read
593 operations in malware.c.
595 PH/12 Applied patch from Magnus Holmgren to include the "h" tag in Domain Keys
598 PH/13 If write_rejectlog was set false when logging was sent to syslog with
599 syslog_duplication set false, log lines that would normally be written
600 both the the main log and to the reject log were not written to syslog at
603 PH/14 In the default configuration, change the use of "message" in ACL warn
604 statements to "add_header".
606 PH/15 Diagnose a filter syntax error for "seen", "unseen", or "noerror" if not
607 not followed by a command (e.g. "seen endif").
609 PH/16 Recognize SMTP codes at the start of "message" in ACLs and after :fail:
610 and :defer: in a redirect router. Add forbid_smtp_code to suppress the
613 PH/17 Added extra conditions to the default value of delay_warning_condition
617 { !eq{$h_list-id:$h_list-post:$h_list-subscribe:}{} } \
618 { match{$h_precedence:}{(?i)bulk|list|junk} } \
619 { match{$h_auto-submitted:}{(?i)auto-generated|auto-replied} } \
622 The Auto-Submitted: and various List- headers are standardised, whereas I
623 don't think Precedence: ever was.
625 PH/18 Refactored debugging code in route_finduser() to show more information,
626 in particular, the error code if getpwnam() issues one.
628 PH/19 Added PQsetClientEncoding(conn, "SQL_ASCII") to the pgsql code module.
629 This is apparently needed in addition to the PH/07 change above to avoid
630 any possible encoding problems.
632 PH/20 Perl can change the locale. Exim was resetting it after a ${perl call,
633 but not after initializing Perl.
635 PH/21 Added a call to PQsetNoticeProcessor() to catch pgsql "notices" and
636 output them only if debugging. By default they are written stderr,
637 apparently, which is not desirable.
639 PH/22 Added Alain Williams' LDAP patch to support setting REFERRALS=off on
642 JJ/01 exipick: added --reverse (and -R synonym), --random, --size, --sort and
645 JJ/02 exipick: rewrote --help documentation to hopefully make more clear.
647 PH/23 Made -oMaa and -oMt work with -bh and -bs to pretend the connection is
648 authenticated or an ident call has been made. Suppress the default
649 values for $authenticated_id and $authenticated_sender (but permit -oMai
650 and -oMas) when testing with -bh.
652 PH/24 Re-jigged the order of the tests in the default configuration so that the
653 tests for valid domains and recipients precede the DNS black list and CSA
654 tests, on the grounds that those ones are more expensive.
656 PH/25 Exim was not testing for a space following SMTP commands such as EHLO
657 that require one. Thus, EHLORHUBARB was interpreted as a valid command.
658 This bug exists in every version of Exim that I still have, right back to
661 PH/26 (n)wildlsearch lookups are documented as being done case-insensitively.
662 However, an attempt to turn on case-sensitivity in a regex key by
663 including (?-i) didn't work because the subject string was already
664 lowercased, and the effects were non-intuitive. It turns out that a
665 one-line patch can be used to allow (?-i) to work as expected.
671 TF/01 Fix the add_header change below (4.61 PH/55) which had a bug that (amongst
672 other effects) broke the use of negated acl sub-conditions.
674 PH/01 ${readsocket now supports Internet domain sockets (modified John Jetmore
677 PH/02 When tcp-wrappers is called from Exim, it returns only "deny" or "allow".
678 "Deny" causes Exim to reject the incoming connection with a 554 error.
679 Unfortunately, if there is a major crisis, such as a disk failure,
680 tcp-wrappers gives "deny", whereas what one would like would be some
681 kind of temporary error. A kludge has been added to help with this.
682 Before calling hosts_ctl(), errno is set zero. If the result is "deny", a
683 554 error is used if errno is still zero or contains ENOENT (which occurs
684 if either of the /etc/hosts.{allow,deny} files is missing). Otherwise, a
687 PH/03 Add -lutil to the default FreeBSD LIBS setting.
689 PH/04 Change PH/19 for 4.61 was too wide. It should not be applied to host
690 errors. Otherwise a message that provokes a temporary error (when other
691 messages do not) can cause a whole host to time out.
693 PH/05 Batch deliveries by appendfile and pipe transports did not work when the
694 addresses were routed directly to files or pipes from a redirect router.
695 File deliveries just didn't batch; pipe deliveries might have suffered
698 PH/06 A failure to get a lock for a hints database would erroneously always say
699 "Failed to get write lock", even when it was really a read lock.
701 PH/07 The appendfile transport was creating MBX lock files with a fixed mode
702 of 0600. This has been changed to use the value of the lockfile_mode
703 option (which defaults to 0600).
705 PH/08 Applied small patch from the Sieve maintainer.
707 PH/09 If maildir_quota_directory_regex was set to exclude (say) the .Trash
708 folder from quota calculations, a direct delivery into this folder messed
709 up the contents of the maildirsize file. This was because the regex was
710 used only to exclude .Trash (or whatever) when the size of the mailbox
711 was calculated. There was no check that a delivery was happening into an
712 excluded directory. This bug has been fixed by ignoring all quota
713 processing for deliveries into excluded directories.
715 PH/10 Added the maildirfolder_create_regex option to appendfile.
721 PH/01 The code for finding all the local interface addresses on a FreeBSD
722 system running IPv6 was broken. This may well have applied to all BSD
723 systems, as well as to others that have similar system calls. The broken
724 code found IPv4 interfaces correctly, but gave incorrect values for the
725 IPv6 interfaces. In particular, ::1 was not found. The effect in Exim was
726 that it would not match correctly against @[] and not recognize the IPv6
729 PH/02 The ipliteral router was not recognizing addresses of the form user@
730 [ipv6:....] because it didn't know about the "ipv6:" prefix.
732 PH/03 Added disable_ipv6.
734 PH/04 Changed $reply_address to use the raw form of the headers instead of the
735 decoded form, because it is most often used to construct To: headers
736 lines in autoreplies, and the decoded form may well be syntactically
737 invalid. However, $reply_address has leading white space removed, and all
738 newlines turned into spaces so that the autoreply transport does not
741 PH/05 If group was specified without a user on a router, and no group or user
742 was specified on a transport, the group from the router was ignored.
744 PH/06 Increased the number of ACL variables to 20 of each type, and arranged
745 for visible compile-time settings that can be used to change these
746 numbers, for those that want even more. Backwards compatibility with old
747 spool files has been maintained. However, going back to a previous Exim
748 release will lost any variables that are in spool files.
750 PH/07 Two small changes when running in the test harness: increase delay when
751 passing a TCP/IP connection to a new process, in case the original
752 process has to generate a bounce, and remove special handling of
753 127.0.0.2 (sic), which is no longer necessary.
755 PH/08 Changed debug output of dbfn_open() flags from numbers to names, so as to
756 be the same on different OS.
758 PH/09 Moved a debug statement in filter processing to avoid a race problem when
761 JJ/01 exipick: fixed bug where -b (brief) output option showed "Vars:"
762 whether --show-vars was specified or not
764 JJ/02 exipick: Added support for new ACL variable spool format introduced
767 PH/10 Fixed another bug related to PH/04 above: if an incoming message had a
768 syntactically invalid From: or Reply-to: line, and a filter used this to
769 generate an autoreply, and therefore failed to obtain an address for the
770 autoreply, Exim could try to deliver to a non-existent relative file
771 name, causing unrelated and misleading errors. What now happens is that
772 it logs this as a hard delivery error, but does not attempt to create a
775 PH/11 The exinext utility has a -C option for testing purposes, but although
776 the given file was scanned by exinext itself; it wasn't being passed on
777 when Exim was called.
779 PH/12 In the smtp transport, treat an explicit ECONNRESET error the same as
780 an end-of-file indication when reading a command response.
782 PH/13 Domain literals for IPv6 were not recognized unless IPv6 support was
783 compiled. In many other places in Exim, IPv6 addresses are always
784 recognized, so I have changed this. It also means that IPv4 domain
785 literals of the form [IPV4:n.n.n.n] are now always recognized.
787 PH/14 When a uid/gid is specified for the queryprogram router, it cannot be
788 used if the router is not running as root, for example, when verifying at
789 ACL time, or when using -bh. The debugging output from this situation was
790 non-existent - all you got was a failure to exec. I have made two
793 (a) Failures to set uid/gid, the current directory, or a process leader
794 in a subprocess such as that created by queryprogram now generate
795 suitable debugging ouput when -d is set.
797 (b) The queryprogram router detects when it is not running as root,
798 outputs suitable debugging information if -d is set, and then runs
799 the subprocess without attempting to change uid/gid.
801 PH/15 Minor change to Makefile for building test_host (undocumented testing
804 PH/16 As discussed on the list in Nov/Dec: Exim no longer looks at the
805 additional section of a DNS packet that returns MX or SRV records.
806 Instead, it always explicitly searches for A/AAAA records. This avoids
807 major problems that occur when a DNS server includes only records of one
808 type (A or AAAA) in an MX/SRV packet. A byproduct of this change has
809 fixed another bug: if SRV records were looked up and the corresponding
810 address records were *not* found in the additional section, the port
811 values from the SRV records were lost.
813 PH/17 If a delivery to a pipe, file, or autoreply was deferred, Exim was not
814 using the correct key (the original address) when searching the retry
815 rules in order to find which one to use for generating the retry hint.
817 PH/18 If quota_warn_message contains a From: header, Exim now refrains from
818 adding the default one. Similarly, if it contains a Reply-To: header, the
819 errors_reply_to option, if set, is not used.
821 PH/19 When calculating a retry time, Exim used to measure the "time since
822 failure" by looking at the "first failed" field in the retry record. Now
823 it does not use this if it is later than than the arrival time of the
824 message. Instead it uses the arrival time. This makes for better
825 behaviour in cases where some deliveries succeed, thus re-setting the
826 "first failed" field. An example is a quota failure for a huge message
827 when small messages continue to be delivered. Without this change, the
828 "time since failure" will always be short, possible causing more frequent
829 delivery attempts for the huge message than are intended.
830 [Note: This change was subsequently modified - see PH/04 for 4.62.]
832 PH/20 Added $auth1, $auth2, $auth3 to contain authentication data (as well as
833 $1, $2, $3) because the numerical variables can be reset during some
834 expansion items (e.g. "match"), thereby losing the authentication data.
836 PH/21 Make -bV show the size of off_t variables so that the test suite can
837 decide whether to run tests for quotas > 2G.
839 PH/22 Test the values given for quota, quota_filecount, quota_warn_threshold,
840 mailbox_size, and mailbox_filecount in the appendfile transport. If a
841 filecount value is greater than 2G or if a quota value is greater than 2G
842 on a system where the size of off_t is not greater than 4, a panic error
845 PH/23 When a malformed item such as 1.2.3/24 appears in a host list, it can
846 never match. The debug and -bh output now contains an explicit error
847 message indicating a malformed IPv4 address or mask.
849 PH/24 An host item such as 1.2.3.4/abc was being treated as the IP address
850 1.2.3.4 without a mask. Now it is not recognized as an IP address, and
853 PH/25 Do not write to syslog when running in the test harness. The only
854 occasion when this arises is a failure to open the main or panic logs
855 (for which there is an explicit test).
857 PH/26 Added the /no_tell option to "control=freeze".
859 PH/27 If a host name lookup failed very early in a connection, for example, if
860 the IP address matched host_lookup and the reverse lookup yielded a name
861 that did not have a forward lookup, an error message of the form "no IP
862 address found for host xxx.xxx.xxx (during SMTP connection from NULL)"
863 could be logged. Now it outputs the IP address instead of "NULL".
865 PH/28 An enabling patch from MH: add new function child_open_exim2() which
866 allows the sender and the authenticated sender to be set when
867 submitting a message from within Exim. Since child_open_exim() is
868 documented for local_scan(), the new function should be too.
870 PH/29 In GnuTLS, a forced expansion failure for tls_privatekey was not being
871 ignored. In both GnuTLS and OpenSSL, an expansion of tls_privatekey that
872 results in an empty string is now treated as unset.
874 PH/30 Fix eximon buffer overflow bug (Bugzilla #73).
876 PH/31 Added sender_verify_fail logging option.
878 PH/32 In November 2003, the code in Exim that added an empty Bcc: header when
879 needed by RFC 822 but not by RFC 2822 was commented out. I have now
880 tidied the source and removed it altogether.
882 PH/33 When a queue run was abandoned because the load average was too high, a
883 log line was always written; now it is written only if the queue_run log
884 selector is set. In addition, the log line for abandonment now contains
885 information about the queue run such as the pid. This is always present
886 in "start" and "stop" lines but was omitted from the "abandon" line.
888 PH/34 Omit spaces between a header name and the colon in the error message that
889 is given when verify = headers_syntax fails (if there are lots of them,
890 the message gets confusing).
892 PH/35 Change the default for dns_check_names_pattern to allow slashes within
893 names, as there are now some PTR records that contain slashes. This check
894 is only to protect against broken name servers that fall over on strange
895 characters, so the fact that it applies to all lookups doesn't matter.
897 PH/36 Now that the new test suite is complete, we can remove some of the
898 special code in Exim that was needed for the old test suite. For example,
899 sorting DNS records because real resolvers return them in an arbitrary
900 order. The new test suite's fake resolver always returns records in the
903 PH/37 When running in the test harness, use -odi for submitted messages (e.g.
904 bounces) except when queue_only is set, to avoid logging races between
905 the different processes.
907 PH/38 Panic-die if .include specifies a non-absolute path.
909 PH/39 A tweak to the "H" retry rule from its user.
911 JJ/03 exipick: Removed parentheses from 'next' and 'last' calls that specified
912 a label. They prevented compilation on older perls.
914 JJ/04 exipick: Refactored code to prevent implicit split to @_ which caused
915 a warning to be raised on newish perls.
917 JJ/05 exipick: Fixed bug where -bpc always showed a count of all messages
918 on queue. Changes to match documented behaviour of showing count of
919 messages matching specified criteria.
921 PH/40 Changed the default ident timeout from 30s to 5s.
923 PH/41 Added support for the use of login_cap features, on those BSD systems
924 that have them, for controlling the resources used by pipe deliveries.
926 PH/42 The content-scanning code uses fopen() to create files in which to put
927 message data. Previously it was not paying any attention to the mode of
928 the files. Exim runs with umask(0) because the rest of the code creates
929 files with open(), and sets the required mode explicitly. Thus, these
930 files were ending up world-writeable. This was not a big issue, because,
931 being within the spool directory, they were not world-accessible. I have
932 created a function called modefopen, which takes an additional mode
933 argument. It sets umask(777), creates the file, chmods it to the required
934 mode, then resets the umask. All the relevant calls to fopen() in the
935 content scanning code have been changed to use this function.
937 PH/43 If retry_interval_max is set greater than 24 hours, it is quietly reset
938 to 24 hours. This avoids potential overflow problems when processing G
939 and H retry rules. I suspect nobody ever tinkers with this value.
941 PH/44 Added STRIP_COMMAND=/usr/bin/strip to the FreeBSD Makefile.
943 PH/45 When the plaintext authenticator is running as a client, the server's
944 challenges are checked to ensure they are valid base64 strings. By
945 default, the authentication attempt is cancelled if an invalid string is
946 received. Setting client_ignore_invalid_base64 true ignores these errors.
947 The decoded challenge strings are now placed in $auth1, $auth2, etc. as
948 they are received. Thus, the responses can be made to depend on the
949 challenges. If an invalid string is ignored, an empty string is placed in
952 PH/46 Messages that are created by the autoreply transport now contains a
953 References: header, in accordance with RFCs 2822 and 3834.
955 PH/47 Added authenticated_sender_force to the smtp transport.
957 PH/48 The ${prvs expansion was broken on systems where time_t was long long.
959 PH/49 Installed latest patch from the Sieve maintainer.
961 PH/50 When an Exim quota was set without a file count quota, and mailbox_size
962 was also set, the appendfile transport was unnecessarily scanning a
963 directory of message files (e.g. for maildir delivery) to find the count
964 of files (along with the size), even though it did not need this
965 information. It now does the scan only if it needs to find either the
966 size of the count of files.
968 PH/51 Added ${time_eval: to convert Exim time strings into seconds.
970 PH/52 Two bugs concerned with error handling when the smtp transport is
973 (i) Exim was not creating retry information for temporary errors given
974 for individual recipients after the DATA command when the smtp transport
975 was used in LMTP mode. This meant that they could be retried too
976 frequently, and not timed out correctly.
978 (ii) Exim was setting the flag that allows error details to be returned
979 for LMTP errors on RCPT commands, but not for LMTP errors for individual
980 recipients that were returned after the DATA command.
982 PH/53 This is related to PH/52, but is more general: for any failing address,
983 when detailed error information was permitted to be returned to the
984 sender, but the error was temporary, then after the final timeout, only
985 "retry timeout exceeded" was returned. Now it returns the full error as
986 well as "retry timeout exceeded".
988 PH/54 Added control=allow_auth_unadvertised, as it seems there are clients that
989 do this, and (what is worse) MTAs that accept it.
991 PH/55 Added the add_header modified to ACLs. The use of "message" with "warn"
992 will now be deprecated.
994 PH/56 New os.c-cygwin from the Cygwin maintainer.
996 JJ/06 exipick: added --unsorted option to allow unsorted output in all output
997 formats (previously only available in exim formats via -bpr, -bpru,
998 and -bpra. Now also available in native and exiqgrep formats)
1000 JJ/07 exipick: added --freeze and --thaw options to allow faster interaction
1001 with very large, slow to parse queues
1003 JJ/08 exipick: added ! as generic prefix to negate any criteria format
1005 JJ/09 exipick: miscellaneous performance enhancements (~24% improvements)
1007 PH/57 Tidies in SMTP dialogue display in debug output: (i) It was not showing
1008 responses to authentication challenges, though it was showing the
1009 challenges; (ii) I've removed the CR characters from the debug output for
1012 PH/58 Allow for the insertion of a newline as well as a space when a string
1013 is turned into more than one encoded-word during RFC 2047 encoding. The
1014 Sieve code now uses this.
1016 PH/59 Added the following errors that can be detected in retry rules: mail_4xx,
1017 data_4xx, lost_connection, tls_required.
1019 PH/60 When a VRFY deferred or FAILED, the log message rather than the user
1020 message was being sent as an SMTP response.
1022 PH/61 Add -l and -k options to exicyclog.
1024 PH/62 When verifying, if an address was redirected to one new address, so that
1025 verification continued, and the new address failed or deferred after
1026 having set something in $address_data, the value of $address_data was not
1027 passed back to the ACL. This was different to the case when no
1028 redirection occurred. The value is now passed back in both cases.
1030 PH/63 Changed the macro HAVE_LOGIN_CAP (see PH/41 for this release above) to
1031 HAVE_SETCLASSRESOURCES because there are different APIs in use that all
1032 use login_cap.h, so on its own it isn't the distinguishing feature. The
1033 new name refers directly to the setclassresources() function.
1035 PH/65 Added configuration files for NetBSD3.
1037 PH/66 Updated OS/Makefile-HP-UX for gcc 4.1.0 with HP-UX 11.
1039 PH/67 Fixed minor infelicity in the sorting of addresses to ensure that IPv6
1040 is preferred over IPv4.
1042 PH/68 The bounce_return_message and bounce_return_body options were not being
1043 honoured for bounces generated during the reception of non-SMTP messages.
1044 In particular, this applied to messages rejected by the ACL. This bug has
1045 been fixed. However, if bounce_return_message is true and bounce_return_
1046 body is false, the headers that are returned for a non-SMTP message
1047 include only those that have been read before the error was detected.
1048 (In the case of an ACL rejection, they have all been read.)
1050 PH/69 The HTML version of the specification is now built in a directory called
1051 spec_html instead of spec.html, because the latter looks like a path with
1052 a MIME-type, and this confuses some software.
1054 PH/70 Catch two compiler warnings in sieve.c.
1056 PH/71 Fixed an obscure and subtle bug (thanks Alexander & Matthias). The
1057 function verify_get_ident() calls ip_connect() to connect a socket, but
1058 if the "connect()" function timed out, ip_connect() used to close the
1059 socket. However, verify_get_ident() also closes the socket later, and in
1060 between Exim writes to the log, which may get opened at this point. When
1061 the socket was closed in ip_connect(), the log could get the same file
1062 descriptor number as the socket. This naturally causes chaos. The fix is
1063 not to close the socket in ip_connect(); the socket should be closed by
1064 the function that creates it. There was only one place in the code where
1065 this was missing, in the iplookup router, which I don't think anybody now
1066 uses, but I've fixed it anyway.
1068 PH/72 Make dns_again_means_nonexist apply to lookups using gethostbyname() as
1069 well as to direct DNS lookups. Otherwise the handling of names in host
1070 lists is inconsistent and therefore confusing.
1076 PH/01 Two changes to the default runtime configuration:
1078 (1) Move the checks for relay_from_hosts and authenticated clients from
1079 after to before the (commented out) DNS black list checks.
1081 (2) Add control=submission to the relay_from_hosts and authenticated
1082 clients checks, on the grounds that messages accepted by these
1083 statements are most likely to be submissions.
1085 PH/02 Several tidies to the handling of ${prvs and ${prvscheck:
1087 (1) Generate an error if the third argument for the ${prvs expansion is
1090 (2) Treat a missing third argument of ${prvscheck as if it were an empty
1093 (3) Reset the variables that are obtained from the first argument of
1094 ${prvscheck and used in the second argument before leaving the code,
1095 because their memory is reclaimed, so using them afterwards may do
1098 (4) Tidy up the code for expanding the arguments of ${prvscheck one by
1099 one (it's much easier than Tom thought :-).
1101 (5) Because of (4), we can now allow for the use of $prvscheck_result
1102 inside the third argument.
1104 PH/03 For some reason, the default setting of PATH when running a command from
1105 a pipe transport was just "/usr/bin". I have changed it to
1108 PH/04 SUPPORT_TRANSLATE_IP_ADDRESS and MOVE_FROZEN_MESSAGES did not cause
1109 anything to be listed in the output from -bV.
1111 PH/05 When a filter generated an autoreply, the entire To: header line was
1112 quoted in the delivery log line, like this:
1114 => >A.N.Other <ano@some.domain> <original@ddress> ...
1116 This has been changed so that it extracts the operative address. There
1117 may be more than one such address. If so, they are comma-separated, like
1120 => >ano@some.domain,ona@other.domain <original@ddress> ...
1122 PH/06 When a client host used a correct literal IP address in a HELO or EHLO
1123 command, (for example, EHLO [1.2.3.4]) and the client's IP address was
1124 not being looked up in the rDNS to get a host name, Exim was showing the
1125 IP address twice in Received: lines, even though the IP addresses were
1126 identical. For example:
1128 Received: from [1.2.3.4] (helo=[1.2.3.4])
1130 However, if the real host name was known, it was omitting the HELO data
1131 if it matched the actual IP address. This has been tidied up so that it
1132 doesn't show the same IP address twice.
1134 PH/07 When both +timestamp and +memory debugging was on, the value given by
1135 $tod_xxx expansions could be wrong, because the tod_stamp() function was
1136 called by the debug printing, thereby overwriting the timestamp buffer.
1137 Debugging no longer uses the tod_stamp() function when +timestamp is set.
1139 PH/08 When the original message was included in an autoreply transport, it
1140 always said "this is a copy of the message, including all the headers",
1141 even if body_only or headers_only was set. It now gives an appropriate
1144 PH/09 Applied a patch from the Sieve maintainer which:
1146 o fixes some comments
1147 o adds the (disabled) notify extension core
1148 o adds some debug output for the result of if/elsif tests
1149 o points to the current vacation draft in the documentation
1150 and documents the missing references header update
1154 o fixes a bug in processing the envelope test (when testing
1155 multiple envelope elements, the last element determinted the
1158 PH/10 Exim was violating RFC 3834 ("Recommendations for Automatic Responses to
1159 Electronic Mail") by including:
1161 Auto-submitted: auto-generated
1163 in the messages that it generates (bounce messages and others, such as
1164 warnings). In the case of bounce messages for non-SMTP mesages, there was
1165 also a typo: it was using "Auto_submitted" (underscore instead of
1166 hyphen). Since every message generated by Exim is necessarily in response
1167 to another message, thes have all been changed to:
1169 Auto-Submitted: auto-replied
1171 in accordance with these statements in the RFC:
1173 The auto-replied keyword:
1175 - SHOULD be used on messages sent in direct response to another
1176 message by an automatic process,
1178 - MUST NOT be used on manually-generated messages,
1180 - MAY be used on Delivery Status Notifications (DSNs) and Message
1181 Disposition Notifications (MDNs),
1183 - MUST NOT be used on messages generated by automatic or periodic
1184 processes, except for messages which are automatic responses to
1187 PH/11 Added "${if def:sender_address {(envelope-from <$sender_address>)\n\t}}"
1188 to the default Received: header definition.
1190 PH/12 Added log selector acl_warn_skipped (default on).
1192 PH/13 After a successful wildlsearch lookup, discard the values of numeric
1193 variables because (a) they are in the wrong storage pool and (b) even if
1194 they were copied, it wouldn't work properly because of the caching.
1196 PH/14 Add check_rfc2047_length to disable enforcement of RFC 2047 length
1197 checking when decoding. Apparently there are clients that generate
1198 overlong encoded strings. Why am I not surprised?
1200 PH/15 If the first argument of "${if match_address" was not empty, but did not
1201 contain an "@" character, Exim crashed. Now it writes a panic log message
1202 and treats the condition as false.
1204 PH/16 In autoreply, treat an empty string for "once" the same as unset.
1206 PH/17 A further patch from the Sieve maintainer: "Introduce the new Sieve
1207 extension "envelope-auth". The code is finished and in agreement with
1208 other implementations, but there is no documentation so far and in fact,
1209 nobody wrote the draft yet. This extension is currently #undef'ed, thus
1210 not changing the active code.
1212 Print executed "if" and "elsif" statements when debugging is used. This
1213 helps a great deal to understand what a filter does.
1215 Document more things not specified clearly in RFC3028. I had all this
1216 sorted out, when out of a sudden new issues came to my mind. Oops."
1218 PH/18 Exim was not recognizing the "net-" search type prefix in match_ip lists
1221 PH/19 Exim expands the IPv6 address given to -bh to its full non-abbreviated
1222 canonical form (as documented). However, after a host name lookup from
1223 the IP address, check_host() was doing a simple string comparison with
1224 addresses acquired from the DNS when checking that the found name did
1225 have the original IP as one of its addresses. Since any found IPv6
1226 addresses are likely to be in abbreviated form, the comparison could
1227 fail. Luckily, there already exists a function for doing the comparison
1228 by converting both addresses to binary, so now that is used instead of
1229 the text comparison.
1231 PH/20 There was another similar case to PH/19, when a complete host name was
1232 given in a host list; looking up its IP address could give an abbreviated
1233 form, whereas the current host's name might or might not be abbreviated.
1234 The same fix has been applied.
1240 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
1241 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
1244 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
1245 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
1247 PH/03 Typo: missing ".o" in src/pcre/Makefile.
1249 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
1250 header line, restrict the check to what is listed in RFCs 2369 and 2929.
1251 Also, for "Auto-Submitted", treat anything other than "no" as
1252 non-personal, in accordance with RFC 3834. (Previously it treated
1253 anything starting "auto-" as non-personal.)
1255 TF/01 The control=submission/name=... option had a problem with syntax
1256 errors if the name included a slash character. The /name= option
1257 now slurps the rest of the string, so it can include any characters
1258 but it must come last in the list of options (after /sender_retain
1261 PH/05 Some modifications to the interface to the fake nameserver for the new
1269 TK/01 Added the "success_on_redirect" address verification option. See
1270 NewStuff for rationale and an example.
1272 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
1274 PH/02 Patch to exigrep to allow it to work on syslog lines.
1276 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
1277 fread() to skip over the body file's header line, because in Cygwin the
1278 header line is locked and is inaccessible.
1280 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
1281 co-exist for some time) to make it clear that it is the Exim ID that is
1282 referenced, not the Message-ID: header line.
1284 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
1285 string_format() function, because snprintf() does not exist on all
1288 PH/06 The use of forbid_filter_existstest now also locks out the use of the
1289 ${stat: expansion item.
1291 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
1292 protocol synchronization error", to keep the pedants happy.
1294 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
1295 well as for IRIX systems, when gcc is being used. See the host.c source
1298 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
1300 PH/10 Named domain lists were not working if used in a queue_smtp_domains
1303 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
1304 transport and to the smtp transport in LMTP mode.
1306 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
1308 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
1309 run a filter in a subprocess. This could lead to confusion in subsequent
1310 lookups in the parent process. There should also be a search_tidyup() at
1311 the end of the subprocess.
1313 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
1314 only if the host matched helo_try_verify_hosts, which caused the
1315 verification to occur when the EHLO/HELO command was issued. The ACL just
1316 tested the remembered result. Now, if a previous verification attempt has
1317 not happened, "verify = helo" does it there and then.
1319 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
1321 TK/03 Fix log output including CR from clamd.
1323 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
1324 exist provoked a memory error which could cause a segfault.
1326 PH/15 Installed PCRE 6.2
1328 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
1330 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
1331 of the problem. Specifically, suggested +O2 rather than +O1 for the
1334 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
1336 PH/20 If a delivery was routed to a non-standard port by means of an SRV
1337 record, the port was not correctly logged when the outgoing_port log
1338 selector was set (it logged the transort's default port).
1340 PH/21 Added support for host-specific ports to manualroute, queryprogram,
1341 fallback_hosts, and "hosts" in the smtp transport.
1343 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
1344 host errors such as "Connection refused".
1346 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
1347 authentication with radiusclient 0.4.9:
1349 - Error returned from rc_read_config was caught wrongly
1350 - Username/password not passed on to radius server due to wrong length.
1352 The presumption is that some radiusclient API changes for 4.51/PH/17
1353 were not taken care of correctly. The code is still untested by me (my
1354 Linux distribution still has 0.3.2 of radiusclient), but it was
1355 contributed by a Radius user.
1357 PH/24 When doing a callout, the value of $domain wasn't set correctly when
1358 expanding the "port" option of the smtp transport.
1360 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
1361 while reading a MIME header. Thanks to Tom Hughes for a patch.
1363 PH/24 Include config.h inside local_scan.h so that configuration settings are
1366 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
1367 that in an ACL for RCPT (for example), you can examine exactly what was
1370 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
1371 commands, but it was not correctly comparing the address with the actual
1372 client host address. Thus, it would show the EHLO address in Received:
1373 header lines when this was not necessary.
1375 PH/27 Added the % operator to ${eval:}.
1377 PH/28 Exim tries to create and chdir to its spool directory when it starts;
1378 it should be ignoring failures (because with -C, for example, it has lost
1379 privilege). It wasn't ignoring creation failures other than "already
1382 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
1385 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
1386 because an input line was too long, either on its own, or by virtue of
1387 too many continuations, the temporary file was not being removed, and the
1388 return code was incorrect.
1390 PH/31 Missing "BOOL" in function definition in filtertest.c.
1392 PH/32 Applied Sieve patches from the maintainer.
1394 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
1396 PH/33 Added "verify = not_blind".
1398 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
1399 Local/Makefile (with some defaults set). These are used in built scripts
1400 such as exicyclog, but they have never been used in the exim_install
1401 script (though there are many overriding facilities there). I have
1402 arranged that the exim_install script now takes note of these two
1405 PH/35 Installed configuration files for Dragonfly.
1407 PH/36 When a locally submitted message by a trusted user did not contain a
1408 From: header, and the sender address was obtained from -f or from an SMTP
1409 MAIL command, and the trusted user did not use -F to supply a sender
1410 name, $originator_name was incorrectly used when constructing a From:
1411 header. Furthermore, $originator_name was used for submission mode
1412 messages from external hosts without From: headers in a similar way,
1413 which is clearly wrong.
1415 PH/37 Added control=suppress_local_fixups.
1417 PH/38 When log_selector = +received_sender was set, and the addition of the
1418 sender made the log line's construction buffer exactly full, or one byte
1419 less than full, an overflow happened when the terminating "\n" was
1422 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
1423 when the result of a list match is failure because a DNS lookup failed.
1425 PH/40 RM_COMMAND is now used in the building process.
1427 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
1428 the "build-* directories that it finds.
1430 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
1431 address in a domain literal was a prefix of an interface address.
1433 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
1434 when verifying a sender address, unless rewrite_headers is false.
1436 PH/44 Wrote a long comment about why errors_to addresses are verified as
1437 recipients, not senders.
1439 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
1440 the ratelimit ACL was added.
1442 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
1444 PH/46 Added extra information about PostgreSQL errors to the error string.
1446 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
1447 avoiding the need to install special zones in a real server. This is
1448 backwards compatible; if it can't find the fake resolver, it drops back.
1449 Thus, both old and new test suites can be run.
1451 TF/02 Added util/ratelimit.pl
1453 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
1454 clock is set back in time.
1456 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
1457 Candler <B.Candler@pobox.com>.
1459 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
1460 OK for addresses that are the result of redirections.
1462 PH/48 A number of further additions for the benefit of the new test suite,
1463 including a fake gethostbyname() that interfaces to the fake DNS resolver
1466 TF/06 The fix for widen_domains has also been applied to qualify_single and
1467 search_parents which are the other dnslookup options that can cause
1470 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
1473 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
1475 TF/07 Exim produced the error message "an SRV record indicated no SMTP
1476 service" if it encountered an MX record with an empty target hostname.
1477 The message is now "an MX or SRV record indicated no SMTP service".
1479 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
1480 if the DNS of the sending site is misconfigured. This is quite a
1481 common situation. This change restores the behaviour of treating a
1482 helo verification defer as a failure.
1484 PH/51 If self=fail was set on a router, the bounce message did not include the
1485 actual error message.
1491 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
1493 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
1494 command itself ended in error, the underlying message about the transport
1495 filter timeout was being overwritten with the pipe command error. Now the
1496 underlying error message should be appended to the second error message.
1498 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
1500 PH/02 Reduce the amount of output that "make" produces by default. Full output
1501 can still be requested.
1503 PH/03 The warning log line about a condition test deferring for a "warn" verb
1504 was being output only once per connection, rather than after each
1505 occurrence (because it was using the same function as for successful
1506 "warn" verbs). This seems wrong, so I have changed it.
1508 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
1509 it should not have, which might have caused a crash in the right
1510 circumstances, but probably never did.
1512 PH/04 Installed a modified version of Tony Finch's patch to make submission
1513 mode fix the return path as well as the Sender: header line, and to
1514 add a /name= option so that you can make the user's friendly name appear
1517 TF/03 Added the control = fakedefer ACL modifier.
1519 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
1520 Mark Lowes for thorough testing.
1522 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
1524 TK/03 Merged latest SRS patch from Miles Wilton.
1526 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
1527 with the definition in sysexits.h (which is #included earlier).
1528 Fortunately, Exim does not actually use EX_OK. The code used to try to
1529 preserve the sysexits.h value, by assumimg that macro definitions were
1530 scanned for macro replacements. I have been disabused of this notion,
1531 so now the code just undefines EX_OK before #including unistd.h.
1533 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
1534 in the smtp transport. When a block could not be written in a single
1535 write() function, the timeout was being re-applied to each part-write.
1536 This seems wrong - if the receiver was accepting one byte at a time it
1537 would take for ever. The timeout is now adjusted when this happens. It
1538 doesn't have to be particularly precise.
1540 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
1541 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
1543 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
1544 without a domain if the check to <postmaster@domain> fails.
1546 SC/01 Eximstats: added -xls and the ability to specify output files
1547 (patch written by Frank Heydlauf).
1549 SC/02 Eximstats: use FileHandles for outputing results.
1551 SC/03 Eximstats: allow any combination of xls, txt, and html output.
1553 SC/04 Eximstats: fixed display of large numbers with -nvr option
1555 SC/05 Eximstats: fixed merging of reports with empty tables.
1557 SC/06 Eximstats: added the -include_original_destination flag
1559 SC/07 Eximstats: removed tabs and trailing whitespace.
1561 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
1563 TK/06 MBOX spool code: Add real "From " MBOX separator line
1564 so the .eml file is really in mbox format (even though
1565 most programs do not really care). Patch from Alex Miller.
1567 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
1568 The latter is generated from $received_to and is only set if the
1569 message has one envelope recipient. SA can use these headers,
1570 obviously out-of-the-box. Patch from Alex Miller.
1572 PH/08 The ${def test on a variable was returning false if the variable's
1573 value was "0", contrary to what the specification has always said!
1574 The result should be true unless the variable is empty.
1576 PH/09 The syntax error of a character other than { following "${if
1577 def:variable_name" (after optional whitespace) was not being diagnosed.
1578 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
1579 accidental colon was present, for example, could give incorrect results.
1581 PH/10 Tidied the code in a number of places where the st_size field of a stat()
1582 result is used (not including appendfile, where other changes are about
1585 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
1586 This involved changing a lot of size variables from int to off_t. It
1587 should work with maildirs and everything.
1589 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
1590 spamd dying while we are connected to it.
1592 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
1595 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
1596 function for BDB 4.3.
1598 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
1600 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
1601 the sources for PCRE have been re-arranged and more files are now
1604 PH/15 The code I had for printing potentially long long variables in PH/11
1605 above was not the best (it lost precision). The length of off_t variables
1606 is now inspected at build time, and an appropriate printing format (%ld
1607 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
1608 to be "long long int" or "long int". This is needed for the internal
1609 formatting function string_vformat().
1611 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
1612 the configuration file to be ":syslog", then the script "guesses" where
1613 the logs files are, rather than using the compiled in default. In our
1614 case the guess is not the same as the compiled default, so the script
1615 suddenly stopped working when I started to use syslog. The patch checks
1616 to see if log_file_path is "". If so, it attempts to read it from exim
1617 with no configuration file to get the compiled in version, before it
1618 falls back to the previous guessing code."
1620 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
1621 implementing BATV in an Exim configuration. See NewStuff for the gory
1624 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
1625 Makefile that are specific to HP-UX.
1627 PH/18 If the "use_postmaster" option was set for a recipient callout together
1628 with the "random" option, the postmaster address was used as the MAIL
1629 FROM address for the random test, but not for the subsequent recipient
1630 test. It is now used for both.
1632 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
1633 patch removes a few documentation additions to RFC 3028, because the
1634 latest draft now contains them. It adds the new en;ascii-case comparator
1635 and a new error check for 8bit text in MIME parts. Comparator and
1636 require names are now matched exactly. I enabled the subaddress
1637 extension, but it is not well tested yet (read: it works for me)."
1639 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
1640 rework some of the code of TK/09 above to avoid the hardwired use of
1641 "%lld" and "long long". Replaced the call to snprintf() with a call to
1644 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
1645 records point to non-existent hosts", "retry timeout exceeded", and
1646 "retry time not reached for any host after a long failure period".
1648 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
1649 experimental DomainKeys support:
1651 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
1652 (2) On an error such as an illegally used "control", the wrong name for
1653 the control was given.
1655 These problems did NOT occur unless DomainKeys support was compiled.
1657 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
1659 PH/24 Added ${if match_ip condition.
1661 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
1662 databases so that it will be absolutely obvious if a crash occurs in the
1663 DB library. This is a regular occurrence (often caused by mis-matched
1666 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
1667 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
1668 picked up on a user's system that detects such things. There doesn't seem
1669 to be a gcc warning option for this - only an attribute that has to be
1670 put on the function's prototype. It seems that in Fedora Core 4 they have
1671 set this on a number of new functions. No doubt there will be more in due
1674 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
1675 specify a transport. However, if an address that was verified by such a
1676 router was the subject of a callout, Exim crashed because it tried to
1677 read the rcpt_include_affixes from the non-existent transport. Now it
1678 just assumes that the setting of that option is false. This bug was
1679 introduced by 4.51/PH/31.
1681 PH/28 Changed -d+all to exclude +memory, because that information is very
1682 rarely of interest, but it makes the output a lot bigger. People tend to
1683 do -d+all out of habit.
1685 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
1686 code in os-type was giving problems when libc.so lives in lib64, like on
1689 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
1690 aren't the modern standard, and it seems that some systems' include files
1691 don't always have them. Exim was already checking for some of the newer
1692 ones like T_AAAA, and defining it itself. I've added checks for all the
1693 record types that Exim uses.
1695 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
1696 not automatically generating a new one, as it is supposed to. This
1697 prevented TLS from working. If the file did exist, but contained invalid
1698 data, a new version was generated, as expected. It was only the case of a
1699 non-existent file that was broken.
1701 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
1702 with a change in libdomainkeys > 0.64.
1704 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
1705 from DNS. If the selector record carries the flag, it now has
1706 precedence over the domain-wide flag.
1708 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
1710 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
1711 the use of a port name that isn't defined in /etc/services) occurred, the
1712 message was deferred as in a normal delivery, and thus remained on the
1713 spool, instead of being failed because of the mua_wrapper setting. This
1714 is now fixed, and I tidied up some of the mua_wrapper messages at the
1717 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
1718 the messages in a hash of arrays rather than using individual hashes.
1719 This is a bit cleaner and results in dramatic memory savings, albeit
1720 at a slight CPU cost.
1722 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
1723 as requested by Marc Sherman.
1725 SC/10 Eximstats: added histograms for user specified patterns as requested
1728 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
1730 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
1731 fopen() in the content-scanning modules that did not already have it.
1737 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
1738 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
1740 TK/02 Fix ACL "control" statement not being available in MIME ACL.
1742 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
1744 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
1745 to test Sieve filters that use "vacation".
1747 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
1748 that changes the way the GnuTLS parameters are stored in the cache file.
1749 The new format can be generated externally. For backward compatibility,
1750 if the data in the cache doesn't make sense, Exim assumes it has read an
1751 old-format file, and it generates new data and writes a new file. This
1752 means that you can't go back to an older release without removing the
1755 PH/03 A redirect router that has both "unseen" and "one_time" set does not
1756 work if there are any delivery delays because "one_time" forces the
1757 parent to be marked "delivered", so its unseen clone is never tried
1758 again. For this reason, Exim now forbids the simultaneous setting of
1761 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
1762 redirected to themselves ("homonym" addresses). Read the long ChangeLog
1763 entry if you want to know the details. The fix, however, neglected to
1764 consider the case when local delivery batching is involved. The test for
1765 "previously delivered" was not happening when checking to see if an
1766 address could be batched with a previous (undelivered) one; under
1767 certain circumstances this could lead to multiple deliveries to the same
1770 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
1771 in its include files, and this causes problems building Exim.
1773 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
1774 header_syntax) but Exim was just ignoring anything given after a slash.
1775 In particular, this caused confusion with an attempt to use "verify =
1776 reverse_host_lookup/defer_ok". An error is now given when options are
1777 supplied for verify items that do not have them. (Maybe reverse_host_
1778 lookup should have a defer_ok option, but that's a different point.)
1780 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
1781 defined by RFC 821) to 2048, because there were problems with some AUTH
1782 commands, and RFC 1869 says the size should be increased for extended
1783 SMTP commands that take arguments.
1785 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
1788 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
1789 "unknown" error; now it says that the functionality isn't in the binary.
1791 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
1792 an address' error message when a string expansion fails (syntax or
1793 whatever). Otherwise the password may appear in the log. Following change
1794 PH/42 below, there is no longer a chance of it appearing in a bounce
1797 PH/11 Installed exipick version 20050225.0 from John Jetmore.
1799 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
1800 of its addresses was ever tried. (Bugzilla bug #2.)
1802 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
1803 the result incorrectly in the debug output. (It correctly added a newline
1804 to what was transported.)
1806 TF/01 Added $received_time.
1808 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
1809 commented out examples of how to interface to a virus scanner and to
1810 SpamAssassin. Also added commented examples of av_scanner and
1811 spamd_address settings.
1813 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
1814 and controls are allowed in which ACLs. There were a couple of minor
1815 errors. Some of the entries in the conditions table (which is a table of
1816 where they are NOT allowed) were getting very unwieldy; rewrote them as a
1817 negation of where the condition IS allowed.
1819 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
1821 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
1822 header file does not have a version number, so I've had to invent a new
1823 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
1824 API. The code is untested by me (my Linux distribution still has 0.3.2 of
1825 radiusclient), but it was contributed by a Radius user.
1827 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
1828 files or directories, for OpenSSL.
1830 PH/19 When an Exim process that is running as root has to create an Exim log
1831 file, it does so in a subprocess that runs as exim:exim so as to get the
1832 ownership right at creation (otherwise, other Exim processes might see
1833 the file with the wrong ownership). There was no test for failure of this
1834 fork() call, which would lead to the process getting stuck as it waited
1835 for a non-existent subprocess. Forks do occasionally fail when resources
1836 run out. I reviewed all the other calls to fork(); they all seem to check
1839 PH/20 When checking for unexpected SMTP input at connect time (before writing
1840 the banner), Exim was not dealing correctly with a non-positive return
1841 from the read() function. If the client had disconnected by this time,
1842 the result was a log entry for a synchronization error with an empty
1843 string after "input=" when read() returned zero. If read() returned -1
1844 (an event I could not check), uninitialized data bytes were printed.
1845 There were reports of junk text (parts of files, etc) appearing after
1848 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
1850 PH/22 Added support for macro redefinition, and (re)definition in between
1851 driver and ACL definitions.
1853 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
1854 forgetting to use the resulting value; it was using the unexpanded value.
1856 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
1857 hadn't been configured. The fix is from Juergen Kreileder, who
1858 understands it better than I do:
1860 "Here's what I see happening with three configured cyrus_sasl
1861 authenticators configured (plain, login, cram-md5):
1863 On startup auth_cyrus_sasl_init() gets called for each of these.
1864 This means three calls to sasl_listmech() without a specified mech_list.
1865 => SASL tests which mechs of all available mechs actually work
1866 => three warnings about OTP not working
1867 => the returned list contains: plain, login, cram-md5, digest-md5, ...
1869 With the patch, sasl_listmech() also gets called three times. But now
1870 SASL's mech_list option is set to the server_mech specified in the the
1871 authenticator. Or in other words, the answer from sasl_listmech()
1872 gets limited to just the mech you're testing for (which is different
1874 => the return list contains just 'plain' or 'login', 'cram-md5' or
1875 nothing depending on the value of ob->server_mech.
1877 I've just tested the patch: Authentication still works fine,
1878 unavailable mechs specified in the exim configuration are still
1879 caught, and the auth.log warnings about OTP are gone."
1881 PH/25 When debugging is enabled, the contents of the command line are added
1882 to the debugging output, even when log_selector=+arguments is not
1885 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
1886 answer is "GNU", and only if the return is "GNU/something" is the answer
1889 PH/27 $acl_verify_message is now set immediately after the failure of a
1890 verification in an ACL, and so is available in subsequent modifiers. In
1891 particular, the message can be preserved by coding like this:
1893 warn !verify = sender
1894 set acl_m0 = $acl_verify_message
1896 Previously, $acl_verify_message was set only while expanding "message"
1897 and "log_message" when a very denied access.
1899 PH/28 Modified OS/os.c-Linux with
1901 -#ifndef OS_LOAD_AVERAGE
1902 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
1904 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
1905 nomenclature these days.)
1907 PH/29 Installed patch from the Sieve maintainer that adds the options
1908 sieve_useraddress and sieve_subaddress to the redirect router.
1910 PH/30 In these circumstances:
1911 . Two addresses routed to the same list of hosts;
1912 . First host does not offer TLS;
1913 . First host accepts first address;
1914 . First host gives temporary error to second address;
1915 . Second host offers TLS and a TLS session is established;
1916 . Second host accepts second address.
1917 Exim incorrectly logged both deliveries with the TLS parameters (cipher
1918 and peerdn, if requested) that were in fact used only for the second
1921 PH/31 When doing a callout as part of verifying an address, Exim was not paying
1922 attention to any local part prefix or suffix that was matched by the
1923 router that accepted the address. It now behaves in the same way as it
1924 does for delivery: the affixes are removed from the local part unless
1925 rcpt_include_affixes is set on the transport.
1927 PH/32 Add the sender address, as F=<...>, to the log line when logging a
1928 timeout during the DATA phase of an incoming message.
1930 PH/33 Sieve envelope tests were broken for match types other than :is. I have
1931 applied a patch sanctioned by the Sieve maintainer.
1933 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
1934 the uid or gid is negative. A case of a negative gid caused this to be
1935 noticed. The fix allows for either to be negative.
1937 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
1938 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
1941 PH/36 The change PH/12 above was broken. Fixed it.
1943 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
1944 the grounds that routing the same address twice would always produce the
1945 same answer. This might have been true once, but it is certainly no
1946 longer true now. Routing a child address may depend on the previous
1947 routing that produced that child. Some complicated redirection strategies
1948 went wrong when messages had multiple recipients, and made Exim's
1949 behaviour dependent on the order in which the addresses were given.
1951 I have moved the duplicate checking until after the routing is complete.
1952 Exim scans the addresses that are assigned to local and remote
1953 transports, and removes any duplicates. This means that more work will be
1954 done, as duplicates will always all be routed, but duplicates are
1955 presumably rare, so I don't expect this is of any significance.
1957 For deliveries to pipes, files, and autoreplies, the duplicate checking
1958 still happens during the routing process, since they are not going to be
1961 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
1962 It corrects a timeout issue with spamd. This is Ian's comment: "The
1963 background is that sometimes spamd either never reads data from a
1964 connection it has accepted, or it never writes response data. The exiscan
1965 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
1966 blindly assumes that writes won't block so it may never time out."
1968 PH/39 Allow G after quota size as well as K and M.
1970 PH/40 The value set for $authenticated_id in an authenticator may not contain
1971 binary zeroes or newlines because the value is written to log lines and
1972 to spool files. There was no check on this. Now the value is run through
1973 the string_printing() function so that such characters are converted to
1974 printable escape sequences.
1976 PH/41 $message_linecount is a new variable that contains the total number of
1977 lines in the message. Compare $body_linecount, which is the count for the
1980 PH/42 Exim no longer gives details of delivery errors for specific addresses in
1981 bounce and delay warning messages, except in certain special cases, which
1984 (a) An SMTP error message from a remote host;
1985 (b) A message specified in a :fail: redirection;
1986 (c) A message specified in a "fail" command in a system filter;
1987 (d) A message specified in a FAIL return from the queryprogram router;
1988 (e) A message specified by the cannot_route_message router option.
1990 In these cases only, Exim does include the error details in bounce and
1991 warning messages. There are also a few cases where bland messages such
1992 as "unrouteable address" or "local delivery error" are given.
1994 PH/43 $value is now also set for the "else" part of a ${run expansion.
1996 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
1997 being worked on, but at least Exim now implements the latest version to
2000 PH/45 In a pipe transport, although a timeout while waiting for the pipe
2001 process to complete was treated as a delivery failure, a timeout while
2002 writing the message to the pipe was logged, but erroneously treated as a
2003 successful delivery. Such timeouts include transport filter timeouts. For
2004 consistency with the overall process timeout, these timeouts are now
2005 treated as errors, giving rise to delivery failures by default. However,
2006 there is now a new Boolean option for the pipe transport called
2007 timeout_defer, which, if set TRUE, converts the failures into defers for
2008 both kinds of timeout. A transport filter timeout is now identified in
2011 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
2012 systems where "make" and "gmake" are different, calling "gmake" at top
2013 level broke things. I've arranged for the value of $(MAKE) to be passed
2014 from the Makefile to this script so that it can call the same version of
2018 A note about Exim versions 4.44 and 4.50
2019 ----------------------------------------
2021 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
2022 changes of various kinds. As a consequence, a big documentation update was
2023 needed. This delayed the release for rather longer than seemed good, especially
2024 in the light of a couple of (minor) security issues. Therefore, the changes
2025 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
2026 release. So 4.44 and 4.50 are in effect two different branches that both start
2029 I have left the 4.50 change log unchanged; it contains all the changes since
2030 4.43. The change log for 4.44 is below; many of its items are identical to
2031 those for 4.50. This seems to be the most sensible way to preserve the
2032 historical information.
2038 1. Minor wording change to the doc/README.SIEVE file.
2040 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
2041 computation of the current number of files was incorrect.
2043 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
2044 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
2047 4. Give more explanation in the error message when the command for a transport
2048 filter fails to execute.
2050 5. There are several places where Exim runs a non-Exim command in a
2051 subprocess. The SIGUSR1 signal should be disabled for these processes. This
2052 was being done only for the command run by the queryprogram router. It is
2053 now done for all such subprocesses. The other cases are: ${run, transport
2054 filters, and the commands run by the lmtp and pipe transports.
2056 6. Added CONFIGURE_GROUP build-time option.
2058 7. Some older OS have a limit of 256 on the maximum number of file
2059 descriptors. Exim was using setrlimit() to set 1000 as a large value
2060 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
2061 systems. I've change it so that if it can't get 1000, it tries for 256.
2063 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
2064 was an oversight, and furthermore, ever since the addition of extra
2065 controls (e.g. 4.43/32), the checks on when to allow different forms of
2066 "control" were broken. There should now be diagnostics for all cases when a
2067 control that does not make sense is encountered.
2069 9. Added the /retain_sender option to "control=submission".
2071 10. $recipients is now available in the predata ACL (oversight).
2073 11. Tidy the search cache before the fork to do a delivery from a message
2074 received from the command line. Otherwise the child will trigger a lookup
2075 failure and thereby defer the delivery if it tries to use (for example) a
2076 cached ldap connection that the parent has called unbind on.
2078 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
2079 of $address_data from the recipient verification was clobbered by the
2080 sender verification.
2082 13. The value of address_data from a sender verification is now available in
2083 $sender_address_data in subsequent conditions in the ACL statement.
2085 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
2087 15. Added a new option "connect=<time>" to callout options, to set a different
2090 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
2091 was its contents. (It was OK if the option was not defined at all.)
2093 17. A "Completed" log line is now written for messages that are removed from
2094 the spool by the -Mrm option.
2096 18. New variables $sender_verify_failure and $recipient_verify_failure contain
2097 information about exactly what failed.
2099 19. Added -dd to debug only the daemon process.
2101 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
2102 handles timeouts, both on the server side and network timeouts. Renamed the
2103 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
2105 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
2106 It is now set to "smtps".
2108 22. $host_address is now set to the target address during the checking of
2109 ignore_target_hosts.
2111 23. When checking ignore_target_hosts for an ipliteral router, no host name was
2112 being passed; this would have caused $sender_host_name to have been used if
2113 matching the list had actually called for a host name (not very likely,
2114 since this list is usually IP addresses). A host name is now passed as
2117 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
2118 code that specifies a non-restarting handler (typically sigaction() in
2119 modern systems) in an attempt to fix a rare and obscure crash bug.
2121 25. Narrowed the window for a race in the daemon that could cause it to ignore
2122 SIGCHLD signals. This is not a major problem, because they are used only to
2123 wake it up if nothing else does.
2125 26. A malformed maildirsize file could cause Exim to calculate negative values
2126 for the mailbox size or file count. Odd effects could occur as a result.
2127 The maildirsize information is now recalculated if the size or filecount
2130 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
2131 support for a long time. Removed HAVE_SYS_VFS_H.
2133 28. Installed the latest version of exipick from John Jetmore.
2135 29. In an address list, if the pattern was not a regular expression, an empty
2136 subject address (from a bounce message) matched only if the pattern was an
2137 empty string. Non-empty patterns were not even tested. This was the wrong
2138 because it is perfectly reasonable to use an empty address as part of a
2139 database query. An empty address is now tested by patterns that are
2140 lookups. However, all the other forms of pattern expect the subject to
2141 contain a local part and a domain, and therefore, for them, an empty
2142 address still always fails if the pattern is not itself empty.
2144 30. Exim went into a mad DNS loop when attempting to do a callout where the
2145 host was specified on an smtp transport, and looking it up yielded more
2146 than one IP address.
2148 31. Re-factored the code for checking spool and log partition space into a
2149 function that finds that data and another that does the check. The former
2150 is then used to implement four new variables: $spool_space, $log_space,
2151 $spool_inodes, and $log_inodes.
2153 32. The RFC2047 encoding function was originally intended for short strings
2154 such as real names; it was not keeping to the 75-character limit for
2155 encoded words that the RFC imposes. It now respects the limit, and
2156 generates multiple encoded words if necessary. To be on the safe side, I
2157 have increased the buffer size for the ${rfc2047: expansion operator from
2160 33. It is now permitted to omit both strings after an "if" condition; if the
2161 condition is true, the result is "true". As before, when the second string
2162 is omitted, a false condition yields an empty string. This makes it less
2163 cumbersome to write custom ACL and router conditions.
2165 34. Failure to deliver a bounce message always caused it to be frozen, even if
2166 there was an errors_to setting on the router. The errors_to setting is now
2169 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
2170 canonical form (fully expanded) before being placed in
2171 $sender_host_address.
2173 36. The table in the code that translates DNS record types into text (T_A to
2174 "A" for instance) was missing entries for NS and CNAME. It is just possible
2175 that this could have caused confusion if both these types were looked up
2176 for the same domain, because the text type is used as part of Exim's
2177 per-process caching. But the chance of anyone hitting this buglet seems
2180 37. The dnsdb lookup has been extended in a number of ways.
2182 (1) There is a new type, "zns", which walks up the domain tree until it
2183 finds some nameserver records. It should be used with care.
2185 (2) There is a new type, "mxh", which is like "mx" except that it returns
2186 just the host names, not the priorities.
2188 (3) It is now possible to give a list of domains (or IP addresses) to be
2189 looked up. The behaviour when one of the lookups defers can be
2190 controlled by a keyword.
2192 (4) It is now possible to specify the separator character for use when
2193 multiple records are returned.
2195 38. The dnslists ACL condition has been extended: it is now possible to supply
2196 a list of IP addresses and/or domains to be looked up in a particular DNS
2199 39. Added log_selector=+queue_time_overall.
2201 40. When running the queue in the test harness, wait just a tad after forking a
2202 delivery process, to get repeatability of debugging output.
2204 41. Include certificate and key file names in error message when GnuTLS fails
2205 to set them up, because the GnuTLS error message doesn't include the name
2206 of the failing file when there is a problem reading it.
2208 42. Allow both -bf and -bF in the same test run.
2210 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
2212 44. The "Exiscan patch" is now merged into the mainline Exim source.
2214 45. Sometimes the final signoff response after QUIT could fail to get
2215 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
2216 < 0 before doing a fflush(). This bug looks as though it goes back to the
2217 introduction of TLS in release 3.20, but "sometimes" must have been rare
2218 because the tests only now provoked it.
2220 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
2221 (this can affect the format of dates).
2223 47. exim_tidydb, when checking for the continued existence of a message for
2224 which it has found a message-specific retry record, was not finding
2225 messages that were in split spool directories. Consequently, it was
2226 deleting retry records that should have stayed in existence.
2228 48. Steve fixed some bugs in eximstats.
2230 49. The SPA authentication driver was not abandoning authentication and moving
2231 on to the next authenticator when an expansion was forced to fail,
2232 contradicting the general specification for all authenticators. Instead it
2233 was generating a temporary error. It now behaves as specified.
2235 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
2236 (the order specifies the preference for clients). The order is now AES256,
2237 AES128, 3DES, ARCFOUR128.
2239 51. Small patch to Sieve code - explicitly set From: when generating an
2242 52. Exim crashed if a remote delivery caused a very long error message to be
2243 recorded - for instance if somebody sent an entire SpamAssassin report back
2244 as a large number of 550 error lines. This bug was coincidentally fixed by
2245 increasing the size of one of Exim's internal buffers (big_buffer) that
2246 happened as part of the Exiscan merge. However, to be on the safe side, I
2247 have made the code more robust (and fixed the comments that describe what
2250 53. Now that there can be additional text after "Completed" in log lines (if
2251 the queue_time_overall log selector is set), a one-byte patch to exigrep
2252 was needed to allow it to recognize "Completed" as not the last thing in
2255 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
2256 patch that reportedly fixes this has been added. I am not expert enough to
2257 create a test for it. This is what the patch creator wrote:
2259 "I found a little strange behaviour of ldap code when working with
2260 Windows 2003 AD Domain, where users was placed in more than one
2261 Organization Units. When I tried to give exim partial DN, the exit code
2262 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
2263 But simultaneously result of request was absolutely normal ldap result,
2264 so I produce this patch..."
2266 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
2267 REFERENCE, so I have modified the code to exclude the patch when that macro
2270 55. Some experimental protocols are using DNS PTR records for new purposes. The
2271 keys for these records are domain names, not reversed IP addresses. The
2272 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
2273 leaves it alone. Component reversal etc. now happens only for IP addresses.
2276 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
2278 57. Double the size of the debug message buffer (to 2048) so that more of very
2279 long debug lines gets shown.
2281 58. The exicyclog utility now does better if the number of log files to keep
2282 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
2284 59. Two changes related to the smtp_active_hostname option:
2286 (1) $smtp_active_hostname is now available as a variable.
2287 (2) The default for smtp_banner uses $smtp_active_hostname instead
2288 of $primary_hostname.
2290 60. The host_aton() function is supposed to be passed a string that is known
2291 to be a valid IP address. However, in the case of IPv6 addresses, it was
2292 not checking this. This is a hostage to fortune. Exim now panics and dies
2293 if the condition is not met. A case was found where this could be provoked
2294 from a dnsdb PTR lookup with an IPv6 address that had more than 8
2295 components; fortuitously, this particular loophole had already been fixed
2296 by change 4.50/55 above.
2298 If there are any other similar loopholes, the new check in host_aton()
2299 itself should stop them being exploited. The report I received stated that
2300 data on the command line could provoke the exploit when Exim was running as
2301 exim, but did not say which command line option was involved. All I could
2302 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
2303 running as the user.
2306 61. There was a buffer overflow vulnerability in the SPA authentication code
2307 (which came originally from the Samba project). I have added a test to the
2308 spa_base64_to_bits() function which I hope fixes it.
2311 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
2312 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
2314 63. The daemon start-up calls getloadavg() while still root for those OS that
2315 need the first call to be done as root, but it missed one case: when
2316 deliver_queue_load_max is set with deliver_drop_privilege. This is
2317 necessary for the benefit of the queue runner, because there is no re-exec
2318 when deliver_drop_privilege is set.
2320 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
2321 This has been fixed.
2323 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
2324 was in use, was not putting the data itself into the right store pool;
2325 consequently, it could be overwritten for a subsequent message in the same
2326 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
2329 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
2331 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
2332 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
2333 the function were treating the return as a boolean value, which happened to
2334 work because 0=false and not-0=true, but is not correct code.
2336 68. The host_aton() function was not handling scoped IPv6 addresses (those
2337 with, for example, "%eth0" on the end) correctly.
2339 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
2340 negated items (that is, ~something) in unsigned ints. Some compilers
2341 apparently mutter when there is no cast.
2343 70. If an address verification called from an ACL failed, and did not produce a
2344 user-specific message (i.e. there was only a "system" message), nothing was
2345 put in $acl_verify_message. In this situation, it now puts the system
2348 71. Change 4.23/11 added synchronization checking at the start of an SMTP
2349 session; change 4.31/43 added the unwanted input to the log line - except
2350 that it did not do this in the start of session case. It now does.
2352 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
2353 This is wrong and can cause the other end to generate a synchronization
2354 error if it is another Exim or anything else that does the synchronization
2355 check. A QUIT command is no longer sent after a timeout.
2357 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
2358 during host lookups.
2360 74. The defer_ok option of callout verification was not working if it was used
2361 when verifying addresses in header lines, that is, for this case:
2363 verify = header_sender/callout=defer_ok
2365 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
2366 those file descriptors could be used for SMTP connections. If anything
2367 wrote to stderr (the example that came up was "warn" in embedded Perl), it
2368 could be sent to the SMTP client, causing chaos. The daemon now opens
2369 stdin, stdout, and stderr to /dev/null when it puts itself into the
2372 76. Arrange for output from Perl's "warn" command to be written to Exim's main
2373 log by default. The user can override this with suitable Perl magic.
2375 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
2376 the log message when discard triggers, was not working for the DATA ACL or
2377 for the non-SMTP ACL.
2379 78. Error message wording change in sieve.c.
2381 79. If smtp_accept_max_per_host was set, the number of connections could be
2382 restricted to fewer than expected, because the daemon was trying to set up
2383 a new connection before checking whether the processes handling previous
2384 connections had finished. The check for completed processes is now done
2385 earlier. On busy systems, this bug wouldn't be noticed because something
2386 else would have woken the daemon, and it would have reaped the completed
2389 80. If a message was submitted locally by a user whose login name contained one
2390 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
2391 It caused a spool format error. I have fixed the spool reading code. A
2392 related problem was that the "from" clause in the Received: line became
2393 illegal because of the space(s). It is now covered by ${quote_local_part.
2395 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
2398 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
2400 83. Patch from Sieve maintainer to fix unterminated string problem in
2401 "vacation" handling.
2403 84. Some minor changes to the Linux configuration files to help with other
2404 OS variants using glibc.
2406 85. One more patch for Sieve to update vacation handling to latest spec.
2409 ----------------------------------------------------
2410 See the note above about the 4.44 and 4.50 releases.
2411 ----------------------------------------------------
2417 1. Change 4.43/35 introduced a bug that caused file counts to be
2418 incorrectly computed when quota_filecount was set in an appendfile
2421 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
2422 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
2425 3. Give more explanation in the error message when the command for a transport
2426 filter fails to execute.
2428 4. There are several places where Exim runs a non-Exim command in a
2429 subprocess. The SIGUSR1 signal should be disabled for these processes. This
2430 was being done only for the command run by the queryprogram router. It is
2431 now done for all such subprocesses. The other cases are: ${run, transport
2432 filters, and the commands run by the lmtp and pipe transports.
2434 5. Some older OS have a limit of 256 on the maximum number of file
2435 descriptors. Exim was using setrlimit() to set 1000 as a large value
2436 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
2437 systems. I've change it so that if it can't get 1000, it tries for 256.
2439 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
2440 was an oversight, and furthermore, ever since the addition of extra
2441 controls (e.g. 4.43/32), the checks on when to allow different forms of
2442 "control" were broken. There should now be diagnostics for all cases when a
2443 control that does not make sense is encountered.
2445 7. $recipients is now available in the predata ACL (oversight).
2447 8. Tidy the search cache before the fork to do a delivery from a message
2448 received from the command line. Otherwise the child will trigger a lookup
2449 failure and thereby defer the delivery if it tries to use (for example) a
2450 cached ldap connection that the parent has called unbind on.
2452 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
2453 of $address_data from the recipient verification was clobbered by the
2454 sender verification.
2456 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
2457 was its contents. (It was OK if the option was not defined at all.)
2459 11. A "Completed" log line is now written for messages that are removed from
2460 the spool by the -Mrm option.
2462 12. $host_address is now set to the target address during the checking of
2463 ignore_target_hosts.
2465 13. When checking ignore_target_hosts for an ipliteral router, no host name was
2466 being passed; this would have caused $sender_host_name to have been used if
2467 matching the list had actually called for a host name (not very likely,
2468 since this list is usually IP addresses). A host name is now passed as
2471 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
2472 code that specifies a non-restarting handler (typically sigaction() in
2473 modern systems) in an attempt to fix a rare and obscure crash bug.
2475 15. Narrowed the window for a race in the daemon that could cause it to ignore
2476 SIGCHLD signals. This is not a major problem, because they are used only to
2477 wake it up if nothing else does.
2479 16. A malformed maildirsize file could cause Exim to calculate negative values
2480 for the mailbox size or file count. Odd effects could occur as a result.
2481 The maildirsize information is now recalculated if the size or filecount
2484 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
2485 support for a long time. Removed HAVE_SYS_VFS_H.
2487 18. Updated exipick to current release from John Jetmore.
2489 19. Allow an empty sender to be matched against a lookup in an address list.
2490 Previously the only cases considered were a regular expression, or an
2493 20. Exim went into a mad DNS lookup loop when doing a callout where the
2494 host was specified on the transport, if the DNS lookup yielded more than
2497 21. The RFC2047 encoding function was originally intended for short strings
2498 such as real names; it was not keeping to the 75-character limit for
2499 encoded words that the RFC imposes. It now respects the limit, and
2500 generates multiple encoded words if necessary. To be on the safe side, I
2501 have increased the buffer size for the ${rfc2047: expansion operator from
2504 22. Failure to deliver a bounce message always caused it to be frozen, even if
2505 there was an errors_to setting on the router. The errors_to setting is now
2508 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
2509 canonical form (fully expanded) before being placed in
2510 $sender_host_address.
2512 24. Updated eximstats to version 1.33
2514 25. Include certificate and key file names in error message when GnuTLS fails
2515 to set them up, because the GnuTLS error message doesn't include the name
2516 of the failing file when there is a problem reading it.
2518 26. Expand error message when OpenSSL has problems setting up cert/key files.
2521 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
2522 (this can affect the format of dates).
2524 28. exim_tidydb, when checking for the continued existence of a message for
2525 which it has found a message-specific retry record, was not finding
2526 messages that were in split spool directories. Consequently, it was
2527 deleting retry records that should have stayed in existence.
2529 29. eximstats updated to version 1.35
2530 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
2531 1.35 - bugfix such that pie charts by volume are generated correctly
2533 30. The SPA authentication driver was not abandoning authentication and moving
2534 on to the next authenticator when an expansion was forced to fail,
2535 contradicting the general specification for all authenticators. Instead it
2536 was generating a temporary error. It now behaves as specified.
2538 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
2539 (the order specifies the preference for clients). The order is now AES256,
2540 AES128, 3DES, ARCFOUR128.
2542 31. Small patch to Sieve code - explicitly set From: when generating an
2545 32. Exim crashed if a remote delivery caused a very long error message to be
2546 recorded - for instance if somebody sent an entire SpamAssassin report back
2547 as a large number of 550 error lines. This bug was coincidentally fixed by
2548 increasing the size of one of Exim's internal buffers (big_buffer) that
2549 happened as part of the Exiscan merge. However, to be on the safe side, I
2550 have made the code more robust (and fixed the comments that describe what
2553 33. Some experimental protocols are using DNS PTR records for new purposes. The
2554 keys for these records are domain names, not reversed IP addresses. The
2555 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
2556 leaves it alone. Component reversal etc. now happens only for IP addresses.
2559 34. The host_aton() function is supposed to be passed a string that is known
2560 to be a valid IP address. However, in the case of IPv6 addresses, it was
2561 not checking this. This is a hostage to fortune. Exim now panics and dies
2562 if the condition is not met. A case was found where this could be provoked
2563 from a dnsdb PTR lookup with an IPv6 address that had more than 8
2564 components; fortuitously, this particular loophole had already been fixed
2565 by change 4.50/55 or 4.44/33 above.
2567 If there are any other similar loopholes, the new check in host_aton()
2568 itself should stop them being exploited. The report I received stated that
2569 data on the command line could provoke the exploit when Exim was running as
2570 exim, but did not say which command line option was involved. All I could
2571 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
2572 running as the user.
2575 35. There was a buffer overflow vulnerability in the SPA authentication code
2576 (which came originally from the Samba project). I have added a test to the
2577 spa_base64_to_bits() function which I hope fixes it.
2580 36. The daemon start-up calls getloadavg() while still root for those OS that
2581 need the first call to be done as root, but it missed one case: when
2582 deliver_queue_load_max is set with deliver_drop_privilege. This is
2583 necessary for the benefit of the queue runner, because there is no re-exec
2584 when deliver_drop_privilege is set.
2586 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
2587 was in use, was not putting the data itself into the right store pool;
2588 consequently, it could be overwritten for a subsequent message in the same
2589 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
2592 38. Sometimes the final signoff response after QUIT could fail to get
2593 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
2594 < 0 before doing a fflush(). This bug looks as though it goes back to the
2595 introduction of TLS in release 3.20, but "sometimes" must have been rare
2596 because the tests only now provoked it.
2602 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
2603 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
2604 Now it returns an integer. A number of calls were still expecting a T/F
2605 return. Fortuitously, in all cases, the tests worked in OK situations,
2606 which is the norm. However, things would have gone wrong on any write
2607 failures on the smtp file descriptor. This function is used when sending
2608 messages over SMTP and also when doing verify callouts.
2610 2. When Exim is called to do synchronous delivery of a locally submitted
2611 message (the -odf or -odi options), it no longer closes stderr before doing
2614 3. Implemented the mua_wrapper option.
2616 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
2618 5. Implemented the functions header_remove(), header_testname(),
2619 header_add_at_position(), and receive_remove_recipient(), and exported them
2622 6. If an ACL "warn" statement specified the addition of headers, Exim already
2623 inserted X-ACL-Warn: at the start if there was no header name. However, it
2624 was not making this test for the second and subsequent header lines if
2625 there were newlines in the string. This meant that an invalid header could
2626 be inserted if Exim was badly configured.
2628 7. Allow an ACL "warn" statement to add header lines at the start or after all
2629 the Received: headers, as well as at the end.
2631 8. Added the rcpt_4xx retry error code.
2633 9. Added postmaster_mailfrom=xxx to callout verification option.
2635 10. Added mailfrom=xxxx to the callout verification option, for verify=
2638 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
2639 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
2641 12. Inserted some casts to stop certain compilers warning when using pointer
2642 differences as field lengths or precisions in printf-type calls (mostly
2643 affecting debugging statements).
2645 13. Added optional readline() support for -be (dynamically loaded).
2647 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
2648 same clock tick as a message's arrival, so that its received time was the
2649 same as the "first fail" time on the retry record, and that message
2650 remained on the queue past the ultimate address timeout, every queue runner
2651 would try a delivery (because it was past the ultimate address timeout) but
2652 after another failure, the ultimate address timeout, which should have then
2653 bounced the address, did not kick in. This was a "< instead of <=" error;
2654 in most cases the first failure would have been in the next clock tick
2655 after the received time, and all would be well.
2657 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
2658 being recognized when the domain list was tested by the match_domain
2659 condition in an expansion string.
2661 16. Added the ${str2b64: operator.
2663 17. Exim was always calling setrlimit() to set a large limit for the number of
2664 processes, without checking whether the existing limit was already
2665 adequate. (It did check for the limit on file descriptors.) Furthermore,
2666 errors from getrlimit() and setrlimit() were being ignored. Now they are
2667 logged to the main and panic logs, but Exim does carry on, to try to do its
2668 job under whatever limits there are.
2670 18. Imported PCRE 5.0.
2672 19. Trivial typo in log message " temporarily refused connection" (the leading
2675 20. If the log selector return_path_on_delivery was set and an address was
2676 redirected to /dev/null, the delivery process crashed because it assumed
2677 that a return path would always be set for a "successful" delivery. In this
2678 case, the whole delivery is bypassed as an optimization, and therefore no
2681 21. Internal re-arrangement: the function for sending a challenge and reading
2682 a response while authentication was assuming a zero-terminated challenge
2683 string. It's now changed to take a pointer and a length, to allow for
2684 binary data in such strings.
2686 22. Added the cyrus_sasl authenticator (code supplied by MBM).
2688 23. Exim was not respecting finduser_retries when seeking the login of the
2689 uid under which it was called; it was always trying 10 times. (The default
2690 setting of finduser_retries is zero.) Also, it was sleeping after the final
2691 failure, which is pointless.
2693 24. Implemented tls_on_connect_ports.
2695 25. Implemented acl_smtp_predata.
2697 26. If the domain in control=submission is set empty, Exim assumes that the
2698 authenticated id is a complete email address when it generates From: or
2699 Sender: header lines.
2701 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
2702 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
2703 chown and chgrp in /bin and hostname in /usr/bin.
2705 28. Exim was keeping the "process log" file open after each use, just as it
2706 does for the main log. This opens the possibility of it remaining open for
2707 long periods when the USR1 signal hits a daemon. Occasional processlog
2708 errors were reported, that could have been caused by this. Anyway, it seems
2709 much more sensible not to leave this file open at all, so that is what now
2712 29. The long-running daemon process does not normally write to the log once it
2713 has entered its main loop, and it closes the log before doing so. This is
2714 so that log files can straightforwardly be renamed and moved. However,
2715 there are a couple of unusual error situations where the daemon does write
2716 log entries, and I had neglected to close the log afterwards.
2718 30. The text of an SMTP error response that was received during a remote
2719 delivery was being truncated at 512 bytes. This is too short for some of
2720 the long messages that one sometimes sees. I've increased the limit to
2723 31. It is now possible to make retry rules that apply only when a message has a
2724 specific sender, in particular, an empty sender.
2726 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
2727 it possible to be selective about when SMTP synchronization is enforced.
2729 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
2731 32. Implemented hosts_connection_nolog.
2733 33. Added an ACL for QUIT.
2735 34. Setting "delay_warning=" to disable warnings was not working; it gave a
2738 35. Added mailbox_size and mailbox_filecount to appendfile.
2740 36. Added control = no_multiline_responses to ACLs.
2742 37. There was a bug in the logic of the code that waits for the clock to tick
2743 in the case where the clock went backwards by a substantial amount such
2744 that the microsecond fraction of "now" was more than the microsecond
2745 fraction of "then" (but the whole seconds number was less).
2747 38. Added support for the libradius Radius client library this is found on
2748 FreeBSD (previously only the radiusclient library was supported).
2754 1. When certain lookups returned multiple values in the form name=value, the
2755 quoting of the values was not always being done properly. Specifically:
2756 (a) If the value started with a double quote, but contained no whitespace,
2758 (b) If the value contained whitespace other than a space character (i.e.
2759 tabs or newlines or carriage returns) it was not quoted.
2760 This fix has been applied to the mysql and pgsql lookups by writing a
2761 separate quoting function and calling it from the lookup code. The fix
2762 should probably also be applied to nisplus, ibase and oracle lookups, but
2763 since I cannot test any of those, I have not disturbed their existing code.
2765 2. A hit in the callout cache for a specific address caused a log line with no
2766 reason for rejecting RCPT. Now it says "Previous (cached) callout
2767 verification failure".
2769 3. There was an off-by-one bug in the queryprogram router. An over-long
2770 return line was truncated at 256 instead of 255 characters, thereby
2771 overflowing its buffer with the terminating zero. As well as fixing this, I
2772 have increased the buffer size to 1024 (and made a note to document this).
2774 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
2775 when Exim is waiting for an SMTP response from a remote server, Exim
2776 restarts its select() call on the socket, thereby resetting its timeout.
2777 This is not a problem when such interrupts are rare. Somebody set up a cron
2778 job to run exiwhat every 2 minutes, which is less than the normal select()
2779 timeout (5 or 10 minutes). This meant that the select() timeout never
2780 kicked in because it was always reset. I have fixed this by comparing the
2781 time when an interrupt arrives with the time at the start of the first call
2782 to select(). If more time than the timeout has elapsed, the interrupt is
2783 treated as a timeout.
2785 5. Some internal re-factoring in preparation for the addition of Sieve
2786 extensions (by MH). In particular, the "personal" test is moved to a
2787 separate function, and given an option for scanning Cc: and Bcc: (which is
2788 not set for Exim filters).
2790 6. When Exim created an email address using the login of the caller as the
2791 local part (e.g. when creating a From: or Sender: header line), it was not
2792 quoting the local part when it contained special characters such as @.
2794 7. Installed new OpenBSD configuration files.
2796 8. Reworded some messages for syntax errors in "and" and "or" conditions to
2797 try to make them clearer.
2799 9. Callout options, other than the timeout value, were being ignored when
2800 verifying sender addresses in header lines. For example, when using
2802 verify = header_sender/callout=no_cache
2804 the cache was (incorrectly) being used.
2806 10. Added a missing instance of ${EXE} to the exim_install script; this affects
2807 only the Cygwin environment.
2809 11. When return_path_on_delivery was set as a log selector, if different remote
2810 addresses in the same message used different return paths and parallel
2811 remote delivery occurred, the wrong values would sometimes be logged.
2812 (Whenever a remote delivery process finished, the return path value from
2813 the most recently started remote delivery process was logged.)
2815 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
2816 lines when AUTH and/or TLS are in use. This is the "received protocol"
2817 field. Exim used to use "asmtp" for authenticated SMTP, without any
2818 indication (in the protocol name) for TLS use. Now it follows the RFC and
2819 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
2820 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
2821 names appear in log lines as well as in Received: header lines.
2823 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
2824 extensions, and comparison tests, and to fix some bugs.
2826 14. Changes to the "personal" filter test:
2828 (1) The test was buggy in that it was just doing the equivalent of
2829 "contains" tests on header lines. For example, if a user's address was
2830 anne@some.where, the "personal" test would incorrectly be true for
2832 To: susanne@some.where
2834 This test is now done by extracting each address from the header in turn,
2835 and checking the entire address. Other tests that are part of "personal"
2836 are now done using regular expressions (for example, to check local parts
2837 of addresses in From: header lines).
2839 (2) The list of non-personal local parts in From: addresses has been
2840 extended to include "listserv", "majordomo", "*-request", and "owner-*",
2841 taken from the Sieve specification recommendations.
2843 (3) If the message contains any header line starting with "List-" it is
2844 treated as non-personal.
2846 (4) The test for "circular" in the Subject: header line has been removed
2847 because it now seems ill-conceived.
2849 15. Minor typos in src/EDITME comments corrected.
2851 16. Installed latest exipick from John Jetmore.
2853 17. If headers_add on a router specified a text string that was too long for
2854 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
2855 of string_sprintf() is now avoided.
2857 18. $message_body_size was not set (it was always zero) when running the DATA
2858 ACL and the local_scan() function.
2860 19. For the "mail" command in an Exim filter, no default was being set for
2861 the once_repeat time, causing a random time value to be used if "once" was
2862 specified. (If the value happened to be <= 0, no repeat happened.) The
2863 default is now 0s, meaning "never repeat". The "vacation" command was OK
2864 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
2865 (I found it when inspecting the code).
2867 20. There is now an overall timeout for performing a callout verification. It
2868 defaults to 4 times the callout timeout, which applies to individual SMTP
2869 commands during the callout. The overall timeout applies when there is more
2870 than one host that can be tried. The timeout is checked before trying the
2871 next host. This prevents very long delays if there are a large number of
2872 hosts and all are timing out (e.g. when the network connections are timing
2873 out). The value of the overall timeout can be changed by specifying an
2874 additional sub-option for "callout", called "maxwait". For example:
2876 verify = sender/callout=5s,maxwait=20s
2878 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
2879 to the end before writing, but this should make it even safer).
2881 22. Exim was forgetting that it had advertised PIPELINING for the second and
2882 subsequent messages on an SMTP connection. It was also not resetting its
2883 memory on STARTTLS and an internal HELO.
2885 23. When Exim logs an SMTP synchronization error within a session, it now
2886 records whether PIPELINING has been advertised or not.
2888 24. Added 3 instances of "(long int)" casts to time_t variables that were being
2889 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
2890 rather than long int.
2892 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
2894 26. Added the never_mail option to autoreply.
2900 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
2901 crash if the getsockname() call failed; this can happen if a connection is
2902 closed very soon after it is established. The problem was simply in the
2903 order in which certain operations were done, causing Exim to try to write
2904 to the SMTP stream before it had set up the file descriptor. The bug has
2905 been fixed by making things happen in the correct order.
2911 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
2912 before the connection was closed, thus losing the rejection response.
2914 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
2915 some early Solaris releases, but causes trouble in current releases where
2916 socklen_t is defined.
2918 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
2921 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
2924 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
2925 that releasing the top of it at the end releases what was used for sub-
2926 expansions (unless the block got too big). However, discard this block if
2927 the first thing is a variable or header, so that we can use its block when
2928 it is dynamic (useful for very large $message_headers, for example).
2930 6. Lookups now cache *every* query, not just the most recent. A new, separate
2931 store pool is used for this. It can be recovered when all lookup caches are
2932 flushed. Lookups now release memory at the end of their result strings.
2933 This has involved some general refactoring of the lookup sources.
2935 7. Some code has been added to the store_xxx() functions to reduce the amount
2936 of flapping under certain conditions.
2938 8. log_incoming_interface used to affect only the <= reception log lines. Now
2939 it causes the local interface and port to be added to several more SMTP log
2940 lines, for example "SMTP connection from", and rejection lines.
2942 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
2944 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
2946 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
2947 could be overwritten at the end of the current message (or the start of a
2948 new message if it was set in a HELO ACL). The value is now preserved for
2949 the duration of the SMTP connection.
2951 12. If a transport had a headers_rewrite setting, and a matching header line
2952 contained an unqualified address, that address was qualified, even if it
2953 did not match any rewriting rules. The underlying bug was that the values
2954 of the flags that permit the existence of unqualified sender and recipient
2955 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
2956 non-local messages, and by -bnq for local messages) were not being
2957 preserved with the message after it was received.
2959 13. When Exim was logging an SMTP synchronization error, it could sometimes log
2960 "next input=" as part of the text comprising the host identity instead of
2961 the correct text. The code was using the same buffer for two different
2962 strings. However, depending on which order the printing function evaluated
2963 its arguments, the bug did not always show up. Under Linux, for example, my
2964 test suite worked just fine.
2966 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
2967 doesn't work with some older versions of Perl. It has been changed to "my",
2968 which in any case is probably the better facility to use.
2970 15. A really picky compiler found some instances of statements for creating
2971 error messages that either had too many or two few arguments for the format
2974 16. The size of the buffer for calls to the DNS resolver has been increased
2975 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
2976 for addresses that have a lot of PTR records. This alleviates a problem; it
2977 does not fully solve it.
2979 17. A dnsdb lookup for PTR records that receives more data than will fit in the
2980 buffer now truncates the list and logs the incident, which is the same
2981 action as happens when Exim is looking up a host name and its aliases.
2982 Previously in this situation something unpredictable would happen;
2983 sometimes it was "internal error: store_reset failed".
2985 18. If a server dropped the connection unexpectedly when an Exim client was
2986 using GnuTLS and trying to read a response, the client delivery process
2987 crashed while trying to generate an error log message.
2989 19. If a "warn" verb in an ACL added multiple headers to a message in a single
2990 string, for example:
2992 warn message = H1: something\nH2: something
2994 the text was added as a single header line from Exim's point of view
2995 though it ended up OK in the delivered message. However, searching for the
2996 second and subsequent header lines using $h_h2: did not work. This has been
2997 fixed. Similarly, if a system filter added multiple headers in this way,
2998 the routers could not see them.
3000 20. Expanded the error message when iplsearch is called with an invalid key to
3001 suggest using net-iplsearch in a host list.
3003 21. When running tests using -bh, any delays imposed by "delay" modifiers in
3004 ACLs are no longer actually imposed (and a message to that effect is
3007 22. If a "gecos" field in a passwd entry contained escaped characters, in
3008 particular, if it contained a \" sequence, Exim got it wrong when building
3009 a From: or a Sender: header from that name. A second bug also caused
3010 incorrect handling when an unquoted " was present following a character
3011 that needed quoting.
3013 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
3014 was not being matched caselessly.
3016 24. Arranged for all hyphens in the exim.8 source to be escaped with
3019 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
3020 properly. Recipient callout cache records were still being keyed to include
3021 the sender, even when use_sender was set false. This led to far more
3022 callouts that were necessary. The sender is no longer included in the key
3023 when use_sender is false.
3025 26. Added "control = submission" modifier to ACLs.
3027 27. Added the ${base62d: operator to decode base 62 numbers.
3029 28. dnsdb lookups can now access SRV records.
3031 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
3032 the configuration file.
3034 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
3035 (-v) mode. This makes the output for a verbose queue run more intelligible.
3037 31. Added a use_postmaster feature to recipient callouts.
3039 32. Added the $body_zerocount variable, containing the number of binary zero
3040 bytes in the message body.
3042 33. The time of last modification of the "new" subdirectory is now used as the
3043 "mailbox time last read" when there is a quota error for a maildir
3046 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
3048 35. Added +ignore_unknown as a special item in host lists.
3050 36. Code for decoding IPv6 addresses in host lists is now included, even if
3051 IPv6 support is not being compiled. This fixes a bug in which an IPv6
3052 address was recognized as an IP address, but was then not correctly decoded
3053 into binary, causing unexpected and incorrect effects when compared with
3060 1. Very minor rewording of debugging text in manualroute to say "list of
3061 hosts" instead of "hostlist".
3063 2. If verify=header_syntax was set, and a header line with an unqualified
3064 address (no domain) and a large number of spaces between the end of the
3065 name and the colon was received, the reception process suffered a buffer
3066 overflow, and (when I tested it) crashed. This was caused by some obsolete
3067 code that should have been removed. The fix is to remove it!
3069 3. When running in the test harness, delay a bit after writing a bounce
3070 message to get a bit more predictability in the log output.
3072 4. Added a call to search_tidyup() just before forking a reception process. In
3073 theory, someone could use a lookup in the expansion of smtp_accept_max_
3074 per_host which, without the tidyup, could leave open a database connection.
3076 5. Added the variables $recipient_data and $sender_data which get set from a
3077 lookup success in an ACL "recipients" or "senders" condition, or a router
3078 "senders" option, similar to $domain_data and $local_part_data.
3080 6. Moved the writing of debug_print from before to after the "senders" test
3083 7. Change 4.31/66 (moving the time when the Received: is generated) caused
3084 problems for message scanning, either using a data ACL, or using
3085 local_scan() because the Received: header was not generated till after they
3086 were called (in order to set the time as the time of reception completion).
3087 I have revised the way this works. The header is now generated after the
3088 body is received, but before the ACL or local_scan() are called. After they
3089 are run, the timestamp in the header is updated.
3095 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
3096 before starting a queue runner without re-exec. This happened only when
3097 deliver_drop_privilege was set or when the Exim user was set to root. The
3098 effect of the bug was that timeouts during subsequent deliveries caused
3099 crashes instead of being properly handled. The handler is now left at its
3100 default (and expected) setting.
3102 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
3103 message, again when deliver_drop_privilege is set or Exim is run as root.
3104 The bug described in (1) was not present in this case, but the tidying up
3105 of the other signals was missing. I have made the two cases consistent.
3107 3. The ignore_target_hosts setting on a manualroute router was being ignored
3108 for hosts that were looked up using the /MX notation.
3110 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
3113 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
3114 operated on the sender address. After changing the $sender_address to <>
3115 for the sender address verify, Exim was re-instated it as the original
3116 (before rewriting) address, but remembering that it had rewritten it, so it
3117 wasn't rewriting it again. This bug also had the effect of breaking the
3118 sender address verification caching when the sender address was rewritten.
3120 6. The ignore_target_hosts option was being ignored by the ipliteral router.
3121 This has been changed so that if the ip literal address matches
3122 ignore_target_hosts, the router declines.
3124 7. Added expansion conditions match_domain, match_address, and match_local_
3125 part (NOT match_host).
3127 8. The placeholder for the Received: header didn't have a length field set.
3129 9. Added code to Exim itself and to exim_lock to test for a specific race
3130 condition that could lead to file corruption when using MBX delivery. The
3131 issue is with the lockfile that is created in /tmp. If this file is removed
3132 after a process has opened it but before that process has acquired a lock,
3133 there is the potential for a second process to recreate the file and also
3134 acquire a lock. This could lead to two Exim processes writing to the file
3135 at the same time. The added code performs the same test as UW imapd; it
3136 checks after acquiring the lock that its file descriptor still refers to
3137 the same named file.
3139 10. The buffer for building added header lines was of fixed size, 8192 bytes.
3140 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
3143 11. Added the smtp_active_hostname option. If used, this will typically be made
3144 to depend on the incoming interface address. Because $interface_address is
3145 not set up until the daemon has forked a reception process, error responses
3146 that can happen earlier (such as "too many connections") no longer contain
3149 12. If an expansion in a condition on a "warn" statement fails because a lookup
3150 defers, the "warn" statement is abandoned, and the next ACL statement is
3151 processed. Previously this caused the whole ACL to be aborted.
3153 13. Added the iplsearch lookup type.
3155 14. Added ident_timeout as a log selector.
3157 15. Added tls_certificate_verified as a log selector.
3159 16. Added a global option tls_require_ciphers (compare the smtp transport
3160 option of the same name). This controls incoming TLS connections.
3162 17. I finally figured out how to make tls_require_ciphers do a similar thing
3163 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
3164 before starting the TLS session.
3166 18. Tabs are now shown as \t in -bP output.
3168 19. If the log selector return_path_on_delivery was set, Exim crashed when
3169 bouncing a message because it had too many Received: header lines.
3171 20. If two routers both had headers_remove settings, and the first one included
3172 a superfluous trailing colon, the final name in the first list and the
3173 first name in the second list were incorrectly joined into one item (with a
3174 colon in the middle).
3180 1. Added -C and -D options to the exinext utility, mainly to make it easier
3181 to include in the automated testing, but these could be helpful when
3182 multiple configurations are in use.
3184 2. The exinext utility was not formatting the output nicely when there was
3185 an alternate port involved in the retry record key, nor when there was a
3186 message id as well (for retries that were specific to a specific message
3187 and a specific host). It was also confused by IPv6 addresses, because of
3188 the additional colons they contain. I have fixed the IPv4 problem, and
3189 patched it up to do a reasonable job for IPv6.
3191 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
3192 delivery, the log line now contains "pipelined" if PIPELINING was used.
3194 4. An SMTP transport process used to panic and die if the bind() call to set
3195 an explicit outgoing interface failed. This has been changed; it is now
3196 treated in the same way as a connect() failure.
3198 5. A reference to $sender_host_name in the part of a conditional expansion
3199 that was being skipped was still causing a DNS lookup. This no longer
3202 6. The def: expansion condition was not recognizing references to header lines
3203 that used bh_ and bheader_.
3205 7. Added the _cache feature to named lists.
3207 8. The code for checking quota_filecount in the appendfile transport was
3208 allowing one more file than it should have been.
3210 9. For compatibility with Sendmail, the command line option
3218 and sets the incoming protocol and host name (for trusted callers). The
3219 host name and its colon can be omitted when only the protocol is to be set.
3220 Note the Exim already has two private options, -pd and -ps, that refer to
3221 embedded Perl. It is therefore impossible to set a protocol value of "d" or
3222 "s", but I don't think that's a major issue.
3224 10. A number of refactoring changes to the code, none of which should affect
3227 (a) The number of logging options was getting close to filling up the
3228 32-bit word that was used as a bit map. I have split them into two classes:
3229 those that are passed in the argument to log_write(), and those that are
3230 only ever tested independently outside of that function. These are now in
3231 separate 32-bit words, so there is plenty of room for expansion again.
3232 There is no change in the user interface or the logging behaviour.
3234 (b) When building, for example, log lines, the code previously used a
3235 macro that called string_cat() twice, in order to add two strings. This is
3236 not really sufficiently general. Furthermore, there was one instance where
3237 it was actually wrong because one of the argument was used twice, and in
3238 one call a function was used. (As it happened, calling the function twice
3239 did not affect the overall behaviour.) The macro has been replaced by a
3240 function that can join an arbitrary number of extra strings onto a growing
3243 (c) The code for expansion conditions now uses a table and a binary chop
3244 instead of a serial search (which was left over from when there were very
3245 few conditions). Also, it now recognizes conditions like "pam" even when
3246 the relevant support is not compiled in: a suitably worded error message is
3247 given if an attempt is made to use such a condition.
3249 11. Added ${time_interval:xxxxx}.
3251 12. A bug was causing one of the ddress fields not to be passed back correctly
3252 from remote delivery subprocesses. The field in question was not being
3253 subsequently used, so this caused to problems in practice.
3255 13. Added new log selectors queue_time and deliver_time.
3257 14. Might have fixed a bug in maildirsizefile handling that threw up
3258 "unexpected character" debug warnings, and recalculated the data
3259 unnecessarily. In any case, I expanded the warning message to give more
3262 15. Added the message "Restricted characters in address" to the statements in
3263 the default ACL that block characters like @ and % in local parts.
3265 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
3266 Three changes have been made:
3268 (a) There was a serious bug; a negative response to MAIL caused the whole
3269 recipient domain to be cached as invalid, thereby blocking all messages
3270 to all local parts at the same domain, from all senders. This bug has
3271 been fixed. The domain is no longer cached after a negative response to
3272 MAIL if the sender used is not empty.
3274 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
3277 (c) A new callout option, "use_sender" has been added for people who want
3278 the modified behaviour.
3284 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
3287 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
3288 indeed breaks things for older releases.
3290 3. Added additional logging to the case where there is a problem reading data
3291 from a filter that is running in a subprocess using a pipe, in order to
3292 try to track down a specific problem.
3294 4. Testing facility fudge: when running in the test harness and attempting
3295 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
3296 getting "No route to host". Convert this to a timeout.
3298 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
3301 6. Some OS don't have socklen_t but use size_t instead. This affects the
3302 fifth argument of getsockopt() amongst other things. This is now
3303 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
3304 can be set for individual OS. I have set it for SunOS5, OSF1, and
3305 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
3306 some earlier ones do not.
3308 7. Change 4.30/15 was not doing the test caselessly.
3310 8. The standard form for an IPv6 address literal was being rejected by address
3311 parsing in, for example, MAIL and RCPT commands. An example of this kind of
3312 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
3313 this, as well as the form without the "IPv6" on the front (but only when
3314 address literals are enabled, of course).
3316 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
3318 10. Exim crashed if a message with an empty sender address specified by -f
3319 encountered a router with an errors_to setting. This could be provoked only
3320 by a command such as
3324 where an empty string was supplied; "<>" did not hit this bug.
3326 11. Installed PCRE release 4.5.
3328 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
3329 remained set. It is now erased.
3331 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
3332 times from message ids (which are base 36 rather than the normal 62).
3334 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
3335 were being counted as actual protocol errors, and logged if the log
3336 selector +smtp_protocol_error was set. One cannot be perfect in this test,
3337 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
3338 and DATA following a set of rejected RCPTs do not count as protocol errors.
3339 In other words, Exim assumes they were pipelined, though this may not
3340 actually be the case. Of course, in all cases the client gets an
3341 appropriate error code.
3343 15. If a lookup fails in an ACL condition, a message about the failure may
3344 be available; it is used if testing the ACL cannot continue, because most
3345 such messages specify what the cause of the deferral is. However, some
3346 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
3347 that caused an old message to be retained and used if a later statement
3348 caused a defer, replacing the real cause of the deferral.
3350 16. If an IP address had so many PTR records that the DNS lookup buffer
3351 was not large enough to hold them, Exim could crash while trying to process
3352 the truncated data. It now detects and logs this case.
3354 17. Further to 4.21/58, another change has been made: if (and only if) the
3355 first line of a message (the first header line) ends with CRLF, a bare LF
3356 in a subsequent header line has a space inserted after it, so as not to
3357 terminate the header.
3359 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
3360 unnecessarily, used atoi() instead of strtol(), and didn't check the
3361 termination when getting file sizes from file names by regex.
3363 19. Completely re-implemented the support for maildirsize files, in the light
3364 of a number of problems with the previous contributed implementation
3365 (4.30/29). In particular:
3367 . If the quota is zero, the maildirsize file is maintained, but no quota is
3370 . If the maildir directory does not exist, it is created before any attempt
3371 to write a maildirsize file.
3373 . The quota value in the file is just a cache; if the quota is changed in
3374 the transport, the new value overrides.
3376 . A regular expression is available for excluding directories from the
3379 20. The autoreply transport checks the characters in options that define the
3380 message's headers; it allows continued headers, but it was checking with
3381 isspace() after an embedded newline instead of explicitly looking for a
3384 21. If all the "regular" hosts to which an address was routed had passed their
3385 expiry times, and had not reached their retry times, the address was
3386 bounced, even if fallback hosts were defined. Now Exim should go on to try
3389 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
3390 equivalent code in the SMTP transport. Some hosts send humungous responses
3391 to HELO/EHLO, more than 1024 it seems.
3393 23. Refactoring: code in filter.c used (void *) for "any old type" but this
3394 gives compiler warnings in some environments. I've now done it "properly",
3397 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
3398 (because of problems with the built-in one) was declared to return uschar *
3399 instead of char *, causing compiler failure.
3401 25. Fixed a file descriptor leak when processing alias/forward files.
3403 26. Fixed a minor format string issue in dbfn.c.
3405 27. Typo in exim.c: ("dmbnz" for "dbmnz").
3407 28. If a filter file refered to $h_xxx or $message_headers, and the headers
3408 contained RFC 2047 "words", Exim's memory could, under certain conditions,
3411 29. When a sender address is verified, it is cached, to save repeating the test
3412 when there is more than one recipient in a message. However, when the
3413 verification involves a callout, it is possible for different callout
3414 options to be set for different recipients. It is too complicated to keep
3415 track of this in the cache, so now Exim always runs a verification when a
3416 callout is required, relying on the callout cache for the optimization.
3417 The overhead is duplication of the address routing, but this should not be
3420 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
3421 to be verified with callout=postmaster, and the main callout worked but the
3422 postmaster check failed, the verification correctly failed. However, if a
3423 subsequent RCPT command asked for sender verification *without* the
3424 postmaster check, incorrect caching caused this verification also to fail,
3427 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
3428 it was not caching the DNS options (qualify_single, search_parents) that
3429 were used when the lookup failed. A subsequent lookup with different
3430 options therefore always gave the same answer, though there were cases
3431 where it should not have. (Example: a "domains = !$mx_any" option on a
3432 dnslookup router: the "domains" option is always processed without any
3433 widening, but the router might have qualify_single set.) Now Exim uses the
3434 cached value only when the same options are set.
3436 32. Added John Jetmore's "exipick" utility to the distribution.
3438 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
3439 than a timeout (e.g. a certificate is required, and is not provided), an
3440 Exim server now closes the connection immediately. Previously it waited for
3441 the client to close - but if the client is SSL, it seems that they each
3442 wait for each other, leading to a delay before one of them times out.
3444 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
3445 maintained 0.8.x compatibility because I don't think many are using it, and
3446 it is clearly obsolete.
3448 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
3451 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
3452 client certificate was expired. A simple patch fixes this, though I don't
3453 understand the full logic of why the verify callback is called multiple
3456 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
3457 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
3458 which causes problems with some clients (such as the Certicom SSL Plus
3459 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
3460 disables the coutermeasure allowing Eudora to connect."
3462 38. Exim was not checking that a write() to a log file succeeded. This could
3463 lead to Bad Things if a log got too big, in particular if it hit a file
3464 size limit. Exim now panics and dies if it cannot write to a log file, just
3465 as it does if it cannot open a log file.
3467 39. Modified OS/Makefile-Linux so that it now contains
3469 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
3471 The two -D definitions ensure that Exim is compiled with large file
3472 support, which makes it possible to handle log files that are bigger than
3475 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
3476 instance) a domain was checked against a named list that involved a lookup,
3477 causing $domain_data to be set, then another domain was checked against the
3478 same list, then the first domain was re-checked, the value of $domain_data
3479 after the final check could be wrong. In particular, if the second check
3480 failed, it could be set empty. This bug probably also applied to
3483 41. The strip_trailing_dot option was not being applied to the address given
3484 with the -f command-line option.
3486 42. The code for reading a message's header from the spool was incrementing
3487 $received_count, but never initializing it. This meant that the value was
3488 incorrect (doubled) while delivering a message in the same process in which
3489 it was received. In the most common configuration of Exim, this never
3490 happens - a fresh exec is done - but it can happen when
3491 deliver_drop_privilege is set.
3493 43. When Exim logs an SMTP synchronization error - client data sent too soon -
3494 it now includes up to 150 characters of the unexpected data in the log
3497 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
3498 and building data strings. The size of both of these buffers was 10 000
3499 bytes - far larger than anybody would *ever* want, thought I. Needless to
3500 say, somebody hit the limit. I have increased the maximum line length to
3501 20 000 and the maximum data length of concatenated lines to 100 000. I have
3502 also fixed two bugs, because there was no checking on these buffers. Tsk,
3503 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
3504 buffer is too small.
3506 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
3507 lsearch lookups. Now it does.
3509 46. When parsing a route_list item in a manualroute router, a fixed-length
3510 buffer was used for the list of hosts. I made this 1024 bytes long,
3511 thinking that nobody would ever have a list of hosts that long. Wrong.
3512 Somebody had a whole pile of complicated expansion conditions, and the
3513 string was silently truncated, leading to an expansion error. It turns out
3514 that it is easier to change to an unlimited length (owing to other changes
3515 that have happened since this code was originally written) than to build
3516 structure for giving a limitation error. The length of the item that
3517 expands into the list of hosts is now unlimited.
3519 47. The lsearch lookup could not handle data where the length of text line was
3520 more than 4095 characters. Such lines were truncated, leading to shortened
3521 data being returned. It should now handle lines of any length.
3523 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
3524 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
3527 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
3528 replacing strings such as BIN_DIRECTORY in a source file: the replacement
3529 no longer happens in comment lines. A list of replacements is now placed
3530 at the head of all of the source files, except those whose only change is
3531 to replace PERL_COMMAND in the very first #! line.
3533 50. Replaced the slow insertion sort in queue.c, for sorting the list of
3534 messages on the queue, with a bottom-up merge sort, using code contributed
3535 by Michael Haardt. This should make operations like -bp somewhat faster on
3536 large queues. It won't affect queue runners, except when queue_run_in_order
3539 51. Installed eximstats 1.31 in the distribution.
3541 52. Added support for SRV lookups to the dnslookup router.
3543 53. If an ACL referred to $message_body or $message_body_end, the value was not
3544 reset for any messages that followed in the same SMTP session.
3546 54. The store-handling optimization for building very long strings was not
3547 differentiating between the different store pools. I don't think this
3548 actually made any difference in practice, but I've tidied it.
3550 55. While running the routers to verify a sender address, $sender_address
3551 was still set to the sender address. This is wrong, because when routing to
3552 send a bounce to the sender, it would be empty. Therefore, I have changed
3553 it so that, while verifying a sender address, $sender_address is set to <>.
3554 (There is no change to what happens when verifying a recipient address.)
3556 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
3557 target A or AAAA records (if not already returned) without resetting the
3558 qualify_single or search_parents options of the DNS resolver. These are
3559 inappropriate in this case because the targets of MX and SRV records must
3560 be FQDNs. A broken DNS record could cause trouble if it happened to have a
3561 target that, when qualified, matched something in the local domain. These
3562 two options are now turned off when doing these lookups.
3564 57. It seems that at least some releases of Reiserfs (which does not have the
3565 concept of a fixed number of inodes) returns zero and not -1 for the
3566 number of available inodes. This interacted badly with check_spool_inodes,
3567 which assumed that -1 was the "no such thing" setting. What I have done is
3568 to check that the total number of inodes is greater than zero before doing
3569 the test of how many are available.
3571 58. When a "warn" ACL statement has a log_message modifier, the message is
3572 remembered, and not repeated. This is to avoid a lot of repetition when a
3573 message has many recipients that cause the same warning to be written.
3574 Howewer, Exim was preserving the list of already written lines for an
3575 entire SMTP session, which doesn't seem right. The memory is now reset if a
3576 new message is started.
3578 59. The "rewrite" debugging flag was not showing the result of rewriting in the
3579 debugging output unless log_rewrite was also set.
3581 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
3582 of (int)(handle) when we know that handle contains (void *)(-1).
3584 61. The Exim daemon panic-logs an error return when it closes the incoming
3585 connection. However "connection reset by peer" seems to be common, and
3586 isn't really an error worthy of noting specially, so that particular error
3589 62. When Exim is trying to find all the local interfaces, it used to panic and
3590 die if the ioctl to get the interface flags failed. However, it seems that
3591 on at least one OS (Solaris 9) it is possible to have an interface that is
3592 included in the list of interfaces, but for which you get a failure error
3593 for this call. This happens when the interface is not "plumbed" into a
3594 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
3595 failure of the "get flags" call assumes that the interface is down.
3597 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
3598 makes life easier for people who are doing arithmetic on fields extracted
3599 from dates, where you often get leading zeros that should not be
3600 interpreted as octal.
3602 64. Added qualify_domain to the redirect router, to override the global
3605 65. If a pathologically long header line contained very many addresses (the
3606 report of this problem mentioned 10 000) and each of them was rewritten,
3607 Exim could use up a very large amount of memory. (It kept on making new
3608 copies of the header line as it rewrote, and never released the old ones.)
3609 At the expense of a bit more processing, the header rewriting function has
3610 been changed so that it no longer eats memory in this way.
3612 66. The generation of the Received: header has been moved from the time that a
3613 message starts to be received, to the time that it finishes. The timestamp
3614 in the Received: header should now be very close to that of the <= log
3615 line. There are two side-effects of this change:
3617 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
3618 logged header lines no longer include the local Received: line, because
3619 it has not yet been created. The same applies to a copy of the message
3620 that is returned to a non-SMTP sender when a message is rejected.
3622 (b) When a filter file is tested using -bf, no additional Received: header
3623 is added to the test message. After some thought, I decided that this
3626 This change does not affect the value of $received_for. It is still set
3627 after address rewriting, but before local_scan() is called.
3629 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
3631 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
3632 gave an unhelpful panic error message, and a defer error. I have managed to
3633 change this behaviour so that it now rejects any supplied certificate,
3634 which seems right, as the list of acceptable certificates is empty.
3636 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
3637 gave an unhelpful defer error. I have not managed to make this reject any
3638 supplied certificates, but the error message it gives is "no certificate
3639 supplied", which is not helpful.
3641 70. exigrep's output now also includes lines that are not associated with any
3642 message, but which match the given pattern. Implemented by a patch from
3643 Martin Sluka, which also tidied up the Perl a bit.
3645 71. Recipient callout verification, like sender verification, was using <> in
3646 the MAIL FROM command. This isn't really the right thing, since the actual
3647 sender may affect whether the remote host accepts the recipient or not. I
3648 have changed it to use the actual sender in the callout; this means that
3649 the cache record is now keyed on a recipient/sender pair, not just the
3650 recipient address. There doesn't seem to be a real danger of callout loops,
3651 since a callout by the remote host to check the sender would use <>.
3652 [SEE ABOVE: changed after hitting problems.]
3654 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
3655 temporary errors. However, in the case of such a code being given after
3656 the end of a data transmission (i.e. after ".") Exim was failing to write
3657 a retry record for the message. (Yes, there was some broken host that was
3658 actually sending 8xx at this point.)
3660 73. An unknown lookup type in a host list could cause Exim to panic-die when
3661 the list was checked. (An example that provoked this was putting <; in the
3662 middle of a list instead of at the start.) If this happened during a DATA
3663 ACL check, a -D file could be left lying around. This kind of configuration
3664 error no longer causes Exim to die; instead it causes a defer errror. The
3665 incident is still logged to the main and panic logs.
3667 74. Buglet left over from Exim 3 conversion. The message "too many messages
3668 in one connection" was written to the rejectlog but not the mainlog, except
3669 when address rewriting (yes!) was being logged.
3671 75. Added write_rejectlog option.
3673 76. When a system filter was run not as root (that is, when system_filter_user
3674 was set), the values of the $n variables were not being returned to the
3675 main process; thus, they were not subsequently available in the $sn
3678 77. Added +return_path_on_delivery log selector.
3680 78. A connection timeout was being treated differently from recipients deferred
3681 when testing hosts_max_try with a message that was older than the host's
3682 retry timeout. (The host should not be counted, thus allowing all hosts to
3683 be tried at least once before bouncing.) This may have been the cause of an
3684 occasionally reported bug whereby a message would remain on the queue
3685 longer than the retry timeout, but would be bounced if a delivery was
3686 forced. I say "may" because I never totally pinned down the problem;
3687 setting up timeout/retry tests is difficult. See also the next item.
3689 79. The ultimate address timeout was not being applied to errors that involved
3690 a combination of host plus message (for example, a timeout on a MAIL
3691 command). When an address resolved to a number of possible hosts, and they
3692 were not all tried for each delivery (e.g. because of hosts_max_try), a
3693 message could remain on the queue longer than the retry timeout.
3695 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
3698 81. Fixed an obscure SMTP outgoing bug which required at least the following
3699 conditions: (a) there was another message waiting for the same server;
3700 (b) the server returned 5xx to all RCPT commands in the first message so
3701 that the message was not completed; (c) the server dropped the connection
3702 or gave a negative response to the RSET that Exim sends to abort the
3703 transaction. The observed case was a dropped connection after DATA that had
3704 been sent in pipelining mode. That is, the server had advertised PIPELINING
3705 but was not implementing it correctly. The effect of the bug was incorrect
3706 behaviour, such as trying another host, and this could lead to a crash.
3712 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
3713 and daemon.c were passed as pointers to ints; they should have been
3714 pointers to socklen_t variables (which are typically unsigned ints).
3716 2. Some signed/unsigned type warnings in the os.c file for Linux have been
3719 3. Fixed a really odd bug that affected only the testing scheme; patching a
3720 certain fixed string in the binary changed the value of another string that
3721 happened to be identical to the end of the original first string.
3723 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
3724 name", it returns that address as the IP address. On some operating
3725 systems (e.g. Solaris), it also passes back the IP address string as the
3726 "host name". However, on others (e.g. Linux), it passes back an empty
3727 string. Exim wasn't checking for this, and was changing the host name to an
3728 empty string, assuming it had been canonicized.
3730 5. Although rare, it is permitted to have more than one PTR record for a given
3731 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
3732 all the names associated with an address, because they do in Solaris.
3733 However, it seems that they do not in Linux for data that comes from the
3734 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
3735 I found this out when I moved to a new Linux workstation and tried to run
3736 the Exim test suite.
3738 To get round this problem I have changed the code so that it now does its
3739 own call to the DNS to look up PTR records when searching for a host name.
3740 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
3741 addresses that are only in /etc/hosts are still found.
3743 This behaviour is, however, controlled by an option called host_lookup_
3744 order, which defaults to "bydns:byaddr". If people want to use the other
3745 order, or indeed, just use one or the other means of lookup, they can
3746 specify it in this variable.
3748 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
3749 some operating systems, this comes back from gethostbyaddr() as an empty
3750 string, and this is what Exim used to test for. However, it seems that in
3751 other systems, "." is yielded. Exim now tests for this case too.
3753 7. The values of check_spool_space and check_log_space are now held internally
3754 as a number of kilobytes instead of an absolute number of bytes. If a
3755 numbers is specified without 'K' or 'M', it is rounded up to the nearest
3756 kilobyte. This means that much larger values can be stored.
3758 8. Exim monitor: an attempt to get the action menu when not actually pointing
3759 at a message produces an empty menu entitled "No message selected". This
3760 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
3761 no entries in it ("Shell widget menu has zero width and/or height"). So I
3762 have added a single, blank menu entry in this case.
3764 9. Added ${quote_local_part.
3766 10. MIME decoding is now applied to the contents of Subject: header lines when
3769 11. Now that a reference to $sender_host_address automatically causes a reverse
3770 lookup to occur if necessary (4.13/18), there is no need to arrange for a
3771 host lookup before query-style lookups in lists that might use this
3772 variable. This has therefore been abolished, and the "net-" prefix is no
3773 longer necessary for query-style lookups.
3775 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
3776 have been a typo for LFLAGS, so it has been changed.
3778 13. The install script calls Exim with "-C /dev/null" in order to find the
3779 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
3780 to be output. Howeve, since Exim outputs its version number before the
3781 error, it didn't break the script. It just looked ugly. I fixed this by
3782 always allowing "-C /dev/null" if the caller is root.
3784 14. Ignore overlarge ACL variable number when reading spool file - insurance
3785 against a later release with more variables having written the file.
3787 15. The standard form for an IPv6 address literal was being rejected by EHLO.
3788 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
3789 this, as well as the form without the "IPv6" on the front.
3791 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
3792 OS/Makefile-Darwin file.
3794 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
3795 only to LDAP libraries that do not have LDAP_OPT_DEREF.
3797 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
3798 variable. However, some OS use ints for inodes. I've added cast to long int
3799 to get rid of the compiler warning.
3801 19. I had forgotten to lock out "/../" in configuration file names when
3802 ALT_CONFIG_PREFIX was set.
3804 20. Routers used for verification do not need to specify transports. However,
3805 if such a router generated a host list, and callout was configured, Exim
3806 crashed, because it could not find a port number from the (non-existent)
3807 transport. It now assumes port 25 in this circumstance.
3809 21. Added the -t option to exigrep.
3811 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
3812 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
3813 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
3814 them is set, LOOKUP_LSEARCH is forced.
3816 23. "exim -bV" now outputs a list of lookups that are included in the binary.
3818 24. Added sender and host information to the "rejected by local_scan()" log
3819 line; previously there was no indication of these.
3821 25. Added .include_if_exists.
3823 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
3824 Linux. It turns out that not all file systems support this. Apparently some
3825 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
3826 do it.) To cope with this, the error EINVAL, which means that sync-ing is
3827 not supported on the file descriptor, is now ignored when Exim is trying to
3828 sync a directory. This applies only to Linux.
3830 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
3832 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
3833 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
3836 29. Added support for maildirsize files from supplied patch (modified a bit).
3838 30. The use of :fail: followed by an empty string could lead Exim to respond to
3839 sender verification failures with (e.g.):
3841 550 Verification failed for <xxx>
3842 550 Sender verify failed
3844 where the first response line was missing the '-' that indicates it is not
3845 the final line of the response.
3847 31. The loop for finding the name of the user that called Exim had a hardwired
3848 limit of 10; it now uses the value of finduser_retries, which is used for
3849 all other user lookups.
3851 32. Added $received_count variable, available in data and not_smtp ACLs, and at
3854 33. Exim was neglecting to zero errno before one call of strtol() when
3855 expanding a string and expecting an integer value. On some systems this
3856 resulted in spurious "integer overflow" errors. Also, it was casting the
3857 result into an int without checking.
3859 34. Testing for a connection timeout using "timeout_connect" in the retry rules
3860 did not work. The code looks as if it has *never* worked, though it appears
3861 to have been documented since at least releast 1.62. I have made it work.
3863 35. The "timeout_DNS" error in retry rules, also documented since at least
3864 1.62, also never worked. As it isn't clear exactly what this means, and
3865 clearly it isn't a major issue, I have abolished the feature by treating it
3866 as "timeout", and writing a warning to the main and panic logs.
3868 36. The display of retry rules for -brt wasn't always showing the error code
3871 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
3872 timeout_connect_A, timeout_connect_MX.
3874 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
3875 to the empty sender.
3877 39. The daemon was not analysing the content of -oX till after it had closed
3878 stderr and disconnected from the controlling terminal. This meant that any
3879 syntax errors were only noted on the panic log, and the return code from
3880 the command was 0. By re-arranging the code a little, I've made the
3881 decoding happen first, so such errors now appear on stderr, and the return
3882 code is 1. However, the actual setting up of the sockets still happens in
3883 the disconnected process, so errors there are still only recorded on the
3886 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
3887 connections (as happens on some IP stacks) was logged at start up time as
3888 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
3889 it from "IPv6 and IPv4", which means that two separate sockets are being
3892 41. The debug output for gethostbyname2() or getipnodebyname() failures now
3893 says whether AF_INET or AF_INET6 was passed as an argument.
3895 42. Exiwhat output was messed up when time zones were included in log
3898 43. Exiwhat now gives more information about the daemon's listening ports,
3899 and whether -tls-on-connect was used.
3901 44. The "port" option of the smtp transport is now expanded.
3903 45. A "message" modifier in a "warn" statement in a non-message ACL was being
3904 silently ignored. Now an error message is written to the main and panic
3907 46. There's a new ACL modifier called "logwrite" which writes to a log file
3908 as soon as it is encountered.
3910 47. Added $local_user_uid and $local_user_gid at routing time.
3912 48. Exim crashed when trying to verify a sender address that was being
3915 49. Exim was recognizing only a space character after ".include". It now also
3916 recognizes a tab character.
3918 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
3919 extracting the relevant information from the specification. The man page no
3920 longer contains scrambled data for the -d option, and I've added a section
3921 at the front about calling Exim under different names.
3923 51. Added "extra_headers" argument to the "mail" command in filter files.
3925 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
3928 53. Installed eximstats 1.29.
3930 54. Added transport_filter_timeout as a generic transport option.
3932 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
3933 Cc: header lines. This was required by RFC 822, but it not required by RFC
3936 56. Exim used to add From:, Date:, and Message-Id: header lines to any
3937 incoming messages that did not have them. Now it does so only if the
3938 message originates locally, that is, if there is no associated remote host
3939 address. When Resent- header lines are present, this applies to the Resent-
3940 lines rather than the non-Resent- lines.
3942 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
3943 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
3945 58. Messages for configuration errors now include the name of the main
3946 configuration file - useful now that there may be more than one file in a
3947 list (.included file names were always shown).
3949 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
3950 for those rare installations that do not start the daemon as root or run it
3951 setuid root. I've cut out the call to initgroups() if the daemon is not
3954 60. The Exim user and group can now be bound into the binary as text strings
3955 that are looked up at the start of Exim's processing.
3957 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
3959 62. Added $mailstore_basename variable.
3961 63. Installed patch to sieve.c from Michael Haardt.
3963 64. When Exim failed to open the panic log after failing to open the main log,
3964 the original message it was trying to log was written to stderr and debug
3965 output, but if they were not available (the usual case in production), it
3966 was lost. Now it is written to syslog before the two lines that record the
3967 failures to open the logs.
3969 65. Users' Exim filters run in subprocesses under the user's uid. It is
3970 possible for a "deliver" command or an alias in a "personal" command to
3971 provoke an address rewrite. If logging of address rewriting is configured,
3972 this fails because the process is not running as root or exim. There may be
3973 a better way of dealing with this, but for the moment (because 4.30 needs
3974 to be released), I have disabled address rewrite logging when running a
3975 filter in a non-root, non-exim process.
3981 1. The buildconfig auxiliary program wasn't quoting the value set for
3982 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
3983 not defined. This bug was masked in 4.22 by the effect that was fixed in
3986 2. Some messages that were rejected after a message id was allocated were
3987 shown as "incomplete" by exigrep. It no longer does this for messages that
3988 are rejected by local_scan() or the DATA or non-SMTP ACLs.
3990 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
3991 have allow_domain_literals set, the ID did not get logged in the <= line.
3992 Domain literals are now always recognized in Message-ID: header lines.
3994 4. The first argument for a ${extract expansion item is the key name or field
3995 number. Leading and trailing spaces in this item were not being ignored,
3996 causing some misleading effects.
3998 5. When deliver_drop_privilege was set, single queue runner processes started
3999 manually (i.e. by the command "exim -q") or by the daemon (which uses the
4000 same command in the process it spins off) were not dropping privilege.
4002 6. When the daemon running as "exim" started a queue runner, it always
4003 re-executed Exim in the spun-off process. This is a waste of effort when
4004 deliver_drop_privilege is set. The new process now just calls the
4005 queue-runner function directly.
4011 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
4014 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
4015 ignored. Though the use of -oP was forcing the writing of a pid file, it
4016 was always written to the default place.
4018 3. If the message "no IP address found for host xxxx" is generated during
4019 incoming verification, it is now followed by identification of the incoming
4020 connection (so you can more easily find what provoked it).
4022 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
4024 5. Added some features to "harden" Exim a bit more against certain attacks:
4026 (a) There is now a build-time option called FIXED_NEVER_USERS that can
4027 be put in Local/Makefile. This is like the never_users runtime option,
4028 but it cannot be overridden. The default setting is "root".
4030 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
4031 prefix string with which any file named in a -C command line option
4034 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
4035 is retained for -C and -D only if the caller of Exim is root. Without
4036 it, the exim user may also use -C and -D and retain privilege.
4038 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
4039 command line option is disabled.
4041 6. Macro names set by the -D option must start with an upper case letter, just
4042 like macro names defined in the configuration file.
4044 7. Added "dereference=" facility to LDAP.
4046 8. Two instances of the typo "uknown" in the source files are fixed.
4048 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
4049 the Configure-Makefile script screwed up while processing it.
4051 10. Incorporated PCRE 4.4.
4053 11. The SMTP synchronization check was not operating right at the start of an
4054 SMTP session. For example, it could not catch a HELO sent before the client
4055 waited for the greeting. There is now a check for outstanding input at the
4056 point when the greeting is written. Because of the duplex, asynchronous
4057 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
4058 way, but not yet received, when the check is performed.
4060 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
4061 on TCP/IP sockets, because this apparently causes some broken clients to
4064 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
4065 unchanged) from the Cygwin maintainer.
4067 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
4068 was supported instead of testing for mbx. Effectively a typo.
4070 15. The spa authenticator server code was not checking that the input it
4071 received was valid base64.
4073 16. The debug output line for the "set" modifier in ACLs was not showing the
4074 name of the variable that was being set.
4076 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
4078 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
4079 lookup on its own. Something else was needed to trigger the lookup. For
4080 example, a match in host_lookup or the need for a host name in a host list.
4081 Now, if $sender_host_name is referenced and the host name has not yet been
4082 looked up, a lookup is performed. If the lookup fails, the variable remains
4083 empty, and $host_lookup_failed is set to "1".
4085 19. Added "eqi" as a case-independent comparison operator.
4087 20. The saslauthd authentication condition could segfault if neither service
4088 nor realm was specified.
4090 21. If an overflowing value such as "2048M" was set for message_size_limit, the
4091 error message that was logged was misleading, and incoming SMTP
4092 connections were dropped. The message is now more accurate, and temporary
4093 errors are given to SMTP connections.
4095 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
4096 (except RSET) with a 421 error, until QUIT is received. However, it was
4097 failing to send a response to QUIT.
4099 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
4100 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
4101 ACL is now run after the helo_try_verify_hosts code.
4103 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
4104 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
4105 case-independent, so other case variants are also recognized). Apparently
4106 some systems use these upper case variants.
4108 25. If more than two messages were waiting for the same host, and a transport
4109 filter was specified for the transport, Exim sent two messages over the
4110 same TCP/IP connection, and then failed with "socket operation on non-
4111 socket" when it tried to send the third.
4113 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
4115 27. The extern definition of crypt16() in expand.c was not being excluded when
4116 the OS had its own crypt16() function.
4118 28. Added bounce_return_body as a new option, and bounce_return_size_limit
4119 as a preferred synonym for return_size_limit, both as an option and as an
4122 29. Added LIBS=-liconv to OS/Makefile-OSF1.
4124 30. Changed the default configuration ACL to relax the local part checking rule
4125 for addresses that are not in any local domains. For these addresses,
4126 slashes and pipe symbols are allowed within local parts, but the sequence
4127 /../ is explicitly forbidden.
4129 31. SPA server authentication was not clearing the challenge buffer before
4132 32. log_message in a "warn" ACL statement was writing to the reject log as
4133 well as to the main log, which contradicts the documentation and doesn't
4134 seem right (because no rejection is happening). So I have stopped it.
4136 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
4137 However, I am unable to do any testing of this.
4139 34. Fixed an infelicity in the appendfile transport. When checking directories
4140 for a mailbox, to see if any needed to be created, it was accidentally
4141 using path names with one or more superfluous leading slashes; tracing
4142 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
4144 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
4145 contents are added to the log line that is written for every discarded
4146 recipient. (Previously a log_message setting was ignored.)
4148 36. The ${quote: operator now quotes the string if it is empty.
4150 37. The install script runs exim in order to find its version number. If for
4151 some reason other than non-existence or emptiness, which it checks, it
4152 could not run './exim', it was installing it with an empty version number,
4153 i.e. as "exim-". This error state is now caught, and the installation is
4156 38. An argument was missing from the function that creates an error message
4157 when Exim fails to connect to the socket for saslauthd authentication.
4158 This could cause Exim to crash, or give a corrupted message.
4160 39. Added isip, isip4, and isip6 to ${if conditions.
4162 40. The ACL variables $acl_xx are now saved with the message, and can be
4163 accessed later in routers, transports, and filters.
4165 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
4166 strings in the file are not string-expanded.
4168 42. If a MAIL command specified a SIZE value that was too large to fit into an
4169 int variable, the check against message_size_limit failed. Such values are
4170 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
4171 day this will have to be increased, but I don't think I want to be around
4172 when emails are that large.
4179 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
4180 iconv() is not standard in FreeBSD.
4182 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
4183 IPv6 enabled. The observed symptom was a segmentation fault on return from
4184 the function os_common_find_running_interfaces() in src/os.c.
4186 3. In the check_special_case() function in daemon.c I had used "errno" as an
4187 argument name, which causes warnings on some systems. This was basically a
4188 typo, since it was named "eno" in the comments!
4190 4. The code that waits for the clock to tick (at a resolution of some fraction
4191 of a second) so as to ensure message-id uniqueness was always waiting for
4192 at least one whole tick, when it could have waited for less. [This is
4193 almost certainly not relevant at current processor speeds, where it is
4194 unlikely to ever wait at all. But we try to future-proof.]
4196 5. The function that sleeps for a time interval that includes fractions of a
4197 second contained a race. It did not block SIGALRM between setting the
4198 timer, and suspending (a couple of lines later). If the interval was short
4199 and the sigsuspend() was delayed until after it had expired, the suspension
4200 never ended. On busy systems this could lead to processes getting stuck for
4203 6. Some uncommon configurations may cause a lookup to happen in a queue runner
4204 process, before it forks any delivery processes. The open lookup caching
4205 mechanism meant that the open file or database connection was passed into
4206 the delivery process. The problem was that delivery processes always tidy
4207 up cached lookup data. This could cause a problem for the next delivery
4208 process started by the queue runner, because the external queue runner
4209 process does not know about the closure. So the next delivery process
4210 still has data in the lookup cache. In the case of a file lookup, there was
4211 no problem because closing a file descriptor in a subprocess doesn't affect
4212 the parent. However, if the lookup was caching a connection to a database,
4213 the connection was closed, and the second delivery process was likely to
4214 see errors such as "PGSQL: query failed: server closed the connection
4215 unexpectedly". The problem has been fixed by closing all cached lookups
4216 in a queue runner before running a delivery process.
4218 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
4219 seem to have the "const" qualifier which it has on other OS. I've
4222 8. Change 4.21/2 was too strict. It is only if there are two authenticators
4223 *of the same type* (client or server) with the same public name that an
4224 error should be diagnosed.
4226 9. When Exim looked up a host name for an IP address, but failed to find the
4227 original IP address when looking up the host name (a safety check), it
4228 output the message "<ip address> does not match any IP for NULL", which was
4229 confusing, to say the least. The bug was that the host name should have
4230 appeared instead of "NULL".
4232 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
4233 user that is built into the binary, and the -C or -D options is used, root
4234 privilege is dropped before the configuration file is read. In addition,
4235 logging is switched to stderr instead of the normal log files. If the
4236 configuration then re-defines the Exim user, the unprivileged environment
4237 is probably not what is expected, so Exim logs a panic warning message (but
4240 However, if deliver_drop_privilege is set, the unprivileged state may well
4241 be exactly what is intended, so the warning has been cut out in that case,
4242 and Exim is allowed to try to write to its normal log files.
4248 1. smtp_return_error_details was not giving details for temporary sender
4249 or receiver verification errors.
4251 2. Diagnose a configuration error if two authenticators have the same public
4254 3. Exim used not to create the message log file for a message until the first
4255 delivery attempt. This could be confusing when incoming messages were held
4256 for policy or load reasons. The message log file is now created at the time
4257 the message is received, and an initial "Received" line is written to it.
4259 4. The automatically generated man page for command line options had a minor
4260 bug that caused no ill effects; however, a more serious problem was that
4261 the procedure for building the man page automatically didn't always
4262 operate. Consequently, release 4.20 contains an out-of-date version. This
4263 shouldn't happen again.
4265 5. When building Exim with embedded Perl support, the script that builds the
4266 Makefile was calling 'perl' to find its compile-time parameters, ignoring
4267 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
4269 6. The freeze_tell option was not being used for messages that were frozen on
4270 arrival, either by an ACL or by local_scan().
4272 7. Added the smtp_incomplete_transaction log selector.
4274 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
4275 was accepting AUTH without a new EHLO.
4277 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
4278 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
4281 10. Logging of TCP/IP connections (when configured) now happens in the main
4282 daemon process instead of the child process, so that the TCP/IP connection
4283 count is more accurate (but it can never be perfect).
4285 11. The use of "drop" in a nested ACL was not being handled correctly in the
4286 outer ACL. Now, if condition failure induced by the nested "drop" causes
4287 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
4288 or "require"), the connection is dropped.
4290 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
4291 that yield "discard" can now be used with an "accept" or a "discard" verb,
4292 but an error is generated for any others (because I can't see a useful way
4293 to define what should happen).
4295 13. When an ACL is read dynamically from a file (or anywhere else), the lines
4296 are now processed in the same way as lines in the Exim configuration file.
4297 In particular, continuation lines are supported.
4299 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
4301 15. Added -ti meaning -t -i.
4303 16. Check for letters, digits, hyphens, and dots in the names of dnslist
4304 domains, and warn by logging if others are found.
4306 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
4307 returned from GIFCONF when Exim is trying to find the list of interfaces on
4308 a host. The code in os.c has been modified to copy each ifreq to an aligned
4309 structure in all cases.
4311 Also, in some cases, the returned ifreq's were being copied to a 'struct
4312 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
4313 means the last couple of bytes of an IPv6 address could be chopped if the
4314 ifreq contained only a normal sockaddr (14 bytes storage).
4316 18. Named domain lists were not supported in the hosts_treat_as_local option.
4317 An entry such as +xxxx was not recognized, and was treated as a literal
4320 19. Ensure that header lines added by a DATA ACL are included in the reject log
4321 if the ACL subsequently rejects the message.
4323 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
4324 MD5 (which is deprecated).
4326 21. When testing a filter file using -bf, Exim was writing a message when it
4327 took the sender from a "From " line in the message, but it was not doing so
4328 when it took $return_path from a Return-Path: header line. It now does.
4330 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
4331 with a valid header line field name (a series of printing characters
4332 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
4334 23. Changed "disc" in the source to "disk" to conform to the documentation and
4335 the book and for uniformity.
4337 24. Ignore Sendmail's -Ooption=value command line item.
4339 25. When execve() failed while trying to run a command in a pipe transport,
4340 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
4341 could be confused with a return value of 69 from the command itself. This
4342 has been changed to 127, the value the shell returns if it is asked to run
4343 a non-existent command. The wording for the related log line suggests a
4344 non-existent command as the problem.
4346 26. If received_header_text expands to an empty string, do not add a Received:
4347 header line to the message. (Well, it adds a token one on the spool, but
4348 marks it "old" so that it doesn't get used or transmitted.)
4350 27. Installed eximstats 1.28 (addition of -nt option).
4352 28. There was no check for failure on the call to getsockname() in the daemon
4353 code. This can fail if there is a shortage of resources on the system, with
4354 ENOMEM, for example. A temporary error is now given on failure.
4356 29. Contrary to the C standard, it seems that in some environments, the
4357 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
4358 program. Exim now does this explicitly; it affects the formatting of
4359 timestamps using strftime().
4361 30. If exiqsumm was given junk data, it threw up some uninitialized variable
4362 complaints. I've now initialized all the variables, to avoid this.
4364 32. Header lines added by a system filter were not being "seen" during
4365 transport-time rewrites.
4367 33. The info_callback() function passed to OpenSSL is set up with type void
4368 (*)(SSL *, int, int), as described somewhere. However, when calling the
4369 function (actually a macro) that sets it up, the type void(*)() is
4370 expected. I've put in a cast to prevent warnings from picky compilers.
4372 34. If a DNS black list lookup found a CNAME record, but there were no A
4373 records associated with the domain it pointed at, Exim crashed.
4375 35. If a DNS black list lookup returned more than one A record, Exim ignored
4376 all but the first. It now scans all returned addresses if a particular IP
4377 value is being sought. In this situation, the contents of the
4378 $dnslist_value variable are a list of all the addresses, separated by a
4381 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
4382 to accept a host name and all its aliases if the forward lookup for any of
4383 them yielded the IP address of the incoming connection. Now it accepts only
4384 those names whose forward lookup yields the correct IP address. Any other
4385 names are discarded. This closes a loophole whereby a rogue DNS
4386 administrator could create reverse DNS records to break through a
4387 wildcarded host restriction in an ACL.
4389 37. If a user filter or a system filter that ran in a subprocess used any of
4390 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
4391 the wrong values were passed to the pipe command ($thisaddress had the
4392 value of $0, $0 had the value of $1, etc). This bug was introduced by
4393 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
4395 38. Improved the line breaking for long SMTP error messages from ACLs.
4396 Previously, if there was no break point between 40 and 75 characters, Exim
4397 left the rest of the message alone. Two changes have been made: (a) I've
4398 reduced the minimum length to 35 characters; (b) if it can't find a break
4399 point between 35 and 75 characters, it looks ahead and uses the first one
4400 that it finds. This may give the occasional overlong line, but at least the
4401 remaining text gets split now.
4403 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
4404 file descriptors might be low, and that setting 1000 would always raise it.
4405 It turns out that in some environments, the limit is already over 1000 and
4406 that lowering it causes trouble. So now Exim takes care not to decrease it.
4408 40. When delivering a message, the value of $return_path is set to $sender_
4409 address at the start of routing (routers may change the value). By an
4410 oversight, this default was not being set up when an address was tested by
4411 -bt or -bv, which affected the outcome if any router or filter referred to
4414 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
4415 log only when "message" or "log_message" are set. However, if one of the
4416 conditions was an address verification, or a call to a nested ACL, the
4417 messages generated by the underlying test were being passed through. This
4418 no longer happens. The underlying message is available in $acl_verify_
4419 message for both "message" and "log_message" expansions, so it can be
4420 passed through if needed.
4422 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
4423 new expansion $bh_ to give the encoded byte string without charset
4424 translation. Translation happens only if iconv() is available; HAVE_ICONV
4425 indicates this at build time. HEADERS_CHARSET gives the charset to
4426 translate to; headers_charset can change it in the configuration, and
4427 "headers charset" can change it in an individual filter file.
4429 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
4430 that creates RFC 2047 encoded "words" labels them as that charset instead
4431 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
4432 expansion operator; (ii) when Exim creates a From: line for a local
4433 message; (iii) when a header line is rewritten to include a "phrase" part.
4435 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
4436 buggy, causing it to skip the first lines of messages whose message ID
4437 ended in 'D'. This would not have bitten before Exim release 4.14, because
4438 message IDs were unlikely to end in 'D' before then. The effect was to have
4439 incorrect size information for certain domains.
4441 45. #include "config.h" was missing at the start of the crypt16.c module. This
4442 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
4445 46. If there was a timeout during a "random" callout check, Exim treated it as
4446 a failure of the random address, and carried on sending RSET and the real
4447 address. If the delay was just some slowness somewhere, the response to the
4448 original RCPT would be taken as a response to RSET and so on, causing
4449 mayhem of various kinds.
4451 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
4452 when I implemented it. It didn't allow for the fact that some option values
4453 may legitimatetly be negative (e.g. size_addition), and it didn't even do
4454 the right test for positive values.
4456 48. Domain names in DNS records are case-independent. Exim always looks them up
4457 in lower case. Some resolvers return domain names in exactly the case they
4458 appear in the zone file, that is, they may contain uppercase letters. Not
4459 all resolvers do this - some return always lower case. Exim was treating a
4460 change of case by a resolver as a change of domain, similar to a widening
4461 of a domain abbreviation. This triggered its re-routing code and so it was
4462 trying to route what was effectively the same domain again. This normally
4463 caused routing to fail (because the router wouldn't handle the domain
4464 twice). Now Exim checks for this case specially, and just changes the
4465 casing of the domain that it ultimately uses when it transmits the message
4468 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
4471 50. If a filter generated a file delivery with a non-absolute name (possible if
4472 no home directory exists for the router), the forbid_file option was not
4475 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
4476 the existing equality matching.
4478 52. Exim was using ints instead of ino_t variables in some places where it was
4479 dealing with inode numbers.
4481 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
4482 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
4483 called TMPDIR, and if it finds it is different, it changes its value.
4485 54. The smtp_printf() function is now made available to local_scan() so
4486 additional output lines can be written before returning. There is also an
4487 smtp_fflush() function to enable the detection of a dropped connection.
4488 The variables smtp_input and smtp_batched_input are exported to
4491 55. Changed the default runtime configuration: the message "Unknown user"
4492 has been removed from the ACL, and instead placed on the localuser router,
4493 using the cannot_route_message feature. This means that any verification
4494 failures that generate their own messages won't get overridden. Similarly,
4495 the "Unrouteable address" message that was in the ACL for unverifiable
4496 relay addresses has also been removed.
4498 56. Added hosts_avoid_esmtp to the smtp transport.
4500 57. The exicyclog script was not checking for the esoteric option
4501 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
4502 will work only if exicyclog is run under the appropriate euid.
4504 58. Following a discussion on the list, the rules by which Exim recognises line
4505 endings on incoming messages have been changed. The -dropcr and drop_cr
4506 options are now no-ops, retained only for backwards compatibility. The
4507 following line terminators are recognized: LF CRLF CR. However, special
4508 processing applies to CR:
4510 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
4511 nor a local message in the state where . is a terminator.
4513 (ii) If a bare CR is encountered in a header line, an extra space is added
4514 after the line terminator so as not to end the header. The reasoning
4515 behind this is that bare CRs in header lines are most likely either
4516 to be mistakes, or people trying to play silly games.
4518 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
4519 was being incorrectly given as 18 bytes larger than it should have been.
4520 This is a VOB (very old bug).
4522 60. This may never have affected anything current, but just in case it has:
4523 When the local host is found other than at the start of a list of hosts,
4524 the local host, those with the same MX, and any that follow, are discarded.
4525 When the list in question was part of a longer list of hosts, the following
4526 hosts (not currently being processed) were also being discarded. This no
4527 longer happens. I'm not sure if this situation could ever has previously
4530 61. Added the "/MX" feature to lists of hosts in the manualroute and query
4533 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
4534 header. This is something that is recommended in a new Internet Draft, and
4535 is something that is documented as being done by Sendmail. There are two
4536 possible values. For messages generated by the autoreply transport, Exim
4539 Auto-Submitted: auto-replied
4541 whereas for all other generated messages (e.g. bounces) it adds
4543 Auto-Submitted: auto-generated
4545 63. The "personal" condition in filters now includes a test for the
4546 Auto-Submitted: header. If it contains the string "auto-" the message it
4547 not considered personal.
4549 64. Added rcpt_include_affixes as a generic transport option.
4551 65. Added queue_only_override (default true).
4553 66. Added the syslog_duplication option.
4555 67. If what should have been the first header line of a message consisted of
4556 a space followed by a colon, Exim was mis-interpreting it as a header line.
4557 It isn't of course - it is syntactically invalid and should therefore be
4558 treated as the start of the message body. The misbehaviour could have
4559 caused a number of strange effects, including loss of data in subsequent
4560 header lines, and spool format errors.
4562 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
4563 client host had authenticated. This control can now be exercised by an ACL
4564 for more flexibility.
4566 69. By default, callouts do not happen when testing with -bh. There is now a
4567 variant, -bhc, which does actually run the callout code, including
4568 consulting and updating the callout cache.
4570 70. Added support for saslauthd authentication, courtesy of Alexander
4573 71. If statvfs() failed on the spool or log directories while checking their
4574 size for availability, Exim confusingly gave the error "space shortage".
4575 Furthermore, in debugging mode it crashed with a floating point exception.
4576 These checks are done if check_{spool,log}_{space,inodes} are set, and when
4577 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
4578 serious problem, Exim now writes to the main and panic logs when this
4579 happens, with details of the failure. It then refuses to accept the
4580 incoming message, giving the message "spool directory problem" or "log
4581 directory problem" with a 421 code for SMTP messages.
4583 72. When Exim is about to re-exec itself, it ensures that the file descriptors
4584 0, 1, and 2 exist, because some OS complain for execs without them (see
4585 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
4586 descriptors. However, the code omitted to check that the open succeeded,
4587 causing mysterious errors if for some reason the permissions on /dev/null
4588 got screwed. Now Exim writes a message to the main and panic logs, and
4589 bombs out if it can't open /dev/null.
4591 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
4592 interact so that it is all more flexible. It is supposed to remain
4593 backwards compatible. Also added extra_local_interfaces.
4595 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
4596 to bomb out with an assertion failure - to the client this appears as a
4597 connection drop. This problem occurs in the part of the code that was taken
4598 from the Samba project. Fortunately, the assertion is in a very simple
4599 function, so I have fixed this by reproducing the function inline in the
4600 one place where it is called, and arranging for authentication to fail
4601 instead of killing the process with assert().
4603 75. The SPA client code was not working when the server requested OEM rather
4604 than Unicode encoding.
4606 76. Added code to make require_files with a specific uid setting more usable in
4607 the case where statting the file as root fails - usually a non-root-mounted
4608 NFS file system. When this happens and the failure is EACCES, Exim now
4609 forks a subprocess and does the per-uid checking as the relevant uid.
4611 77. Added process_log_path.
4613 78. If log_file_path was not explicitly set, a setting of check_log_space or
4614 check_log_inodes was ignored.
4616 79. If a space check for the spool or log partitions fails, the incident is now
4617 logged. Of course, in the latter case the data may get lost...
4619 80. Added the %p formatting code to string_format() so that it can be used to
4620 print addresses in debug_print(). Adjusted all the address printing in the
4621 debugging in store.c to use %p rather than %d.
4623 81. There was a concern that a line of code in smtp_in.c could overflow a
4624 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
4625 initially expressed, the concern was not well-founded, because trailing
4626 spaces are removed early. However, if the trailing spaces were followed by
4627 a NULL, they did not get removed, so the overflow was possible. Two fixes
4630 (a) I re-wrote the offending code in a cleaner fashion.
4631 (b) If an incoming SMTP command contains a NULL character, it is rejected
4634 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
4635 runs initgroups(), so that if the Exim user is in any additional groups,
4636 they will be used during message reception.
4642 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob