1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.632 2010/06/12 15:21:25 jetmore Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 NM/01 Workround for PCRE version dependancy in version reporting
16 TF/01 Failure to get a lock on a hints database can have serious
17 consequences so log it to the panic log.
19 TF/02 Log LMTP confirmation messages in the same way as SMTP,
20 controlled using the smtp_confirmation log selector.
22 TF/03 Include the error message when we fail to unlink a spool file.
24 DW/01 Bugzilla 139: Support dynamically loaded lookups as modules.
25 With thanks to Steve Haslam, Johannes Berg & Serge Demonchaux
26 for maintaining out-of-tree patches for some time.
28 PP/01 Bugzilla 139: Documentation and portability issues.
29 Avoid GNU Makefile-isms, let Exim continue to build on BSD.
30 Handle per-OS dynamic-module compilation flags.
32 PP/02 Let /dev/null have normal permissions.
33 The 4.73 fixes were a little too stringent and complained about the
34 permissions on /dev/null. Exempt it from some checks.
35 Reported by Andreas M. Kirchwitz.
37 PP/03 Report version information for many libraries, including
38 Exim version information for dynamically loaded libraries. Created
39 version.h, now support a version extension string for distributors
40 who patch heavily. Dynamic module ABI change.
42 PP/04 CVE-2011-0017 - check return value of setuid/setgid. This is a
43 privilege escalation vulnerability whereby the Exim run-time user
44 can cause root to append content of the attacker's choosing to
47 PP/05 Bugzilla 1041: merged DCC maintainer's fixes for return code.
50 PP/06 Bugzilla 1071: fix delivery logging with untrusted macros.
51 If dropping privileges for untrusted macros, we disabled normal logging
52 on the basis that it would fail; for the Exim run-time user, this is not
53 the case, and it resulted in successful deliveries going unlogged.
54 Fixed. Reported by Andreas Metzler.
60 PP/01 Date: & Message-Id: revert to normally being appended to a message,
61 only prepend for the Resent-* case. Fixes regression introduced in
62 Exim 4.70 by NM/22 for Bugzilla 607.
64 PP/02 Include check_rfc2047_length in configure.default because we're seeing
65 increasing numbers of administrators be bitten by this.
67 JJ/01 Added DISABLE_DKIM and comment to src/EDITME
69 PP/03 Bugzilla 994: added openssl_options main configuration option.
71 PP/04 Bugzilla 995: provide better SSL diagnostics on failed reads.
73 PP/05 Bugzilla 834: provide a permit_coredump option for pipe transports.
75 PP/06 Adjust NTLM authentication to handle SASL Initial Response.
77 PP/07 If TLS negotiated an anonymous cipher, we could end up with SSL but
78 without a peer certificate, leading to a segfault because of an
79 assumption that peers always have certificates. Be a little more
80 paranoid. Problem reported by Martin Tscholak.
82 PP/08 Bugzilla 926: switch ClamAV to use the new zINSTREAM API for content
83 filtering; old API available if built with WITH_OLD_CLAMAV_STREAM=yes
84 NB: ClamAV planning to remove STREAM in "middle of 2010".
85 CL also introduces -bmalware, various -d+acl logging additions and
86 more caution in buffer sizes.
88 PP/09 Implemented reverse_ip expansion operator.
90 PP/10 Bugzilla 937: provide a "debug" ACL control.
92 PP/11 Bugzilla 922: Documentation dusting, patch provided by John Horne.
94 PP/12 Bugzilla 973: Implement --version.
96 PP/13 Bugzilla 752: Refuse to build/run if Exim user is root/0.
98 PP/14 Build without WITH_CONTENT_SCAN. Path from Andreas Metzler.
100 PP/15 Bugzilla 816: support multiple condition rules on Routers.
102 PP/16 Add bool_lax{} expansion operator and use that for combining multiple
103 condition rules, instead of bool{}. Make both bool{} and bool_lax{}
104 ignore trailing whitespace.
106 JJ/02 prevent non-panic DKIM error from being sent to paniclog
108 JJ/03 added tcp_wrappers_daemon_name to allow host entries other than
111 PP/17 Fix malware regression for cmdline scanner introduced in PP/08.
112 Notification from Dr Andrew Aitchison.
114 PP/18 Change ClamAV response parsing to be more robust and to handle ClamAV's
115 ExtendedDetectionInfo response format.
116 Notification from John Horne.
118 PP/19 OpenSSL 1.0.0a compatibility const-ness change, should be backwards
121 PP/20 Added a CONTRIBUTING file. Fixed the documentation build to use http:
122 XSL and documented dependency on system catalogs, with examples of how
125 DW/21 Added Valgrind hooks in store.c to help it capture out-of-bounds store
128 DW/22 Bugzilla 1044: CVE-2010-4345 - partial fix: restrict default behaviour
129 of CONFIGURE_OWNER and CONFIGURE_GROUP options to no longer allow a
130 configuration file which is writeable by the Exim user or group.
132 DW/23 Bugzilla 1044: CVE-2010-4345 - part two: extend checks for writeability
133 of configuration files to cover files specified with the -C option if
134 they are going to be used with root privileges, not just the default
137 DW/24 Bugzilla 1044: CVE-2010-4345 - part three: remove ALT_CONFIG_ROOT_ONLY
138 option (effectively making it always true).
140 DW/25 Add TRUSTED_CONFIG_PREFIX_FILE option to allow alternative configuration
141 files to be used while preserving root privileges.
143 DW/26 Set FD_CLOEXEC on SMTP sockets after forking in the daemon, to ensure
144 that rogue child processes cannot use them.
146 PP/27 Bugzilla 1047: change the default for system_filter_user to be the Exim
147 run-time user, instead of root.
149 PP/28 Add WHITELIST_D_MACROS option to let some macros be overriden by the
150 Exim run-time user without dropping privileges.
152 DW/29 Remove use of va_copy() which breaks pre-C99 systems. Duplicate the
153 result string, instead of calling string_vformat() twice with the same
156 DW/30 Allow TRUSTED_CONFIG_PREFIX_FILE only for Exim or CONFIGURE_OWNER, not
157 for other users. Others should always drop root privileges if they use
158 -C on the command line, even for a whitelisted configure file.
160 DW/31 Turn TRUSTED_CONFIG_PREFIX_FILE into TRUSTED_CONFIG_FILE. No prefixes.
162 NM/01 Fixed bug #1002 - Message loss when using multiple deliveries
168 JJ/01 installed exipick 20100104.1, adding $max_received_linelength,
169 $data_path, and $header_path variables; fixed documentation bugs and
172 JJ/02 installed exipick 20100222.0, added --input-dir and --finput to allow
173 exipick to access non-standard spools, including the "frozen" queue
176 NM/01 Bugzilla 965: Support mysql stored procedures.
177 Patch from Alain Williams
179 NM/02 Bugzilla 961: Spacing fix (syntax error) on Makefile directives for NetBSD
181 NM/03 Bugzilla 955: Documentation fix for max_rcpts.
182 Patch from Andreas Metzler
184 NM/04 Bugzilla 954: Fix for unknown responses from Dovecot authenticator.
185 Patch from Kirill Miazine
187 NM/05 Bugzilla 671: Added umask to procmail example.
189 JJ/03 installed exipick 20100323.0, fixing doc bug
191 NM/06 Bugzilla 988: CVE-2010-2023 - prevent hardlink attack on sticky mail
192 directory. Notification and patch from Dan Rosenberg.
194 TK/01 PDKIM: Upgrade PolarSSL files to upstream version 0.12.1.
196 TK/02 Improve log output when DKIM signing operation fails.
198 MH/01 Treat the transport option dkim_domain as a colon separated
199 list, not as a single string, and sign the message with each element,
200 omitting multiple occurences of the same signer.
202 NM/07 Null terminate DKIM strings, Null initialise DKIM variable
203 Bugzilla 985, 986. Patch by Simon Arlott
205 NM/08 Bugzilla 967. dnsdb DNS TXT record bug fix (DKIM-related)
206 Patch by Simon Arlott
208 PP/01 Bugzilla 989: CVE-2010-2024 - work round race condition on
209 MBX locking. Notification from Dan Rosenberg.
215 TK/01 Bugzilla 912: Fix DKIM segfault on empty headers/body.
217 NM/01 Bugzilla 913: Documentation fix for gnutls_* options.
219 NM/02 Bugzilla 722: Documentation for randint. Better randomness defaults.
221 NM/03 Bugzilla 847: Enable DNSDB lookup by default.
223 NM/04 Bugzilla 915: Flag broken perl installation during build.
229 TK/01 Added patch by Johannes Berg that expands the main option
230 "spamd_address" if it starts with a dollar sign.
232 TK/02 Write list of recipients to X-Envelope-Sender header when building
233 the mbox-format spool file for content scanning (suggested by Jakob
236 TK/03 Added patch by Wolfgang Breyha that adds experimental DCC
237 (http://www.dcc-servers.net/) support via dccifd. Activated by
238 setting EXPERIMENTAL_DCC=yes in Local/Makefile.
240 TK/04 Bugzilla 673: Add f-protd malware scanner support. Patch submitted
241 by Mark Daniel Reidel <mr@df.eu>.
243 NM/01 Bugzilla 657: Embedded PCRE removed from the exim source tree.
244 When building exim an external PCRE library is now needed -
245 PCRE is a system library on the majority of modern systems.
246 See entry on PCRE_LIBS in EDITME file.
248 NM/02 Bugzilla 646: Removed unwanted C/R in Dovecot authenticator
249 conversation. Added nologin parameter to request.
250 Patch contributed by Kirill Miazine.
252 TF/01 Do not log submission mode rewrites if they do not change the address.
254 TF/02 Bugzilla 662: Fix stack corruption before exec() in daemon.c.
256 NM/03 Bugzilla 602: exicyclog now handles panic log, and creates empty
257 log files in place. Contributed by Roberto Lima.
259 NM/04 Bugzilla 667: Close socket used by dovecot authenticator.
261 TF/03 Bugzilla 615: When checking the local_parts router precondition
262 after a local_part_suffix or local_part_prefix option, Exim now
263 does not use the address's named list lookup cache, since this
264 contains cached lookups for the whole local part.
266 NM/05 Bugzilla 521: Integrated SPF Best Guess support contributed by
267 Robert Millan. Documentation is in experimental-spec.txt.
269 TF/04 Bugzilla 668: Fix parallel build (make -j).
271 NM/05.2 Bugzilla 437: Prevent Maildir aux files being created with mode 000.
273 NM/05.3 Bugzilla 598: Improvement to Dovecot authenticator handling.
274 Patch provided by Jan Srzednicki.
276 TF/05 Leading white space used to be stripped from $spam_report which
277 wrecked the formatting. Now it is preserved.
279 TF/06 Save $spam_score, $spam_bar, and $spam_report in spool files, so
280 that they are available at delivery time.
282 TF/07 Fix the way ${extract is skipped in the untaken branch of a conditional.
284 TF/08 TLS error reporting now respects the incoming_interface and
285 incoming_port log selectors.
287 TF/09 Produce a more useful error message if an SMTP transport's hosts
288 setting expands to an empty string.
290 NM/06 Bugzilla 744: EXPN did not work under TLS.
291 Patch provided by Phil Pennock.
293 NM/07 Bugzilla 769: Extraneous comma in usage fprintf
294 Patch provided by Richard Godbee.
296 NM/08 Fixed erroneous documentation references to smtp_notquit_acl to be
297 acl_smtp_notquit, added index entry.
299 NM/09 Bugzilla 787: Potential buffer overflow in string_format.
300 Patch provided by Eugene Bujak.
302 NM/10 Bugzilla 770: Problem on some platforms modifying the len parameter to
303 accept(). Patch provided by Maxim Dounin.
305 NM/11 Bugzilla 749: Preserve old behaviour of blanks comparing equal to zero.
306 Patch provided by Phil Pennock.
308 NM/12 Bugzilla 497: Correct behaviour of exiwhat when no config exists.
310 NM/13 Bugzilla 590: Correct handling of Resent-Date headers.
311 Patch provided by Brad "anomie" Jorsch.
313 NM/14 Bugzilla 622: Added timeout setting to transport filter.
314 Patch provided by Dean Brooks.
316 TK/05 Add native DKIM support (does not depend on external libraries).
318 NM/15 Bugzilla 854: Removed code that symlinks to pcre as its no longer useful.
319 Patch provided by Graeme Fowler.
321 NM/16 Bugzilla 851: Documentation example syntax fix.
323 NM/17 Changed NOTICE file to remove references to embedded PCRE.
325 NM/18 Bugzilla 894: Fix issue with very long lines including comments in
328 NM/19 Bugzilla 745: TLS version reporting.
329 Patch provided by Phil Pennock.
331 NM/20 Bugzilla 167: bool: condition support.
332 Patch provided by Phil Pennock.
334 NM/21 Bugzilla 665: gnutls_compat_mode to allow compatibility with broken
335 clients. Patch provided by Phil Pennock.
337 NM/22 Bugzilla 607: prepend (not append) Resent-Message-ID and Resent-Date.
338 Patch provided by Brad "anomie" Jorsch.
340 NM/23 Bugzilla 687: Fix misparses in eximstats.
341 Patch provided by Heiko Schlittermann.
343 NM/24 Bugzilla 688: Fix exiwhat to handle log_selector = +pid.
344 Patch provided by Heiko Schlittermann.
346 NM/25 Bugzilla 727: Use transport mode as default mode for maildirsize file.
347 plus update to original patch.
349 NM/26 Bugzilla 799: Documentation correction for ratelimit.
351 NM/27 Bugzilla 802: Improvements to local interface IP addr detection.
352 Patch provided by David Brownlee.
354 NM/28 Bugzilla 807: Improvements to LMTP delivery logging.
356 NM/29 Bugzilla 862, 866, 875: Documentation bugfixes.
358 NM/30 Bugzilla 888: TLS documentation bugfixes.
360 NM/31 Bugzilla 896: Dovecot buffer overrun fix.
362 NM/32 Bugzilla 889: Change all instances of "expr" in shell scripts to "expr --"
363 Unlike the original bugzilla I have changed all shell scripts in src tree.
365 NM/33 Bugzilla 898: Transport filter timeout fix.
366 Patch by Todd Rinaldo.
368 NM/34 Bugzilla 901: Fix sign/unsigned and UTF mistmatches.
369 Patch by Serge Demonchaux.
371 NM/35 Bugzilla 39: Base64 decode bug fixes.
372 Patch by Jakob Hirsch.
374 NM/36 Bugzilla 909: Correct connect() call in dcc code.
376 NM/37 Bugzilla 910: Correct issue with relaxed/simple handling.
378 NM/38 Bugzilla 908: Removed NetBSD3 support as no longer needed.
380 NM/39 Bugzilla 911: Fixed MakeLinks build script.
386 TK/01 Add preliminary DKIM support. Currently requires a forked version of
387 ALT-N's libdkim that I have put here:
388 http://duncanthrax.net/exim-experimental/
390 Note to Michael Haardt: I had to rename some vars in sieve.c. They
391 were called 'true' and it seems that C99 defines that as a reserved
392 keyword to be used with 'bool' variable types. That means you could
393 not include C99-style headers which use bools without triggering
394 build errors in sieve.c.
396 NM/01 Bugzilla 592: --help option is handled incorrectly if exim is invoked
397 as mailq or other aliases. Changed the --help handling significantly
398 to do whats expected. exim_usage() emits usage/help information.
400 SC/01 Added the -bylocaldomain option to eximstats.
402 NM/02 Bugzilla 619: Defended against bad data coming back from gethostbyaddr.
404 NM/03 Bugzilla 613: Documentation fix for acl_not_smtp.
406 NM/04 Bugzilla 628: PCRE update to 7.4 (work done by John Hall).
412 PH/01 Another patch from the Sieve maintainer.
414 PH/02 When an IPv6 address is converted to a string for single-key lookup
415 in an address list (e.g. for an item such as "net24-dbm;/net/works"),
416 dots are used instead of colons so that keys in lsearch files need not
417 contain colons. This was done some time before quoting was made available
418 in lsearch files. However, iplsearch files do require colons in IPv6 keys
419 (notated using the quote facility) so as to distinguish them from IPv4
420 keys. This meant that lookups for IP addresses in host lists did not work
421 for iplsearch lookups.
423 This has been fixed by arranging for IPv6 addresses to be expressed with
424 colons if the lookup type is iplsearch. This is not incompatible, because
425 previously such lookups could never work.
427 The situation is now rather anomolous, since one *can* have colons in
428 ordinary lsearch keys. However, making the change in all cases is
429 incompatible and would probably break a number of configurations.
431 TK/01 Change PRVS address formatting scheme to reflect latests BATV draft
434 MH/01 The "spam" ACL condition code contained a sscanf() call with a %s
435 conversion specification without a maximum field width, thereby enabling
436 a rogue spamd server to cause a buffer overflow. While nobody in their
437 right mind would setup Exim to query an untrusted spamd server, an
438 attacker that gains access to a server running spamd could potentially
439 exploit this vulnerability to run arbitrary code as the Exim user.
441 TK/02 Bugzilla 502: Apply patch to make the SPF-Received: header use
442 $primary_hostname instead of what libspf2 thinks the hosts name is.
444 MH/02 The dsearch lookup now uses lstat(2) instead of stat(2) to look for
445 a directory entry by the name of the lookup key. Previously, if a
446 symlink pointed to a non-existing file or a file in a directory that
447 Exim lacked permissions to read, a lookup for a key matching that
448 symlink would fail. Now it is enough that a matching directory entry
449 exists, symlink or not. (Bugzilla 503.)
451 PH/03 The body_linecount and body_zerocount variables are now exported in the
454 PH/04 Added the $dnslist_matched variable.
456 PH/05 Unset $tls_cipher and $tls_peerdn before making a connection as a client.
457 This means they are set thereafter only if the connection becomes
460 PH/06 Added the client_condition to authenticators so that some can be skipped
461 by clients under certain conditions.
463 PH/07 The error message for a badly-placed control=no_multiline_responses left
464 "_responses" off the end of the name.
466 PH/08 Added -Mvc to output a copy of a message in RFC 2822 format.
468 PH/09 Tidied the code for creating ratelimiting keys, creating them explicitly
469 (without spaces) instead of just copying the configuration text.
471 PH/10 Added the /noupdate option to the ratelimit ACL condition.
473 PH/11 Added $max_received_linelength.
475 PH/12 Added +ignore_defer and +include_defer to host lists.
477 PH/13 Installed PCRE version 7.2. This needed some changes because of the new
478 way in which PCRE > 7.0 is built.
480 PH/14 Implemented queue_only_load_latch.
482 PH/15 Removed an incorrect (int) cast when reading the value of SIZE in a
483 MAIL command. The effect was to mangle the value on 64-bit systems.
485 PH/16 Another patch from the Sieve maintainer.
487 PH/17 Added the NOTQUIT ACL, based on a patch from Ted Cooper.
489 PH/18 If a system quota error occurred while trying to create the file for
490 a maildir delivery, the message "Mailbox is full" was not appended to the
491 bounce if the delivery eventually timed out. Change 4.67/27 below applied
492 only to a quota excession during the actual writing of the file.
494 PH/19 It seems that peer DN values may contain newlines (and other non-printing
495 characters?) which causes problems in log lines. The DN values are now
496 passed through string_printing() before being added to log lines.
498 PH/20 Added the "servers=" facility to MySQL and PostgreSQL lookups. (Oracle
499 and InterBase are left for another time.)
501 PH/21 Added message_body_newlines option.
503 PH/22 Guard against possible overflow in moan_check_errorcopy().
505 PH/23 POSIX allows open() to be a macro; guard against that.
507 PH/24 If the recipient of an error message contained an @ in the local part
508 (suitably quoted, of course), incorrect values were put in $domain and
509 $local_part during the evaluation of errors_copy.
515 MH/01 Fix for bug #448, segfault in Dovecot authenticator when interface_address
516 is unset (happens when testing with -bh and -oMi isn't used). Thanks to
519 PH/01 Added a new log selector smtp_no_mail, to log SMTP sessions that do not
520 issue a MAIL command.
522 PH/02 In an ACL statement such as
524 deny dnslists = X!=127.0.0.2 : X=127.0.0.2
526 if a client was not listed at all, or was listed with a value other than
527 127.0.0.2, in the X list, but was listed with 127.0.0.2 in the Y list,
528 the condition was not true (as it should be), so access was not denied.
529 The bug was that the ! inversion was incorrectly passed on to the second
530 item. This has been fixed.
532 PH/03 Added additional dnslists conditions == and =& which are different from
533 = and & when the dns lookup returns more than one IP address.
535 PH/04 Added gnutls_require_{kx,mac,protocols} to give more control over the
536 cipher suites used by GnuTLS. These options are ignored by OpenSSL.
538 PH/05 After discussion on the list, added a compile time option ENABLE_DISABLE_
539 FSYNC, which compiles an option called disable_fsync that allows for
540 bypassing fsync(). The documentation is heavily laced with warnings.
542 SC/01 Updated eximstats to collate all SpamAssassin rejects into one bucket.
544 PH/06 Some tidies to the infrastructure of the Test Suite that is concerned
545 with the auxiliary C programs that it uses: (1) Arrange for BIND_8_COMPAT
546 to be defined when compiling on OSX (Darwin); (2) Tidies to the Makefile,
547 including adding "make clean"; (3) Added -fPIC when compiling the test
548 dynamically loaded module, to get rid of a warning.
550 MH/02 Fix for bug #451, causing paniclog entries to be written if a bounce
551 message fails, move_frozen_messages = true and ignore_bounce_errors_after
552 = 0s. The bug is otherwise harmless.
554 PH/07 There was a bug in the dovecot authenticator such that the value of
555 $auth1 could be overwritten, and so not correctly preserved, after a
556 successful authentication. This usually meant that the value preserved by
557 the server_setid option was incorrect.
559 PH/08 Added $smtp_count_at_connection_start, deliberately with a long name.
561 PH/09 Installed PCRE release 7.0.
563 PH/10 The acl_not_smtp_start ACL was, contrary to the documentation, not being
564 run for batched SMTP input. It is now run at the start of every message
565 in the batch. While fixing this I discovered that the process information
566 (output by running exiwhat) was not always getting set for -bs and -bS
567 input. This is fixed, and it now also says "batched" for BSMTP.
569 PH/11 Added control=no_pipelining.
571 PH/12 Added $sending_ip_address and $sending_port (mostly Magnus Holmgren's
572 patch, slightly modified), and move the expansion of helo_data till after
573 the connection is made in the smtp transport (so it can use these
576 PH/13 Added ${rfc2047d: to decoded RFC 2047 strings.
578 PH/14 Added log_selector = +pid.
580 PH/15 Flush SMTP output before delaying, unless control=no_delay_flush is set.
582 PH/16 Add ${if forany and ${if forall.
584 PH/17 Added dsn_from option to vary the From: line in DSNs.
586 PH/18 Flush SMTP output before performing a callout, unless control =
587 no_callout_flush is set.
589 PH/19 Change 4.64/PH/36 introduced a bug: when address_retry_include_sender
590 was true (the default) a successful delivery failed to delete the retry
591 item, thus causing premature timeout of the address. The bug is now
594 PH/20 Added hosts_avoid_pipelining to the smtp transport.
596 PH/21 Long custom messages for fakedefer and fakereject are now split up
597 into multiline reponses in the same way that messages for "deny" and
598 other ACL rejections are.
600 PH/22 Applied Jori Hamalainen's speed-up changes and typo fixes to exigrep,
601 with slight modification.
603 PH/23 Applied sieve patches from the maintainer "tracking the latest notify
604 draft, changing the syntax and factoring some duplicate code".
606 PH/24 When the log selector "outgoing_port" was set, the port was shown as -1
607 for deliveries of the second and subsequent messages over the same SMTP
610 PH/25 Applied Magnus Holmgren's patch for ${addresses, ${map, ${filter, and
611 ${reduce, with only minor "tidies".
613 SC/02 Applied Daniel Tiefnig's patch to improve the '($parent) =' pattern match.
615 PH/26 Added a "continue" ACL modifier that does nothing, for the benefit of its
616 expansion side effects.
618 PH/27 When a message times out after an over-quota error from an Exim-imposed
619 quota, the bounce message says "mailbox is full". This message was not
620 being given when it was a system quota that was exceeded. It now should
623 MH/03 Made $recipients available in local_scan(). local_scan() already has
624 better access to the recipient list through recipients_list[], but
625 $recipients can be useful in postmaster-provided expansion strings.
627 PH/28 The $smtp_command and $smtp_command_argument variables were not correct
628 in the case of a MAIL command with additional options following the
629 address, for example: MAIL FROM:<foo@bar> SIZE=1234. The option settings
630 were accidentally chopped off.
632 PH/29 SMTP synchronization checks are implemented when a command is read -
633 there is a check that no more input is waiting when there shouldn't be
634 any. However, for some commands, a delay in an ACL can mean that it is
635 some time before the response is written. In this time, more input might
636 arrive, invalidly. So now there are extra checks after an ACL has run for
637 HELO/EHLO and after the predata ACL, and likewise for MAIL and RCPT when
638 pipelining has not been advertised.
640 PH/30 MH's patch to allow iscntrl() characters to be list separators.
642 PH/31 Unlike :fail:, a custom message specified with :defer: was not being
643 returned in the SMTP response when smtp_return_error_details was false.
646 PH/32 Change the Dovecot authenticator to use read() and write() on the socket
647 instead of the C I/O that was originally supplied, because problems were
650 PH/33 Compile failed with OpenSSL 0.9.8e. This was due to a coding error in
651 Exim which did not show up earlier: it was assuming that a call to
652 SSL_CTX_set_info_callback() might give an error value. In fact, there is
653 no error. In previous releases of OpenSSL, SSL_CTX_set_info_callback()
654 was a macro that became an assignment, so it seemed to work. This has
655 changed to a proper function call with a void return, hence the compile
656 error. Exim's code has been fixed.
658 PH/34 Change HDA_SIZE in oracle.c from 256 to 512. This is needed for 64-bit
661 PH/35 Applied a patch from the Sieve maintainer which fixes a bug in "notify".
663 PH/36 Applied John Jetmore's patch to add -v functionality to exigrep.
665 PH/37 If a message is not accepted after it has had an id assigned (e.g.
666 because it turns out to be too big or there is a timeout) there is no
667 "Completed" line in the log. When some messages of this type were
668 selected by exigrep, they were listed as "not completed". Others were
669 picked up by some special patterns. I have improved the selection
670 criteria to be more general.
672 PH/38 The host_find_failed option in the manualroute router can now be set
673 to "ignore", to completely ignore a host whose IP address cannot be
674 found. If all hosts are ignored, the behaviour is controlled by the new
675 host_all_ignored option.
677 PH/39 In a list of hosts for manualroute, if one item (either because of multi-
678 homing or because of multiple MX records with /mx) generated more than
679 one IP address, and the following item turned out to be the local host,
680 all the secondary addresses of the first item were incorrectly removed
681 from the list, along with the local host and any following hosts (which
682 is what is supposed to happen).
684 PH/40 When Exim receives a message, it writes the login name, uid, and gid of
685 whoever called Exim into the -H file. In the case of the daemon it was
686 behaving confusingly. When first started, it used values for whoever
687 started the daemon, but after a SIGHUP it used the Exim user (because it
688 calls itself on a restart). I have changed the code so that it now always
691 PH/41 (Following a suggestion from Tony Finch) If all the RCPT commands in a
692 message are rejected with the same error (e.g. no authentication or bad
693 sender address), and a DATA command is nevertheless sent (as can happen
694 with PIPELINING or a stupid MUA), the error message that was given to the
695 RCPT commands is included in the rejection of the DATA command. This is
696 intended to be helpful for MUAs that show only the final error to their
699 PH/42 Another patch from the Sieve maintainer.
701 SC/02 Eximstats - Differentiate between permanent and temporary rejects.
702 Eximstats - Fixed some broken HTML links and added missing column headers
704 Eximstats - Fixed Grand Total Summary Domains, Edomains, and Email
705 columns for Rejects, Temp Rejects, Ham, and Spam rows.
707 SC/03 Eximstats - V1.58 Fix to get <> and blackhole to show in edomain tables.
709 PH/43 Yet another patch from the Sieve maintainer.
711 PH/44 I found a way to check for a TCP/IP connection going away before sending
712 the response to the final '.' that terminates a message, but only in the
713 case where the client has not sent further data following the '.'
714 (unfortunately, this is allowed). However, in many cases there won't be
715 any further data because there won't be any more messages to send. A call
716 to select() can be used: if it shows that the input is "ready", there is
717 either input waiting, or the socket has been closed. An attempt to read
718 the next input character can distinguish the two cases. Previously, Exim
719 would have sent an OK response which the client would never have see.
720 This could lead to message repetition. This fix should cure that, at
721 least in a lot of common cases.
723 PH/45 Do not advertise STARTTLS in response to HELP unless it would be
724 advertised in response to EHLO.
730 PH/01 Two more bugs that were introduced by 4.64/PH/07, in addition to the one
731 fixed by 4.65/MH/01 (is this a record?) are fixed:
733 (i) An empty string was always treated as zero by the numeric comparison
734 operators. This behaviour has been restored.
736 (ii) It is documented that the numeric comparison operators always treat
737 their arguments as decimal numbers. This was broken in that numbers
738 starting with 0 were being interpreted as octal.
740 While fixing these problems I realized that there was another issue that
741 hadn't been noticed. Values of message_size_limit (both the global option
742 and the transport option) were treated as octal if they started with 0.
743 The documentation was vague. These values are now always treated as
744 decimal, and I will make that clear in the documentation.
750 TK/01 Disable default definition of HAVE_LINUX_SENDFILE. Clashes with
751 Linux large file support (_FILE_OFFSET_BITS=64) on older glibc
754 MH/01 Don't check that the operands of numeric comparison operators are
755 integers when their expansion is in "skipping" mode (fixes bug
756 introduced by 4.64-PH/07).
758 PH/01 If a system filter or a router generates more than SHRT_MAX (32767)
759 child addresses, Exim now panics and dies. Previously, because the count
760 is held in a short int, deliveries were likely to be lost. As such a
761 large number of recipients for a single message is ridiculous
762 (performance will be very, very poor), I have chosen to impose a limit
763 rather than extend the field.
769 TK/01 Bugzilla #401. Fix DK spooling code so that it can overwrite a
770 leftover -K file (the existence of which was triggered by #402).
771 While we were at it, introduced process PID as part of the -K
772 filename. This should rule out race conditions when creating
775 TK/02 Bugzilla #402. Apply patch from Simon Arlott, speeding up DK signing
776 processing considerably. Previous code took too long for large mails,
777 triggering a timeout which in turn triggers #401.
779 TK/03 Introduced HAVE_LINUX_SENDFILE to os.h-Linux. Currently only used
780 in the DK code in transports.c. sendfile() is not really portable,
781 hence the _LINUX specificness.
783 TF/01 In the add_headers option to the mail command in an Exim filter,
784 there was a bug that Exim would claim a syntax error in any
785 header after the first one which had an odd number of characters
788 PH/01 If a server that rejects MAIL FROM:<> was the target of a sender
789 callout verification, Exim cached a "reject" for the entire domain. This
790 is correct for most verifications, but it is not correct for a recipient
791 verification with use_sender or use_postmaster set, because in that case
792 the callout does not use MAIL FROM:<>. Exim now distinguishes the special
793 case of MAIL FROM:<> rejection from other early rejections (e.g.
794 rejection of HELO). When verifying a recipient using a non-null MAIL
795 address, the cache is ignored if it shows MAIL FROM:<> rejection.
796 Whatever the result of the callout, the value of the domain cache is
797 left unchanged (for any other kind of callout, getting as far as trying
798 RCPT means that the domain itself is ok).
800 PH/02 Tidied a number of unused variable and signed/unsigned warnings that
803 PH/03 On Solaris, an unexpectedly close socket (dropped connection) can
804 manifest itself as EPIPE rather than ECONNECT. When tidying away a
805 session, the daemon ignores ECONNECT errors and logs others; it now
806 ignores EPIPE as well.
808 PH/04 Applied Nico Erfurth's refactoring patch to tidy up mime.c
809 (quoted-printable decoding).
811 PH/05 Applied Nico Erfurth's refactoring patch to tidy up spool_mbox.c, and
812 later the small subsequent patch to fix an introduced bug.
814 PH/06 Installed the latest Cygwin Makefile from the Cygwin maintainer.
816 PH/07 There was no check for overflow in expansions such as ${if >{1}{4096M}}.
818 PH/08 An error is now given if message_size_limit is specified negative.
820 PH/09 Applied and tidied up Jakob Hirsch's patch for allowing ACL variables
821 to be given (somewhat) arbitrary names.
823 JJ/01 exipick 20060919.0, allow for arbitrary acl_ variables introduced
826 JJ/02 exipick 20060919.0, --show-vars args can now be regular expressions,
827 miscellaneous code fixes
829 PH/10 Added the log_reject_target ACL modifier to specify where to log
832 PH/11 Callouts were setting the name used for EHLO/HELO from $smtp_active_
833 hostname. This is wrong, because it relates to the incoming message (and
834 probably the interface on which it is arriving) and not to the outgoing
835 callout (which could be using a different interface). This has been
836 changed to use the value of the helo_data option from the smtp transport
837 instead - this is what is used when a message is actually being sent. If
838 there is no remote transport (possible with a router that sets up host
839 addresses), $smtp_active_hostname is used.
841 PH/12 Installed Andrey Panin's patch to add a dovecot authenticator. Various
842 tweaks were necessary in order to get it to work (see also 21 below):
843 (a) The code assumed that strncpy() returns a negative number on buffer
844 overflow, which isn't the case. Replaced with Exim's string_format()
846 (b) There were several signed/unsigned issues. I just did the minimum
847 hacking in of casts. There is scope for a larger refactoring.
848 (c) The code used strcasecmp() which is not a standard C function.
849 Replaced with Exim's strcmpic() function.
850 (d) The code set only $1; it now sets $auth1 as well.
851 (e) A simple test gave the error "authentication client didn't specify
852 service in request". It would seem that Dovecot has changed its
853 interface. Fortunately there's a specification; I followed it and
854 changed what the client sends and it appears to be working now.
856 PH/13 Added $message_headers_raw to provide the headers without RFC 2047
859 PH/14 Corrected misleading output from -bv when -v was also used. Suppose the
860 address A is aliased to B and C, where B exists and C does not. Without
861 -v the output is "A verified" because verification stops after a
862 successful redirection if more than one address is generated. However,
863 with -v the child addresses are also verified. Exim was outputting "A
864 failed to verify" and then showing the successful verification for C,
865 with its parentage. It now outputs "B failed to verify", showing B's
866 parentage before showing the successful verification of C.
868 PH/15 Applied Michael Deutschmann's patch to allow DNS black list processing to
869 look up a TXT record in a specific list after matching in a combined
872 PH/16 It seems that the options setting for the resolver (RES_DEFNAMES and
873 RES_DNSRCH) can affect the behaviour of gethostbyname() and friends when
874 they consult the DNS. I had assumed they would set it the way they
875 wanted; and indeed my experiments on Linux seem to show that in some
876 cases they do (I could influence IPv6 lookups but not IPv4 lookups).
877 To be on the safe side, however, I have now made the interface to
878 host_find_byname() similar to host_find_bydns(), with an argument
879 containing the DNS resolver options. The host_find_byname() function now
880 sets these options at its start, just as host_find_bydns() does. The smtp
881 transport options dns_qualify_single and dns_search_parents are passed to
882 host_find_byname() when gethostbyname=TRUE in this transport. Other uses
883 of host_find_byname() use the default settings of RES_DEFNAMES
884 (qualify_single) but not RES_DNSRCH (search_parents).
886 PH/17 Applied (a modified version of) Nico Erfurth's patch to make
887 spool_read_header() do less string testing, by means of a preliminary
888 switch on the second character of optional "-foo" lines. (This is
889 overdue, caused by the large number of possibilities that now exist.
890 Originally there were few.) While I was there, I also converted the
891 str(n)cmp tests so they don't re-test the leading "-" and the first
892 character, in the hope this might squeeze out yet more improvement.
894 PH/18 Two problems with "group" syntax in header lines when verifying: (1) The
895 flag allowing group syntax was set by the header_syntax check but not
896 turned off, possible causing trouble later; (2) The flag was not being
897 set at all for the header_verify test, causing "group"-style headers to
898 be rejected. I have now set it in this case, and also caused header_
899 verify to ignore an empty address taken from a group. While doing this, I
900 came across some other cases where the code for allowing group syntax
901 while scanning a header line wasn't quite right (mostly, not resetting
902 the flag correctly in the right place). These bugs could have caused
903 trouble for malformed header lines. I hope it is now all correct.
905 PH/19 The functions {pwcheck,saslauthd}_verify_password() are always called
906 with the "reply" argument non-NULL. The code, however (which originally
907 came from elsewhere) had *some* tests for NULL when it wrote to *reply,
908 but it didn't always do it. This confused somebody who was copying the
909 code for some other use. I have removed all the tests.
911 PH/20 It was discovered that the GnuTLS code had support for RSA_EXPORT, a
912 feature that was used to support insecure browsers during the U.S. crypto
913 embargo. It requires special client support, and Exim is probably the
914 only MTA that supported it -- and would never use it because real RSA is
915 always available. This code has been removed, because it had the bad
916 effect of slowing Exim down by computing (never used) parameters for the
917 RSA_EXPORT functionality.
919 PH/21 On the advice of Timo Sirainen, added a check to the dovecot
920 authenticator to fail if there's a tab character in the incoming data
921 (there should never be unless someone is messing about, as it's supposed
922 to be base64-encoded). Also added, on Timo's advice, the "secured" option
923 if the connection is using TLS or if the remote IP is the same as the
924 local IP, and the "valid-client-cert option" if a client certificate has
927 PH/22 As suggested by Dennis Davis, added a server_condition option to *all*
928 authenticators. This can be used for authorization after authentication
929 succeeds. (In the case of plaintext, it servers for both authentication
932 PH/23 Testing for tls_required and lost_connection in a retry rule didn't work
933 if any retry times were supplied.
935 PH/24 Exim crashed if verify=helo was activated during an incoming -bs
936 connection, where there is no client IP address to check. In this
937 situation, the verify now always succeeds.
939 PH/25 Applied John Jetmore's -Mset patch.
941 PH/26 Added -bem to be like -Mset, but loading a message from a file.
943 PH/27 In a string expansion for a processed (not raw) header when multiple
944 headers of the same name were present, leading whitespace was being
945 removed from all of them, but trailing whitespace was being removed only
946 from the last one. Now trailing whitespace is removed from each header
947 before concatenation. Completely empty headers in a concatenation (as
950 PH/28 Fixed bug in backwards-compatibility feature of PH/09 (thanks to John
951 Jetmore). It would have mis-read ACL variables from pre-4.61 spool files.
953 PH/29 [Removed. This was a change that I later backed out, and forgot to
954 correct the ChangeLog entry (that I had efficiently created) before
955 committing the later change.]
957 PH/30 Exim was sometimes attempting to deliver messages that had suffered
958 address errors (4xx response to RCPT) over the same connection as other
959 messages routed to the same hosts. Such deliveries are always "forced",
960 so retry times are not inspected. This resulted in far too many retries
961 for the affected addresses. The effect occurred only when there were more
962 hosts than the hosts_max_try setting in the smtp transport when it had
963 the 4xx errors. Those hosts that it had tried were not added to the list
964 of hosts for which the message was waiting, so if all were tried, there
965 was no problem. Two fixes have been applied:
967 (i) If there are any address or message errors in an SMTP delivery, none
968 of the hosts (tried or untried) are now added to the list of hosts
969 for which the message is waiting, so the message should not be a
970 candidate for sending over the same connection that was used for a
971 successful delivery of some other message. This seems entirely
972 reasonable: after all the message is NOT "waiting for some host".
973 This is so "obvious" that I'm not sure why it wasn't done
974 previously. Hope I haven't missed anything, but it can't do any
975 harm, as the worst effect is to miss an optimization.
977 (ii) If, despite (i), such a delivery is accidentally attempted, the
978 routing retry time is respected, so at least it doesn't keep
979 hammering the server.
981 PH/31 Installed Andrew Findlay's patch to close the writing end of the socket
982 in ${readsocket because some servers need this prod.
984 PH/32 Added some extra debug output when updating a wait-xxx database.
986 PH/33 The hint "could be header name not terminated by colon", which has been
987 given for certain expansion errors for a long time, was not being given
988 for the ${if def:h_colon_omitted{... case.
990 PH/34 The spec says: "With one important exception, whenever a domain list is
991 being scanned, $domain contains the subject domain." There was at least
992 one case where this was not true.
994 PH/35 The error "getsockname() failed: connection reset by peer" was being
995 written to the panic log as well as the main log, but it isn't really
996 panic-worthy as it just means the connection died rather early on. I have
997 removed the panic log writing for the ECONNRESET error when getsockname()
1000 PH/36 After a 4xx response to a RCPT error, that address was delayed (in queue
1001 runs only) independently of the message's sender address. This meant
1002 that, if the 4xx error was in fact related to the sender, a different
1003 message to the same recipient with a different sender could confuse
1004 things. In particualar, this can happen when sending to a greylisting
1005 server, but other circumstances could also provoke similar problems.
1006 I have changed the default so that the retry time for these errors is now
1007 based a combination of the sender and recipient addresses. This change
1008 can be overridden by setting address_retry_include_sender=false in the
1011 PH/37 For LMTP over TCP/IP (the smtp transport), error responses from the
1012 remote server are returned as part of bounce messages. This was not
1013 happening for LMTP over a pipe (the lmtp transport), but now it is the
1014 same for both kinds of LMTP.
1016 PH/38 Despite being documented as not happening, Exim was rewriting addresses
1017 in header lines that were in fact CNAMEs. This is no longer the case.
1019 PH/39 If -R or -S was given with -q<time>, the effect of -R or -S was ignored,
1020 and queue runs started by the daemon processed all messages. This has
1021 been fixed so that -R and -S can now usefully be given with -q<time>.
1023 PH/40 Import PCRE release 6.7 (fixes some bugs).
1025 PH/41 Add bitwise logical operations to eval (courtesy Brad Jorsch).
1027 PH/42 Give an error if -q is specified more than once.
1029 PH/43 Renamed the variables $interface_address and $interface_port as
1030 $received_ip_address and $received_port, to make it clear that these
1031 values apply to message reception, and not to the outgoing interface when
1032 a message is delivered. (The old names remain recognized, of course.)
1034 PH/44 There was no timeout on the connect() call when using a Unix domain
1035 socket in the ${readsocket expansion. There now is.
1037 PH/45 Applied a modified version of Brad Jorsch's patch to allow "message" to
1038 be meaningful with "accept".
1040 SC/01 Eximstats V1.43
1041 Bug fix for V1.42 with -h0 specified. Spotted by Chris Lear.
1043 SC/02 Eximstats V1.44
1044 Use a glob alias rather than an array ref in the generated
1045 parser. This improves both readability and performance.
1047 SC/03 Eximstats V1.45 (Marco Gaiarin / Steve Campbell)
1048 Collect SpamAssassin and rejection statistics.
1049 Don't display local sender or destination tables unless
1050 there is data to show.
1051 Added average volumes into the top table text output.
1053 SC/04 Eximstats V1.46
1054 Collect data on the number of addresses (recipients)
1055 as well as the number of messages.
1057 SC/05 Eximstats V1.47
1058 Added 'Message too big' to the list of mail rejection
1059 reasons (thanks to Marco Gaiarin).
1061 SC/06 Eximstats V1.48
1062 Mainlog lines which have GMT offsets and are too short to
1063 have a flag are now skipped.
1065 SC/07 Eximstats V1.49 (Alain Williams)
1066 Added the -emptyok flag.
1068 SC/08 Eximstats V1.50
1069 Fixes for obtaining the IP address from reject messages.
1071 JJ/03 exipick.20061117.2, made header handling as similar to exim as possible
1072 (added [br]h_ prefixes, implemented RFC2047 decoding. Fixed
1073 whitesspace changes from 4.64-PH/27
1075 JJ/04 exipick.20061117.2, fixed format and added $message_headers_raw to
1078 JJ/05 exipick.20061117.2, bug fixes (error out sooner when invalid criteria
1079 are found, allow negative numbers in numeric criteria)
1081 JJ/06 exipick.20061117.2, added new $message_body_missing variable
1083 JJ/07 exipick.20061117.2, added $received_ip_address and $received_port
1084 to match changes made in 4.64-PH/43
1086 PH/46 Applied Jori Hamalainen's patch to add features to exiqsumm.
1088 PH/47 Put in an explicit test for a DNS lookup of an address record where the
1089 "domain" is actually an IP address, and force a failure. This locks out
1090 those revolvers/nameservers that support "A-for-A" lookups, in
1091 contravention of the specifications.
1093 PH/48 When a host name was looked up from an IP address, and the subsequent
1094 forward lookup of the name timed out, the host name was left in
1095 $sender_host_name, contrary to the specification.
1097 PH/49 Although default lookup types such as lsearch* or cdb*@ have always been
1098 restricted to single-key lookups, Exim was not diagnosing an error if
1099 * or *@ was used with a query-style lookup.
1101 PH/50 Increased the value of DH_BITS in tls-gnu.c from 768 to 1024.
1103 MH/01 local_scan ABI version incremented to 1.1. It should have been updated
1104 long ago, but noone interested enough thought of it. Let's just say that
1105 the "1.1" means that there are some new functions that weren't there at
1106 some point in the past.
1108 PH/51 Error processing for expansion failure of helo_data from an smtp
1109 transport during callout processing was broken.
1111 PH/52 Applied John Jetmore's patch to allow tls-on-connect and STARTTLS to be
1112 tested/used via the -bh/-bhc/-bs options.
1114 PH/53 Added missing "#include <time.h>" to pcre/pcretest.c (this was a PCRE
1115 bug, fixed in subsequent PCRE releases).
1117 PH/54 Applied Robert Bannocks' patch to avoid a problem with references that
1118 arises when using the Solaris LDAP libraries (but not with OpenLDAP).
1120 PH/55 Check for a ridiculously long file name in exim_dbmbuild.
1126 SC/01 Use a glob alias rather than an array ref in eximstats generated
1127 parser. This improves both readability and performance.
1129 SC/02 Collect SpamAssassin and rejection statistics in eximstats.
1130 Don't display local sender or destination tables in eximstats unless
1131 there is data to show.
1132 Added average volumes into the eximstats top table text output.
1134 SC/03 Collect data on the number of addresses (recipients) as well
1135 as the number of messages in eximstats.
1137 TF/01 Correct an error in the documentation for the redirect router. Exim
1138 does (usually) call initgroups() when daemonizing.
1140 TF/02 Call initgroups() when dropping privilege in exim.c, so that Exim runs
1141 with consistent privilege compared to when running as a daemon.
1143 TF/03 Note in the spec that $authenticated_id is not set for local
1144 submissions from trusted users.
1146 TF/04 The ratelimit per_rcpt option now works correctly in acl_not_smtp.
1147 Thanks to Dean Brooks <dean@iglou.com> for the patch.
1149 TF/05 Make it easier to get SMTP authentication and TLS/SSL support working
1150 by adding some example configuration directives to the default
1151 configuration file. A little bit of work is required to uncomment the
1152 directives and define how usernames and passwords are checked, but
1153 there is now a framework to start from.
1155 PH/01 Added #define LDAP_DEPRECATED 1 to ldap.c because some of the "old"
1156 functions that Exim currently uses aren't defined in ldap.h for OpenLDAP
1157 without this. I don't know how relevant this is to other LDAP libraries.
1159 PH/02 Add the verb name to the "unknown ACL verb" error.
1161 PH/03 Magnus Holmgren's patch for filter_prepend_home.
1163 PH/03 Fixed Bugzilla #101: macro definition between ACLs doesn't work.
1165 PH/04 Applied Magnus Holmgren's patch to fix Bugzilla #98: transport's home
1166 directory not expanded when it should be if an expanded home directory
1167 was set for the address (which is overridden by the transport).
1169 PH/05 Applied Alex Kiernan's patch to fix Bugzilla #99: a problem with
1172 PH/06 Added acl_not_smtp_start, based on Johannes Berg's patch, and set the
1173 bit to forbid control=suppress_local_fixups in the acl_not_smtp ACL,
1174 because it is too late at that time, and has no effect.
1176 PH/07 Changed ${quote_pgsql to quote ' as '' instead of \' because of a
1177 security issue with \' (bugzilla #107). I could not use the
1178 PQescapeStringConn() function, because it needs a PGconn value as one of
1181 PH/08 When testing addresses using -bt, indicate those final addresses that
1182 are duplicates that would not cause an additional delivery. At least one
1183 person was confused, thinking that -bt output corresponded to deliveries.
1184 (Suppressing duplicates isn't a good idea as you lose the information
1185 about possibly different redirections that led to the duplicates.)
1187 PH/09 Applied patch from Erik to use select() instead of poll() in spam.c on
1188 systems where poll() doesn't work, in particular OS X.
1190 PH/10 Added more information to debugging output for retry time not reached.
1192 PH/11 Applied patch from Arkadiusz Miskiewicz to apply a timeout to read
1193 operations in malware.c.
1195 PH/12 Applied patch from Magnus Holmgren to include the "h" tag in Domain Keys
1198 PH/13 If write_rejectlog was set false when logging was sent to syslog with
1199 syslog_duplication set false, log lines that would normally be written
1200 both the the main log and to the reject log were not written to syslog at
1203 PH/14 In the default configuration, change the use of "message" in ACL warn
1204 statements to "add_header".
1206 PH/15 Diagnose a filter syntax error for "seen", "unseen", or "noerror" if not
1207 not followed by a command (e.g. "seen endif").
1209 PH/16 Recognize SMTP codes at the start of "message" in ACLs and after :fail:
1210 and :defer: in a redirect router. Add forbid_smtp_code to suppress the
1213 PH/17 Added extra conditions to the default value of delay_warning_condition
1217 { !eq{$h_list-id:$h_list-post:$h_list-subscribe:}{} } \
1218 { match{$h_precedence:}{(?i)bulk|list|junk} } \
1219 { match{$h_auto-submitted:}{(?i)auto-generated|auto-replied} } \
1222 The Auto-Submitted: and various List- headers are standardised, whereas I
1223 don't think Precedence: ever was.
1225 PH/18 Refactored debugging code in route_finduser() to show more information,
1226 in particular, the error code if getpwnam() issues one.
1228 PH/19 Added PQsetClientEncoding(conn, "SQL_ASCII") to the pgsql code module.
1229 This is apparently needed in addition to the PH/07 change above to avoid
1230 any possible encoding problems.
1232 PH/20 Perl can change the locale. Exim was resetting it after a ${perl call,
1233 but not after initializing Perl.
1235 PH/21 Added a call to PQsetNoticeProcessor() to catch pgsql "notices" and
1236 output them only if debugging. By default they are written stderr,
1237 apparently, which is not desirable.
1239 PH/22 Added Alain Williams' LDAP patch to support setting REFERRALS=off on
1242 JJ/01 exipick: added --reverse (and -R synonym), --random, --size, --sort and
1245 JJ/02 exipick: rewrote --help documentation to hopefully make more clear.
1247 PH/23 Made -oMaa and -oMt work with -bh and -bs to pretend the connection is
1248 authenticated or an ident call has been made. Suppress the default
1249 values for $authenticated_id and $authenticated_sender (but permit -oMai
1250 and -oMas) when testing with -bh.
1252 PH/24 Re-jigged the order of the tests in the default configuration so that the
1253 tests for valid domains and recipients precede the DNS black list and CSA
1254 tests, on the grounds that those ones are more expensive.
1256 PH/25 Exim was not testing for a space following SMTP commands such as EHLO
1257 that require one. Thus, EHLORHUBARB was interpreted as a valid command.
1258 This bug exists in every version of Exim that I still have, right back to
1261 PH/26 (n)wildlsearch lookups are documented as being done case-insensitively.
1262 However, an attempt to turn on case-sensitivity in a regex key by
1263 including (?-i) didn't work because the subject string was already
1264 lowercased, and the effects were non-intuitive. It turns out that a
1265 one-line patch can be used to allow (?-i) to work as expected.
1271 TF/01 Fix the add_header change below (4.61 PH/55) which had a bug that (amongst
1272 other effects) broke the use of negated acl sub-conditions.
1274 PH/01 ${readsocket now supports Internet domain sockets (modified John Jetmore
1277 PH/02 When tcp-wrappers is called from Exim, it returns only "deny" or "allow".
1278 "Deny" causes Exim to reject the incoming connection with a 554 error.
1279 Unfortunately, if there is a major crisis, such as a disk failure,
1280 tcp-wrappers gives "deny", whereas what one would like would be some
1281 kind of temporary error. A kludge has been added to help with this.
1282 Before calling hosts_ctl(), errno is set zero. If the result is "deny", a
1283 554 error is used if errno is still zero or contains ENOENT (which occurs
1284 if either of the /etc/hosts.{allow,deny} files is missing). Otherwise, a
1287 PH/03 Add -lutil to the default FreeBSD LIBS setting.
1289 PH/04 Change PH/19 for 4.61 was too wide. It should not be applied to host
1290 errors. Otherwise a message that provokes a temporary error (when other
1291 messages do not) can cause a whole host to time out.
1293 PH/05 Batch deliveries by appendfile and pipe transports did not work when the
1294 addresses were routed directly to files or pipes from a redirect router.
1295 File deliveries just didn't batch; pipe deliveries might have suffered
1298 PH/06 A failure to get a lock for a hints database would erroneously always say
1299 "Failed to get write lock", even when it was really a read lock.
1301 PH/07 The appendfile transport was creating MBX lock files with a fixed mode
1302 of 0600. This has been changed to use the value of the lockfile_mode
1303 option (which defaults to 0600).
1305 PH/08 Applied small patch from the Sieve maintainer.
1307 PH/09 If maildir_quota_directory_regex was set to exclude (say) the .Trash
1308 folder from quota calculations, a direct delivery into this folder messed
1309 up the contents of the maildirsize file. This was because the regex was
1310 used only to exclude .Trash (or whatever) when the size of the mailbox
1311 was calculated. There was no check that a delivery was happening into an
1312 excluded directory. This bug has been fixed by ignoring all quota
1313 processing for deliveries into excluded directories.
1315 PH/10 Added the maildirfolder_create_regex option to appendfile.
1321 PH/01 The code for finding all the local interface addresses on a FreeBSD
1322 system running IPv6 was broken. This may well have applied to all BSD
1323 systems, as well as to others that have similar system calls. The broken
1324 code found IPv4 interfaces correctly, but gave incorrect values for the
1325 IPv6 interfaces. In particular, ::1 was not found. The effect in Exim was
1326 that it would not match correctly against @[] and not recognize the IPv6
1329 PH/02 The ipliteral router was not recognizing addresses of the form user@
1330 [ipv6:....] because it didn't know about the "ipv6:" prefix.
1332 PH/03 Added disable_ipv6.
1334 PH/04 Changed $reply_address to use the raw form of the headers instead of the
1335 decoded form, because it is most often used to construct To: headers
1336 lines in autoreplies, and the decoded form may well be syntactically
1337 invalid. However, $reply_address has leading white space removed, and all
1338 newlines turned into spaces so that the autoreply transport does not
1341 PH/05 If group was specified without a user on a router, and no group or user
1342 was specified on a transport, the group from the router was ignored.
1344 PH/06 Increased the number of ACL variables to 20 of each type, and arranged
1345 for visible compile-time settings that can be used to change these
1346 numbers, for those that want even more. Backwards compatibility with old
1347 spool files has been maintained. However, going back to a previous Exim
1348 release will lost any variables that are in spool files.
1350 PH/07 Two small changes when running in the test harness: increase delay when
1351 passing a TCP/IP connection to a new process, in case the original
1352 process has to generate a bounce, and remove special handling of
1353 127.0.0.2 (sic), which is no longer necessary.
1355 PH/08 Changed debug output of dbfn_open() flags from numbers to names, so as to
1356 be the same on different OS.
1358 PH/09 Moved a debug statement in filter processing to avoid a race problem when
1361 JJ/01 exipick: fixed bug where -b (brief) output option showed "Vars:"
1362 whether --show-vars was specified or not
1364 JJ/02 exipick: Added support for new ACL variable spool format introduced
1367 PH/10 Fixed another bug related to PH/04 above: if an incoming message had a
1368 syntactically invalid From: or Reply-to: line, and a filter used this to
1369 generate an autoreply, and therefore failed to obtain an address for the
1370 autoreply, Exim could try to deliver to a non-existent relative file
1371 name, causing unrelated and misleading errors. What now happens is that
1372 it logs this as a hard delivery error, but does not attempt to create a
1375 PH/11 The exinext utility has a -C option for testing purposes, but although
1376 the given file was scanned by exinext itself; it wasn't being passed on
1377 when Exim was called.
1379 PH/12 In the smtp transport, treat an explicit ECONNRESET error the same as
1380 an end-of-file indication when reading a command response.
1382 PH/13 Domain literals for IPv6 were not recognized unless IPv6 support was
1383 compiled. In many other places in Exim, IPv6 addresses are always
1384 recognized, so I have changed this. It also means that IPv4 domain
1385 literals of the form [IPV4:n.n.n.n] are now always recognized.
1387 PH/14 When a uid/gid is specified for the queryprogram router, it cannot be
1388 used if the router is not running as root, for example, when verifying at
1389 ACL time, or when using -bh. The debugging output from this situation was
1390 non-existent - all you got was a failure to exec. I have made two
1393 (a) Failures to set uid/gid, the current directory, or a process leader
1394 in a subprocess such as that created by queryprogram now generate
1395 suitable debugging ouput when -d is set.
1397 (b) The queryprogram router detects when it is not running as root,
1398 outputs suitable debugging information if -d is set, and then runs
1399 the subprocess without attempting to change uid/gid.
1401 PH/15 Minor change to Makefile for building test_host (undocumented testing
1404 PH/16 As discussed on the list in Nov/Dec: Exim no longer looks at the
1405 additional section of a DNS packet that returns MX or SRV records.
1406 Instead, it always explicitly searches for A/AAAA records. This avoids
1407 major problems that occur when a DNS server includes only records of one
1408 type (A or AAAA) in an MX/SRV packet. A byproduct of this change has
1409 fixed another bug: if SRV records were looked up and the corresponding
1410 address records were *not* found in the additional section, the port
1411 values from the SRV records were lost.
1413 PH/17 If a delivery to a pipe, file, or autoreply was deferred, Exim was not
1414 using the correct key (the original address) when searching the retry
1415 rules in order to find which one to use for generating the retry hint.
1417 PH/18 If quota_warn_message contains a From: header, Exim now refrains from
1418 adding the default one. Similarly, if it contains a Reply-To: header, the
1419 errors_reply_to option, if set, is not used.
1421 PH/19 When calculating a retry time, Exim used to measure the "time since
1422 failure" by looking at the "first failed" field in the retry record. Now
1423 it does not use this if it is later than than the arrival time of the
1424 message. Instead it uses the arrival time. This makes for better
1425 behaviour in cases where some deliveries succeed, thus re-setting the
1426 "first failed" field. An example is a quota failure for a huge message
1427 when small messages continue to be delivered. Without this change, the
1428 "time since failure" will always be short, possible causing more frequent
1429 delivery attempts for the huge message than are intended.
1430 [Note: This change was subsequently modified - see PH/04 for 4.62.]
1432 PH/20 Added $auth1, $auth2, $auth3 to contain authentication data (as well as
1433 $1, $2, $3) because the numerical variables can be reset during some
1434 expansion items (e.g. "match"), thereby losing the authentication data.
1436 PH/21 Make -bV show the size of off_t variables so that the test suite can
1437 decide whether to run tests for quotas > 2G.
1439 PH/22 Test the values given for quota, quota_filecount, quota_warn_threshold,
1440 mailbox_size, and mailbox_filecount in the appendfile transport. If a
1441 filecount value is greater than 2G or if a quota value is greater than 2G
1442 on a system where the size of off_t is not greater than 4, a panic error
1445 PH/23 When a malformed item such as 1.2.3/24 appears in a host list, it can
1446 never match. The debug and -bh output now contains an explicit error
1447 message indicating a malformed IPv4 address or mask.
1449 PH/24 An host item such as 1.2.3.4/abc was being treated as the IP address
1450 1.2.3.4 without a mask. Now it is not recognized as an IP address, and
1451 PH/23 above applies.
1453 PH/25 Do not write to syslog when running in the test harness. The only
1454 occasion when this arises is a failure to open the main or panic logs
1455 (for which there is an explicit test).
1457 PH/26 Added the /no_tell option to "control=freeze".
1459 PH/27 If a host name lookup failed very early in a connection, for example, if
1460 the IP address matched host_lookup and the reverse lookup yielded a name
1461 that did not have a forward lookup, an error message of the form "no IP
1462 address found for host xxx.xxx.xxx (during SMTP connection from NULL)"
1463 could be logged. Now it outputs the IP address instead of "NULL".
1465 PH/28 An enabling patch from MH: add new function child_open_exim2() which
1466 allows the sender and the authenticated sender to be set when
1467 submitting a message from within Exim. Since child_open_exim() is
1468 documented for local_scan(), the new function should be too.
1470 PH/29 In GnuTLS, a forced expansion failure for tls_privatekey was not being
1471 ignored. In both GnuTLS and OpenSSL, an expansion of tls_privatekey that
1472 results in an empty string is now treated as unset.
1474 PH/30 Fix eximon buffer overflow bug (Bugzilla #73).
1476 PH/31 Added sender_verify_fail logging option.
1478 PH/32 In November 2003, the code in Exim that added an empty Bcc: header when
1479 needed by RFC 822 but not by RFC 2822 was commented out. I have now
1480 tidied the source and removed it altogether.
1482 PH/33 When a queue run was abandoned because the load average was too high, a
1483 log line was always written; now it is written only if the queue_run log
1484 selector is set. In addition, the log line for abandonment now contains
1485 information about the queue run such as the pid. This is always present
1486 in "start" and "stop" lines but was omitted from the "abandon" line.
1488 PH/34 Omit spaces between a header name and the colon in the error message that
1489 is given when verify = headers_syntax fails (if there are lots of them,
1490 the message gets confusing).
1492 PH/35 Change the default for dns_check_names_pattern to allow slashes within
1493 names, as there are now some PTR records that contain slashes. This check
1494 is only to protect against broken name servers that fall over on strange
1495 characters, so the fact that it applies to all lookups doesn't matter.
1497 PH/36 Now that the new test suite is complete, we can remove some of the
1498 special code in Exim that was needed for the old test suite. For example,
1499 sorting DNS records because real resolvers return them in an arbitrary
1500 order. The new test suite's fake resolver always returns records in the
1503 PH/37 When running in the test harness, use -odi for submitted messages (e.g.
1504 bounces) except when queue_only is set, to avoid logging races between
1505 the different processes.
1507 PH/38 Panic-die if .include specifies a non-absolute path.
1509 PH/39 A tweak to the "H" retry rule from its user.
1511 JJ/03 exipick: Removed parentheses from 'next' and 'last' calls that specified
1512 a label. They prevented compilation on older perls.
1514 JJ/04 exipick: Refactored code to prevent implicit split to @_ which caused
1515 a warning to be raised on newish perls.
1517 JJ/05 exipick: Fixed bug where -bpc always showed a count of all messages
1518 on queue. Changes to match documented behaviour of showing count of
1519 messages matching specified criteria.
1521 PH/40 Changed the default ident timeout from 30s to 5s.
1523 PH/41 Added support for the use of login_cap features, on those BSD systems
1524 that have them, for controlling the resources used by pipe deliveries.
1526 PH/42 The content-scanning code uses fopen() to create files in which to put
1527 message data. Previously it was not paying any attention to the mode of
1528 the files. Exim runs with umask(0) because the rest of the code creates
1529 files with open(), and sets the required mode explicitly. Thus, these
1530 files were ending up world-writeable. This was not a big issue, because,
1531 being within the spool directory, they were not world-accessible. I have
1532 created a function called modefopen, which takes an additional mode
1533 argument. It sets umask(777), creates the file, chmods it to the required
1534 mode, then resets the umask. All the relevant calls to fopen() in the
1535 content scanning code have been changed to use this function.
1537 PH/43 If retry_interval_max is set greater than 24 hours, it is quietly reset
1538 to 24 hours. This avoids potential overflow problems when processing G
1539 and H retry rules. I suspect nobody ever tinkers with this value.
1541 PH/44 Added STRIP_COMMAND=/usr/bin/strip to the FreeBSD Makefile.
1543 PH/45 When the plaintext authenticator is running as a client, the server's
1544 challenges are checked to ensure they are valid base64 strings. By
1545 default, the authentication attempt is cancelled if an invalid string is
1546 received. Setting client_ignore_invalid_base64 true ignores these errors.
1547 The decoded challenge strings are now placed in $auth1, $auth2, etc. as
1548 they are received. Thus, the responses can be made to depend on the
1549 challenges. If an invalid string is ignored, an empty string is placed in
1552 PH/46 Messages that are created by the autoreply transport now contains a
1553 References: header, in accordance with RFCs 2822 and 3834.
1555 PH/47 Added authenticated_sender_force to the smtp transport.
1557 PH/48 The ${prvs expansion was broken on systems where time_t was long long.
1559 PH/49 Installed latest patch from the Sieve maintainer.
1561 PH/50 When an Exim quota was set without a file count quota, and mailbox_size
1562 was also set, the appendfile transport was unnecessarily scanning a
1563 directory of message files (e.g. for maildir delivery) to find the count
1564 of files (along with the size), even though it did not need this
1565 information. It now does the scan only if it needs to find either the
1566 size of the count of files.
1568 PH/51 Added ${time_eval: to convert Exim time strings into seconds.
1570 PH/52 Two bugs concerned with error handling when the smtp transport is
1573 (i) Exim was not creating retry information for temporary errors given
1574 for individual recipients after the DATA command when the smtp transport
1575 was used in LMTP mode. This meant that they could be retried too
1576 frequently, and not timed out correctly.
1578 (ii) Exim was setting the flag that allows error details to be returned
1579 for LMTP errors on RCPT commands, but not for LMTP errors for individual
1580 recipients that were returned after the DATA command.
1582 PH/53 This is related to PH/52, but is more general: for any failing address,
1583 when detailed error information was permitted to be returned to the
1584 sender, but the error was temporary, then after the final timeout, only
1585 "retry timeout exceeded" was returned. Now it returns the full error as
1586 well as "retry timeout exceeded".
1588 PH/54 Added control=allow_auth_unadvertised, as it seems there are clients that
1589 do this, and (what is worse) MTAs that accept it.
1591 PH/55 Added the add_header modified to ACLs. The use of "message" with "warn"
1592 will now be deprecated.
1594 PH/56 New os.c-cygwin from the Cygwin maintainer.
1596 JJ/06 exipick: added --unsorted option to allow unsorted output in all output
1597 formats (previously only available in exim formats via -bpr, -bpru,
1598 and -bpra. Now also available in native and exiqgrep formats)
1600 JJ/07 exipick: added --freeze and --thaw options to allow faster interaction
1601 with very large, slow to parse queues
1603 JJ/08 exipick: added ! as generic prefix to negate any criteria format
1605 JJ/09 exipick: miscellaneous performance enhancements (~24% improvements)
1607 PH/57 Tidies in SMTP dialogue display in debug output: (i) It was not showing
1608 responses to authentication challenges, though it was showing the
1609 challenges; (ii) I've removed the CR characters from the debug output for
1612 PH/58 Allow for the insertion of a newline as well as a space when a string
1613 is turned into more than one encoded-word during RFC 2047 encoding. The
1614 Sieve code now uses this.
1616 PH/59 Added the following errors that can be detected in retry rules: mail_4xx,
1617 data_4xx, lost_connection, tls_required.
1619 PH/60 When a VRFY deferred or FAILED, the log message rather than the user
1620 message was being sent as an SMTP response.
1622 PH/61 Add -l and -k options to exicyclog.
1624 PH/62 When verifying, if an address was redirected to one new address, so that
1625 verification continued, and the new address failed or deferred after
1626 having set something in $address_data, the value of $address_data was not
1627 passed back to the ACL. This was different to the case when no
1628 redirection occurred. The value is now passed back in both cases.
1630 PH/63 Changed the macro HAVE_LOGIN_CAP (see PH/41 for this release above) to
1631 HAVE_SETCLASSRESOURCES because there are different APIs in use that all
1632 use login_cap.h, so on its own it isn't the distinguishing feature. The
1633 new name refers directly to the setclassresources() function.
1635 PH/65 Added configuration files for NetBSD3.
1637 PH/66 Updated OS/Makefile-HP-UX for gcc 4.1.0 with HP-UX 11.
1639 PH/67 Fixed minor infelicity in the sorting of addresses to ensure that IPv6
1640 is preferred over IPv4.
1642 PH/68 The bounce_return_message and bounce_return_body options were not being
1643 honoured for bounces generated during the reception of non-SMTP messages.
1644 In particular, this applied to messages rejected by the ACL. This bug has
1645 been fixed. However, if bounce_return_message is true and bounce_return_
1646 body is false, the headers that are returned for a non-SMTP message
1647 include only those that have been read before the error was detected.
1648 (In the case of an ACL rejection, they have all been read.)
1650 PH/69 The HTML version of the specification is now built in a directory called
1651 spec_html instead of spec.html, because the latter looks like a path with
1652 a MIME-type, and this confuses some software.
1654 PH/70 Catch two compiler warnings in sieve.c.
1656 PH/71 Fixed an obscure and subtle bug (thanks Alexander & Matthias). The
1657 function verify_get_ident() calls ip_connect() to connect a socket, but
1658 if the "connect()" function timed out, ip_connect() used to close the
1659 socket. However, verify_get_ident() also closes the socket later, and in
1660 between Exim writes to the log, which may get opened at this point. When
1661 the socket was closed in ip_connect(), the log could get the same file
1662 descriptor number as the socket. This naturally causes chaos. The fix is
1663 not to close the socket in ip_connect(); the socket should be closed by
1664 the function that creates it. There was only one place in the code where
1665 this was missing, in the iplookup router, which I don't think anybody now
1666 uses, but I've fixed it anyway.
1668 PH/72 Make dns_again_means_nonexist apply to lookups using gethostbyname() as
1669 well as to direct DNS lookups. Otherwise the handling of names in host
1670 lists is inconsistent and therefore confusing.
1676 PH/01 Two changes to the default runtime configuration:
1678 (1) Move the checks for relay_from_hosts and authenticated clients from
1679 after to before the (commented out) DNS black list checks.
1681 (2) Add control=submission to the relay_from_hosts and authenticated
1682 clients checks, on the grounds that messages accepted by these
1683 statements are most likely to be submissions.
1685 PH/02 Several tidies to the handling of ${prvs and ${prvscheck:
1687 (1) Generate an error if the third argument for the ${prvs expansion is
1690 (2) Treat a missing third argument of ${prvscheck as if it were an empty
1693 (3) Reset the variables that are obtained from the first argument of
1694 ${prvscheck and used in the second argument before leaving the code,
1695 because their memory is reclaimed, so using them afterwards may do
1698 (4) Tidy up the code for expanding the arguments of ${prvscheck one by
1699 one (it's much easier than Tom thought :-).
1701 (5) Because of (4), we can now allow for the use of $prvscheck_result
1702 inside the third argument.
1704 PH/03 For some reason, the default setting of PATH when running a command from
1705 a pipe transport was just "/usr/bin". I have changed it to
1708 PH/04 SUPPORT_TRANSLATE_IP_ADDRESS and MOVE_FROZEN_MESSAGES did not cause
1709 anything to be listed in the output from -bV.
1711 PH/05 When a filter generated an autoreply, the entire To: header line was
1712 quoted in the delivery log line, like this:
1714 => >A.N.Other <ano@some.domain> <original@ddress> ...
1716 This has been changed so that it extracts the operative address. There
1717 may be more than one such address. If so, they are comma-separated, like
1720 => >ano@some.domain,ona@other.domain <original@ddress> ...
1722 PH/06 When a client host used a correct literal IP address in a HELO or EHLO
1723 command, (for example, EHLO [1.2.3.4]) and the client's IP address was
1724 not being looked up in the rDNS to get a host name, Exim was showing the
1725 IP address twice in Received: lines, even though the IP addresses were
1726 identical. For example:
1728 Received: from [1.2.3.4] (helo=[1.2.3.4])
1730 However, if the real host name was known, it was omitting the HELO data
1731 if it matched the actual IP address. This has been tidied up so that it
1732 doesn't show the same IP address twice.
1734 PH/07 When both +timestamp and +memory debugging was on, the value given by
1735 $tod_xxx expansions could be wrong, because the tod_stamp() function was
1736 called by the debug printing, thereby overwriting the timestamp buffer.
1737 Debugging no longer uses the tod_stamp() function when +timestamp is set.
1739 PH/08 When the original message was included in an autoreply transport, it
1740 always said "this is a copy of the message, including all the headers",
1741 even if body_only or headers_only was set. It now gives an appropriate
1744 PH/09 Applied a patch from the Sieve maintainer which:
1746 o fixes some comments
1747 o adds the (disabled) notify extension core
1748 o adds some debug output for the result of if/elsif tests
1749 o points to the current vacation draft in the documentation
1750 and documents the missing references header update
1754 o fixes a bug in processing the envelope test (when testing
1755 multiple envelope elements, the last element determinted the
1758 PH/10 Exim was violating RFC 3834 ("Recommendations for Automatic Responses to
1759 Electronic Mail") by including:
1761 Auto-submitted: auto-generated
1763 in the messages that it generates (bounce messages and others, such as
1764 warnings). In the case of bounce messages for non-SMTP mesages, there was
1765 also a typo: it was using "Auto_submitted" (underscore instead of
1766 hyphen). Since every message generated by Exim is necessarily in response
1767 to another message, thes have all been changed to:
1769 Auto-Submitted: auto-replied
1771 in accordance with these statements in the RFC:
1773 The auto-replied keyword:
1775 - SHOULD be used on messages sent in direct response to another
1776 message by an automatic process,
1778 - MUST NOT be used on manually-generated messages,
1780 - MAY be used on Delivery Status Notifications (DSNs) and Message
1781 Disposition Notifications (MDNs),
1783 - MUST NOT be used on messages generated by automatic or periodic
1784 processes, except for messages which are automatic responses to
1787 PH/11 Added "${if def:sender_address {(envelope-from <$sender_address>)\n\t}}"
1788 to the default Received: header definition.
1790 PH/12 Added log selector acl_warn_skipped (default on).
1792 PH/13 After a successful wildlsearch lookup, discard the values of numeric
1793 variables because (a) they are in the wrong storage pool and (b) even if
1794 they were copied, it wouldn't work properly because of the caching.
1796 PH/14 Add check_rfc2047_length to disable enforcement of RFC 2047 length
1797 checking when decoding. Apparently there are clients that generate
1798 overlong encoded strings. Why am I not surprised?
1800 PH/15 If the first argument of "${if match_address" was not empty, but did not
1801 contain an "@" character, Exim crashed. Now it writes a panic log message
1802 and treats the condition as false.
1804 PH/16 In autoreply, treat an empty string for "once" the same as unset.
1806 PH/17 A further patch from the Sieve maintainer: "Introduce the new Sieve
1807 extension "envelope-auth". The code is finished and in agreement with
1808 other implementations, but there is no documentation so far and in fact,
1809 nobody wrote the draft yet. This extension is currently #undef'ed, thus
1810 not changing the active code.
1812 Print executed "if" and "elsif" statements when debugging is used. This
1813 helps a great deal to understand what a filter does.
1815 Document more things not specified clearly in RFC3028. I had all this
1816 sorted out, when out of a sudden new issues came to my mind. Oops."
1818 PH/18 Exim was not recognizing the "net-" search type prefix in match_ip lists
1821 PH/19 Exim expands the IPv6 address given to -bh to its full non-abbreviated
1822 canonical form (as documented). However, after a host name lookup from
1823 the IP address, check_host() was doing a simple string comparison with
1824 addresses acquired from the DNS when checking that the found name did
1825 have the original IP as one of its addresses. Since any found IPv6
1826 addresses are likely to be in abbreviated form, the comparison could
1827 fail. Luckily, there already exists a function for doing the comparison
1828 by converting both addresses to binary, so now that is used instead of
1829 the text comparison.
1831 PH/20 There was another similar case to PH/19, when a complete host name was
1832 given in a host list; looking up its IP address could give an abbreviated
1833 form, whereas the current host's name might or might not be abbreviated.
1834 The same fix has been applied.
1840 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
1841 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
1844 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
1845 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
1847 PH/03 Typo: missing ".o" in src/pcre/Makefile.
1849 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
1850 header line, restrict the check to what is listed in RFCs 2369 and 2929.
1851 Also, for "Auto-Submitted", treat anything other than "no" as
1852 non-personal, in accordance with RFC 3834. (Previously it treated
1853 anything starting "auto-" as non-personal.)
1855 TF/01 The control=submission/name=... option had a problem with syntax
1856 errors if the name included a slash character. The /name= option
1857 now slurps the rest of the string, so it can include any characters
1858 but it must come last in the list of options (after /sender_retain
1861 PH/05 Some modifications to the interface to the fake nameserver for the new
1869 TK/01 Added the "success_on_redirect" address verification option. See
1870 NewStuff for rationale and an example.
1872 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
1874 PH/02 Patch to exigrep to allow it to work on syslog lines.
1876 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
1877 fread() to skip over the body file's header line, because in Cygwin the
1878 header line is locked and is inaccessible.
1880 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
1881 co-exist for some time) to make it clear that it is the Exim ID that is
1882 referenced, not the Message-ID: header line.
1884 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
1885 string_format() function, because snprintf() does not exist on all
1888 PH/06 The use of forbid_filter_existstest now also locks out the use of the
1889 ${stat: expansion item.
1891 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
1892 protocol synchronization error", to keep the pedants happy.
1894 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
1895 well as for IRIX systems, when gcc is being used. See the host.c source
1898 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
1900 PH/10 Named domain lists were not working if used in a queue_smtp_domains
1903 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
1904 transport and to the smtp transport in LMTP mode.
1906 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
1908 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
1909 run a filter in a subprocess. This could lead to confusion in subsequent
1910 lookups in the parent process. There should also be a search_tidyup() at
1911 the end of the subprocess.
1913 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
1914 only if the host matched helo_try_verify_hosts, which caused the
1915 verification to occur when the EHLO/HELO command was issued. The ACL just
1916 tested the remembered result. Now, if a previous verification attempt has
1917 not happened, "verify = helo" does it there and then.
1919 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
1921 TK/03 Fix log output including CR from clamd.
1923 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
1924 exist provoked a memory error which could cause a segfault.
1926 PH/15 Installed PCRE 6.2
1928 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
1930 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
1931 of the problem. Specifically, suggested +O2 rather than +O1 for the
1934 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
1936 PH/20 If a delivery was routed to a non-standard port by means of an SRV
1937 record, the port was not correctly logged when the outgoing_port log
1938 selector was set (it logged the transort's default port).
1940 PH/21 Added support for host-specific ports to manualroute, queryprogram,
1941 fallback_hosts, and "hosts" in the smtp transport.
1943 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
1944 host errors such as "Connection refused".
1946 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
1947 authentication with radiusclient 0.4.9:
1949 - Error returned from rc_read_config was caught wrongly
1950 - Username/password not passed on to radius server due to wrong length.
1952 The presumption is that some radiusclient API changes for 4.51/PH/17
1953 were not taken care of correctly. The code is still untested by me (my
1954 Linux distribution still has 0.3.2 of radiusclient), but it was
1955 contributed by a Radius user.
1957 PH/24 When doing a callout, the value of $domain wasn't set correctly when
1958 expanding the "port" option of the smtp transport.
1960 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
1961 while reading a MIME header. Thanks to Tom Hughes for a patch.
1963 PH/24 Include config.h inside local_scan.h so that configuration settings are
1966 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
1967 that in an ACL for RCPT (for example), you can examine exactly what was
1970 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
1971 commands, but it was not correctly comparing the address with the actual
1972 client host address. Thus, it would show the EHLO address in Received:
1973 header lines when this was not necessary.
1975 PH/27 Added the % operator to ${eval:}.
1977 PH/28 Exim tries to create and chdir to its spool directory when it starts;
1978 it should be ignoring failures (because with -C, for example, it has lost
1979 privilege). It wasn't ignoring creation failures other than "already
1982 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
1985 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
1986 because an input line was too long, either on its own, or by virtue of
1987 too many continuations, the temporary file was not being removed, and the
1988 return code was incorrect.
1990 PH/31 Missing "BOOL" in function definition in filtertest.c.
1992 PH/32 Applied Sieve patches from the maintainer.
1994 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
1996 PH/33 Added "verify = not_blind".
1998 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
1999 Local/Makefile (with some defaults set). These are used in built scripts
2000 such as exicyclog, but they have never been used in the exim_install
2001 script (though there are many overriding facilities there). I have
2002 arranged that the exim_install script now takes note of these two
2005 PH/35 Installed configuration files for Dragonfly.
2007 PH/36 When a locally submitted message by a trusted user did not contain a
2008 From: header, and the sender address was obtained from -f or from an SMTP
2009 MAIL command, and the trusted user did not use -F to supply a sender
2010 name, $originator_name was incorrectly used when constructing a From:
2011 header. Furthermore, $originator_name was used for submission mode
2012 messages from external hosts without From: headers in a similar way,
2013 which is clearly wrong.
2015 PH/37 Added control=suppress_local_fixups.
2017 PH/38 When log_selector = +received_sender was set, and the addition of the
2018 sender made the log line's construction buffer exactly full, or one byte
2019 less than full, an overflow happened when the terminating "\n" was
2022 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
2023 when the result of a list match is failure because a DNS lookup failed.
2025 PH/40 RM_COMMAND is now used in the building process.
2027 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
2028 the "build-* directories that it finds.
2030 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
2031 address in a domain literal was a prefix of an interface address.
2033 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
2034 when verifying a sender address, unless rewrite_headers is false.
2036 PH/44 Wrote a long comment about why errors_to addresses are verified as
2037 recipients, not senders.
2039 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
2040 the ratelimit ACL was added.
2042 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
2044 PH/46 Added extra information about PostgreSQL errors to the error string.
2046 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
2047 avoiding the need to install special zones in a real server. This is
2048 backwards compatible; if it can't find the fake resolver, it drops back.
2049 Thus, both old and new test suites can be run.
2051 TF/02 Added util/ratelimit.pl
2053 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
2054 clock is set back in time.
2056 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
2057 Candler <B.Candler@pobox.com>.
2059 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
2060 OK for addresses that are the result of redirections.
2062 PH/48 A number of further additions for the benefit of the new test suite,
2063 including a fake gethostbyname() that interfaces to the fake DNS resolver
2066 TF/06 The fix for widen_domains has also been applied to qualify_single and
2067 search_parents which are the other dnslookup options that can cause
2070 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
2073 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
2075 TF/07 Exim produced the error message "an SRV record indicated no SMTP
2076 service" if it encountered an MX record with an empty target hostname.
2077 The message is now "an MX or SRV record indicated no SMTP service".
2079 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
2080 if the DNS of the sending site is misconfigured. This is quite a
2081 common situation. This change restores the behaviour of treating a
2082 helo verification defer as a failure.
2084 PH/51 If self=fail was set on a router, the bounce message did not include the
2085 actual error message.
2091 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
2093 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
2094 command itself ended in error, the underlying message about the transport
2095 filter timeout was being overwritten with the pipe command error. Now the
2096 underlying error message should be appended to the second error message.
2098 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
2100 PH/02 Reduce the amount of output that "make" produces by default. Full output
2101 can still be requested.
2103 PH/03 The warning log line about a condition test deferring for a "warn" verb
2104 was being output only once per connection, rather than after each
2105 occurrence (because it was using the same function as for successful
2106 "warn" verbs). This seems wrong, so I have changed it.
2108 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
2109 it should not have, which might have caused a crash in the right
2110 circumstances, but probably never did.
2112 PH/04 Installed a modified version of Tony Finch's patch to make submission
2113 mode fix the return path as well as the Sender: header line, and to
2114 add a /name= option so that you can make the user's friendly name appear
2117 TF/03 Added the control = fakedefer ACL modifier.
2119 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
2120 Mark Lowes for thorough testing.
2122 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
2124 TK/03 Merged latest SRS patch from Miles Wilton.
2126 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
2127 with the definition in sysexits.h (which is #included earlier).
2128 Fortunately, Exim does not actually use EX_OK. The code used to try to
2129 preserve the sysexits.h value, by assumimg that macro definitions were
2130 scanned for macro replacements. I have been disabused of this notion,
2131 so now the code just undefines EX_OK before #including unistd.h.
2133 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
2134 in the smtp transport. When a block could not be written in a single
2135 write() function, the timeout was being re-applied to each part-write.
2136 This seems wrong - if the receiver was accepting one byte at a time it
2137 would take for ever. The timeout is now adjusted when this happens. It
2138 doesn't have to be particularly precise.
2140 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
2141 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
2143 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
2144 without a domain if the check to <postmaster@domain> fails.
2146 SC/01 Eximstats: added -xls and the ability to specify output files
2147 (patch written by Frank Heydlauf).
2149 SC/02 Eximstats: use FileHandles for outputing results.
2151 SC/03 Eximstats: allow any combination of xls, txt, and html output.
2153 SC/04 Eximstats: fixed display of large numbers with -nvr option
2155 SC/05 Eximstats: fixed merging of reports with empty tables.
2157 SC/06 Eximstats: added the -include_original_destination flag
2159 SC/07 Eximstats: removed tabs and trailing whitespace.
2161 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
2163 TK/06 MBOX spool code: Add real "From " MBOX separator line
2164 so the .eml file is really in mbox format (even though
2165 most programs do not really care). Patch from Alex Miller.
2167 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
2168 The latter is generated from $received_to and is only set if the
2169 message has one envelope recipient. SA can use these headers,
2170 obviously out-of-the-box. Patch from Alex Miller.
2172 PH/08 The ${def test on a variable was returning false if the variable's
2173 value was "0", contrary to what the specification has always said!
2174 The result should be true unless the variable is empty.
2176 PH/09 The syntax error of a character other than { following "${if
2177 def:variable_name" (after optional whitespace) was not being diagnosed.
2178 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
2179 accidental colon was present, for example, could give incorrect results.
2181 PH/10 Tidied the code in a number of places where the st_size field of a stat()
2182 result is used (not including appendfile, where other changes are about
2185 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
2186 This involved changing a lot of size variables from int to off_t. It
2187 should work with maildirs and everything.
2189 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
2190 spamd dying while we are connected to it.
2192 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
2195 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
2196 function for BDB 4.3.
2198 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
2200 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
2201 the sources for PCRE have been re-arranged and more files are now
2204 PH/15 The code I had for printing potentially long long variables in PH/11
2205 above was not the best (it lost precision). The length of off_t variables
2206 is now inspected at build time, and an appropriate printing format (%ld
2207 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
2208 to be "long long int" or "long int". This is needed for the internal
2209 formatting function string_vformat().
2211 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
2212 the configuration file to be ":syslog", then the script "guesses" where
2213 the logs files are, rather than using the compiled in default. In our
2214 case the guess is not the same as the compiled default, so the script
2215 suddenly stopped working when I started to use syslog. The patch checks
2216 to see if log_file_path is "". If so, it attempts to read it from exim
2217 with no configuration file to get the compiled in version, before it
2218 falls back to the previous guessing code."
2220 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
2221 implementing BATV in an Exim configuration. See NewStuff for the gory
2224 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
2225 Makefile that are specific to HP-UX.
2227 PH/18 If the "use_postmaster" option was set for a recipient callout together
2228 with the "random" option, the postmaster address was used as the MAIL
2229 FROM address for the random test, but not for the subsequent recipient
2230 test. It is now used for both.
2232 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
2233 patch removes a few documentation additions to RFC 3028, because the
2234 latest draft now contains them. It adds the new en;ascii-case comparator
2235 and a new error check for 8bit text in MIME parts. Comparator and
2236 require names are now matched exactly. I enabled the subaddress
2237 extension, but it is not well tested yet (read: it works for me)."
2239 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
2240 rework some of the code of TK/09 above to avoid the hardwired use of
2241 "%lld" and "long long". Replaced the call to snprintf() with a call to
2244 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
2245 records point to non-existent hosts", "retry timeout exceeded", and
2246 "retry time not reached for any host after a long failure period".
2248 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
2249 experimental DomainKeys support:
2251 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
2252 (2) On an error such as an illegally used "control", the wrong name for
2253 the control was given.
2255 These problems did NOT occur unless DomainKeys support was compiled.
2257 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
2259 PH/24 Added ${if match_ip condition.
2261 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
2262 databases so that it will be absolutely obvious if a crash occurs in the
2263 DB library. This is a regular occurrence (often caused by mis-matched
2266 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
2267 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
2268 picked up on a user's system that detects such things. There doesn't seem
2269 to be a gcc warning option for this - only an attribute that has to be
2270 put on the function's prototype. It seems that in Fedora Core 4 they have
2271 set this on a number of new functions. No doubt there will be more in due
2274 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
2275 specify a transport. However, if an address that was verified by such a
2276 router was the subject of a callout, Exim crashed because it tried to
2277 read the rcpt_include_affixes from the non-existent transport. Now it
2278 just assumes that the setting of that option is false. This bug was
2279 introduced by 4.51/PH/31.
2281 PH/28 Changed -d+all to exclude +memory, because that information is very
2282 rarely of interest, but it makes the output a lot bigger. People tend to
2283 do -d+all out of habit.
2285 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
2286 code in os-type was giving problems when libc.so lives in lib64, like on
2289 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
2290 aren't the modern standard, and it seems that some systems' include files
2291 don't always have them. Exim was already checking for some of the newer
2292 ones like T_AAAA, and defining it itself. I've added checks for all the
2293 record types that Exim uses.
2295 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
2296 not automatically generating a new one, as it is supposed to. This
2297 prevented TLS from working. If the file did exist, but contained invalid
2298 data, a new version was generated, as expected. It was only the case of a
2299 non-existent file that was broken.
2301 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
2302 with a change in libdomainkeys > 0.64.
2304 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
2305 from DNS. If the selector record carries the flag, it now has
2306 precedence over the domain-wide flag.
2308 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
2310 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
2311 the use of a port name that isn't defined in /etc/services) occurred, the
2312 message was deferred as in a normal delivery, and thus remained on the
2313 spool, instead of being failed because of the mua_wrapper setting. This
2314 is now fixed, and I tidied up some of the mua_wrapper messages at the
2317 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
2318 the messages in a hash of arrays rather than using individual hashes.
2319 This is a bit cleaner and results in dramatic memory savings, albeit
2320 at a slight CPU cost.
2322 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
2323 as requested by Marc Sherman.
2325 SC/10 Eximstats: added histograms for user specified patterns as requested
2328 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
2330 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
2331 fopen() in the content-scanning modules that did not already have it.
2337 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
2338 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
2340 TK/02 Fix ACL "control" statement not being available in MIME ACL.
2342 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
2344 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
2345 to test Sieve filters that use "vacation".
2347 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
2348 that changes the way the GnuTLS parameters are stored in the cache file.
2349 The new format can be generated externally. For backward compatibility,
2350 if the data in the cache doesn't make sense, Exim assumes it has read an
2351 old-format file, and it generates new data and writes a new file. This
2352 means that you can't go back to an older release without removing the
2355 PH/03 A redirect router that has both "unseen" and "one_time" set does not
2356 work if there are any delivery delays because "one_time" forces the
2357 parent to be marked "delivered", so its unseen clone is never tried
2358 again. For this reason, Exim now forbids the simultaneous setting of
2361 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
2362 redirected to themselves ("homonym" addresses). Read the long ChangeLog
2363 entry if you want to know the details. The fix, however, neglected to
2364 consider the case when local delivery batching is involved. The test for
2365 "previously delivered" was not happening when checking to see if an
2366 address could be batched with a previous (undelivered) one; under
2367 certain circumstances this could lead to multiple deliveries to the same
2370 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
2371 in its include files, and this causes problems building Exim.
2373 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
2374 header_syntax) but Exim was just ignoring anything given after a slash.
2375 In particular, this caused confusion with an attempt to use "verify =
2376 reverse_host_lookup/defer_ok". An error is now given when options are
2377 supplied for verify items that do not have them. (Maybe reverse_host_
2378 lookup should have a defer_ok option, but that's a different point.)
2380 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
2381 defined by RFC 821) to 2048, because there were problems with some AUTH
2382 commands, and RFC 1869 says the size should be increased for extended
2383 SMTP commands that take arguments.
2385 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
2388 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
2389 "unknown" error; now it says that the functionality isn't in the binary.
2391 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
2392 an address' error message when a string expansion fails (syntax or
2393 whatever). Otherwise the password may appear in the log. Following change
2394 PH/42 below, there is no longer a chance of it appearing in a bounce
2397 PH/11 Installed exipick version 20050225.0 from John Jetmore.
2399 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
2400 of its addresses was ever tried. (Bugzilla bug #2.)
2402 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
2403 the result incorrectly in the debug output. (It correctly added a newline
2404 to what was transported.)
2406 TF/01 Added $received_time.
2408 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
2409 commented out examples of how to interface to a virus scanner and to
2410 SpamAssassin. Also added commented examples of av_scanner and
2411 spamd_address settings.
2413 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
2414 and controls are allowed in which ACLs. There were a couple of minor
2415 errors. Some of the entries in the conditions table (which is a table of
2416 where they are NOT allowed) were getting very unwieldy; rewrote them as a
2417 negation of where the condition IS allowed.
2419 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
2421 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
2422 header file does not have a version number, so I've had to invent a new
2423 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
2424 API. The code is untested by me (my Linux distribution still has 0.3.2 of
2425 radiusclient), but it was contributed by a Radius user.
2427 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
2428 files or directories, for OpenSSL.
2430 PH/19 When an Exim process that is running as root has to create an Exim log
2431 file, it does so in a subprocess that runs as exim:exim so as to get the
2432 ownership right at creation (otherwise, other Exim processes might see
2433 the file with the wrong ownership). There was no test for failure of this
2434 fork() call, which would lead to the process getting stuck as it waited
2435 for a non-existent subprocess. Forks do occasionally fail when resources
2436 run out. I reviewed all the other calls to fork(); they all seem to check
2439 PH/20 When checking for unexpected SMTP input at connect time (before writing
2440 the banner), Exim was not dealing correctly with a non-positive return
2441 from the read() function. If the client had disconnected by this time,
2442 the result was a log entry for a synchronization error with an empty
2443 string after "input=" when read() returned zero. If read() returned -1
2444 (an event I could not check), uninitialized data bytes were printed.
2445 There were reports of junk text (parts of files, etc) appearing after
2448 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
2450 PH/22 Added support for macro redefinition, and (re)definition in between
2451 driver and ACL definitions.
2453 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
2454 forgetting to use the resulting value; it was using the unexpanded value.
2456 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
2457 hadn't been configured. The fix is from Juergen Kreileder, who
2458 understands it better than I do:
2460 "Here's what I see happening with three configured cyrus_sasl
2461 authenticators configured (plain, login, cram-md5):
2463 On startup auth_cyrus_sasl_init() gets called for each of these.
2464 This means three calls to sasl_listmech() without a specified mech_list.
2465 => SASL tests which mechs of all available mechs actually work
2466 => three warnings about OTP not working
2467 => the returned list contains: plain, login, cram-md5, digest-md5, ...
2469 With the patch, sasl_listmech() also gets called three times. But now
2470 SASL's mech_list option is set to the server_mech specified in the the
2471 authenticator. Or in other words, the answer from sasl_listmech()
2472 gets limited to just the mech you're testing for (which is different
2474 => the return list contains just 'plain' or 'login', 'cram-md5' or
2475 nothing depending on the value of ob->server_mech.
2477 I've just tested the patch: Authentication still works fine,
2478 unavailable mechs specified in the exim configuration are still
2479 caught, and the auth.log warnings about OTP are gone."
2481 PH/25 When debugging is enabled, the contents of the command line are added
2482 to the debugging output, even when log_selector=+arguments is not
2485 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
2486 answer is "GNU", and only if the return is "GNU/something" is the answer
2489 PH/27 $acl_verify_message is now set immediately after the failure of a
2490 verification in an ACL, and so is available in subsequent modifiers. In
2491 particular, the message can be preserved by coding like this:
2493 warn !verify = sender
2494 set acl_m0 = $acl_verify_message
2496 Previously, $acl_verify_message was set only while expanding "message"
2497 and "log_message" when a very denied access.
2499 PH/28 Modified OS/os.c-Linux with
2501 -#ifndef OS_LOAD_AVERAGE
2502 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
2504 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
2505 nomenclature these days.)
2507 PH/29 Installed patch from the Sieve maintainer that adds the options
2508 sieve_useraddress and sieve_subaddress to the redirect router.
2510 PH/30 In these circumstances:
2511 . Two addresses routed to the same list of hosts;
2512 . First host does not offer TLS;
2513 . First host accepts first address;
2514 . First host gives temporary error to second address;
2515 . Second host offers TLS and a TLS session is established;
2516 . Second host accepts second address.
2517 Exim incorrectly logged both deliveries with the TLS parameters (cipher
2518 and peerdn, if requested) that were in fact used only for the second
2521 PH/31 When doing a callout as part of verifying an address, Exim was not paying
2522 attention to any local part prefix or suffix that was matched by the
2523 router that accepted the address. It now behaves in the same way as it
2524 does for delivery: the affixes are removed from the local part unless
2525 rcpt_include_affixes is set on the transport.
2527 PH/32 Add the sender address, as F=<...>, to the log line when logging a
2528 timeout during the DATA phase of an incoming message.
2530 PH/33 Sieve envelope tests were broken for match types other than :is. I have
2531 applied a patch sanctioned by the Sieve maintainer.
2533 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
2534 the uid or gid is negative. A case of a negative gid caused this to be
2535 noticed. The fix allows for either to be negative.
2537 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
2538 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
2541 PH/36 The change PH/12 above was broken. Fixed it.
2543 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
2544 the grounds that routing the same address twice would always produce the
2545 same answer. This might have been true once, but it is certainly no
2546 longer true now. Routing a child address may depend on the previous
2547 routing that produced that child. Some complicated redirection strategies
2548 went wrong when messages had multiple recipients, and made Exim's
2549 behaviour dependent on the order in which the addresses were given.
2551 I have moved the duplicate checking until after the routing is complete.
2552 Exim scans the addresses that are assigned to local and remote
2553 transports, and removes any duplicates. This means that more work will be
2554 done, as duplicates will always all be routed, but duplicates are
2555 presumably rare, so I don't expect this is of any significance.
2557 For deliveries to pipes, files, and autoreplies, the duplicate checking
2558 still happens during the routing process, since they are not going to be
2561 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
2562 It corrects a timeout issue with spamd. This is Ian's comment: "The
2563 background is that sometimes spamd either never reads data from a
2564 connection it has accepted, or it never writes response data. The exiscan
2565 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
2566 blindly assumes that writes won't block so it may never time out."
2568 PH/39 Allow G after quota size as well as K and M.
2570 PH/40 The value set for $authenticated_id in an authenticator may not contain
2571 binary zeroes or newlines because the value is written to log lines and
2572 to spool files. There was no check on this. Now the value is run through
2573 the string_printing() function so that such characters are converted to
2574 printable escape sequences.
2576 PH/41 $message_linecount is a new variable that contains the total number of
2577 lines in the message. Compare $body_linecount, which is the count for the
2580 PH/42 Exim no longer gives details of delivery errors for specific addresses in
2581 bounce and delay warning messages, except in certain special cases, which
2584 (a) An SMTP error message from a remote host;
2585 (b) A message specified in a :fail: redirection;
2586 (c) A message specified in a "fail" command in a system filter;
2587 (d) A message specified in a FAIL return from the queryprogram router;
2588 (e) A message specified by the cannot_route_message router option.
2590 In these cases only, Exim does include the error details in bounce and
2591 warning messages. There are also a few cases where bland messages such
2592 as "unrouteable address" or "local delivery error" are given.
2594 PH/43 $value is now also set for the "else" part of a ${run expansion.
2596 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
2597 being worked on, but at least Exim now implements the latest version to
2600 PH/45 In a pipe transport, although a timeout while waiting for the pipe
2601 process to complete was treated as a delivery failure, a timeout while
2602 writing the message to the pipe was logged, but erroneously treated as a
2603 successful delivery. Such timeouts include transport filter timeouts. For
2604 consistency with the overall process timeout, these timeouts are now
2605 treated as errors, giving rise to delivery failures by default. However,
2606 there is now a new Boolean option for the pipe transport called
2607 timeout_defer, which, if set TRUE, converts the failures into defers for
2608 both kinds of timeout. A transport filter timeout is now identified in
2611 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
2612 systems where "make" and "gmake" are different, calling "gmake" at top
2613 level broke things. I've arranged for the value of $(MAKE) to be passed
2614 from the Makefile to this script so that it can call the same version of
2618 A note about Exim versions 4.44 and 4.50
2619 ----------------------------------------
2621 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
2622 changes of various kinds. As a consequence, a big documentation update was
2623 needed. This delayed the release for rather longer than seemed good, especially
2624 in the light of a couple of (minor) security issues. Therefore, the changes
2625 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
2626 release. So 4.44 and 4.50 are in effect two different branches that both start
2629 I have left the 4.50 change log unchanged; it contains all the changes since
2630 4.43. The change log for 4.44 is below; many of its items are identical to
2631 those for 4.50. This seems to be the most sensible way to preserve the
2632 historical information.
2638 1. Minor wording change to the doc/README.SIEVE file.
2640 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
2641 computation of the current number of files was incorrect.
2643 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
2644 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
2647 4. Give more explanation in the error message when the command for a transport
2648 filter fails to execute.
2650 5. There are several places where Exim runs a non-Exim command in a
2651 subprocess. The SIGUSR1 signal should be disabled for these processes. This
2652 was being done only for the command run by the queryprogram router. It is
2653 now done for all such subprocesses. The other cases are: ${run, transport
2654 filters, and the commands run by the lmtp and pipe transports.
2656 6. Added CONFIGURE_GROUP build-time option.
2658 7. Some older OS have a limit of 256 on the maximum number of file
2659 descriptors. Exim was using setrlimit() to set 1000 as a large value
2660 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
2661 systems. I've change it so that if it can't get 1000, it tries for 256.
2663 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
2664 was an oversight, and furthermore, ever since the addition of extra
2665 controls (e.g. 4.43/32), the checks on when to allow different forms of
2666 "control" were broken. There should now be diagnostics for all cases when a
2667 control that does not make sense is encountered.
2669 9. Added the /retain_sender option to "control=submission".
2671 10. $recipients is now available in the predata ACL (oversight).
2673 11. Tidy the search cache before the fork to do a delivery from a message
2674 received from the command line. Otherwise the child will trigger a lookup
2675 failure and thereby defer the delivery if it tries to use (for example) a
2676 cached ldap connection that the parent has called unbind on.
2678 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
2679 of $address_data from the recipient verification was clobbered by the
2680 sender verification.
2682 13. The value of address_data from a sender verification is now available in
2683 $sender_address_data in subsequent conditions in the ACL statement.
2685 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
2687 15. Added a new option "connect=<time>" to callout options, to set a different
2690 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
2691 was its contents. (It was OK if the option was not defined at all.)
2693 17. A "Completed" log line is now written for messages that are removed from
2694 the spool by the -Mrm option.
2696 18. New variables $sender_verify_failure and $recipient_verify_failure contain
2697 information about exactly what failed.
2699 19. Added -dd to debug only the daemon process.
2701 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
2702 handles timeouts, both on the server side and network timeouts. Renamed the
2703 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
2705 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
2706 It is now set to "smtps".
2708 22. $host_address is now set to the target address during the checking of
2709 ignore_target_hosts.
2711 23. When checking ignore_target_hosts for an ipliteral router, no host name was
2712 being passed; this would have caused $sender_host_name to have been used if
2713 matching the list had actually called for a host name (not very likely,
2714 since this list is usually IP addresses). A host name is now passed as
2717 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
2718 code that specifies a non-restarting handler (typically sigaction() in
2719 modern systems) in an attempt to fix a rare and obscure crash bug.
2721 25. Narrowed the window for a race in the daemon that could cause it to ignore
2722 SIGCHLD signals. This is not a major problem, because they are used only to
2723 wake it up if nothing else does.
2725 26. A malformed maildirsize file could cause Exim to calculate negative values
2726 for the mailbox size or file count. Odd effects could occur as a result.
2727 The maildirsize information is now recalculated if the size or filecount
2730 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
2731 support for a long time. Removed HAVE_SYS_VFS_H.
2733 28. Installed the latest version of exipick from John Jetmore.
2735 29. In an address list, if the pattern was not a regular expression, an empty
2736 subject address (from a bounce message) matched only if the pattern was an
2737 empty string. Non-empty patterns were not even tested. This was the wrong
2738 because it is perfectly reasonable to use an empty address as part of a
2739 database query. An empty address is now tested by patterns that are
2740 lookups. However, all the other forms of pattern expect the subject to
2741 contain a local part and a domain, and therefore, for them, an empty
2742 address still always fails if the pattern is not itself empty.
2744 30. Exim went into a mad DNS loop when attempting to do a callout where the
2745 host was specified on an smtp transport, and looking it up yielded more
2746 than one IP address.
2748 31. Re-factored the code for checking spool and log partition space into a
2749 function that finds that data and another that does the check. The former
2750 is then used to implement four new variables: $spool_space, $log_space,
2751 $spool_inodes, and $log_inodes.
2753 32. The RFC2047 encoding function was originally intended for short strings
2754 such as real names; it was not keeping to the 75-character limit for
2755 encoded words that the RFC imposes. It now respects the limit, and
2756 generates multiple encoded words if necessary. To be on the safe side, I
2757 have increased the buffer size for the ${rfc2047: expansion operator from
2760 33. It is now permitted to omit both strings after an "if" condition; if the
2761 condition is true, the result is "true". As before, when the second string
2762 is omitted, a false condition yields an empty string. This makes it less
2763 cumbersome to write custom ACL and router conditions.
2765 34. Failure to deliver a bounce message always caused it to be frozen, even if
2766 there was an errors_to setting on the router. The errors_to setting is now
2769 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
2770 canonical form (fully expanded) before being placed in
2771 $sender_host_address.
2773 36. The table in the code that translates DNS record types into text (T_A to
2774 "A" for instance) was missing entries for NS and CNAME. It is just possible
2775 that this could have caused confusion if both these types were looked up
2776 for the same domain, because the text type is used as part of Exim's
2777 per-process caching. But the chance of anyone hitting this buglet seems
2780 37. The dnsdb lookup has been extended in a number of ways.
2782 (1) There is a new type, "zns", which walks up the domain tree until it
2783 finds some nameserver records. It should be used with care.
2785 (2) There is a new type, "mxh", which is like "mx" except that it returns
2786 just the host names, not the priorities.
2788 (3) It is now possible to give a list of domains (or IP addresses) to be
2789 looked up. The behaviour when one of the lookups defers can be
2790 controlled by a keyword.
2792 (4) It is now possible to specify the separator character for use when
2793 multiple records are returned.
2795 38. The dnslists ACL condition has been extended: it is now possible to supply
2796 a list of IP addresses and/or domains to be looked up in a particular DNS
2799 39. Added log_selector=+queue_time_overall.
2801 40. When running the queue in the test harness, wait just a tad after forking a
2802 delivery process, to get repeatability of debugging output.
2804 41. Include certificate and key file names in error message when GnuTLS fails
2805 to set them up, because the GnuTLS error message doesn't include the name
2806 of the failing file when there is a problem reading it.
2808 42. Allow both -bf and -bF in the same test run.
2810 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
2812 44. The "Exiscan patch" is now merged into the mainline Exim source.
2814 45. Sometimes the final signoff response after QUIT could fail to get
2815 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
2816 < 0 before doing a fflush(). This bug looks as though it goes back to the
2817 introduction of TLS in release 3.20, but "sometimes" must have been rare
2818 because the tests only now provoked it.
2820 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
2821 (this can affect the format of dates).
2823 47. exim_tidydb, when checking for the continued existence of a message for
2824 which it has found a message-specific retry record, was not finding
2825 messages that were in split spool directories. Consequently, it was
2826 deleting retry records that should have stayed in existence.
2828 48. Steve fixed some bugs in eximstats.
2830 49. The SPA authentication driver was not abandoning authentication and moving
2831 on to the next authenticator when an expansion was forced to fail,
2832 contradicting the general specification for all authenticators. Instead it
2833 was generating a temporary error. It now behaves as specified.
2835 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
2836 (the order specifies the preference for clients). The order is now AES256,
2837 AES128, 3DES, ARCFOUR128.
2839 51. Small patch to Sieve code - explicitly set From: when generating an
2842 52. Exim crashed if a remote delivery caused a very long error message to be
2843 recorded - for instance if somebody sent an entire SpamAssassin report back
2844 as a large number of 550 error lines. This bug was coincidentally fixed by
2845 increasing the size of one of Exim's internal buffers (big_buffer) that
2846 happened as part of the Exiscan merge. However, to be on the safe side, I
2847 have made the code more robust (and fixed the comments that describe what
2850 53. Now that there can be additional text after "Completed" in log lines (if
2851 the queue_time_overall log selector is set), a one-byte patch to exigrep
2852 was needed to allow it to recognize "Completed" as not the last thing in
2855 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
2856 patch that reportedly fixes this has been added. I am not expert enough to
2857 create a test for it. This is what the patch creator wrote:
2859 "I found a little strange behaviour of ldap code when working with
2860 Windows 2003 AD Domain, where users was placed in more than one
2861 Organization Units. When I tried to give exim partial DN, the exit code
2862 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
2863 But simultaneously result of request was absolutely normal ldap result,
2864 so I produce this patch..."
2866 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
2867 REFERENCE, so I have modified the code to exclude the patch when that macro
2870 55. Some experimental protocols are using DNS PTR records for new purposes. The
2871 keys for these records are domain names, not reversed IP addresses. The
2872 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
2873 leaves it alone. Component reversal etc. now happens only for IP addresses.
2876 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
2878 57. Double the size of the debug message buffer (to 2048) so that more of very
2879 long debug lines gets shown.
2881 58. The exicyclog utility now does better if the number of log files to keep
2882 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
2884 59. Two changes related to the smtp_active_hostname option:
2886 (1) $smtp_active_hostname is now available as a variable.
2887 (2) The default for smtp_banner uses $smtp_active_hostname instead
2888 of $primary_hostname.
2890 60. The host_aton() function is supposed to be passed a string that is known
2891 to be a valid IP address. However, in the case of IPv6 addresses, it was
2892 not checking this. This is a hostage to fortune. Exim now panics and dies
2893 if the condition is not met. A case was found where this could be provoked
2894 from a dnsdb PTR lookup with an IPv6 address that had more than 8
2895 components; fortuitously, this particular loophole had already been fixed
2896 by change 4.50/55 above.
2898 If there are any other similar loopholes, the new check in host_aton()
2899 itself should stop them being exploited. The report I received stated that
2900 data on the command line could provoke the exploit when Exim was running as
2901 exim, but did not say which command line option was involved. All I could
2902 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
2903 running as the user.
2906 61. There was a buffer overflow vulnerability in the SPA authentication code
2907 (which came originally from the Samba project). I have added a test to the
2908 spa_base64_to_bits() function which I hope fixes it.
2911 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
2912 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
2914 63. The daemon start-up calls getloadavg() while still root for those OS that
2915 need the first call to be done as root, but it missed one case: when
2916 deliver_queue_load_max is set with deliver_drop_privilege. This is
2917 necessary for the benefit of the queue runner, because there is no re-exec
2918 when deliver_drop_privilege is set.
2920 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
2921 This has been fixed.
2923 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
2924 was in use, was not putting the data itself into the right store pool;
2925 consequently, it could be overwritten for a subsequent message in the same
2926 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
2929 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
2931 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
2932 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
2933 the function were treating the return as a boolean value, which happened to
2934 work because 0=false and not-0=true, but is not correct code.
2936 68. The host_aton() function was not handling scoped IPv6 addresses (those
2937 with, for example, "%eth0" on the end) correctly.
2939 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
2940 negated items (that is, ~something) in unsigned ints. Some compilers
2941 apparently mutter when there is no cast.
2943 70. If an address verification called from an ACL failed, and did not produce a
2944 user-specific message (i.e. there was only a "system" message), nothing was
2945 put in $acl_verify_message. In this situation, it now puts the system
2948 71. Change 4.23/11 added synchronization checking at the start of an SMTP
2949 session; change 4.31/43 added the unwanted input to the log line - except
2950 that it did not do this in the start of session case. It now does.
2952 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
2953 This is wrong and can cause the other end to generate a synchronization
2954 error if it is another Exim or anything else that does the synchronization
2955 check. A QUIT command is no longer sent after a timeout.
2957 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
2958 during host lookups.
2960 74. The defer_ok option of callout verification was not working if it was used
2961 when verifying addresses in header lines, that is, for this case:
2963 verify = header_sender/callout=defer_ok
2965 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
2966 those file descriptors could be used for SMTP connections. If anything
2967 wrote to stderr (the example that came up was "warn" in embedded Perl), it
2968 could be sent to the SMTP client, causing chaos. The daemon now opens
2969 stdin, stdout, and stderr to /dev/null when it puts itself into the
2972 76. Arrange for output from Perl's "warn" command to be written to Exim's main
2973 log by default. The user can override this with suitable Perl magic.
2975 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
2976 the log message when discard triggers, was not working for the DATA ACL or
2977 for the non-SMTP ACL.
2979 78. Error message wording change in sieve.c.
2981 79. If smtp_accept_max_per_host was set, the number of connections could be
2982 restricted to fewer than expected, because the daemon was trying to set up
2983 a new connection before checking whether the processes handling previous
2984 connections had finished. The check for completed processes is now done
2985 earlier. On busy systems, this bug wouldn't be noticed because something
2986 else would have woken the daemon, and it would have reaped the completed
2989 80. If a message was submitted locally by a user whose login name contained one
2990 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
2991 It caused a spool format error. I have fixed the spool reading code. A
2992 related problem was that the "from" clause in the Received: line became
2993 illegal because of the space(s). It is now covered by ${quote_local_part.
2995 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
2998 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
3000 83. Patch from Sieve maintainer to fix unterminated string problem in
3001 "vacation" handling.
3003 84. Some minor changes to the Linux configuration files to help with other
3004 OS variants using glibc.
3006 85. One more patch for Sieve to update vacation handling to latest spec.
3009 ----------------------------------------------------
3010 See the note above about the 4.44 and 4.50 releases.
3011 ----------------------------------------------------
3017 1. Change 4.43/35 introduced a bug that caused file counts to be
3018 incorrectly computed when quota_filecount was set in an appendfile
3021 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
3022 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
3025 3. Give more explanation in the error message when the command for a transport
3026 filter fails to execute.
3028 4. There are several places where Exim runs a non-Exim command in a
3029 subprocess. The SIGUSR1 signal should be disabled for these processes. This
3030 was being done only for the command run by the queryprogram router. It is
3031 now done for all such subprocesses. The other cases are: ${run, transport
3032 filters, and the commands run by the lmtp and pipe transports.
3034 5. Some older OS have a limit of 256 on the maximum number of file
3035 descriptors. Exim was using setrlimit() to set 1000 as a large value
3036 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
3037 systems. I've change it so that if it can't get 1000, it tries for 256.
3039 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
3040 was an oversight, and furthermore, ever since the addition of extra
3041 controls (e.g. 4.43/32), the checks on when to allow different forms of
3042 "control" were broken. There should now be diagnostics for all cases when a
3043 control that does not make sense is encountered.
3045 7. $recipients is now available in the predata ACL (oversight).
3047 8. Tidy the search cache before the fork to do a delivery from a message
3048 received from the command line. Otherwise the child will trigger a lookup
3049 failure and thereby defer the delivery if it tries to use (for example) a
3050 cached ldap connection that the parent has called unbind on.
3052 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
3053 of $address_data from the recipient verification was clobbered by the
3054 sender verification.
3056 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
3057 was its contents. (It was OK if the option was not defined at all.)
3059 11. A "Completed" log line is now written for messages that are removed from
3060 the spool by the -Mrm option.
3062 12. $host_address is now set to the target address during the checking of
3063 ignore_target_hosts.
3065 13. When checking ignore_target_hosts for an ipliteral router, no host name was
3066 being passed; this would have caused $sender_host_name to have been used if
3067 matching the list had actually called for a host name (not very likely,
3068 since this list is usually IP addresses). A host name is now passed as
3071 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
3072 code that specifies a non-restarting handler (typically sigaction() in
3073 modern systems) in an attempt to fix a rare and obscure crash bug.
3075 15. Narrowed the window for a race in the daemon that could cause it to ignore
3076 SIGCHLD signals. This is not a major problem, because they are used only to
3077 wake it up if nothing else does.
3079 16. A malformed maildirsize file could cause Exim to calculate negative values
3080 for the mailbox size or file count. Odd effects could occur as a result.
3081 The maildirsize information is now recalculated if the size or filecount
3084 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
3085 support for a long time. Removed HAVE_SYS_VFS_H.
3087 18. Updated exipick to current release from John Jetmore.
3089 19. Allow an empty sender to be matched against a lookup in an address list.
3090 Previously the only cases considered were a regular expression, or an
3093 20. Exim went into a mad DNS lookup loop when doing a callout where the
3094 host was specified on the transport, if the DNS lookup yielded more than
3097 21. The RFC2047 encoding function was originally intended for short strings
3098 such as real names; it was not keeping to the 75-character limit for
3099 encoded words that the RFC imposes. It now respects the limit, and
3100 generates multiple encoded words if necessary. To be on the safe side, I
3101 have increased the buffer size for the ${rfc2047: expansion operator from
3104 22. Failure to deliver a bounce message always caused it to be frozen, even if
3105 there was an errors_to setting on the router. The errors_to setting is now
3108 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
3109 canonical form (fully expanded) before being placed in
3110 $sender_host_address.
3112 24. Updated eximstats to version 1.33
3114 25. Include certificate and key file names in error message when GnuTLS fails
3115 to set them up, because the GnuTLS error message doesn't include the name
3116 of the failing file when there is a problem reading it.
3118 26. Expand error message when OpenSSL has problems setting up cert/key files.
3121 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
3122 (this can affect the format of dates).
3124 28. exim_tidydb, when checking for the continued existence of a message for
3125 which it has found a message-specific retry record, was not finding
3126 messages that were in split spool directories. Consequently, it was
3127 deleting retry records that should have stayed in existence.
3129 29. eximstats updated to version 1.35
3130 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
3131 1.35 - bugfix such that pie charts by volume are generated correctly
3133 30. The SPA authentication driver was not abandoning authentication and moving
3134 on to the next authenticator when an expansion was forced to fail,
3135 contradicting the general specification for all authenticators. Instead it
3136 was generating a temporary error. It now behaves as specified.
3138 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
3139 (the order specifies the preference for clients). The order is now AES256,
3140 AES128, 3DES, ARCFOUR128.
3142 31. Small patch to Sieve code - explicitly set From: when generating an
3145 32. Exim crashed if a remote delivery caused a very long error message to be
3146 recorded - for instance if somebody sent an entire SpamAssassin report back
3147 as a large number of 550 error lines. This bug was coincidentally fixed by
3148 increasing the size of one of Exim's internal buffers (big_buffer) that
3149 happened as part of the Exiscan merge. However, to be on the safe side, I
3150 have made the code more robust (and fixed the comments that describe what
3153 33. Some experimental protocols are using DNS PTR records for new purposes. The
3154 keys for these records are domain names, not reversed IP addresses. The
3155 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
3156 leaves it alone. Component reversal etc. now happens only for IP addresses.
3159 34. The host_aton() function is supposed to be passed a string that is known
3160 to be a valid IP address. However, in the case of IPv6 addresses, it was
3161 not checking this. This is a hostage to fortune. Exim now panics and dies
3162 if the condition is not met. A case was found where this could be provoked
3163 from a dnsdb PTR lookup with an IPv6 address that had more than 8
3164 components; fortuitously, this particular loophole had already been fixed
3165 by change 4.50/55 or 4.44/33 above.
3167 If there are any other similar loopholes, the new check in host_aton()
3168 itself should stop them being exploited. The report I received stated that
3169 data on the command line could provoke the exploit when Exim was running as
3170 exim, but did not say which command line option was involved. All I could
3171 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
3172 running as the user.
3175 35. There was a buffer overflow vulnerability in the SPA authentication code
3176 (which came originally from the Samba project). I have added a test to the
3177 spa_base64_to_bits() function which I hope fixes it.
3180 36. The daemon start-up calls getloadavg() while still root for those OS that
3181 need the first call to be done as root, but it missed one case: when
3182 deliver_queue_load_max is set with deliver_drop_privilege. This is
3183 necessary for the benefit of the queue runner, because there is no re-exec
3184 when deliver_drop_privilege is set.
3186 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
3187 was in use, was not putting the data itself into the right store pool;
3188 consequently, it could be overwritten for a subsequent message in the same
3189 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
3192 38. Sometimes the final signoff response after QUIT could fail to get
3193 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
3194 < 0 before doing a fflush(). This bug looks as though it goes back to the
3195 introduction of TLS in release 3.20, but "sometimes" must have been rare
3196 because the tests only now provoked it.
3202 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
3203 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
3204 Now it returns an integer. A number of calls were still expecting a T/F
3205 return. Fortuitously, in all cases, the tests worked in OK situations,
3206 which is the norm. However, things would have gone wrong on any write
3207 failures on the smtp file descriptor. This function is used when sending
3208 messages over SMTP and also when doing verify callouts.
3210 2. When Exim is called to do synchronous delivery of a locally submitted
3211 message (the -odf or -odi options), it no longer closes stderr before doing
3214 3. Implemented the mua_wrapper option.
3216 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
3218 5. Implemented the functions header_remove(), header_testname(),
3219 header_add_at_position(), and receive_remove_recipient(), and exported them
3222 6. If an ACL "warn" statement specified the addition of headers, Exim already
3223 inserted X-ACL-Warn: at the start if there was no header name. However, it
3224 was not making this test for the second and subsequent header lines if
3225 there were newlines in the string. This meant that an invalid header could
3226 be inserted if Exim was badly configured.
3228 7. Allow an ACL "warn" statement to add header lines at the start or after all
3229 the Received: headers, as well as at the end.
3231 8. Added the rcpt_4xx retry error code.
3233 9. Added postmaster_mailfrom=xxx to callout verification option.
3235 10. Added mailfrom=xxxx to the callout verification option, for verify=
3238 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
3239 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
3241 12. Inserted some casts to stop certain compilers warning when using pointer
3242 differences as field lengths or precisions in printf-type calls (mostly
3243 affecting debugging statements).
3245 13. Added optional readline() support for -be (dynamically loaded).
3247 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
3248 same clock tick as a message's arrival, so that its received time was the
3249 same as the "first fail" time on the retry record, and that message
3250 remained on the queue past the ultimate address timeout, every queue runner
3251 would try a delivery (because it was past the ultimate address timeout) but
3252 after another failure, the ultimate address timeout, which should have then
3253 bounced the address, did not kick in. This was a "< instead of <=" error;
3254 in most cases the first failure would have been in the next clock tick
3255 after the received time, and all would be well.
3257 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
3258 being recognized when the domain list was tested by the match_domain
3259 condition in an expansion string.
3261 16. Added the ${str2b64: operator.
3263 17. Exim was always calling setrlimit() to set a large limit for the number of
3264 processes, without checking whether the existing limit was already
3265 adequate. (It did check for the limit on file descriptors.) Furthermore,
3266 errors from getrlimit() and setrlimit() were being ignored. Now they are
3267 logged to the main and panic logs, but Exim does carry on, to try to do its
3268 job under whatever limits there are.
3270 18. Imported PCRE 5.0.
3272 19. Trivial typo in log message " temporarily refused connection" (the leading
3275 20. If the log selector return_path_on_delivery was set and an address was
3276 redirected to /dev/null, the delivery process crashed because it assumed
3277 that a return path would always be set for a "successful" delivery. In this
3278 case, the whole delivery is bypassed as an optimization, and therefore no
3281 21. Internal re-arrangement: the function for sending a challenge and reading
3282 a response while authentication was assuming a zero-terminated challenge
3283 string. It's now changed to take a pointer and a length, to allow for
3284 binary data in such strings.
3286 22. Added the cyrus_sasl authenticator (code supplied by MBM).
3288 23. Exim was not respecting finduser_retries when seeking the login of the
3289 uid under which it was called; it was always trying 10 times. (The default
3290 setting of finduser_retries is zero.) Also, it was sleeping after the final
3291 failure, which is pointless.
3293 24. Implemented tls_on_connect_ports.
3295 25. Implemented acl_smtp_predata.
3297 26. If the domain in control=submission is set empty, Exim assumes that the
3298 authenticated id is a complete email address when it generates From: or
3299 Sender: header lines.
3301 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
3302 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
3303 chown and chgrp in /bin and hostname in /usr/bin.
3305 28. Exim was keeping the "process log" file open after each use, just as it
3306 does for the main log. This opens the possibility of it remaining open for
3307 long periods when the USR1 signal hits a daemon. Occasional processlog
3308 errors were reported, that could have been caused by this. Anyway, it seems
3309 much more sensible not to leave this file open at all, so that is what now
3312 29. The long-running daemon process does not normally write to the log once it
3313 has entered its main loop, and it closes the log before doing so. This is
3314 so that log files can straightforwardly be renamed and moved. However,
3315 there are a couple of unusual error situations where the daemon does write
3316 log entries, and I had neglected to close the log afterwards.
3318 30. The text of an SMTP error response that was received during a remote
3319 delivery was being truncated at 512 bytes. This is too short for some of
3320 the long messages that one sometimes sees. I've increased the limit to
3323 31. It is now possible to make retry rules that apply only when a message has a
3324 specific sender, in particular, an empty sender.
3326 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
3327 it possible to be selective about when SMTP synchronization is enforced.
3329 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
3331 32. Implemented hosts_connection_nolog.
3333 33. Added an ACL for QUIT.
3335 34. Setting "delay_warning=" to disable warnings was not working; it gave a
3338 35. Added mailbox_size and mailbox_filecount to appendfile.
3340 36. Added control = no_multiline_responses to ACLs.
3342 37. There was a bug in the logic of the code that waits for the clock to tick
3343 in the case where the clock went backwards by a substantial amount such
3344 that the microsecond fraction of "now" was more than the microsecond
3345 fraction of "then" (but the whole seconds number was less).
3347 38. Added support for the libradius Radius client library this is found on
3348 FreeBSD (previously only the radiusclient library was supported).
3354 1. When certain lookups returned multiple values in the form name=value, the
3355 quoting of the values was not always being done properly. Specifically:
3356 (a) If the value started with a double quote, but contained no whitespace,
3358 (b) If the value contained whitespace other than a space character (i.e.
3359 tabs or newlines or carriage returns) it was not quoted.
3360 This fix has been applied to the mysql and pgsql lookups by writing a
3361 separate quoting function and calling it from the lookup code. The fix
3362 should probably also be applied to nisplus, ibase and oracle lookups, but
3363 since I cannot test any of those, I have not disturbed their existing code.
3365 2. A hit in the callout cache for a specific address caused a log line with no
3366 reason for rejecting RCPT. Now it says "Previous (cached) callout
3367 verification failure".
3369 3. There was an off-by-one bug in the queryprogram router. An over-long
3370 return line was truncated at 256 instead of 255 characters, thereby
3371 overflowing its buffer with the terminating zero. As well as fixing this, I
3372 have increased the buffer size to 1024 (and made a note to document this).
3374 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
3375 when Exim is waiting for an SMTP response from a remote server, Exim
3376 restarts its select() call on the socket, thereby resetting its timeout.
3377 This is not a problem when such interrupts are rare. Somebody set up a cron
3378 job to run exiwhat every 2 minutes, which is less than the normal select()
3379 timeout (5 or 10 minutes). This meant that the select() timeout never
3380 kicked in because it was always reset. I have fixed this by comparing the
3381 time when an interrupt arrives with the time at the start of the first call
3382 to select(). If more time than the timeout has elapsed, the interrupt is
3383 treated as a timeout.
3385 5. Some internal re-factoring in preparation for the addition of Sieve
3386 extensions (by MH). In particular, the "personal" test is moved to a
3387 separate function, and given an option for scanning Cc: and Bcc: (which is
3388 not set for Exim filters).
3390 6. When Exim created an email address using the login of the caller as the
3391 local part (e.g. when creating a From: or Sender: header line), it was not
3392 quoting the local part when it contained special characters such as @.
3394 7. Installed new OpenBSD configuration files.
3396 8. Reworded some messages for syntax errors in "and" and "or" conditions to
3397 try to make them clearer.
3399 9. Callout options, other than the timeout value, were being ignored when
3400 verifying sender addresses in header lines. For example, when using
3402 verify = header_sender/callout=no_cache
3404 the cache was (incorrectly) being used.
3406 10. Added a missing instance of ${EXE} to the exim_install script; this affects
3407 only the Cygwin environment.
3409 11. When return_path_on_delivery was set as a log selector, if different remote
3410 addresses in the same message used different return paths and parallel
3411 remote delivery occurred, the wrong values would sometimes be logged.
3412 (Whenever a remote delivery process finished, the return path value from
3413 the most recently started remote delivery process was logged.)
3415 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
3416 lines when AUTH and/or TLS are in use. This is the "received protocol"
3417 field. Exim used to use "asmtp" for authenticated SMTP, without any
3418 indication (in the protocol name) for TLS use. Now it follows the RFC and
3419 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
3420 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
3421 names appear in log lines as well as in Received: header lines.
3423 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
3424 extensions, and comparison tests, and to fix some bugs.
3426 14. Changes to the "personal" filter test:
3428 (1) The test was buggy in that it was just doing the equivalent of
3429 "contains" tests on header lines. For example, if a user's address was
3430 anne@some.where, the "personal" test would incorrectly be true for
3432 To: susanne@some.where
3434 This test is now done by extracting each address from the header in turn,
3435 and checking the entire address. Other tests that are part of "personal"
3436 are now done using regular expressions (for example, to check local parts
3437 of addresses in From: header lines).
3439 (2) The list of non-personal local parts in From: addresses has been
3440 extended to include "listserv", "majordomo", "*-request", and "owner-*",
3441 taken from the Sieve specification recommendations.
3443 (3) If the message contains any header line starting with "List-" it is
3444 treated as non-personal.
3446 (4) The test for "circular" in the Subject: header line has been removed
3447 because it now seems ill-conceived.
3449 15. Minor typos in src/EDITME comments corrected.
3451 16. Installed latest exipick from John Jetmore.
3453 17. If headers_add on a router specified a text string that was too long for
3454 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
3455 of string_sprintf() is now avoided.
3457 18. $message_body_size was not set (it was always zero) when running the DATA
3458 ACL and the local_scan() function.
3460 19. For the "mail" command in an Exim filter, no default was being set for
3461 the once_repeat time, causing a random time value to be used if "once" was
3462 specified. (If the value happened to be <= 0, no repeat happened.) The
3463 default is now 0s, meaning "never repeat". The "vacation" command was OK
3464 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
3465 (I found it when inspecting the code).
3467 20. There is now an overall timeout for performing a callout verification. It
3468 defaults to 4 times the callout timeout, which applies to individual SMTP
3469 commands during the callout. The overall timeout applies when there is more
3470 than one host that can be tried. The timeout is checked before trying the
3471 next host. This prevents very long delays if there are a large number of
3472 hosts and all are timing out (e.g. when the network connections are timing
3473 out). The value of the overall timeout can be changed by specifying an
3474 additional sub-option for "callout", called "maxwait". For example:
3476 verify = sender/callout=5s,maxwait=20s
3478 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
3479 to the end before writing, but this should make it even safer).
3481 22. Exim was forgetting that it had advertised PIPELINING for the second and
3482 subsequent messages on an SMTP connection. It was also not resetting its
3483 memory on STARTTLS and an internal HELO.
3485 23. When Exim logs an SMTP synchronization error within a session, it now
3486 records whether PIPELINING has been advertised or not.
3488 24. Added 3 instances of "(long int)" casts to time_t variables that were being
3489 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
3490 rather than long int.
3492 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
3494 26. Added the never_mail option to autoreply.
3500 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
3501 crash if the getsockname() call failed; this can happen if a connection is
3502 closed very soon after it is established. The problem was simply in the
3503 order in which certain operations were done, causing Exim to try to write
3504 to the SMTP stream before it had set up the file descriptor. The bug has
3505 been fixed by making things happen in the correct order.
3511 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
3512 before the connection was closed, thus losing the rejection response.
3514 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
3515 some early Solaris releases, but causes trouble in current releases where
3516 socklen_t is defined.
3518 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
3521 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
3524 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
3525 that releasing the top of it at the end releases what was used for sub-
3526 expansions (unless the block got too big). However, discard this block if
3527 the first thing is a variable or header, so that we can use its block when
3528 it is dynamic (useful for very large $message_headers, for example).
3530 6. Lookups now cache *every* query, not just the most recent. A new, separate
3531 store pool is used for this. It can be recovered when all lookup caches are
3532 flushed. Lookups now release memory at the end of their result strings.
3533 This has involved some general refactoring of the lookup sources.
3535 7. Some code has been added to the store_xxx() functions to reduce the amount
3536 of flapping under certain conditions.
3538 8. log_incoming_interface used to affect only the <= reception log lines. Now
3539 it causes the local interface and port to be added to several more SMTP log
3540 lines, for example "SMTP connection from", and rejection lines.
3542 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
3544 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
3546 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
3547 could be overwritten at the end of the current message (or the start of a
3548 new message if it was set in a HELO ACL). The value is now preserved for
3549 the duration of the SMTP connection.
3551 12. If a transport had a headers_rewrite setting, and a matching header line
3552 contained an unqualified address, that address was qualified, even if it
3553 did not match any rewriting rules. The underlying bug was that the values
3554 of the flags that permit the existence of unqualified sender and recipient
3555 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
3556 non-local messages, and by -bnq for local messages) were not being
3557 preserved with the message after it was received.
3559 13. When Exim was logging an SMTP synchronization error, it could sometimes log
3560 "next input=" as part of the text comprising the host identity instead of
3561 the correct text. The code was using the same buffer for two different
3562 strings. However, depending on which order the printing function evaluated
3563 its arguments, the bug did not always show up. Under Linux, for example, my
3564 test suite worked just fine.
3566 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
3567 doesn't work with some older versions of Perl. It has been changed to "my",
3568 which in any case is probably the better facility to use.
3570 15. A really picky compiler found some instances of statements for creating
3571 error messages that either had too many or two few arguments for the format
3574 16. The size of the buffer for calls to the DNS resolver has been increased
3575 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
3576 for addresses that have a lot of PTR records. This alleviates a problem; it
3577 does not fully solve it.
3579 17. A dnsdb lookup for PTR records that receives more data than will fit in the
3580 buffer now truncates the list and logs the incident, which is the same
3581 action as happens when Exim is looking up a host name and its aliases.
3582 Previously in this situation something unpredictable would happen;
3583 sometimes it was "internal error: store_reset failed".
3585 18. If a server dropped the connection unexpectedly when an Exim client was
3586 using GnuTLS and trying to read a response, the client delivery process
3587 crashed while trying to generate an error log message.
3589 19. If a "warn" verb in an ACL added multiple headers to a message in a single
3590 string, for example:
3592 warn message = H1: something\nH2: something
3594 the text was added as a single header line from Exim's point of view
3595 though it ended up OK in the delivered message. However, searching for the
3596 second and subsequent header lines using $h_h2: did not work. This has been
3597 fixed. Similarly, if a system filter added multiple headers in this way,
3598 the routers could not see them.
3600 20. Expanded the error message when iplsearch is called with an invalid key to
3601 suggest using net-iplsearch in a host list.
3603 21. When running tests using -bh, any delays imposed by "delay" modifiers in
3604 ACLs are no longer actually imposed (and a message to that effect is
3607 22. If a "gecos" field in a passwd entry contained escaped characters, in
3608 particular, if it contained a \" sequence, Exim got it wrong when building
3609 a From: or a Sender: header from that name. A second bug also caused
3610 incorrect handling when an unquoted " was present following a character
3611 that needed quoting.
3613 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
3614 was not being matched caselessly.
3616 24. Arranged for all hyphens in the exim.8 source to be escaped with
3619 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
3620 properly. Recipient callout cache records were still being keyed to include
3621 the sender, even when use_sender was set false. This led to far more
3622 callouts that were necessary. The sender is no longer included in the key
3623 when use_sender is false.
3625 26. Added "control = submission" modifier to ACLs.
3627 27. Added the ${base62d: operator to decode base 62 numbers.
3629 28. dnsdb lookups can now access SRV records.
3631 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
3632 the configuration file.
3634 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
3635 (-v) mode. This makes the output for a verbose queue run more intelligible.
3637 31. Added a use_postmaster feature to recipient callouts.
3639 32. Added the $body_zerocount variable, containing the number of binary zero
3640 bytes in the message body.
3642 33. The time of last modification of the "new" subdirectory is now used as the
3643 "mailbox time last read" when there is a quota error for a maildir
3646 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
3648 35. Added +ignore_unknown as a special item in host lists.
3650 36. Code for decoding IPv6 addresses in host lists is now included, even if
3651 IPv6 support is not being compiled. This fixes a bug in which an IPv6
3652 address was recognized as an IP address, but was then not correctly decoded
3653 into binary, causing unexpected and incorrect effects when compared with
3660 1. Very minor rewording of debugging text in manualroute to say "list of
3661 hosts" instead of "hostlist".
3663 2. If verify=header_syntax was set, and a header line with an unqualified
3664 address (no domain) and a large number of spaces between the end of the
3665 name and the colon was received, the reception process suffered a buffer
3666 overflow, and (when I tested it) crashed. This was caused by some obsolete
3667 code that should have been removed. The fix is to remove it!
3669 3. When running in the test harness, delay a bit after writing a bounce
3670 message to get a bit more predictability in the log output.
3672 4. Added a call to search_tidyup() just before forking a reception process. In
3673 theory, someone could use a lookup in the expansion of smtp_accept_max_
3674 per_host which, without the tidyup, could leave open a database connection.
3676 5. Added the variables $recipient_data and $sender_data which get set from a
3677 lookup success in an ACL "recipients" or "senders" condition, or a router
3678 "senders" option, similar to $domain_data and $local_part_data.
3680 6. Moved the writing of debug_print from before to after the "senders" test
3683 7. Change 4.31/66 (moving the time when the Received: is generated) caused
3684 problems for message scanning, either using a data ACL, or using
3685 local_scan() because the Received: header was not generated till after they
3686 were called (in order to set the time as the time of reception completion).
3687 I have revised the way this works. The header is now generated after the
3688 body is received, but before the ACL or local_scan() are called. After they
3689 are run, the timestamp in the header is updated.
3695 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
3696 before starting a queue runner without re-exec. This happened only when
3697 deliver_drop_privilege was set or when the Exim user was set to root. The
3698 effect of the bug was that timeouts during subsequent deliveries caused
3699 crashes instead of being properly handled. The handler is now left at its
3700 default (and expected) setting.
3702 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
3703 message, again when deliver_drop_privilege is set or Exim is run as root.
3704 The bug described in (1) was not present in this case, but the tidying up
3705 of the other signals was missing. I have made the two cases consistent.
3707 3. The ignore_target_hosts setting on a manualroute router was being ignored
3708 for hosts that were looked up using the /MX notation.
3710 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
3713 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
3714 operated on the sender address. After changing the $sender_address to <>
3715 for the sender address verify, Exim was re-instated it as the original
3716 (before rewriting) address, but remembering that it had rewritten it, so it
3717 wasn't rewriting it again. This bug also had the effect of breaking the
3718 sender address verification caching when the sender address was rewritten.
3720 6. The ignore_target_hosts option was being ignored by the ipliteral router.
3721 This has been changed so that if the ip literal address matches
3722 ignore_target_hosts, the router declines.
3724 7. Added expansion conditions match_domain, match_address, and match_local_
3725 part (NOT match_host).
3727 8. The placeholder for the Received: header didn't have a length field set.
3729 9. Added code to Exim itself and to exim_lock to test for a specific race
3730 condition that could lead to file corruption when using MBX delivery. The
3731 issue is with the lockfile that is created in /tmp. If this file is removed
3732 after a process has opened it but before that process has acquired a lock,
3733 there is the potential for a second process to recreate the file and also
3734 acquire a lock. This could lead to two Exim processes writing to the file
3735 at the same time. The added code performs the same test as UW imapd; it
3736 checks after acquiring the lock that its file descriptor still refers to
3737 the same named file.
3739 10. The buffer for building added header lines was of fixed size, 8192 bytes.
3740 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
3743 11. Added the smtp_active_hostname option. If used, this will typically be made
3744 to depend on the incoming interface address. Because $interface_address is
3745 not set up until the daemon has forked a reception process, error responses
3746 that can happen earlier (such as "too many connections") no longer contain
3749 12. If an expansion in a condition on a "warn" statement fails because a lookup
3750 defers, the "warn" statement is abandoned, and the next ACL statement is
3751 processed. Previously this caused the whole ACL to be aborted.
3753 13. Added the iplsearch lookup type.
3755 14. Added ident_timeout as a log selector.
3757 15. Added tls_certificate_verified as a log selector.
3759 16. Added a global option tls_require_ciphers (compare the smtp transport
3760 option of the same name). This controls incoming TLS connections.
3762 17. I finally figured out how to make tls_require_ciphers do a similar thing
3763 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
3764 before starting the TLS session.
3766 18. Tabs are now shown as \t in -bP output.
3768 19. If the log selector return_path_on_delivery was set, Exim crashed when
3769 bouncing a message because it had too many Received: header lines.
3771 20. If two routers both had headers_remove settings, and the first one included
3772 a superfluous trailing colon, the final name in the first list and the
3773 first name in the second list were incorrectly joined into one item (with a
3774 colon in the middle).
3780 1. Added -C and -D options to the exinext utility, mainly to make it easier
3781 to include in the automated testing, but these could be helpful when
3782 multiple configurations are in use.
3784 2. The exinext utility was not formatting the output nicely when there was
3785 an alternate port involved in the retry record key, nor when there was a
3786 message id as well (for retries that were specific to a specific message
3787 and a specific host). It was also confused by IPv6 addresses, because of
3788 the additional colons they contain. I have fixed the IPv4 problem, and
3789 patched it up to do a reasonable job for IPv6.
3791 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
3792 delivery, the log line now contains "pipelined" if PIPELINING was used.
3794 4. An SMTP transport process used to panic and die if the bind() call to set
3795 an explicit outgoing interface failed. This has been changed; it is now
3796 treated in the same way as a connect() failure.
3798 5. A reference to $sender_host_name in the part of a conditional expansion
3799 that was being skipped was still causing a DNS lookup. This no longer
3802 6. The def: expansion condition was not recognizing references to header lines
3803 that used bh_ and bheader_.
3805 7. Added the _cache feature to named lists.
3807 8. The code for checking quota_filecount in the appendfile transport was
3808 allowing one more file than it should have been.
3810 9. For compatibility with Sendmail, the command line option
3818 and sets the incoming protocol and host name (for trusted callers). The
3819 host name and its colon can be omitted when only the protocol is to be set.
3820 Note the Exim already has two private options, -pd and -ps, that refer to
3821 embedded Perl. It is therefore impossible to set a protocol value of "d" or
3822 "s", but I don't think that's a major issue.
3824 10. A number of refactoring changes to the code, none of which should affect
3827 (a) The number of logging options was getting close to filling up the
3828 32-bit word that was used as a bit map. I have split them into two classes:
3829 those that are passed in the argument to log_write(), and those that are
3830 only ever tested independently outside of that function. These are now in
3831 separate 32-bit words, so there is plenty of room for expansion again.
3832 There is no change in the user interface or the logging behaviour.
3834 (b) When building, for example, log lines, the code previously used a
3835 macro that called string_cat() twice, in order to add two strings. This is
3836 not really sufficiently general. Furthermore, there was one instance where
3837 it was actually wrong because one of the argument was used twice, and in
3838 one call a function was used. (As it happened, calling the function twice
3839 did not affect the overall behaviour.) The macro has been replaced by a
3840 function that can join an arbitrary number of extra strings onto a growing
3843 (c) The code for expansion conditions now uses a table and a binary chop
3844 instead of a serial search (which was left over from when there were very
3845 few conditions). Also, it now recognizes conditions like "pam" even when
3846 the relevant support is not compiled in: a suitably worded error message is
3847 given if an attempt is made to use such a condition.
3849 11. Added ${time_interval:xxxxx}.
3851 12. A bug was causing one of the ddress fields not to be passed back correctly
3852 from remote delivery subprocesses. The field in question was not being
3853 subsequently used, so this caused to problems in practice.
3855 13. Added new log selectors queue_time and deliver_time.
3857 14. Might have fixed a bug in maildirsizefile handling that threw up
3858 "unexpected character" debug warnings, and recalculated the data
3859 unnecessarily. In any case, I expanded the warning message to give more
3862 15. Added the message "Restricted characters in address" to the statements in
3863 the default ACL that block characters like @ and % in local parts.
3865 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
3866 Three changes have been made:
3868 (a) There was a serious bug; a negative response to MAIL caused the whole
3869 recipient domain to be cached as invalid, thereby blocking all messages
3870 to all local parts at the same domain, from all senders. This bug has
3871 been fixed. The domain is no longer cached after a negative response to
3872 MAIL if the sender used is not empty.
3874 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
3877 (c) A new callout option, "use_sender" has been added for people who want
3878 the modified behaviour.
3884 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
3887 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
3888 indeed breaks things for older releases.
3890 3. Added additional logging to the case where there is a problem reading data
3891 from a filter that is running in a subprocess using a pipe, in order to
3892 try to track down a specific problem.
3894 4. Testing facility fudge: when running in the test harness and attempting
3895 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
3896 getting "No route to host". Convert this to a timeout.
3898 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
3901 6. Some OS don't have socklen_t but use size_t instead. This affects the
3902 fifth argument of getsockopt() amongst other things. This is now
3903 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
3904 can be set for individual OS. I have set it for SunOS5, OSF1, and
3905 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
3906 some earlier ones do not.
3908 7. Change 4.30/15 was not doing the test caselessly.
3910 8. The standard form for an IPv6 address literal was being rejected by address
3911 parsing in, for example, MAIL and RCPT commands. An example of this kind of
3912 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
3913 this, as well as the form without the "IPv6" on the front (but only when
3914 address literals are enabled, of course).
3916 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
3918 10. Exim crashed if a message with an empty sender address specified by -f
3919 encountered a router with an errors_to setting. This could be provoked only
3920 by a command such as
3924 where an empty string was supplied; "<>" did not hit this bug.
3926 11. Installed PCRE release 4.5.
3928 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
3929 remained set. It is now erased.
3931 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
3932 times from message ids (which are base 36 rather than the normal 62).
3934 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
3935 were being counted as actual protocol errors, and logged if the log
3936 selector +smtp_protocol_error was set. One cannot be perfect in this test,
3937 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
3938 and DATA following a set of rejected RCPTs do not count as protocol errors.
3939 In other words, Exim assumes they were pipelined, though this may not
3940 actually be the case. Of course, in all cases the client gets an
3941 appropriate error code.
3943 15. If a lookup fails in an ACL condition, a message about the failure may
3944 be available; it is used if testing the ACL cannot continue, because most
3945 such messages specify what the cause of the deferral is. However, some
3946 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
3947 that caused an old message to be retained and used if a later statement
3948 caused a defer, replacing the real cause of the deferral.
3950 16. If an IP address had so many PTR records that the DNS lookup buffer
3951 was not large enough to hold them, Exim could crash while trying to process
3952 the truncated data. It now detects and logs this case.
3954 17. Further to 4.21/58, another change has been made: if (and only if) the
3955 first line of a message (the first header line) ends with CRLF, a bare LF
3956 in a subsequent header line has a space inserted after it, so as not to
3957 terminate the header.
3959 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
3960 unnecessarily, used atoi() instead of strtol(), and didn't check the
3961 termination when getting file sizes from file names by regex.
3963 19. Completely re-implemented the support for maildirsize files, in the light
3964 of a number of problems with the previous contributed implementation
3965 (4.30/29). In particular:
3967 . If the quota is zero, the maildirsize file is maintained, but no quota is
3970 . If the maildir directory does not exist, it is created before any attempt
3971 to write a maildirsize file.
3973 . The quota value in the file is just a cache; if the quota is changed in
3974 the transport, the new value overrides.
3976 . A regular expression is available for excluding directories from the
3979 20. The autoreply transport checks the characters in options that define the
3980 message's headers; it allows continued headers, but it was checking with
3981 isspace() after an embedded newline instead of explicitly looking for a
3984 21. If all the "regular" hosts to which an address was routed had passed their
3985 expiry times, and had not reached their retry times, the address was
3986 bounced, even if fallback hosts were defined. Now Exim should go on to try
3989 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
3990 equivalent code in the SMTP transport. Some hosts send humungous responses
3991 to HELO/EHLO, more than 1024 it seems.
3993 23. Refactoring: code in filter.c used (void *) for "any old type" but this
3994 gives compiler warnings in some environments. I've now done it "properly",
3997 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
3998 (because of problems with the built-in one) was declared to return uschar *
3999 instead of char *, causing compiler failure.
4001 25. Fixed a file descriptor leak when processing alias/forward files.
4003 26. Fixed a minor format string issue in dbfn.c.
4005 27. Typo in exim.c: ("dmbnz" for "dbmnz").
4007 28. If a filter file refered to $h_xxx or $message_headers, and the headers
4008 contained RFC 2047 "words", Exim's memory could, under certain conditions,
4011 29. When a sender address is verified, it is cached, to save repeating the test
4012 when there is more than one recipient in a message. However, when the
4013 verification involves a callout, it is possible for different callout
4014 options to be set for different recipients. It is too complicated to keep
4015 track of this in the cache, so now Exim always runs a verification when a
4016 callout is required, relying on the callout cache for the optimization.
4017 The overhead is duplication of the address routing, but this should not be
4020 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
4021 to be verified with callout=postmaster, and the main callout worked but the
4022 postmaster check failed, the verification correctly failed. However, if a
4023 subsequent RCPT command asked for sender verification *without* the
4024 postmaster check, incorrect caching caused this verification also to fail,
4027 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
4028 it was not caching the DNS options (qualify_single, search_parents) that
4029 were used when the lookup failed. A subsequent lookup with different
4030 options therefore always gave the same answer, though there were cases
4031 where it should not have. (Example: a "domains = !$mx_any" option on a
4032 dnslookup router: the "domains" option is always processed without any
4033 widening, but the router might have qualify_single set.) Now Exim uses the
4034 cached value only when the same options are set.
4036 32. Added John Jetmore's "exipick" utility to the distribution.
4038 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
4039 than a timeout (e.g. a certificate is required, and is not provided), an
4040 Exim server now closes the connection immediately. Previously it waited for
4041 the client to close - but if the client is SSL, it seems that they each
4042 wait for each other, leading to a delay before one of them times out.
4044 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
4045 maintained 0.8.x compatibility because I don't think many are using it, and
4046 it is clearly obsolete.
4048 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
4051 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
4052 client certificate was expired. A simple patch fixes this, though I don't
4053 understand the full logic of why the verify callback is called multiple
4056 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
4057 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
4058 which causes problems with some clients (such as the Certicom SSL Plus
4059 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
4060 disables the coutermeasure allowing Eudora to connect."
4062 38. Exim was not checking that a write() to a log file succeeded. This could
4063 lead to Bad Things if a log got too big, in particular if it hit a file
4064 size limit. Exim now panics and dies if it cannot write to a log file, just
4065 as it does if it cannot open a log file.
4067 39. Modified OS/Makefile-Linux so that it now contains
4069 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
4071 The two -D definitions ensure that Exim is compiled with large file
4072 support, which makes it possible to handle log files that are bigger than
4075 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
4076 instance) a domain was checked against a named list that involved a lookup,
4077 causing $domain_data to be set, then another domain was checked against the
4078 same list, then the first domain was re-checked, the value of $domain_data
4079 after the final check could be wrong. In particular, if the second check
4080 failed, it could be set empty. This bug probably also applied to
4083 41. The strip_trailing_dot option was not being applied to the address given
4084 with the -f command-line option.
4086 42. The code for reading a message's header from the spool was incrementing
4087 $received_count, but never initializing it. This meant that the value was
4088 incorrect (doubled) while delivering a message in the same process in which
4089 it was received. In the most common configuration of Exim, this never
4090 happens - a fresh exec is done - but it can happen when
4091 deliver_drop_privilege is set.
4093 43. When Exim logs an SMTP synchronization error - client data sent too soon -
4094 it now includes up to 150 characters of the unexpected data in the log
4097 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
4098 and building data strings. The size of both of these buffers was 10 000
4099 bytes - far larger than anybody would *ever* want, thought I. Needless to
4100 say, somebody hit the limit. I have increased the maximum line length to
4101 20 000 and the maximum data length of concatenated lines to 100 000. I have
4102 also fixed two bugs, because there was no checking on these buffers. Tsk,
4103 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
4104 buffer is too small.
4106 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
4107 lsearch lookups. Now it does.
4109 46. When parsing a route_list item in a manualroute router, a fixed-length
4110 buffer was used for the list of hosts. I made this 1024 bytes long,
4111 thinking that nobody would ever have a list of hosts that long. Wrong.
4112 Somebody had a whole pile of complicated expansion conditions, and the
4113 string was silently truncated, leading to an expansion error. It turns out
4114 that it is easier to change to an unlimited length (owing to other changes
4115 that have happened since this code was originally written) than to build
4116 structure for giving a limitation error. The length of the item that
4117 expands into the list of hosts is now unlimited.
4119 47. The lsearch lookup could not handle data where the length of text line was
4120 more than 4095 characters. Such lines were truncated, leading to shortened
4121 data being returned. It should now handle lines of any length.
4123 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
4124 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
4127 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
4128 replacing strings such as BIN_DIRECTORY in a source file: the replacement
4129 no longer happens in comment lines. A list of replacements is now placed
4130 at the head of all of the source files, except those whose only change is
4131 to replace PERL_COMMAND in the very first #! line.
4133 50. Replaced the slow insertion sort in queue.c, for sorting the list of
4134 messages on the queue, with a bottom-up merge sort, using code contributed
4135 by Michael Haardt. This should make operations like -bp somewhat faster on
4136 large queues. It won't affect queue runners, except when queue_run_in_order
4139 51. Installed eximstats 1.31 in the distribution.
4141 52. Added support for SRV lookups to the dnslookup router.
4143 53. If an ACL referred to $message_body or $message_body_end, the value was not
4144 reset for any messages that followed in the same SMTP session.
4146 54. The store-handling optimization for building very long strings was not
4147 differentiating between the different store pools. I don't think this
4148 actually made any difference in practice, but I've tidied it.
4150 55. While running the routers to verify a sender address, $sender_address
4151 was still set to the sender address. This is wrong, because when routing to
4152 send a bounce to the sender, it would be empty. Therefore, I have changed
4153 it so that, while verifying a sender address, $sender_address is set to <>.
4154 (There is no change to what happens when verifying a recipient address.)
4156 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
4157 target A or AAAA records (if not already returned) without resetting the
4158 qualify_single or search_parents options of the DNS resolver. These are
4159 inappropriate in this case because the targets of MX and SRV records must
4160 be FQDNs. A broken DNS record could cause trouble if it happened to have a
4161 target that, when qualified, matched something in the local domain. These
4162 two options are now turned off when doing these lookups.
4164 57. It seems that at least some releases of Reiserfs (which does not have the
4165 concept of a fixed number of inodes) returns zero and not -1 for the
4166 number of available inodes. This interacted badly with check_spool_inodes,
4167 which assumed that -1 was the "no such thing" setting. What I have done is
4168 to check that the total number of inodes is greater than zero before doing
4169 the test of how many are available.
4171 58. When a "warn" ACL statement has a log_message modifier, the message is
4172 remembered, and not repeated. This is to avoid a lot of repetition when a
4173 message has many recipients that cause the same warning to be written.
4174 Howewer, Exim was preserving the list of already written lines for an
4175 entire SMTP session, which doesn't seem right. The memory is now reset if a
4176 new message is started.
4178 59. The "rewrite" debugging flag was not showing the result of rewriting in the
4179 debugging output unless log_rewrite was also set.
4181 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
4182 of (int)(handle) when we know that handle contains (void *)(-1).
4184 61. The Exim daemon panic-logs an error return when it closes the incoming
4185 connection. However "connection reset by peer" seems to be common, and
4186 isn't really an error worthy of noting specially, so that particular error
4189 62. When Exim is trying to find all the local interfaces, it used to panic and
4190 die if the ioctl to get the interface flags failed. However, it seems that
4191 on at least one OS (Solaris 9) it is possible to have an interface that is
4192 included in the list of interfaces, but for which you get a failure error
4193 for this call. This happens when the interface is not "plumbed" into a
4194 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
4195 failure of the "get flags" call assumes that the interface is down.
4197 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
4198 makes life easier for people who are doing arithmetic on fields extracted
4199 from dates, where you often get leading zeros that should not be
4200 interpreted as octal.
4202 64. Added qualify_domain to the redirect router, to override the global
4205 65. If a pathologically long header line contained very many addresses (the
4206 report of this problem mentioned 10 000) and each of them was rewritten,
4207 Exim could use up a very large amount of memory. (It kept on making new
4208 copies of the header line as it rewrote, and never released the old ones.)
4209 At the expense of a bit more processing, the header rewriting function has
4210 been changed so that it no longer eats memory in this way.
4212 66. The generation of the Received: header has been moved from the time that a
4213 message starts to be received, to the time that it finishes. The timestamp
4214 in the Received: header should now be very close to that of the <= log
4215 line. There are two side-effects of this change:
4217 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
4218 logged header lines no longer include the local Received: line, because
4219 it has not yet been created. The same applies to a copy of the message
4220 that is returned to a non-SMTP sender when a message is rejected.
4222 (b) When a filter file is tested using -bf, no additional Received: header
4223 is added to the test message. After some thought, I decided that this
4226 This change does not affect the value of $received_for. It is still set
4227 after address rewriting, but before local_scan() is called.
4229 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
4231 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
4232 gave an unhelpful panic error message, and a defer error. I have managed to
4233 change this behaviour so that it now rejects any supplied certificate,
4234 which seems right, as the list of acceptable certificates is empty.
4236 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
4237 gave an unhelpful defer error. I have not managed to make this reject any
4238 supplied certificates, but the error message it gives is "no certificate
4239 supplied", which is not helpful.
4241 70. exigrep's output now also includes lines that are not associated with any
4242 message, but which match the given pattern. Implemented by a patch from
4243 Martin Sluka, which also tidied up the Perl a bit.
4245 71. Recipient callout verification, like sender verification, was using <> in
4246 the MAIL FROM command. This isn't really the right thing, since the actual
4247 sender may affect whether the remote host accepts the recipient or not. I
4248 have changed it to use the actual sender in the callout; this means that
4249 the cache record is now keyed on a recipient/sender pair, not just the
4250 recipient address. There doesn't seem to be a real danger of callout loops,
4251 since a callout by the remote host to check the sender would use <>.
4252 [SEE ABOVE: changed after hitting problems.]
4254 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
4255 temporary errors. However, in the case of such a code being given after
4256 the end of a data transmission (i.e. after ".") Exim was failing to write
4257 a retry record for the message. (Yes, there was some broken host that was
4258 actually sending 8xx at this point.)
4260 73. An unknown lookup type in a host list could cause Exim to panic-die when
4261 the list was checked. (An example that provoked this was putting <; in the
4262 middle of a list instead of at the start.) If this happened during a DATA
4263 ACL check, a -D file could be left lying around. This kind of configuration
4264 error no longer causes Exim to die; instead it causes a defer errror. The
4265 incident is still logged to the main and panic logs.
4267 74. Buglet left over from Exim 3 conversion. The message "too many messages
4268 in one connection" was written to the rejectlog but not the mainlog, except
4269 when address rewriting (yes!) was being logged.
4271 75. Added write_rejectlog option.
4273 76. When a system filter was run not as root (that is, when system_filter_user
4274 was set), the values of the $n variables were not being returned to the
4275 main process; thus, they were not subsequently available in the $sn
4278 77. Added +return_path_on_delivery log selector.
4280 78. A connection timeout was being treated differently from recipients deferred
4281 when testing hosts_max_try with a message that was older than the host's
4282 retry timeout. (The host should not be counted, thus allowing all hosts to
4283 be tried at least once before bouncing.) This may have been the cause of an
4284 occasionally reported bug whereby a message would remain on the queue
4285 longer than the retry timeout, but would be bounced if a delivery was
4286 forced. I say "may" because I never totally pinned down the problem;
4287 setting up timeout/retry tests is difficult. See also the next item.
4289 79. The ultimate address timeout was not being applied to errors that involved
4290 a combination of host plus message (for example, a timeout on a MAIL
4291 command). When an address resolved to a number of possible hosts, and they
4292 were not all tried for each delivery (e.g. because of hosts_max_try), a
4293 message could remain on the queue longer than the retry timeout.
4295 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
4298 81. Fixed an obscure SMTP outgoing bug which required at least the following
4299 conditions: (a) there was another message waiting for the same server;
4300 (b) the server returned 5xx to all RCPT commands in the first message so
4301 that the message was not completed; (c) the server dropped the connection
4302 or gave a negative response to the RSET that Exim sends to abort the
4303 transaction. The observed case was a dropped connection after DATA that had
4304 been sent in pipelining mode. That is, the server had advertised PIPELINING
4305 but was not implementing it correctly. The effect of the bug was incorrect
4306 behaviour, such as trying another host, and this could lead to a crash.
4312 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
4313 and daemon.c were passed as pointers to ints; they should have been
4314 pointers to socklen_t variables (which are typically unsigned ints).
4316 2. Some signed/unsigned type warnings in the os.c file for Linux have been
4319 3. Fixed a really odd bug that affected only the testing scheme; patching a
4320 certain fixed string in the binary changed the value of another string that
4321 happened to be identical to the end of the original first string.
4323 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
4324 name", it returns that address as the IP address. On some operating
4325 systems (e.g. Solaris), it also passes back the IP address string as the
4326 "host name". However, on others (e.g. Linux), it passes back an empty
4327 string. Exim wasn't checking for this, and was changing the host name to an
4328 empty string, assuming it had been canonicized.
4330 5. Although rare, it is permitted to have more than one PTR record for a given
4331 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
4332 all the names associated with an address, because they do in Solaris.
4333 However, it seems that they do not in Linux for data that comes from the
4334 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
4335 I found this out when I moved to a new Linux workstation and tried to run
4336 the Exim test suite.
4338 To get round this problem I have changed the code so that it now does its
4339 own call to the DNS to look up PTR records when searching for a host name.
4340 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
4341 addresses that are only in /etc/hosts are still found.
4343 This behaviour is, however, controlled by an option called host_lookup_
4344 order, which defaults to "bydns:byaddr". If people want to use the other
4345 order, or indeed, just use one or the other means of lookup, they can
4346 specify it in this variable.
4348 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
4349 some operating systems, this comes back from gethostbyaddr() as an empty
4350 string, and this is what Exim used to test for. However, it seems that in
4351 other systems, "." is yielded. Exim now tests for this case too.
4353 7. The values of check_spool_space and check_log_space are now held internally
4354 as a number of kilobytes instead of an absolute number of bytes. If a
4355 numbers is specified without 'K' or 'M', it is rounded up to the nearest
4356 kilobyte. This means that much larger values can be stored.
4358 8. Exim monitor: an attempt to get the action menu when not actually pointing
4359 at a message produces an empty menu entitled "No message selected". This
4360 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
4361 no entries in it ("Shell widget menu has zero width and/or height"). So I
4362 have added a single, blank menu entry in this case.
4364 9. Added ${quote_local_part.
4366 10. MIME decoding is now applied to the contents of Subject: header lines when
4369 11. Now that a reference to $sender_host_address automatically causes a reverse
4370 lookup to occur if necessary (4.13/18), there is no need to arrange for a
4371 host lookup before query-style lookups in lists that might use this
4372 variable. This has therefore been abolished, and the "net-" prefix is no
4373 longer necessary for query-style lookups.
4375 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
4376 have been a typo for LFLAGS, so it has been changed.
4378 13. The install script calls Exim with "-C /dev/null" in order to find the
4379 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
4380 to be output. Howeve, since Exim outputs its version number before the
4381 error, it didn't break the script. It just looked ugly. I fixed this by
4382 always allowing "-C /dev/null" if the caller is root.
4384 14. Ignore overlarge ACL variable number when reading spool file - insurance
4385 against a later release with more variables having written the file.
4387 15. The standard form for an IPv6 address literal was being rejected by EHLO.
4388 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
4389 this, as well as the form without the "IPv6" on the front.
4391 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
4392 OS/Makefile-Darwin file.
4394 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
4395 only to LDAP libraries that do not have LDAP_OPT_DEREF.
4397 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
4398 variable. However, some OS use ints for inodes. I've added cast to long int
4399 to get rid of the compiler warning.
4401 19. I had forgotten to lock out "/../" in configuration file names when
4402 ALT_CONFIG_PREFIX was set.
4404 20. Routers used for verification do not need to specify transports. However,
4405 if such a router generated a host list, and callout was configured, Exim
4406 crashed, because it could not find a port number from the (non-existent)
4407 transport. It now assumes port 25 in this circumstance.
4409 21. Added the -t option to exigrep.
4411 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
4412 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
4413 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
4414 them is set, LOOKUP_LSEARCH is forced.
4416 23. "exim -bV" now outputs a list of lookups that are included in the binary.
4418 24. Added sender and host information to the "rejected by local_scan()" log
4419 line; previously there was no indication of these.
4421 25. Added .include_if_exists.
4423 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
4424 Linux. It turns out that not all file systems support this. Apparently some
4425 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
4426 do it.) To cope with this, the error EINVAL, which means that sync-ing is
4427 not supported on the file descriptor, is now ignored when Exim is trying to
4428 sync a directory. This applies only to Linux.
4430 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
4432 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
4433 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
4436 29. Added support for maildirsize files from supplied patch (modified a bit).
4438 30. The use of :fail: followed by an empty string could lead Exim to respond to
4439 sender verification failures with (e.g.):
4441 550 Verification failed for <xxx>
4442 550 Sender verify failed
4444 where the first response line was missing the '-' that indicates it is not
4445 the final line of the response.
4447 31. The loop for finding the name of the user that called Exim had a hardwired
4448 limit of 10; it now uses the value of finduser_retries, which is used for
4449 all other user lookups.
4451 32. Added $received_count variable, available in data and not_smtp ACLs, and at
4454 33. Exim was neglecting to zero errno before one call of strtol() when
4455 expanding a string and expecting an integer value. On some systems this
4456 resulted in spurious "integer overflow" errors. Also, it was casting the
4457 result into an int without checking.
4459 34. Testing for a connection timeout using "timeout_connect" in the retry rules
4460 did not work. The code looks as if it has *never* worked, though it appears
4461 to have been documented since at least releast 1.62. I have made it work.
4463 35. The "timeout_DNS" error in retry rules, also documented since at least
4464 1.62, also never worked. As it isn't clear exactly what this means, and
4465 clearly it isn't a major issue, I have abolished the feature by treating it
4466 as "timeout", and writing a warning to the main and panic logs.
4468 36. The display of retry rules for -brt wasn't always showing the error code
4471 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
4472 timeout_connect_A, timeout_connect_MX.
4474 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
4475 to the empty sender.
4477 39. The daemon was not analysing the content of -oX till after it had closed
4478 stderr and disconnected from the controlling terminal. This meant that any
4479 syntax errors were only noted on the panic log, and the return code from
4480 the command was 0. By re-arranging the code a little, I've made the
4481 decoding happen first, so such errors now appear on stderr, and the return
4482 code is 1. However, the actual setting up of the sockets still happens in
4483 the disconnected process, so errors there are still only recorded on the
4486 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
4487 connections (as happens on some IP stacks) was logged at start up time as
4488 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
4489 it from "IPv6 and IPv4", which means that two separate sockets are being
4492 41. The debug output for gethostbyname2() or getipnodebyname() failures now
4493 says whether AF_INET or AF_INET6 was passed as an argument.
4495 42. Exiwhat output was messed up when time zones were included in log
4498 43. Exiwhat now gives more information about the daemon's listening ports,
4499 and whether -tls-on-connect was used.
4501 44. The "port" option of the smtp transport is now expanded.
4503 45. A "message" modifier in a "warn" statement in a non-message ACL was being
4504 silently ignored. Now an error message is written to the main and panic
4507 46. There's a new ACL modifier called "logwrite" which writes to a log file
4508 as soon as it is encountered.
4510 47. Added $local_user_uid and $local_user_gid at routing time.
4512 48. Exim crashed when trying to verify a sender address that was being
4515 49. Exim was recognizing only a space character after ".include". It now also
4516 recognizes a tab character.
4518 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
4519 extracting the relevant information from the specification. The man page no
4520 longer contains scrambled data for the -d option, and I've added a section
4521 at the front about calling Exim under different names.
4523 51. Added "extra_headers" argument to the "mail" command in filter files.
4525 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
4528 53. Installed eximstats 1.29.
4530 54. Added transport_filter_timeout as a generic transport option.
4532 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
4533 Cc: header lines. This was required by RFC 822, but it not required by RFC
4536 56. Exim used to add From:, Date:, and Message-Id: header lines to any
4537 incoming messages that did not have them. Now it does so only if the
4538 message originates locally, that is, if there is no associated remote host
4539 address. When Resent- header lines are present, this applies to the Resent-
4540 lines rather than the non-Resent- lines.
4542 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
4543 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
4545 58. Messages for configuration errors now include the name of the main
4546 configuration file - useful now that there may be more than one file in a
4547 list (.included file names were always shown).
4549 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
4550 for those rare installations that do not start the daemon as root or run it
4551 setuid root. I've cut out the call to initgroups() if the daemon is not
4554 60. The Exim user and group can now be bound into the binary as text strings
4555 that are looked up at the start of Exim's processing.
4557 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
4559 62. Added $mailstore_basename variable.
4561 63. Installed patch to sieve.c from Michael Haardt.
4563 64. When Exim failed to open the panic log after failing to open the main log,
4564 the original message it was trying to log was written to stderr and debug
4565 output, but if they were not available (the usual case in production), it
4566 was lost. Now it is written to syslog before the two lines that record the
4567 failures to open the logs.
4569 65. Users' Exim filters run in subprocesses under the user's uid. It is
4570 possible for a "deliver" command or an alias in a "personal" command to
4571 provoke an address rewrite. If logging of address rewriting is configured,
4572 this fails because the process is not running as root or exim. There may be
4573 a better way of dealing with this, but for the moment (because 4.30 needs
4574 to be released), I have disabled address rewrite logging when running a
4575 filter in a non-root, non-exim process.
4581 1. The buildconfig auxiliary program wasn't quoting the value set for
4582 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
4583 not defined. This bug was masked in 4.22 by the effect that was fixed in
4586 2. Some messages that were rejected after a message id was allocated were
4587 shown as "incomplete" by exigrep. It no longer does this for messages that
4588 are rejected by local_scan() or the DATA or non-SMTP ACLs.
4590 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
4591 have allow_domain_literals set, the ID did not get logged in the <= line.
4592 Domain literals are now always recognized in Message-ID: header lines.
4594 4. The first argument for a ${extract expansion item is the key name or field
4595 number. Leading and trailing spaces in this item were not being ignored,
4596 causing some misleading effects.
4598 5. When deliver_drop_privilege was set, single queue runner processes started
4599 manually (i.e. by the command "exim -q") or by the daemon (which uses the
4600 same command in the process it spins off) were not dropping privilege.
4602 6. When the daemon running as "exim" started a queue runner, it always
4603 re-executed Exim in the spun-off process. This is a waste of effort when
4604 deliver_drop_privilege is set. The new process now just calls the
4605 queue-runner function directly.
4611 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
4614 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
4615 ignored. Though the use of -oP was forcing the writing of a pid file, it
4616 was always written to the default place.
4618 3. If the message "no IP address found for host xxxx" is generated during
4619 incoming verification, it is now followed by identification of the incoming
4620 connection (so you can more easily find what provoked it).
4622 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
4624 5. Added some features to "harden" Exim a bit more against certain attacks:
4626 (a) There is now a build-time option called FIXED_NEVER_USERS that can
4627 be put in Local/Makefile. This is like the never_users runtime option,
4628 but it cannot be overridden. The default setting is "root".
4630 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
4631 prefix string with which any file named in a -C command line option
4634 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
4635 is retained for -C and -D only if the caller of Exim is root. Without
4636 it, the exim user may also use -C and -D and retain privilege.
4638 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
4639 command line option is disabled.
4641 6. Macro names set by the -D option must start with an upper case letter, just
4642 like macro names defined in the configuration file.
4644 7. Added "dereference=" facility to LDAP.
4646 8. Two instances of the typo "uknown" in the source files are fixed.
4648 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
4649 the Configure-Makefile script screwed up while processing it.
4651 10. Incorporated PCRE 4.4.
4653 11. The SMTP synchronization check was not operating right at the start of an
4654 SMTP session. For example, it could not catch a HELO sent before the client
4655 waited for the greeting. There is now a check for outstanding input at the
4656 point when the greeting is written. Because of the duplex, asynchronous
4657 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
4658 way, but not yet received, when the check is performed.
4660 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
4661 on TCP/IP sockets, because this apparently causes some broken clients to
4664 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
4665 unchanged) from the Cygwin maintainer.
4667 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
4668 was supported instead of testing for mbx. Effectively a typo.
4670 15. The spa authenticator server code was not checking that the input it
4671 received was valid base64.
4673 16. The debug output line for the "set" modifier in ACLs was not showing the
4674 name of the variable that was being set.
4676 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
4678 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
4679 lookup on its own. Something else was needed to trigger the lookup. For
4680 example, a match in host_lookup or the need for a host name in a host list.
4681 Now, if $sender_host_name is referenced and the host name has not yet been
4682 looked up, a lookup is performed. If the lookup fails, the variable remains
4683 empty, and $host_lookup_failed is set to "1".
4685 19. Added "eqi" as a case-independent comparison operator.
4687 20. The saslauthd authentication condition could segfault if neither service
4688 nor realm was specified.
4690 21. If an overflowing value such as "2048M" was set for message_size_limit, the
4691 error message that was logged was misleading, and incoming SMTP
4692 connections were dropped. The message is now more accurate, and temporary
4693 errors are given to SMTP connections.
4695 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
4696 (except RSET) with a 421 error, until QUIT is received. However, it was
4697 failing to send a response to QUIT.
4699 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
4700 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
4701 ACL is now run after the helo_try_verify_hosts code.
4703 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
4704 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
4705 case-independent, so other case variants are also recognized). Apparently
4706 some systems use these upper case variants.
4708 25. If more than two messages were waiting for the same host, and a transport
4709 filter was specified for the transport, Exim sent two messages over the
4710 same TCP/IP connection, and then failed with "socket operation on non-
4711 socket" when it tried to send the third.
4713 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
4715 27. The extern definition of crypt16() in expand.c was not being excluded when
4716 the OS had its own crypt16() function.
4718 28. Added bounce_return_body as a new option, and bounce_return_size_limit
4719 as a preferred synonym for return_size_limit, both as an option and as an
4722 29. Added LIBS=-liconv to OS/Makefile-OSF1.
4724 30. Changed the default configuration ACL to relax the local part checking rule
4725 for addresses that are not in any local domains. For these addresses,
4726 slashes and pipe symbols are allowed within local parts, but the sequence
4727 /../ is explicitly forbidden.
4729 31. SPA server authentication was not clearing the challenge buffer before
4732 32. log_message in a "warn" ACL statement was writing to the reject log as
4733 well as to the main log, which contradicts the documentation and doesn't
4734 seem right (because no rejection is happening). So I have stopped it.
4736 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
4737 However, I am unable to do any testing of this.
4739 34. Fixed an infelicity in the appendfile transport. When checking directories
4740 for a mailbox, to see if any needed to be created, it was accidentally
4741 using path names with one or more superfluous leading slashes; tracing
4742 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
4744 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
4745 contents are added to the log line that is written for every discarded
4746 recipient. (Previously a log_message setting was ignored.)
4748 36. The ${quote: operator now quotes the string if it is empty.
4750 37. The install script runs exim in order to find its version number. If for
4751 some reason other than non-existence or emptiness, which it checks, it
4752 could not run './exim', it was installing it with an empty version number,
4753 i.e. as "exim-". This error state is now caught, and the installation is
4756 38. An argument was missing from the function that creates an error message
4757 when Exim fails to connect to the socket for saslauthd authentication.
4758 This could cause Exim to crash, or give a corrupted message.
4760 39. Added isip, isip4, and isip6 to ${if conditions.
4762 40. The ACL variables $acl_xx are now saved with the message, and can be
4763 accessed later in routers, transports, and filters.
4765 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
4766 strings in the file are not string-expanded.
4768 42. If a MAIL command specified a SIZE value that was too large to fit into an
4769 int variable, the check against message_size_limit failed. Such values are
4770 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
4771 day this will have to be increased, but I don't think I want to be around
4772 when emails are that large.
4779 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
4780 iconv() is not standard in FreeBSD.
4782 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
4783 IPv6 enabled. The observed symptom was a segmentation fault on return from
4784 the function os_common_find_running_interfaces() in src/os.c.
4786 3. In the check_special_case() function in daemon.c I had used "errno" as an
4787 argument name, which causes warnings on some systems. This was basically a
4788 typo, since it was named "eno" in the comments!
4790 4. The code that waits for the clock to tick (at a resolution of some fraction
4791 of a second) so as to ensure message-id uniqueness was always waiting for
4792 at least one whole tick, when it could have waited for less. [This is
4793 almost certainly not relevant at current processor speeds, where it is
4794 unlikely to ever wait at all. But we try to future-proof.]
4796 5. The function that sleeps for a time interval that includes fractions of a
4797 second contained a race. It did not block SIGALRM between setting the
4798 timer, and suspending (a couple of lines later). If the interval was short
4799 and the sigsuspend() was delayed until after it had expired, the suspension
4800 never ended. On busy systems this could lead to processes getting stuck for
4803 6. Some uncommon configurations may cause a lookup to happen in a queue runner
4804 process, before it forks any delivery processes. The open lookup caching
4805 mechanism meant that the open file or database connection was passed into
4806 the delivery process. The problem was that delivery processes always tidy
4807 up cached lookup data. This could cause a problem for the next delivery
4808 process started by the queue runner, because the external queue runner
4809 process does not know about the closure. So the next delivery process
4810 still has data in the lookup cache. In the case of a file lookup, there was
4811 no problem because closing a file descriptor in a subprocess doesn't affect
4812 the parent. However, if the lookup was caching a connection to a database,
4813 the connection was closed, and the second delivery process was likely to
4814 see errors such as "PGSQL: query failed: server closed the connection
4815 unexpectedly". The problem has been fixed by closing all cached lookups
4816 in a queue runner before running a delivery process.
4818 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
4819 seem to have the "const" qualifier which it has on other OS. I've
4822 8. Change 4.21/2 was too strict. It is only if there are two authenticators
4823 *of the same type* (client or server) with the same public name that an
4824 error should be diagnosed.
4826 9. When Exim looked up a host name for an IP address, but failed to find the
4827 original IP address when looking up the host name (a safety check), it
4828 output the message "<ip address> does not match any IP for NULL", which was
4829 confusing, to say the least. The bug was that the host name should have
4830 appeared instead of "NULL".
4832 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
4833 user that is built into the binary, and the -C or -D options is used, root
4834 privilege is dropped before the configuration file is read. In addition,
4835 logging is switched to stderr instead of the normal log files. If the
4836 configuration then re-defines the Exim user, the unprivileged environment
4837 is probably not what is expected, so Exim logs a panic warning message (but
4840 However, if deliver_drop_privilege is set, the unprivileged state may well
4841 be exactly what is intended, so the warning has been cut out in that case,
4842 and Exim is allowed to try to write to its normal log files.
4848 1. smtp_return_error_details was not giving details for temporary sender
4849 or receiver verification errors.
4851 2. Diagnose a configuration error if two authenticators have the same public
4854 3. Exim used not to create the message log file for a message until the first
4855 delivery attempt. This could be confusing when incoming messages were held
4856 for policy or load reasons. The message log file is now created at the time
4857 the message is received, and an initial "Received" line is written to it.
4859 4. The automatically generated man page for command line options had a minor
4860 bug that caused no ill effects; however, a more serious problem was that
4861 the procedure for building the man page automatically didn't always
4862 operate. Consequently, release 4.20 contains an out-of-date version. This
4863 shouldn't happen again.
4865 5. When building Exim with embedded Perl support, the script that builds the
4866 Makefile was calling 'perl' to find its compile-time parameters, ignoring
4867 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
4869 6. The freeze_tell option was not being used for messages that were frozen on
4870 arrival, either by an ACL or by local_scan().
4872 7. Added the smtp_incomplete_transaction log selector.
4874 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
4875 was accepting AUTH without a new EHLO.
4877 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
4878 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
4881 10. Logging of TCP/IP connections (when configured) now happens in the main
4882 daemon process instead of the child process, so that the TCP/IP connection
4883 count is more accurate (but it can never be perfect).
4885 11. The use of "drop" in a nested ACL was not being handled correctly in the
4886 outer ACL. Now, if condition failure induced by the nested "drop" causes
4887 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
4888 or "require"), the connection is dropped.
4890 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
4891 that yield "discard" can now be used with an "accept" or a "discard" verb,
4892 but an error is generated for any others (because I can't see a useful way
4893 to define what should happen).
4895 13. When an ACL is read dynamically from a file (or anywhere else), the lines
4896 are now processed in the same way as lines in the Exim configuration file.
4897 In particular, continuation lines are supported.
4899 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
4901 15. Added -ti meaning -t -i.
4903 16. Check for letters, digits, hyphens, and dots in the names of dnslist
4904 domains, and warn by logging if others are found.
4906 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
4907 returned from GIFCONF when Exim is trying to find the list of interfaces on
4908 a host. The code in os.c has been modified to copy each ifreq to an aligned
4909 structure in all cases.
4911 Also, in some cases, the returned ifreq's were being copied to a 'struct
4912 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
4913 means the last couple of bytes of an IPv6 address could be chopped if the
4914 ifreq contained only a normal sockaddr (14 bytes storage).
4916 18. Named domain lists were not supported in the hosts_treat_as_local option.
4917 An entry such as +xxxx was not recognized, and was treated as a literal
4920 19. Ensure that header lines added by a DATA ACL are included in the reject log
4921 if the ACL subsequently rejects the message.
4923 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
4924 MD5 (which is deprecated).
4926 21. When testing a filter file using -bf, Exim was writing a message when it
4927 took the sender from a "From " line in the message, but it was not doing so
4928 when it took $return_path from a Return-Path: header line. It now does.
4930 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
4931 with a valid header line field name (a series of printing characters
4932 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
4934 23. Changed "disc" in the source to "disk" to conform to the documentation and
4935 the book and for uniformity.
4937 24. Ignore Sendmail's -Ooption=value command line item.
4939 25. When execve() failed while trying to run a command in a pipe transport,
4940 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
4941 could be confused with a return value of 69 from the command itself. This
4942 has been changed to 127, the value the shell returns if it is asked to run
4943 a non-existent command. The wording for the related log line suggests a
4944 non-existent command as the problem.
4946 26. If received_header_text expands to an empty string, do not add a Received:
4947 header line to the message. (Well, it adds a token one on the spool, but
4948 marks it "old" so that it doesn't get used or transmitted.)
4950 27. Installed eximstats 1.28 (addition of -nt option).
4952 28. There was no check for failure on the call to getsockname() in the daemon
4953 code. This can fail if there is a shortage of resources on the system, with
4954 ENOMEM, for example. A temporary error is now given on failure.
4956 29. Contrary to the C standard, it seems that in some environments, the
4957 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
4958 program. Exim now does this explicitly; it affects the formatting of
4959 timestamps using strftime().
4961 30. If exiqsumm was given junk data, it threw up some uninitialized variable
4962 complaints. I've now initialized all the variables, to avoid this.
4964 32. Header lines added by a system filter were not being "seen" during
4965 transport-time rewrites.
4967 33. The info_callback() function passed to OpenSSL is set up with type void
4968 (*)(SSL *, int, int), as described somewhere. However, when calling the
4969 function (actually a macro) that sets it up, the type void(*)() is
4970 expected. I've put in a cast to prevent warnings from picky compilers.
4972 34. If a DNS black list lookup found a CNAME record, but there were no A
4973 records associated with the domain it pointed at, Exim crashed.
4975 35. If a DNS black list lookup returned more than one A record, Exim ignored
4976 all but the first. It now scans all returned addresses if a particular IP
4977 value is being sought. In this situation, the contents of the
4978 $dnslist_value variable are a list of all the addresses, separated by a
4981 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
4982 to accept a host name and all its aliases if the forward lookup for any of
4983 them yielded the IP address of the incoming connection. Now it accepts only
4984 those names whose forward lookup yields the correct IP address. Any other
4985 names are discarded. This closes a loophole whereby a rogue DNS
4986 administrator could create reverse DNS records to break through a
4987 wildcarded host restriction in an ACL.
4989 37. If a user filter or a system filter that ran in a subprocess used any of
4990 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
4991 the wrong values were passed to the pipe command ($thisaddress had the
4992 value of $0, $0 had the value of $1, etc). This bug was introduced by
4993 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
4995 38. Improved the line breaking for long SMTP error messages from ACLs.
4996 Previously, if there was no break point between 40 and 75 characters, Exim
4997 left the rest of the message alone. Two changes have been made: (a) I've
4998 reduced the minimum length to 35 characters; (b) if it can't find a break
4999 point between 35 and 75 characters, it looks ahead and uses the first one
5000 that it finds. This may give the occasional overlong line, but at least the
5001 remaining text gets split now.
5003 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
5004 file descriptors might be low, and that setting 1000 would always raise it.
5005 It turns out that in some environments, the limit is already over 1000 and
5006 that lowering it causes trouble. So now Exim takes care not to decrease it.
5008 40. When delivering a message, the value of $return_path is set to $sender_
5009 address at the start of routing (routers may change the value). By an
5010 oversight, this default was not being set up when an address was tested by
5011 -bt or -bv, which affected the outcome if any router or filter referred to
5014 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
5015 log only when "message" or "log_message" are set. However, if one of the
5016 conditions was an address verification, or a call to a nested ACL, the
5017 messages generated by the underlying test were being passed through. This
5018 no longer happens. The underlying message is available in $acl_verify_
5019 message for both "message" and "log_message" expansions, so it can be
5020 passed through if needed.
5022 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
5023 new expansion $bh_ to give the encoded byte string without charset
5024 translation. Translation happens only if iconv() is available; HAVE_ICONV
5025 indicates this at build time. HEADERS_CHARSET gives the charset to
5026 translate to; headers_charset can change it in the configuration, and
5027 "headers charset" can change it in an individual filter file.
5029 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
5030 that creates RFC 2047 encoded "words" labels them as that charset instead
5031 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
5032 expansion operator; (ii) when Exim creates a From: line for a local
5033 message; (iii) when a header line is rewritten to include a "phrase" part.
5035 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
5036 buggy, causing it to skip the first lines of messages whose message ID
5037 ended in 'D'. This would not have bitten before Exim release 4.14, because
5038 message IDs were unlikely to end in 'D' before then. The effect was to have
5039 incorrect size information for certain domains.
5041 45. #include "config.h" was missing at the start of the crypt16.c module. This
5042 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
5045 46. If there was a timeout during a "random" callout check, Exim treated it as
5046 a failure of the random address, and carried on sending RSET and the real
5047 address. If the delay was just some slowness somewhere, the response to the
5048 original RCPT would be taken as a response to RSET and so on, causing
5049 mayhem of various kinds.
5051 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
5052 when I implemented it. It didn't allow for the fact that some option values
5053 may legitimatetly be negative (e.g. size_addition), and it didn't even do
5054 the right test for positive values.
5056 48. Domain names in DNS records are case-independent. Exim always looks them up
5057 in lower case. Some resolvers return domain names in exactly the case they
5058 appear in the zone file, that is, they may contain uppercase letters. Not
5059 all resolvers do this - some return always lower case. Exim was treating a
5060 change of case by a resolver as a change of domain, similar to a widening
5061 of a domain abbreviation. This triggered its re-routing code and so it was
5062 trying to route what was effectively the same domain again. This normally
5063 caused routing to fail (because the router wouldn't handle the domain
5064 twice). Now Exim checks for this case specially, and just changes the
5065 casing of the domain that it ultimately uses when it transmits the message
5068 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
5071 50. If a filter generated a file delivery with a non-absolute name (possible if
5072 no home directory exists for the router), the forbid_file option was not
5075 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
5076 the existing equality matching.
5078 52. Exim was using ints instead of ino_t variables in some places where it was
5079 dealing with inode numbers.
5081 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
5082 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
5083 called TMPDIR, and if it finds it is different, it changes its value.
5085 54. The smtp_printf() function is now made available to local_scan() so
5086 additional output lines can be written before returning. There is also an
5087 smtp_fflush() function to enable the detection of a dropped connection.
5088 The variables smtp_input and smtp_batched_input are exported to
5091 55. Changed the default runtime configuration: the message "Unknown user"
5092 has been removed from the ACL, and instead placed on the localuser router,
5093 using the cannot_route_message feature. This means that any verification
5094 failures that generate their own messages won't get overridden. Similarly,
5095 the "Unrouteable address" message that was in the ACL for unverifiable
5096 relay addresses has also been removed.
5098 56. Added hosts_avoid_esmtp to the smtp transport.
5100 57. The exicyclog script was not checking for the esoteric option
5101 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
5102 will work only if exicyclog is run under the appropriate euid.
5104 58. Following a discussion on the list, the rules by which Exim recognises line
5105 endings on incoming messages have been changed. The -dropcr and drop_cr
5106 options are now no-ops, retained only for backwards compatibility. The
5107 following line terminators are recognized: LF CRLF CR. However, special
5108 processing applies to CR:
5110 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
5111 nor a local message in the state where . is a terminator.
5113 (ii) If a bare CR is encountered in a header line, an extra space is added
5114 after the line terminator so as not to end the header. The reasoning
5115 behind this is that bare CRs in header lines are most likely either
5116 to be mistakes, or people trying to play silly games.
5118 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
5119 was being incorrectly given as 18 bytes larger than it should have been.
5120 This is a VOB (very old bug).
5122 60. This may never have affected anything current, but just in case it has:
5123 When the local host is found other than at the start of a list of hosts,
5124 the local host, those with the same MX, and any that follow, are discarded.
5125 When the list in question was part of a longer list of hosts, the following
5126 hosts (not currently being processed) were also being discarded. This no
5127 longer happens. I'm not sure if this situation could ever has previously
5130 61. Added the "/MX" feature to lists of hosts in the manualroute and query
5133 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
5134 header. This is something that is recommended in a new Internet Draft, and
5135 is something that is documented as being done by Sendmail. There are two
5136 possible values. For messages generated by the autoreply transport, Exim
5139 Auto-Submitted: auto-replied
5141 whereas for all other generated messages (e.g. bounces) it adds
5143 Auto-Submitted: auto-generated
5145 63. The "personal" condition in filters now includes a test for the
5146 Auto-Submitted: header. If it contains the string "auto-" the message it
5147 not considered personal.
5149 64. Added rcpt_include_affixes as a generic transport option.
5151 65. Added queue_only_override (default true).
5153 66. Added the syslog_duplication option.
5155 67. If what should have been the first header line of a message consisted of
5156 a space followed by a colon, Exim was mis-interpreting it as a header line.
5157 It isn't of course - it is syntactically invalid and should therefore be
5158 treated as the start of the message body. The misbehaviour could have
5159 caused a number of strange effects, including loss of data in subsequent
5160 header lines, and spool format errors.
5162 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
5163 client host had authenticated. This control can now be exercised by an ACL
5164 for more flexibility.
5166 69. By default, callouts do not happen when testing with -bh. There is now a
5167 variant, -bhc, which does actually run the callout code, including
5168 consulting and updating the callout cache.
5170 70. Added support for saslauthd authentication, courtesy of Alexander
5173 71. If statvfs() failed on the spool or log directories while checking their
5174 size for availability, Exim confusingly gave the error "space shortage".
5175 Furthermore, in debugging mode it crashed with a floating point exception.
5176 These checks are done if check_{spool,log}_{space,inodes} are set, and when
5177 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
5178 serious problem, Exim now writes to the main and panic logs when this
5179 happens, with details of the failure. It then refuses to accept the
5180 incoming message, giving the message "spool directory problem" or "log
5181 directory problem" with a 421 code for SMTP messages.
5183 72. When Exim is about to re-exec itself, it ensures that the file descriptors
5184 0, 1, and 2 exist, because some OS complain for execs without them (see
5185 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
5186 descriptors. However, the code omitted to check that the open succeeded,
5187 causing mysterious errors if for some reason the permissions on /dev/null
5188 got screwed. Now Exim writes a message to the main and panic logs, and
5189 bombs out if it can't open /dev/null.
5191 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
5192 interact so that it is all more flexible. It is supposed to remain
5193 backwards compatible. Also added extra_local_interfaces.
5195 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
5196 to bomb out with an assertion failure - to the client this appears as a
5197 connection drop. This problem occurs in the part of the code that was taken
5198 from the Samba project. Fortunately, the assertion is in a very simple
5199 function, so I have fixed this by reproducing the function inline in the
5200 one place where it is called, and arranging for authentication to fail
5201 instead of killing the process with assert().
5203 75. The SPA client code was not working when the server requested OEM rather
5204 than Unicode encoding.
5206 76. Added code to make require_files with a specific uid setting more usable in
5207 the case where statting the file as root fails - usually a non-root-mounted
5208 NFS file system. When this happens and the failure is EACCES, Exim now
5209 forks a subprocess and does the per-uid checking as the relevant uid.
5211 77. Added process_log_path.
5213 78. If log_file_path was not explicitly set, a setting of check_log_space or
5214 check_log_inodes was ignored.
5216 79. If a space check for the spool or log partitions fails, the incident is now
5217 logged. Of course, in the latter case the data may get lost...
5219 80. Added the %p formatting code to string_format() so that it can be used to
5220 print addresses in debug_print(). Adjusted all the address printing in the
5221 debugging in store.c to use %p rather than %d.
5223 81. There was a concern that a line of code in smtp_in.c could overflow a
5224 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
5225 initially expressed, the concern was not well-founded, because trailing
5226 spaces are removed early. However, if the trailing spaces were followed by
5227 a NULL, they did not get removed, so the overflow was possible. Two fixes
5230 (a) I re-wrote the offending code in a cleaner fashion.
5231 (b) If an incoming SMTP command contains a NULL character, it is rejected
5234 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
5235 runs initgroups(), so that if the Exim user is in any additional groups,
5236 they will be used during message reception.
5242 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / exim.git / blob