1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.246 2005/10/11 09:30:41 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 PH/01 In the default runtime configuration, move the checks for
10 relay_from_hosts and authenticated clients from after to before the
11 (commented out) DNS black list checks.
17 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
18 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
21 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
22 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
24 PH/03 Typo: missing ".o" in src/pcre/Makefile.
26 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
27 header line, restrict the check to what is listed in RFCs 2369 and 2929.
28 Also, for "Auto-Submitted", treat anything other than "no" as
29 non-personal, in accordance with RFC 3834. (Previously it treated
30 anything starting "auto-" as non-personal.)
32 TF/01 The control=submission/name=... option had a problem with syntax
33 errors if the name included a slash character. The /name= option
34 now slurps the rest of the string, so it can include any characters
35 but it must come last in the list of options (after /sender_retain
38 PH/05 Some modifications to the interface to the fake nameserver for the new
45 TK/01 Added the "success_on_redirect" address verification option. See
46 NewStuff for rationale and an example.
48 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
50 PH/02 Patch to exigrep to allow it to work on syslog lines.
52 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
53 fread() to skip over the body file's header line, because in Cygwin the
54 header line is locked and is inaccessible.
56 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
57 co-exist for some time) to make it clear that it is the Exim ID that is
58 referenced, not the Message-ID: header line.
60 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
61 string_format() function, because snprintf() does not exist on all
64 PH/06 The use of forbid_filter_existstest now also locks out the use of the
65 ${stat: expansion item.
67 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
68 protocol synchronization error", to keep the pedants happy.
70 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
71 well as for IRIX systems, when gcc is being used. See the host.c source
74 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
76 PH/10 Named domain lists were not working if used in a queue_smtp_domains
79 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
80 transport and to the smtp transport in LMTP mode.
82 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
84 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
85 run a filter in a subprocess. This could lead to confusion in subsequent
86 lookups in the parent process. There should also be a search_tidyup() at
87 the end of the subprocess.
89 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
90 only if the host matched helo_try_verify_hosts, which caused the
91 verification to occur when the EHLO/HELO command was issued. The ACL just
92 tested the remembered result. Now, if a previous verification attempt has
93 not happened, "verify = helo" does it there and then.
95 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
97 TK/03 Fix log output including CR from clamd.
99 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
100 exist provoked a memory error which could cause a segfault.
102 PH/15 Installed PCRE 6.2
104 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
106 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
107 of the problem. Specifically, suggested +O2 rather than +O1 for the
110 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
112 PH/20 If a delivery was routed to a non-standard port by means of an SRV
113 record, the port was not correctly logged when the outgoing_port log
114 selector was set (it logged the transort's default port).
116 PH/21 Added support for host-specific ports to manualroute, queryprogram,
117 fallback_hosts, and "hosts" in the smtp transport.
119 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
120 host errors such as "Connection refused".
122 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
123 authentication with radiusclient 0.4.9:
125 - Error returned from rc_read_config was caught wrongly
126 - Username/password not passed on to radius server due to wrong length.
128 The presumption is that some radiusclient API changes for 4.51/PH/17
129 were not taken care of correctly. The code is still untested by me (my
130 Linux distribution still has 0.3.2 of radiusclient), but it was
131 contributed by a Radius user.
133 PH/24 When doing a callout, the value of $domain wasn't set correctly when
134 expanding the "port" option of the smtp transport.
136 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
137 while reading a MIME header. Thanks to Tom Hughes for a patch.
139 PH/24 Include config.h inside local_scan.h so that configuration settings are
142 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
143 that in an ACL for RCPT (for example), you can examine exactly what was
146 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
147 commands, but it was not correctly comparing the address with the actual
148 client host address. Thus, it would show the EHLO address in Received:
149 header lines when this was not necessary.
151 PH/27 Added the % operator to ${eval:}.
153 PH/28 Exim tries to create and chdir to its spool directory when it starts;
154 it should be ignoring failures (because with -C, for example, it has lost
155 privilege). It wasn't ignoring creation failures other than "already
158 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
161 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
162 because an input line was too long, either on its own, or by virtue of
163 too many continuations, the temporary file was not being removed, and the
164 return code was incorrect.
166 PH/31 Missing "BOOL" in function definition in filtertest.c.
168 PH/32 Applied Sieve patches from the maintainer.
170 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
172 PH/33 Added "verify = not_blind".
174 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
175 Local/Makefile (with some defaults set). These are used in built scripts
176 such as exicyclog, but they have never been used in the exim_install
177 script (though there are many overriding facilities there). I have
178 arranged that the exim_install script now takes note of these two
181 PH/35 Installed configuration files for Dragonfly.
183 PH/36 When a locally submitted message by a trusted user did not contain a
184 From: header, and the sender address was obtained from -f or from an SMTP
185 MAIL command, and the trusted user did not use -F to supply a sender
186 name, $originator_name was incorrectly used when constructing a From:
187 header. Furthermore, $originator_name was used for submission mode
188 messages from external hosts without From: headers in a similar way,
189 which is clearly wrong.
191 PH/37 Added control=suppress_local_fixups.
193 PH/38 When log_selector = +received_sender was set, and the addition of the
194 sender made the log line's construction buffer exactly full, or one byte
195 less than full, an overflow happened when the terminating "\n" was
198 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
199 when the result of a list match is failure because a DNS lookup failed.
201 PH/40 RM_COMMAND is now used in the building process.
203 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
204 the "build-* directories that it finds.
206 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
207 address in a domain literal was a prefix of an interface address.
209 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
210 when verifying a sender address, unless rewrite_headers is false.
212 PH/44 Wrote a long comment about why errors_to addresses are verified as
213 recipients, not senders.
215 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
216 the ratelimit ACL was added.
218 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
220 PH/46 Added extra information about PostgreSQL errors to the error string.
222 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
223 avoiding the need to install special zones in a real server. This is
224 backwards compatible; if it can't find the fake resolver, it drops back.
225 Thus, both old and new test suites can be run.
227 TF/02 Added util/ratelimit.pl
229 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
230 clock is set back in time.
232 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
233 Candler <B.Candler@pobox.com>.
235 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
236 OK for addresses that are the result of redirections.
238 PH/48 A number of further additions for the benefit of the new test suite,
239 including a fake gethostbyname() that interfaces to the fake DNS resolver
242 TF/06 The fix for widen_domains has also been applied to qualify_single and
243 search_parents which are the other dnslookup options that can cause
246 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
249 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
251 TF/07 Exim produced the error message "an SRV record indicated no SMTP
252 service" if it encountered an MX record with an empty target hostname.
253 The message is now "an MX or SRV record indicated no SMTP service".
255 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
256 if the DNS of the sending site is misconfigured. This is quite a
257 common situation. This change restores the behaviour of treating a
258 helo verification defer as a failure.
260 PH/51 If self=fail was set on a router, the bounce message did not include the
261 actual error message.
267 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
269 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
270 command itself ended in error, the underlying message about the transport
271 filter timeout was being overwritten with the pipe command error. Now the
272 underlying error message should be appended to the second error message.
274 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
276 PH/02 Reduce the amount of output that "make" produces by default. Full output
277 can still be requested.
279 PH/03 The warning log line about a condition test deferring for a "warn" verb
280 was being output only once per connection, rather than after each
281 occurrence (because it was using the same function as for successful
282 "warn" verbs). This seems wrong, so I have changed it.
284 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
285 it should not have, which might have caused a crash in the right
286 circumstances, but probably never did.
288 PH/04 Installed a modified version of Tony Finch's patch to make submission
289 mode fix the return path as well as the Sender: header line, and to
290 add a /name= option so that you can make the user's friendly name appear
293 TF/03 Added the control = fakedefer ACL modifier.
295 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
296 Mark Lowes for thorough testing.
298 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
300 TK/03 Merged latest SRS patch from Miles Wilton.
302 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
303 with the definition in sysexits.h (which is #included earlier).
304 Fortunately, Exim does not actually use EX_OK. The code used to try to
305 preserve the sysexits.h value, by assumimg that macro definitions were
306 scanned for macro replacements. I have been disabused of this notion,
307 so now the code just undefines EX_OK before #including unistd.h.
309 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
310 in the smtp transport. When a block could not be written in a single
311 write() function, the timeout was being re-applied to each part-write.
312 This seems wrong - if the receiver was accepting one byte at a time it
313 would take for ever. The timeout is now adjusted when this happens. It
314 doesn't have to be particularly precise.
316 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
317 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
319 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
320 without a domain if the check to <postmaster@domain> fails.
322 SC/01 Eximstats: added -xls and the ability to specify output files
323 (patch written by Frank Heydlauf).
325 SC/02 Eximstats: use FileHandles for outputing results.
327 SC/03 Eximstats: allow any combination of xls, txt, and html output.
329 SC/04 Eximstats: fixed display of large numbers with -nvr option
331 SC/05 Eximstats: fixed merging of reports with empty tables.
333 SC/06 Eximstats: added the -include_original_destination flag
335 SC/07 Eximstats: removed tabs and trailing whitespace.
337 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
339 TK/06 MBOX spool code: Add real "From " MBOX separator line
340 so the .eml file is really in mbox format (even though
341 most programs do not really care). Patch from Alex Miller.
343 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
344 The latter is generated from $received_to and is only set if the
345 message has one envelope recipient. SA can use these headers,
346 obviously out-of-the-box. Patch from Alex Miller.
348 PH/08 The ${def test on a variable was returning false if the variable's
349 value was "0", contrary to what the specification has always said!
350 The result should be true unless the variable is empty.
352 PH/09 The syntax error of a character other than { following "${if
353 def:variable_name" (after optional whitespace) was not being diagnosed.
354 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
355 accidental colon was present, for example, could give incorrect results.
357 PH/10 Tidied the code in a number of places where the st_size field of a stat()
358 result is used (not including appendfile, where other changes are about
361 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
362 This involved changing a lot of size variables from int to off_t. It
363 should work with maildirs and everything.
365 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
366 spamd dying while we are connected to it.
368 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
371 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
372 function for BDB 4.3.
374 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
376 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
377 the sources for PCRE have been re-arranged and more files are now
380 PH/15 The code I had for printing potentially long long variables in PH/11
381 above was not the best (it lost precision). The length of off_t variables
382 is now inspected at build time, and an appropriate printing format (%ld
383 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
384 to be "long long int" or "long int". This is needed for the internal
385 formatting function string_vformat().
387 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
388 the configuration file to be ":syslog", then the script "guesses" where
389 the logs files are, rather than using the compiled in default. In our
390 case the guess is not the same as the compiled default, so the script
391 suddenly stopped working when I started to use syslog. The patch checks
392 to see if log_file_path is "". If so, it attempts to read it from exim
393 with no configuration file to get the compiled in version, before it
394 falls back to the previous guessing code."
396 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
397 implementing BATV in an Exim configuration. See NewStuff for the gory
400 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
401 Makefile that are specific to HP-UX.
403 PH/18 If the "use_postmaster" option was set for a recipient callout together
404 with the "random" option, the postmaster address was used as the MAIL
405 FROM address for the random test, but not for the subsequent recipient
406 test. It is now used for both.
408 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
409 patch removes a few documentation additions to RFC 3028, because the
410 latest draft now contains them. It adds the new en;ascii-case comparator
411 and a new error check for 8bit text in MIME parts. Comparator and
412 require names are now matched exactly. I enabled the subaddress
413 extension, but it is not well tested yet (read: it works for me)."
415 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
416 rework some of the code of TK/09 above to avoid the hardwired use of
417 "%lld" and "long long". Replaced the call to snprintf() with a call to
420 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
421 records point to non-existent hosts", "retry timeout exceeded", and
422 "retry time not reached for any host after a long failure period".
424 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
425 experimental DomainKeys support:
427 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
428 (2) On an error such as an illegally used "control", the wrong name for
429 the control was given.
431 These problems did NOT occur unless DomainKeys support was compiled.
433 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
435 PH/24 Added ${if match_ip condition.
437 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
438 databases so that it will be absolutely obvious if a crash occurs in the
439 DB library. This is a regular occurrence (often caused by mis-matched
442 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
443 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
444 picked up on a user's system that detects such things. There doesn't seem
445 to be a gcc warning option for this - only an attribute that has to be
446 put on the function's prototype. It seems that in Fedora Core 4 they have
447 set this on a number of new functions. No doubt there will be more in due
450 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
451 specify a transport. However, if an address that was verified by such a
452 router was the subject of a callout, Exim crashed because it tried to
453 read the rcpt_include_affixes from the non-existent transport. Now it
454 just assumes that the setting of that option is false. This bug was
455 introduced by 4.51/PH/31.
457 PH/28 Changed -d+all to exclude +memory, because that information is very
458 rarely of interest, but it makes the output a lot bigger. People tend to
459 do -d+all out of habit.
461 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
462 code in os-type was giving problems when libc.so lives in lib64, like on
465 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
466 aren't the modern standard, and it seems that some systems' include files
467 don't always have them. Exim was already checking for some of the newer
468 ones like T_AAAA, and defining it itself. I've added checks for all the
469 record types that Exim uses.
471 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
472 not automatically generating a new one, as it is supposed to. This
473 prevented TLS from working. If the file did exist, but contained invalid
474 data, a new version was generated, as expected. It was only the case of a
475 non-existent file that was broken.
477 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
478 with a change in libdomainkeys > 0.64.
480 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
481 from DNS. If the selector record carries the flag, it now has
482 precedence over the domain-wide flag.
484 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
486 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
487 the use of a port name that isn't defined in /etc/services) occurred, the
488 message was deferred as in a normal delivery, and thus remained on the
489 spool, instead of being failed because of the mua_wrapper setting. This
490 is now fixed, and I tidied up some of the mua_wrapper messages at the
493 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
494 the messages in a hash of arrays rather than using individual hashes.
495 This is a bit cleaner and results in dramatic memory savings, albeit
496 at a slight CPU cost.
498 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
499 as requested by Marc Sherman.
501 SC/10 Eximstats: added histograms for user specified patterns as requested
504 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
506 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
507 fopen() in the content-scanning modules that did not already have it.
513 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
514 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
516 TK/02 Fix ACL "control" statement not being available in MIME ACL.
518 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
520 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
521 to test Sieve filters that use "vacation".
523 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
524 that changes the way the GnuTLS parameters are stored in the cache file.
525 The new format can be generated externally. For backward compatibility,
526 if the data in the cache doesn't make sense, Exim assumes it has read an
527 old-format file, and it generates new data and writes a new file. This
528 means that you can't go back to an older release without removing the
531 PH/03 A redirect router that has both "unseen" and "one_time" set does not
532 work if there are any delivery delays because "one_time" forces the
533 parent to be marked "delivered", so its unseen clone is never tried
534 again. For this reason, Exim now forbids the simultaneous setting of
537 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
538 redirected to themselves ("homonym" addresses). Read the long ChangeLog
539 entry if you want to know the details. The fix, however, neglected to
540 consider the case when local delivery batching is involved. The test for
541 "previously delivered" was not happening when checking to see if an
542 address could be batched with a previous (undelivered) one; under
543 certain circumstances this could lead to multiple deliveries to the same
546 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
547 in its include files, and this causes problems building Exim.
549 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
550 header_syntax) but Exim was just ignoring anything given after a slash.
551 In particular, this caused confusion with an attempt to use "verify =
552 reverse_host_lookup/defer_ok". An error is now given when options are
553 supplied for verify items that do not have them. (Maybe reverse_host_
554 lookup should have a defer_ok option, but that's a different point.)
556 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
557 defined by RFC 821) to 2048, because there were problems with some AUTH
558 commands, and RFC 1869 says the size should be increased for extended
559 SMTP commands that take arguments.
561 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
564 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
565 "unknown" error; now it says that the functionality isn't in the binary.
567 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
568 an address' error message when a string expansion fails (syntax or
569 whatever). Otherwise the password may appear in the log. Following change
570 PH/42 below, there is no longer a chance of it appearing in a bounce
573 PH/11 Installed exipick version 20050225.0 from John Jetmore.
575 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
576 of its addresses was ever tried. (Bugzilla bug #2.)
578 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
579 the result incorrectly in the debug output. (It correctly added a newline
580 to what was transported.)
582 TF/01 Added $received_time.
584 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
585 commented out examples of how to interface to a virus scanner and to
586 SpamAssassin. Also added commented examples of av_scanner and
587 spamd_address settings.
589 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
590 and controls are allowed in which ACLs. There were a couple of minor
591 errors. Some of the entries in the conditions table (which is a table of
592 where they are NOT allowed) were getting very unwieldy; rewrote them as a
593 negation of where the condition IS allowed.
595 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
597 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
598 header file does not have a version number, so I've had to invent a new
599 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
600 API. The code is untested by me (my Linux distribution still has 0.3.2 of
601 radiusclient), but it was contributed by a Radius user.
603 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
604 files or directories, for OpenSSL.
606 PH/19 When an Exim process that is running as root has to create an Exim log
607 file, it does so in a subprocess that runs as exim:exim so as to get the
608 ownership right at creation (otherwise, other Exim processes might see
609 the file with the wrong ownership). There was no test for failure of this
610 fork() call, which would lead to the process getting stuck as it waited
611 for a non-existent subprocess. Forks do occasionally fail when resources
612 run out. I reviewed all the other calls to fork(); they all seem to check
615 PH/20 When checking for unexpected SMTP input at connect time (before writing
616 the banner), Exim was not dealing correctly with a non-positive return
617 from the read() function. If the client had disconnected by this time,
618 the result was a log entry for a synchronization error with an empty
619 string after "input=" when read() returned zero. If read() returned -1
620 (an event I could not check), uninitialized data bytes were printed.
621 There were reports of junk text (parts of files, etc) appearing after
624 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
626 PH/22 Added support for macro redefinition, and (re)definition in between
627 driver and ACL definitions.
629 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
630 forgetting to use the resulting value; it was using the unexpanded value.
632 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
633 hadn't been configured. The fix is from Juergen Kreileder, who
634 understands it better than I do:
636 "Here's what I see happening with three configured cyrus_sasl
637 authenticators configured (plain, login, cram-md5):
639 On startup auth_cyrus_sasl_init() gets called for each of these.
640 This means three calls to sasl_listmech() without a specified mech_list.
641 => SASL tests which mechs of all available mechs actually work
642 => three warnings about OTP not working
643 => the returned list contains: plain, login, cram-md5, digest-md5, ...
645 With the patch, sasl_listmech() also gets called three times. But now
646 SASL's mech_list option is set to the server_mech specified in the the
647 authenticator. Or in other words, the answer from sasl_listmech()
648 gets limited to just the mech you're testing for (which is different
650 => the return list contains just 'plain' or 'login', 'cram-md5' or
651 nothing depending on the value of ob->server_mech.
653 I've just tested the patch: Authentication still works fine,
654 unavailable mechs specified in the exim configuration are still
655 caught, and the auth.log warnings about OTP are gone."
657 PH/25 When debugging is enabled, the contents of the command line are added
658 to the debugging output, even when log_selector=+arguments is not
661 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
662 answer is "GNU", and only if the return is "GNU/something" is the answer
665 PH/27 $acl_verify_message is now set immediately after the failure of a
666 verification in an ACL, and so is available in subsequent modifiers. In
667 particular, the message can be preserved by coding like this:
669 warn !verify = sender
670 set acl_m0 = $acl_verify_message
672 Previously, $acl_verify_message was set only while expanding "message"
673 and "log_message" when a very denied access.
675 PH/28 Modified OS/os.c-Linux with
677 -#ifndef OS_LOAD_AVERAGE
678 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
680 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
681 nomenclature these days.)
683 PH/29 Installed patch from the Sieve maintainer that adds the options
684 sieve_useraddress and sieve_subaddress to the redirect router.
686 PH/30 In these circumstances:
687 . Two addresses routed to the same list of hosts;
688 . First host does not offer TLS;
689 . First host accepts first address;
690 . First host gives temporary error to second address;
691 . Second host offers TLS and a TLS session is established;
692 . Second host accepts second address.
693 Exim incorrectly logged both deliveries with the TLS parameters (cipher
694 and peerdn, if requested) that were in fact used only for the second
697 PH/31 When doing a callout as part of verifying an address, Exim was not paying
698 attention to any local part prefix or suffix that was matched by the
699 router that accepted the address. It now behaves in the same way as it
700 does for delivery: the affixes are removed from the local part unless
701 rcpt_include_affixes is set on the transport.
703 PH/32 Add the sender address, as F=<...>, to the log line when logging a
704 timeout during the DATA phase of an incoming message.
706 PH/33 Sieve envelope tests were broken for match types other than :is. I have
707 applied a patch sanctioned by the Sieve maintainer.
709 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
710 the uid or gid is negative. A case of a negative gid caused this to be
711 noticed. The fix allows for either to be negative.
713 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
714 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
717 PH/36 The change PH/12 above was broken. Fixed it.
719 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
720 the grounds that routing the same address twice would always produce the
721 same answer. This might have been true once, but it is certainly no
722 longer true now. Routing a child address may depend on the previous
723 routing that produced that child. Some complicated redirection strategies
724 went wrong when messages had multiple recipients, and made Exim's
725 behaviour dependent on the order in which the addresses were given.
727 I have moved the duplicate checking until after the routing is complete.
728 Exim scans the addresses that are assigned to local and remote
729 transports, and removes any duplicates. This means that more work will be
730 done, as duplicates will always all be routed, but duplicates are
731 presumably rare, so I don't expect this is of any significance.
733 For deliveries to pipes, files, and autoreplies, the duplicate checking
734 still happens during the routing process, since they are not going to be
737 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
738 It corrects a timeout issue with spamd. This is Ian's comment: "The
739 background is that sometimes spamd either never reads data from a
740 connection it has accepted, or it never writes response data. The exiscan
741 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
742 blindly assumes that writes won't block so it may never time out."
744 PH/39 Allow G after quota size as well as K and M.
746 PH/40 The value set for $authenticated_id in an authenticator may not contain
747 binary zeroes or newlines because the value is written to log lines and
748 to spool files. There was no check on this. Now the value is run through
749 the string_printing() function so that such characters are converted to
750 printable escape sequences.
752 PH/41 $message_linecount is a new variable that contains the total number of
753 lines in the message. Compare $body_linecount, which is the count for the
756 PH/42 Exim no longer gives details of delivery errors for specific addresses in
757 bounce and delay warning messages, except in certain special cases, which
760 (a) An SMTP error message from a remote host;
761 (b) A message specified in a :fail: redirection;
762 (c) A message specified in a "fail" command in a system filter;
763 (d) A message specified in a FAIL return from the queryprogram router;
764 (e) A message specified by the cannot_route_message router option.
766 In these cases only, Exim does include the error details in bounce and
767 warning messages. There are also a few cases where bland messages such
768 as "unrouteable address" or "local delivery error" are given.
770 PH/43 $value is now also set for the "else" part of a ${run expansion.
772 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
773 being worked on, but at least Exim now implements the latest version to
776 PH/45 In a pipe transport, although a timeout while waiting for the pipe
777 process to complete was treated as a delivery failure, a timeout while
778 writing the message to the pipe was logged, but erroneously treated as a
779 successful delivery. Such timeouts include transport filter timeouts. For
780 consistency with the overall process timeout, these timeouts are now
781 treated as errors, giving rise to delivery failures by default. However,
782 there is now a new Boolean option for the pipe transport called
783 timeout_defer, which, if set TRUE, converts the failures into defers for
784 both kinds of timeout. A transport filter timeout is now identified in
787 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
788 systems where "make" and "gmake" are different, calling "gmake" at top
789 level broke things. I've arranged for the value of $(MAKE) to be passed
790 from the Makefile to this script so that it can call the same version of
794 A note about Exim versions 4.44 and 4.50
795 ----------------------------------------
797 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
798 changes of various kinds. As a consequence, a big documentation update was
799 needed. This delayed the release for rather longer than seemed good, especially
800 in the light of a couple of (minor) security issues. Therefore, the changes
801 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
802 release. So 4.44 and 4.50 are in effect two different branches that both start
805 I have left the 4.50 change log unchanged; it contains all the changes since
806 4.43. The change log for 4.44 is below; many of its items are identical to
807 those for 4.50. This seems to be the most sensible way to preserve the
808 historical information.
814 1. Minor wording change to the doc/README.SIEVE file.
816 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
817 computation of the current number of files was incorrect.
819 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
820 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
823 4. Give more explanation in the error message when the command for a transport
824 filter fails to execute.
826 5. There are several places where Exim runs a non-Exim command in a
827 subprocess. The SIGUSR1 signal should be disabled for these processes. This
828 was being done only for the command run by the queryprogram router. It is
829 now done for all such subprocesses. The other cases are: ${run, transport
830 filters, and the commands run by the lmtp and pipe transports.
832 6. Added CONFIGURE_GROUP build-time option.
834 7. Some older OS have a limit of 256 on the maximum number of file
835 descriptors. Exim was using setrlimit() to set 1000 as a large value
836 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
837 systems. I've change it so that if it can't get 1000, it tries for 256.
839 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
840 was an oversight, and furthermore, ever since the addition of extra
841 controls (e.g. 4.43/32), the checks on when to allow different forms of
842 "control" were broken. There should now be diagnostics for all cases when a
843 control that does not make sense is encountered.
845 9. Added the /retain_sender option to "control=submission".
847 10. $recipients is now available in the predata ACL (oversight).
849 11. Tidy the search cache before the fork to do a delivery from a message
850 received from the command line. Otherwise the child will trigger a lookup
851 failure and thereby defer the delivery if it tries to use (for example) a
852 cached ldap connection that the parent has called unbind on.
854 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
855 of $address_data from the recipient verification was clobbered by the
858 13. The value of address_data from a sender verification is now available in
859 $sender_address_data in subsequent conditions in the ACL statement.
861 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
863 15. Added a new option "connect=<time>" to callout options, to set a different
866 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
867 was its contents. (It was OK if the option was not defined at all.)
869 17. A "Completed" log line is now written for messages that are removed from
870 the spool by the -Mrm option.
872 18. New variables $sender_verify_failure and $recipient_verify_failure contain
873 information about exactly what failed.
875 19. Added -dd to debug only the daemon process.
877 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
878 handles timeouts, both on the server side and network timeouts. Renamed the
879 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
881 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
882 It is now set to "smtps".
884 22. $host_address is now set to the target address during the checking of
887 23. When checking ignore_target_hosts for an ipliteral router, no host name was
888 being passed; this would have caused $sender_host_name to have been used if
889 matching the list had actually called for a host name (not very likely,
890 since this list is usually IP addresses). A host name is now passed as
893 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
894 code that specifies a non-restarting handler (typically sigaction() in
895 modern systems) in an attempt to fix a rare and obscure crash bug.
897 25. Narrowed the window for a race in the daemon that could cause it to ignore
898 SIGCHLD signals. This is not a major problem, because they are used only to
899 wake it up if nothing else does.
901 26. A malformed maildirsize file could cause Exim to calculate negative values
902 for the mailbox size or file count. Odd effects could occur as a result.
903 The maildirsize information is now recalculated if the size or filecount
906 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
907 support for a long time. Removed HAVE_SYS_VFS_H.
909 28. Installed the latest version of exipick from John Jetmore.
911 29. In an address list, if the pattern was not a regular expression, an empty
912 subject address (from a bounce message) matched only if the pattern was an
913 empty string. Non-empty patterns were not even tested. This was the wrong
914 because it is perfectly reasonable to use an empty address as part of a
915 database query. An empty address is now tested by patterns that are
916 lookups. However, all the other forms of pattern expect the subject to
917 contain a local part and a domain, and therefore, for them, an empty
918 address still always fails if the pattern is not itself empty.
920 30. Exim went into a mad DNS loop when attempting to do a callout where the
921 host was specified on an smtp transport, and looking it up yielded more
924 31. Re-factored the code for checking spool and log partition space into a
925 function that finds that data and another that does the check. The former
926 is then used to implement four new variables: $spool_space, $log_space,
927 $spool_inodes, and $log_inodes.
929 32. The RFC2047 encoding function was originally intended for short strings
930 such as real names; it was not keeping to the 75-character limit for
931 encoded words that the RFC imposes. It now respects the limit, and
932 generates multiple encoded words if necessary. To be on the safe side, I
933 have increased the buffer size for the ${rfc2047: expansion operator from
936 33. It is now permitted to omit both strings after an "if" condition; if the
937 condition is true, the result is "true". As before, when the second string
938 is omitted, a false condition yields an empty string. This makes it less
939 cumbersome to write custom ACL and router conditions.
941 34. Failure to deliver a bounce message always caused it to be frozen, even if
942 there was an errors_to setting on the router. The errors_to setting is now
945 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
946 canonical form (fully expanded) before being placed in
947 $sender_host_address.
949 36. The table in the code that translates DNS record types into text (T_A to
950 "A" for instance) was missing entries for NS and CNAME. It is just possible
951 that this could have caused confusion if both these types were looked up
952 for the same domain, because the text type is used as part of Exim's
953 per-process caching. But the chance of anyone hitting this buglet seems
956 37. The dnsdb lookup has been extended in a number of ways.
958 (1) There is a new type, "zns", which walks up the domain tree until it
959 finds some nameserver records. It should be used with care.
961 (2) There is a new type, "mxh", which is like "mx" except that it returns
962 just the host names, not the priorities.
964 (3) It is now possible to give a list of domains (or IP addresses) to be
965 looked up. The behaviour when one of the lookups defers can be
966 controlled by a keyword.
968 (4) It is now possible to specify the separator character for use when
969 multiple records are returned.
971 38. The dnslists ACL condition has been extended: it is now possible to supply
972 a list of IP addresses and/or domains to be looked up in a particular DNS
975 39. Added log_selector=+queue_time_overall.
977 40. When running the queue in the test harness, wait just a tad after forking a
978 delivery process, to get repeatability of debugging output.
980 41. Include certificate and key file names in error message when GnuTLS fails
981 to set them up, because the GnuTLS error message doesn't include the name
982 of the failing file when there is a problem reading it.
984 42. Allow both -bf and -bF in the same test run.
986 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
988 44. The "Exiscan patch" is now merged into the mainline Exim source.
990 45. Sometimes the final signoff response after QUIT could fail to get
991 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
992 < 0 before doing a fflush(). This bug looks as though it goes back to the
993 introduction of TLS in release 3.20, but "sometimes" must have been rare
994 because the tests only now provoked it.
996 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
997 (this can affect the format of dates).
999 47. exim_tidydb, when checking for the continued existence of a message for
1000 which it has found a message-specific retry record, was not finding
1001 messages that were in split spool directories. Consequently, it was
1002 deleting retry records that should have stayed in existence.
1004 48. Steve fixed some bugs in eximstats.
1006 49. The SPA authentication driver was not abandoning authentication and moving
1007 on to the next authenticator when an expansion was forced to fail,
1008 contradicting the general specification for all authenticators. Instead it
1009 was generating a temporary error. It now behaves as specified.
1011 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
1012 (the order specifies the preference for clients). The order is now AES256,
1013 AES128, 3DES, ARCFOUR128.
1015 51. Small patch to Sieve code - explicitly set From: when generating an
1018 52. Exim crashed if a remote delivery caused a very long error message to be
1019 recorded - for instance if somebody sent an entire SpamAssassin report back
1020 as a large number of 550 error lines. This bug was coincidentally fixed by
1021 increasing the size of one of Exim's internal buffers (big_buffer) that
1022 happened as part of the Exiscan merge. However, to be on the safe side, I
1023 have made the code more robust (and fixed the comments that describe what
1026 53. Now that there can be additional text after "Completed" in log lines (if
1027 the queue_time_overall log selector is set), a one-byte patch to exigrep
1028 was needed to allow it to recognize "Completed" as not the last thing in
1031 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1032 patch that reportedly fixes this has been added. I am not expert enough to
1033 create a test for it. This is what the patch creator wrote:
1035 "I found a little strange behaviour of ldap code when working with
1036 Windows 2003 AD Domain, where users was placed in more than one
1037 Organization Units. When I tried to give exim partial DN, the exit code
1038 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1039 But simultaneously result of request was absolutely normal ldap result,
1040 so I produce this patch..."
1042 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1043 REFERENCE, so I have modified the code to exclude the patch when that macro
1046 55. Some experimental protocols are using DNS PTR records for new purposes. The
1047 keys for these records are domain names, not reversed IP addresses. The
1048 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1049 leaves it alone. Component reversal etc. now happens only for IP addresses.
1052 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1054 57. Double the size of the debug message buffer (to 2048) so that more of very
1055 long debug lines gets shown.
1057 58. The exicyclog utility now does better if the number of log files to keep
1058 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1060 59. Two changes related to the smtp_active_hostname option:
1062 (1) $smtp_active_hostname is now available as a variable.
1063 (2) The default for smtp_banner uses $smtp_active_hostname instead
1064 of $primary_hostname.
1066 60. The host_aton() function is supposed to be passed a string that is known
1067 to be a valid IP address. However, in the case of IPv6 addresses, it was
1068 not checking this. This is a hostage to fortune. Exim now panics and dies
1069 if the condition is not met. A case was found where this could be provoked
1070 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1071 components; fortuitously, this particular loophole had already been fixed
1072 by change 4.50/55 above.
1074 If there are any other similar loopholes, the new check in host_aton()
1075 itself should stop them being exploited. The report I received stated that
1076 data on the command line could provoke the exploit when Exim was running as
1077 exim, but did not say which command line option was involved. All I could
1078 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1079 running as the user.
1082 61. There was a buffer overflow vulnerability in the SPA authentication code
1083 (which came originally from the Samba project). I have added a test to the
1084 spa_base64_to_bits() function which I hope fixes it.
1087 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1088 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1090 63. The daemon start-up calls getloadavg() while still root for those OS that
1091 need the first call to be done as root, but it missed one case: when
1092 deliver_queue_load_max is set with deliver_drop_privilege. This is
1093 necessary for the benefit of the queue runner, because there is no re-exec
1094 when deliver_drop_privilege is set.
1096 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1097 This has been fixed.
1099 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1100 was in use, was not putting the data itself into the right store pool;
1101 consequently, it could be overwritten for a subsequent message in the same
1102 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1105 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1107 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1108 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1109 the function were treating the return as a boolean value, which happened to
1110 work because 0=false and not-0=true, but is not correct code.
1112 68. The host_aton() function was not handling scoped IPv6 addresses (those
1113 with, for example, "%eth0" on the end) correctly.
1115 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1116 negated items (that is, ~something) in unsigned ints. Some compilers
1117 apparently mutter when there is no cast.
1119 70. If an address verification called from an ACL failed, and did not produce a
1120 user-specific message (i.e. there was only a "system" message), nothing was
1121 put in $acl_verify_message. In this situation, it now puts the system
1124 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1125 session; change 4.31/43 added the unwanted input to the log line - except
1126 that it did not do this in the start of session case. It now does.
1128 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1129 This is wrong and can cause the other end to generate a synchronization
1130 error if it is another Exim or anything else that does the synchronization
1131 check. A QUIT command is no longer sent after a timeout.
1133 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1134 during host lookups.
1136 74. The defer_ok option of callout verification was not working if it was used
1137 when verifying addresses in header lines, that is, for this case:
1139 verify = header_sender/callout=defer_ok
1141 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1142 those file descriptors could be used for SMTP connections. If anything
1143 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1144 could be sent to the SMTP client, causing chaos. The daemon now opens
1145 stdin, stdout, and stderr to /dev/null when it puts itself into the
1148 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1149 log by default. The user can override this with suitable Perl magic.
1151 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1152 the log message when discard triggers, was not working for the DATA ACL or
1153 for the non-SMTP ACL.
1155 78. Error message wording change in sieve.c.
1157 79. If smtp_accept_max_per_host was set, the number of connections could be
1158 restricted to fewer than expected, because the daemon was trying to set up
1159 a new connection before checking whether the processes handling previous
1160 connections had finished. The check for completed processes is now done
1161 earlier. On busy systems, this bug wouldn't be noticed because something
1162 else would have woken the daemon, and it would have reaped the completed
1165 80. If a message was submitted locally by a user whose login name contained one
1166 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1167 It caused a spool format error. I have fixed the spool reading code. A
1168 related problem was that the "from" clause in the Received: line became
1169 illegal because of the space(s). It is now covered by ${quote_local_part.
1171 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1174 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1176 83. Patch from Sieve maintainer to fix unterminated string problem in
1177 "vacation" handling.
1179 84. Some minor changes to the Linux configuration files to help with other
1180 OS variants using glibc.
1182 85. One more patch for Sieve to update vacation handling to latest spec.
1185 ----------------------------------------------------
1186 See the note above about the 4.44 and 4.50 releases.
1187 ----------------------------------------------------
1193 1. Change 4.43/35 introduced a bug that caused file counts to be
1194 incorrectly computed when quota_filecount was set in an appendfile
1197 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1198 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1201 3. Give more explanation in the error message when the command for a transport
1202 filter fails to execute.
1204 4. There are several places where Exim runs a non-Exim command in a
1205 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1206 was being done only for the command run by the queryprogram router. It is
1207 now done for all such subprocesses. The other cases are: ${run, transport
1208 filters, and the commands run by the lmtp and pipe transports.
1210 5. Some older OS have a limit of 256 on the maximum number of file
1211 descriptors. Exim was using setrlimit() to set 1000 as a large value
1212 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1213 systems. I've change it so that if it can't get 1000, it tries for 256.
1215 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1216 was an oversight, and furthermore, ever since the addition of extra
1217 controls (e.g. 4.43/32), the checks on when to allow different forms of
1218 "control" were broken. There should now be diagnostics for all cases when a
1219 control that does not make sense is encountered.
1221 7. $recipients is now available in the predata ACL (oversight).
1223 8. Tidy the search cache before the fork to do a delivery from a message
1224 received from the command line. Otherwise the child will trigger a lookup
1225 failure and thereby defer the delivery if it tries to use (for example) a
1226 cached ldap connection that the parent has called unbind on.
1228 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1229 of $address_data from the recipient verification was clobbered by the
1230 sender verification.
1232 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1233 was its contents. (It was OK if the option was not defined at all.)
1235 11. A "Completed" log line is now written for messages that are removed from
1236 the spool by the -Mrm option.
1238 12. $host_address is now set to the target address during the checking of
1239 ignore_target_hosts.
1241 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1242 being passed; this would have caused $sender_host_name to have been used if
1243 matching the list had actually called for a host name (not very likely,
1244 since this list is usually IP addresses). A host name is now passed as
1247 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1248 code that specifies a non-restarting handler (typically sigaction() in
1249 modern systems) in an attempt to fix a rare and obscure crash bug.
1251 15. Narrowed the window for a race in the daemon that could cause it to ignore
1252 SIGCHLD signals. This is not a major problem, because they are used only to
1253 wake it up if nothing else does.
1255 16. A malformed maildirsize file could cause Exim to calculate negative values
1256 for the mailbox size or file count. Odd effects could occur as a result.
1257 The maildirsize information is now recalculated if the size or filecount
1260 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1261 support for a long time. Removed HAVE_SYS_VFS_H.
1263 18. Updated exipick to current release from John Jetmore.
1265 19. Allow an empty sender to be matched against a lookup in an address list.
1266 Previously the only cases considered were a regular expression, or an
1269 20. Exim went into a mad DNS lookup loop when doing a callout where the
1270 host was specified on the transport, if the DNS lookup yielded more than
1273 21. The RFC2047 encoding function was originally intended for short strings
1274 such as real names; it was not keeping to the 75-character limit for
1275 encoded words that the RFC imposes. It now respects the limit, and
1276 generates multiple encoded words if necessary. To be on the safe side, I
1277 have increased the buffer size for the ${rfc2047: expansion operator from
1280 22. Failure to deliver a bounce message always caused it to be frozen, even if
1281 there was an errors_to setting on the router. The errors_to setting is now
1284 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1285 canonical form (fully expanded) before being placed in
1286 $sender_host_address.
1288 24. Updated eximstats to version 1.33
1290 25. Include certificate and key file names in error message when GnuTLS fails
1291 to set them up, because the GnuTLS error message doesn't include the name
1292 of the failing file when there is a problem reading it.
1294 26. Expand error message when OpenSSL has problems setting up cert/key files.
1297 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1298 (this can affect the format of dates).
1300 28. exim_tidydb, when checking for the continued existence of a message for
1301 which it has found a message-specific retry record, was not finding
1302 messages that were in split spool directories. Consequently, it was
1303 deleting retry records that should have stayed in existence.
1305 29. eximstats updated to version 1.35
1306 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1307 1.35 - bugfix such that pie charts by volume are generated correctly
1309 30. The SPA authentication driver was not abandoning authentication and moving
1310 on to the next authenticator when an expansion was forced to fail,
1311 contradicting the general specification for all authenticators. Instead it
1312 was generating a temporary error. It now behaves as specified.
1314 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1315 (the order specifies the preference for clients). The order is now AES256,
1316 AES128, 3DES, ARCFOUR128.
1318 31. Small patch to Sieve code - explicitly set From: when generating an
1321 32. Exim crashed if a remote delivery caused a very long error message to be
1322 recorded - for instance if somebody sent an entire SpamAssassin report back
1323 as a large number of 550 error lines. This bug was coincidentally fixed by
1324 increasing the size of one of Exim's internal buffers (big_buffer) that
1325 happened as part of the Exiscan merge. However, to be on the safe side, I
1326 have made the code more robust (and fixed the comments that describe what
1329 33. Some experimental protocols are using DNS PTR records for new purposes. The
1330 keys for these records are domain names, not reversed IP addresses. The
1331 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1332 leaves it alone. Component reversal etc. now happens only for IP addresses.
1335 34. The host_aton() function is supposed to be passed a string that is known
1336 to be a valid IP address. However, in the case of IPv6 addresses, it was
1337 not checking this. This is a hostage to fortune. Exim now panics and dies
1338 if the condition is not met. A case was found where this could be provoked
1339 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1340 components; fortuitously, this particular loophole had already been fixed
1341 by change 4.50/55 or 4.44/33 above.
1343 If there are any other similar loopholes, the new check in host_aton()
1344 itself should stop them being exploited. The report I received stated that
1345 data on the command line could provoke the exploit when Exim was running as
1346 exim, but did not say which command line option was involved. All I could
1347 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1348 running as the user.
1351 35. There was a buffer overflow vulnerability in the SPA authentication code
1352 (which came originally from the Samba project). I have added a test to the
1353 spa_base64_to_bits() function which I hope fixes it.
1356 36. The daemon start-up calls getloadavg() while still root for those OS that
1357 need the first call to be done as root, but it missed one case: when
1358 deliver_queue_load_max is set with deliver_drop_privilege. This is
1359 necessary for the benefit of the queue runner, because there is no re-exec
1360 when deliver_drop_privilege is set.
1362 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1363 was in use, was not putting the data itself into the right store pool;
1364 consequently, it could be overwritten for a subsequent message in the same
1365 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1368 38. Sometimes the final signoff response after QUIT could fail to get
1369 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1370 < 0 before doing a fflush(). This bug looks as though it goes back to the
1371 introduction of TLS in release 3.20, but "sometimes" must have been rare
1372 because the tests only now provoked it.
1378 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1379 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1380 Now it returns an integer. A number of calls were still expecting a T/F
1381 return. Fortuitously, in all cases, the tests worked in OK situations,
1382 which is the norm. However, things would have gone wrong on any write
1383 failures on the smtp file descriptor. This function is used when sending
1384 messages over SMTP and also when doing verify callouts.
1386 2. When Exim is called to do synchronous delivery of a locally submitted
1387 message (the -odf or -odi options), it no longer closes stderr before doing
1390 3. Implemented the mua_wrapper option.
1392 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1394 5. Implemented the functions header_remove(), header_testname(),
1395 header_add_at_position(), and receive_remove_recipient(), and exported them
1398 6. If an ACL "warn" statement specified the addition of headers, Exim already
1399 inserted X-ACL-Warn: at the start if there was no header name. However, it
1400 was not making this test for the second and subsequent header lines if
1401 there were newlines in the string. This meant that an invalid header could
1402 be inserted if Exim was badly configured.
1404 7. Allow an ACL "warn" statement to add header lines at the start or after all
1405 the Received: headers, as well as at the end.
1407 8. Added the rcpt_4xx retry error code.
1409 9. Added postmaster_mailfrom=xxx to callout verification option.
1411 10. Added mailfrom=xxxx to the callout verification option, for verify=
1414 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1415 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1417 12. Inserted some casts to stop certain compilers warning when using pointer
1418 differences as field lengths or precisions in printf-type calls (mostly
1419 affecting debugging statements).
1421 13. Added optional readline() support for -be (dynamically loaded).
1423 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1424 same clock tick as a message's arrival, so that its received time was the
1425 same as the "first fail" time on the retry record, and that message
1426 remained on the queue past the ultimate address timeout, every queue runner
1427 would try a delivery (because it was past the ultimate address timeout) but
1428 after another failure, the ultimate address timeout, which should have then
1429 bounced the address, did not kick in. This was a "< instead of <=" error;
1430 in most cases the first failure would have been in the next clock tick
1431 after the received time, and all would be well.
1433 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1434 being recognized when the domain list was tested by the match_domain
1435 condition in an expansion string.
1437 16. Added the ${str2b64: operator.
1439 17. Exim was always calling setrlimit() to set a large limit for the number of
1440 processes, without checking whether the existing limit was already
1441 adequate. (It did check for the limit on file descriptors.) Furthermore,
1442 errors from getrlimit() and setrlimit() were being ignored. Now they are
1443 logged to the main and panic logs, but Exim does carry on, to try to do its
1444 job under whatever limits there are.
1446 18. Imported PCRE 5.0.
1448 19. Trivial typo in log message " temporarily refused connection" (the leading
1451 20. If the log selector return_path_on_delivery was set and an address was
1452 redirected to /dev/null, the delivery process crashed because it assumed
1453 that a return path would always be set for a "successful" delivery. In this
1454 case, the whole delivery is bypassed as an optimization, and therefore no
1457 21. Internal re-arrangement: the function for sending a challenge and reading
1458 a response while authentication was assuming a zero-terminated challenge
1459 string. It's now changed to take a pointer and a length, to allow for
1460 binary data in such strings.
1462 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1464 23. Exim was not respecting finduser_retries when seeking the login of the
1465 uid under which it was called; it was always trying 10 times. (The default
1466 setting of finduser_retries is zero.) Also, it was sleeping after the final
1467 failure, which is pointless.
1469 24. Implemented tls_on_connect_ports.
1471 25. Implemented acl_smtp_predata.
1473 26. If the domain in control=submission is set empty, Exim assumes that the
1474 authenticated id is a complete email address when it generates From: or
1475 Sender: header lines.
1477 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1478 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1479 chown and chgrp in /bin and hostname in /usr/bin.
1481 28. Exim was keeping the "process log" file open after each use, just as it
1482 does for the main log. This opens the possibility of it remaining open for
1483 long periods when the USR1 signal hits a daemon. Occasional processlog
1484 errors were reported, that could have been caused by this. Anyway, it seems
1485 much more sensible not to leave this file open at all, so that is what now
1488 29. The long-running daemon process does not normally write to the log once it
1489 has entered its main loop, and it closes the log before doing so. This is
1490 so that log files can straightforwardly be renamed and moved. However,
1491 there are a couple of unusual error situations where the daemon does write
1492 log entries, and I had neglected to close the log afterwards.
1494 30. The text of an SMTP error response that was received during a remote
1495 delivery was being truncated at 512 bytes. This is too short for some of
1496 the long messages that one sometimes sees. I've increased the limit to
1499 31. It is now possible to make retry rules that apply only when a message has a
1500 specific sender, in particular, an empty sender.
1502 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1503 it possible to be selective about when SMTP synchronization is enforced.
1505 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1507 32. Implemented hosts_connection_nolog.
1509 33. Added an ACL for QUIT.
1511 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1514 35. Added mailbox_size and mailbox_filecount to appendfile.
1516 36. Added control = no_multiline_responses to ACLs.
1518 37. There was a bug in the logic of the code that waits for the clock to tick
1519 in the case where the clock went backwards by a substantial amount such
1520 that the microsecond fraction of "now" was more than the microsecond
1521 fraction of "then" (but the whole seconds number was less).
1523 38. Added support for the libradius Radius client library this is found on
1524 FreeBSD (previously only the radiusclient library was supported).
1530 1. When certain lookups returned multiple values in the form name=value, the
1531 quoting of the values was not always being done properly. Specifically:
1532 (a) If the value started with a double quote, but contained no whitespace,
1534 (b) If the value contained whitespace other than a space character (i.e.
1535 tabs or newlines or carriage returns) it was not quoted.
1536 This fix has been applied to the mysql and pgsql lookups by writing a
1537 separate quoting function and calling it from the lookup code. The fix
1538 should probably also be applied to nisplus, ibase and oracle lookups, but
1539 since I cannot test any of those, I have not disturbed their existing code.
1541 2. A hit in the callout cache for a specific address caused a log line with no
1542 reason for rejecting RCPT. Now it says "Previous (cached) callout
1543 verification failure".
1545 3. There was an off-by-one bug in the queryprogram router. An over-long
1546 return line was truncated at 256 instead of 255 characters, thereby
1547 overflowing its buffer with the terminating zero. As well as fixing this, I
1548 have increased the buffer size to 1024 (and made a note to document this).
1550 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1551 when Exim is waiting for an SMTP response from a remote server, Exim
1552 restarts its select() call on the socket, thereby resetting its timeout.
1553 This is not a problem when such interrupts are rare. Somebody set up a cron
1554 job to run exiwhat every 2 minutes, which is less than the normal select()
1555 timeout (5 or 10 minutes). This meant that the select() timeout never
1556 kicked in because it was always reset. I have fixed this by comparing the
1557 time when an interrupt arrives with the time at the start of the first call
1558 to select(). If more time than the timeout has elapsed, the interrupt is
1559 treated as a timeout.
1561 5. Some internal re-factoring in preparation for the addition of Sieve
1562 extensions (by MH). In particular, the "personal" test is moved to a
1563 separate function, and given an option for scanning Cc: and Bcc: (which is
1564 not set for Exim filters).
1566 6. When Exim created an email address using the login of the caller as the
1567 local part (e.g. when creating a From: or Sender: header line), it was not
1568 quoting the local part when it contained special characters such as @.
1570 7. Installed new OpenBSD configuration files.
1572 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1573 try to make them clearer.
1575 9. Callout options, other than the timeout value, were being ignored when
1576 verifying sender addresses in header lines. For example, when using
1578 verify = header_sender/callout=no_cache
1580 the cache was (incorrectly) being used.
1582 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1583 only the Cygwin environment.
1585 11. When return_path_on_delivery was set as a log selector, if different remote
1586 addresses in the same message used different return paths and parallel
1587 remote delivery occurred, the wrong values would sometimes be logged.
1588 (Whenever a remote delivery process finished, the return path value from
1589 the most recently started remote delivery process was logged.)
1591 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1592 lines when AUTH and/or TLS are in use. This is the "received protocol"
1593 field. Exim used to use "asmtp" for authenticated SMTP, without any
1594 indication (in the protocol name) for TLS use. Now it follows the RFC and
1595 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1596 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1597 names appear in log lines as well as in Received: header lines.
1599 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1600 extensions, and comparison tests, and to fix some bugs.
1602 14. Changes to the "personal" filter test:
1604 (1) The test was buggy in that it was just doing the equivalent of
1605 "contains" tests on header lines. For example, if a user's address was
1606 anne@some.where, the "personal" test would incorrectly be true for
1608 To: susanne@some.where
1610 This test is now done by extracting each address from the header in turn,
1611 and checking the entire address. Other tests that are part of "personal"
1612 are now done using regular expressions (for example, to check local parts
1613 of addresses in From: header lines).
1615 (2) The list of non-personal local parts in From: addresses has been
1616 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1617 taken from the Sieve specification recommendations.
1619 (3) If the message contains any header line starting with "List-" it is
1620 treated as non-personal.
1622 (4) The test for "circular" in the Subject: header line has been removed
1623 because it now seems ill-conceived.
1625 15. Minor typos in src/EDITME comments corrected.
1627 16. Installed latest exipick from John Jetmore.
1629 17. If headers_add on a router specified a text string that was too long for
1630 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1631 of string_sprintf() is now avoided.
1633 18. $message_body_size was not set (it was always zero) when running the DATA
1634 ACL and the local_scan() function.
1636 19. For the "mail" command in an Exim filter, no default was being set for
1637 the once_repeat time, causing a random time value to be used if "once" was
1638 specified. (If the value happened to be <= 0, no repeat happened.) The
1639 default is now 0s, meaning "never repeat". The "vacation" command was OK
1640 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1641 (I found it when inspecting the code).
1643 20. There is now an overall timeout for performing a callout verification. It
1644 defaults to 4 times the callout timeout, which applies to individual SMTP
1645 commands during the callout. The overall timeout applies when there is more
1646 than one host that can be tried. The timeout is checked before trying the
1647 next host. This prevents very long delays if there are a large number of
1648 hosts and all are timing out (e.g. when the network connections are timing
1649 out). The value of the overall timeout can be changed by specifying an
1650 additional sub-option for "callout", called "maxwait". For example:
1652 verify = sender/callout=5s,maxwait=20s
1654 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1655 to the end before writing, but this should make it even safer).
1657 22. Exim was forgetting that it had advertised PIPELINING for the second and
1658 subsequent messages on an SMTP connection. It was also not resetting its
1659 memory on STARTTLS and an internal HELO.
1661 23. When Exim logs an SMTP synchronization error within a session, it now
1662 records whether PIPELINING has been advertised or not.
1664 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1665 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1666 rather than long int.
1668 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1670 26. Added the never_mail option to autoreply.
1676 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1677 crash if the getsockname() call failed; this can happen if a connection is
1678 closed very soon after it is established. The problem was simply in the
1679 order in which certain operations were done, causing Exim to try to write
1680 to the SMTP stream before it had set up the file descriptor. The bug has
1681 been fixed by making things happen in the correct order.
1687 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1688 before the connection was closed, thus losing the rejection response.
1690 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1691 some early Solaris releases, but causes trouble in current releases where
1692 socklen_t is defined.
1694 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1697 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1700 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1701 that releasing the top of it at the end releases what was used for sub-
1702 expansions (unless the block got too big). However, discard this block if
1703 the first thing is a variable or header, so that we can use its block when
1704 it is dynamic (useful for very large $message_headers, for example).
1706 6. Lookups now cache *every* query, not just the most recent. A new, separate
1707 store pool is used for this. It can be recovered when all lookup caches are
1708 flushed. Lookups now release memory at the end of their result strings.
1709 This has involved some general refactoring of the lookup sources.
1711 7. Some code has been added to the store_xxx() functions to reduce the amount
1712 of flapping under certain conditions.
1714 8. log_incoming_interface used to affect only the <= reception log lines. Now
1715 it causes the local interface and port to be added to several more SMTP log
1716 lines, for example "SMTP connection from", and rejection lines.
1718 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1720 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1722 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1723 could be overwritten at the end of the current message (or the start of a
1724 new message if it was set in a HELO ACL). The value is now preserved for
1725 the duration of the SMTP connection.
1727 12. If a transport had a headers_rewrite setting, and a matching header line
1728 contained an unqualified address, that address was qualified, even if it
1729 did not match any rewriting rules. The underlying bug was that the values
1730 of the flags that permit the existence of unqualified sender and recipient
1731 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1732 non-local messages, and by -bnq for local messages) were not being
1733 preserved with the message after it was received.
1735 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1736 "next input=" as part of the text comprising the host identity instead of
1737 the correct text. The code was using the same buffer for two different
1738 strings. However, depending on which order the printing function evaluated
1739 its arguments, the bug did not always show up. Under Linux, for example, my
1740 test suite worked just fine.
1742 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1743 doesn't work with some older versions of Perl. It has been changed to "my",
1744 which in any case is probably the better facility to use.
1746 15. A really picky compiler found some instances of statements for creating
1747 error messages that either had too many or two few arguments for the format
1750 16. The size of the buffer for calls to the DNS resolver has been increased
1751 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1752 for addresses that have a lot of PTR records. This alleviates a problem; it
1753 does not fully solve it.
1755 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1756 buffer now truncates the list and logs the incident, which is the same
1757 action as happens when Exim is looking up a host name and its aliases.
1758 Previously in this situation something unpredictable would happen;
1759 sometimes it was "internal error: store_reset failed".
1761 18. If a server dropped the connection unexpectedly when an Exim client was
1762 using GnuTLS and trying to read a response, the client delivery process
1763 crashed while trying to generate an error log message.
1765 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1766 string, for example:
1768 warn message = H1: something\nH2: something
1770 the text was added as a single header line from Exim's point of view
1771 though it ended up OK in the delivered message. However, searching for the
1772 second and subsequent header lines using $h_h2: did not work. This has been
1773 fixed. Similarly, if a system filter added multiple headers in this way,
1774 the routers could not see them.
1776 20. Expanded the error message when iplsearch is called with an invalid key to
1777 suggest using net-iplsearch in a host list.
1779 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1780 ACLs are no longer actually imposed (and a message to that effect is
1783 22. If a "gecos" field in a passwd entry contained escaped characters, in
1784 particular, if it contained a \" sequence, Exim got it wrong when building
1785 a From: or a Sender: header from that name. A second bug also caused
1786 incorrect handling when an unquoted " was present following a character
1787 that needed quoting.
1789 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1790 was not being matched caselessly.
1792 24. Arranged for all hyphens in the exim.8 source to be escaped with
1795 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1796 properly. Recipient callout cache records were still being keyed to include
1797 the sender, even when use_sender was set false. This led to far more
1798 callouts that were necessary. The sender is no longer included in the key
1799 when use_sender is false.
1801 26. Added "control = submission" modifier to ACLs.
1803 27. Added the ${base62d: operator to decode base 62 numbers.
1805 28. dnsdb lookups can now access SRV records.
1807 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1808 the configuration file.
1810 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1811 (-v) mode. This makes the output for a verbose queue run more intelligible.
1813 31. Added a use_postmaster feature to recipient callouts.
1815 32. Added the $body_zerocount variable, containing the number of binary zero
1816 bytes in the message body.
1818 33. The time of last modification of the "new" subdirectory is now used as the
1819 "mailbox time last read" when there is a quota error for a maildir
1822 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1824 35. Added +ignore_unknown as a special item in host lists.
1826 36. Code for decoding IPv6 addresses in host lists is now included, even if
1827 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1828 address was recognized as an IP address, but was then not correctly decoded
1829 into binary, causing unexpected and incorrect effects when compared with
1836 1. Very minor rewording of debugging text in manualroute to say "list of
1837 hosts" instead of "hostlist".
1839 2. If verify=header_syntax was set, and a header line with an unqualified
1840 address (no domain) and a large number of spaces between the end of the
1841 name and the colon was received, the reception process suffered a buffer
1842 overflow, and (when I tested it) crashed. This was caused by some obsolete
1843 code that should have been removed. The fix is to remove it!
1845 3. When running in the test harness, delay a bit after writing a bounce
1846 message to get a bit more predictability in the log output.
1848 4. Added a call to search_tidyup() just before forking a reception process. In
1849 theory, someone could use a lookup in the expansion of smtp_accept_max_
1850 per_host which, without the tidyup, could leave open a database connection.
1852 5. Added the variables $recipient_data and $sender_data which get set from a
1853 lookup success in an ACL "recipients" or "senders" condition, or a router
1854 "senders" option, similar to $domain_data and $local_part_data.
1856 6. Moved the writing of debug_print from before to after the "senders" test
1859 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1860 problems for message scanning, either using a data ACL, or using
1861 local_scan() because the Received: header was not generated till after they
1862 were called (in order to set the time as the time of reception completion).
1863 I have revised the way this works. The header is now generated after the
1864 body is received, but before the ACL or local_scan() are called. After they
1865 are run, the timestamp in the header is updated.
1871 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1872 before starting a queue runner without re-exec. This happened only when
1873 deliver_drop_privilege was set or when the Exim user was set to root. The
1874 effect of the bug was that timeouts during subsequent deliveries caused
1875 crashes instead of being properly handled. The handler is now left at its
1876 default (and expected) setting.
1878 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1879 message, again when deliver_drop_privilege is set or Exim is run as root.
1880 The bug described in (1) was not present in this case, but the tidying up
1881 of the other signals was missing. I have made the two cases consistent.
1883 3. The ignore_target_hosts setting on a manualroute router was being ignored
1884 for hosts that were looked up using the /MX notation.
1886 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1889 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1890 operated on the sender address. After changing the $sender_address to <>
1891 for the sender address verify, Exim was re-instated it as the original
1892 (before rewriting) address, but remembering that it had rewritten it, so it
1893 wasn't rewriting it again. This bug also had the effect of breaking the
1894 sender address verification caching when the sender address was rewritten.
1896 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1897 This has been changed so that if the ip literal address matches
1898 ignore_target_hosts, the router declines.
1900 7. Added expansion conditions match_domain, match_address, and match_local_
1901 part (NOT match_host).
1903 8. The placeholder for the Received: header didn't have a length field set.
1905 9. Added code to Exim itself and to exim_lock to test for a specific race
1906 condition that could lead to file corruption when using MBX delivery. The
1907 issue is with the lockfile that is created in /tmp. If this file is removed
1908 after a process has opened it but before that process has acquired a lock,
1909 there is the potential for a second process to recreate the file and also
1910 acquire a lock. This could lead to two Exim processes writing to the file
1911 at the same time. The added code performs the same test as UW imapd; it
1912 checks after acquiring the lock that its file descriptor still refers to
1913 the same named file.
1915 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1916 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1919 11. Added the smtp_active_hostname option. If used, this will typically be made
1920 to depend on the incoming interface address. Because $interface_address is
1921 not set up until the daemon has forked a reception process, error responses
1922 that can happen earlier (such as "too many connections") no longer contain
1925 12. If an expansion in a condition on a "warn" statement fails because a lookup
1926 defers, the "warn" statement is abandoned, and the next ACL statement is
1927 processed. Previously this caused the whole ACL to be aborted.
1929 13. Added the iplsearch lookup type.
1931 14. Added ident_timeout as a log selector.
1933 15. Added tls_certificate_verified as a log selector.
1935 16. Added a global option tls_require_ciphers (compare the smtp transport
1936 option of the same name). This controls incoming TLS connections.
1938 17. I finally figured out how to make tls_require_ciphers do a similar thing
1939 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1940 before starting the TLS session.
1942 18. Tabs are now shown as \t in -bP output.
1944 19. If the log selector return_path_on_delivery was set, Exim crashed when
1945 bouncing a message because it had too many Received: header lines.
1947 20. If two routers both had headers_remove settings, and the first one included
1948 a superfluous trailing colon, the final name in the first list and the
1949 first name in the second list were incorrectly joined into one item (with a
1950 colon in the middle).
1956 1. Added -C and -D options to the exinext utility, mainly to make it easier
1957 to include in the automated testing, but these could be helpful when
1958 multiple configurations are in use.
1960 2. The exinext utility was not formatting the output nicely when there was
1961 an alternate port involved in the retry record key, nor when there was a
1962 message id as well (for retries that were specific to a specific message
1963 and a specific host). It was also confused by IPv6 addresses, because of
1964 the additional colons they contain. I have fixed the IPv4 problem, and
1965 patched it up to do a reasonable job for IPv6.
1967 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1968 delivery, the log line now contains "pipelined" if PIPELINING was used.
1970 4. An SMTP transport process used to panic and die if the bind() call to set
1971 an explicit outgoing interface failed. This has been changed; it is now
1972 treated in the same way as a connect() failure.
1974 5. A reference to $sender_host_name in the part of a conditional expansion
1975 that was being skipped was still causing a DNS lookup. This no longer
1978 6. The def: expansion condition was not recognizing references to header lines
1979 that used bh_ and bheader_.
1981 7. Added the _cache feature to named lists.
1983 8. The code for checking quota_filecount in the appendfile transport was
1984 allowing one more file than it should have been.
1986 9. For compatibility with Sendmail, the command line option
1994 and sets the incoming protocol and host name (for trusted callers). The
1995 host name and its colon can be omitted when only the protocol is to be set.
1996 Note the Exim already has two private options, -pd and -ps, that refer to
1997 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1998 "s", but I don't think that's a major issue.
2000 10. A number of refactoring changes to the code, none of which should affect
2003 (a) The number of logging options was getting close to filling up the
2004 32-bit word that was used as a bit map. I have split them into two classes:
2005 those that are passed in the argument to log_write(), and those that are
2006 only ever tested independently outside of that function. These are now in
2007 separate 32-bit words, so there is plenty of room for expansion again.
2008 There is no change in the user interface or the logging behaviour.
2010 (b) When building, for example, log lines, the code previously used a
2011 macro that called string_cat() twice, in order to add two strings. This is
2012 not really sufficiently general. Furthermore, there was one instance where
2013 it was actually wrong because one of the argument was used twice, and in
2014 one call a function was used. (As it happened, calling the function twice
2015 did not affect the overall behaviour.) The macro has been replaced by a
2016 function that can join an arbitrary number of extra strings onto a growing
2019 (c) The code for expansion conditions now uses a table and a binary chop
2020 instead of a serial search (which was left over from when there were very
2021 few conditions). Also, it now recognizes conditions like "pam" even when
2022 the relevant support is not compiled in: a suitably worded error message is
2023 given if an attempt is made to use such a condition.
2025 11. Added ${time_interval:xxxxx}.
2027 12. A bug was causing one of the ddress fields not to be passed back correctly
2028 from remote delivery subprocesses. The field in question was not being
2029 subsequently used, so this caused to problems in practice.
2031 13. Added new log selectors queue_time and deliver_time.
2033 14. Might have fixed a bug in maildirsizefile handling that threw up
2034 "unexpected character" debug warnings, and recalculated the data
2035 unnecessarily. In any case, I expanded the warning message to give more
2038 15. Added the message "Restricted characters in address" to the statements in
2039 the default ACL that block characters like @ and % in local parts.
2041 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2042 Three changes have been made:
2044 (a) There was a serious bug; a negative response to MAIL caused the whole
2045 recipient domain to be cached as invalid, thereby blocking all messages
2046 to all local parts at the same domain, from all senders. This bug has
2047 been fixed. The domain is no longer cached after a negative response to
2048 MAIL if the sender used is not empty.
2050 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2053 (c) A new callout option, "use_sender" has been added for people who want
2054 the modified behaviour.
2060 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2063 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2064 indeed breaks things for older releases.
2066 3. Added additional logging to the case where there is a problem reading data
2067 from a filter that is running in a subprocess using a pipe, in order to
2068 try to track down a specific problem.
2070 4. Testing facility fudge: when running in the test harness and attempting
2071 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2072 getting "No route to host". Convert this to a timeout.
2074 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2077 6. Some OS don't have socklen_t but use size_t instead. This affects the
2078 fifth argument of getsockopt() amongst other things. This is now
2079 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2080 can be set for individual OS. I have set it for SunOS5, OSF1, and
2081 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2082 some earlier ones do not.
2084 7. Change 4.30/15 was not doing the test caselessly.
2086 8. The standard form for an IPv6 address literal was being rejected by address
2087 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2088 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2089 this, as well as the form without the "IPv6" on the front (but only when
2090 address literals are enabled, of course).
2092 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2094 10. Exim crashed if a message with an empty sender address specified by -f
2095 encountered a router with an errors_to setting. This could be provoked only
2096 by a command such as
2100 where an empty string was supplied; "<>" did not hit this bug.
2102 11. Installed PCRE release 4.5.
2104 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2105 remained set. It is now erased.
2107 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2108 times from message ids (which are base 36 rather than the normal 62).
2110 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2111 were being counted as actual protocol errors, and logged if the log
2112 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2113 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2114 and DATA following a set of rejected RCPTs do not count as protocol errors.
2115 In other words, Exim assumes they were pipelined, though this may not
2116 actually be the case. Of course, in all cases the client gets an
2117 appropriate error code.
2119 15. If a lookup fails in an ACL condition, a message about the failure may
2120 be available; it is used if testing the ACL cannot continue, because most
2121 such messages specify what the cause of the deferral is. However, some
2122 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2123 that caused an old message to be retained and used if a later statement
2124 caused a defer, replacing the real cause of the deferral.
2126 16. If an IP address had so many PTR records that the DNS lookup buffer
2127 was not large enough to hold them, Exim could crash while trying to process
2128 the truncated data. It now detects and logs this case.
2130 17. Further to 4.21/58, another change has been made: if (and only if) the
2131 first line of a message (the first header line) ends with CRLF, a bare LF
2132 in a subsequent header line has a space inserted after it, so as not to
2133 terminate the header.
2135 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2136 unnecessarily, used atoi() instead of strtol(), and didn't check the
2137 termination when getting file sizes from file names by regex.
2139 19. Completely re-implemented the support for maildirsize files, in the light
2140 of a number of problems with the previous contributed implementation
2141 (4.30/29). In particular:
2143 . If the quota is zero, the maildirsize file is maintained, but no quota is
2146 . If the maildir directory does not exist, it is created before any attempt
2147 to write a maildirsize file.
2149 . The quota value in the file is just a cache; if the quota is changed in
2150 the transport, the new value overrides.
2152 . A regular expression is available for excluding directories from the
2155 20. The autoreply transport checks the characters in options that define the
2156 message's headers; it allows continued headers, but it was checking with
2157 isspace() after an embedded newline instead of explicitly looking for a
2160 21. If all the "regular" hosts to which an address was routed had passed their
2161 expiry times, and had not reached their retry times, the address was
2162 bounced, even if fallback hosts were defined. Now Exim should go on to try
2165 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2166 equivalent code in the SMTP transport. Some hosts send humungous responses
2167 to HELO/EHLO, more than 1024 it seems.
2169 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2170 gives compiler warnings in some environments. I've now done it "properly",
2173 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2174 (because of problems with the built-in one) was declared to return uschar *
2175 instead of char *, causing compiler failure.
2177 25. Fixed a file descriptor leak when processing alias/forward files.
2179 26. Fixed a minor format string issue in dbfn.c.
2181 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2183 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2184 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2187 29. When a sender address is verified, it is cached, to save repeating the test
2188 when there is more than one recipient in a message. However, when the
2189 verification involves a callout, it is possible for different callout
2190 options to be set for different recipients. It is too complicated to keep
2191 track of this in the cache, so now Exim always runs a verification when a
2192 callout is required, relying on the callout cache for the optimization.
2193 The overhead is duplication of the address routing, but this should not be
2196 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2197 to be verified with callout=postmaster, and the main callout worked but the
2198 postmaster check failed, the verification correctly failed. However, if a
2199 subsequent RCPT command asked for sender verification *without* the
2200 postmaster check, incorrect caching caused this verification also to fail,
2203 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2204 it was not caching the DNS options (qualify_single, search_parents) that
2205 were used when the lookup failed. A subsequent lookup with different
2206 options therefore always gave the same answer, though there were cases
2207 where it should not have. (Example: a "domains = !$mx_any" option on a
2208 dnslookup router: the "domains" option is always processed without any
2209 widening, but the router might have qualify_single set.) Now Exim uses the
2210 cached value only when the same options are set.
2212 32. Added John Jetmore's "exipick" utility to the distribution.
2214 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2215 than a timeout (e.g. a certificate is required, and is not provided), an
2216 Exim server now closes the connection immediately. Previously it waited for
2217 the client to close - but if the client is SSL, it seems that they each
2218 wait for each other, leading to a delay before one of them times out.
2220 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2221 maintained 0.8.x compatibility because I don't think many are using it, and
2222 it is clearly obsolete.
2224 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2227 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2228 client certificate was expired. A simple patch fixes this, though I don't
2229 understand the full logic of why the verify callback is called multiple
2232 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2233 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2234 which causes problems with some clients (such as the Certicom SSL Plus
2235 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2236 disables the coutermeasure allowing Eudora to connect."
2238 38. Exim was not checking that a write() to a log file succeeded. This could
2239 lead to Bad Things if a log got too big, in particular if it hit a file
2240 size limit. Exim now panics and dies if it cannot write to a log file, just
2241 as it does if it cannot open a log file.
2243 39. Modified OS/Makefile-Linux so that it now contains
2245 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2247 The two -D definitions ensure that Exim is compiled with large file
2248 support, which makes it possible to handle log files that are bigger than
2251 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2252 instance) a domain was checked against a named list that involved a lookup,
2253 causing $domain_data to be set, then another domain was checked against the
2254 same list, then the first domain was re-checked, the value of $domain_data
2255 after the final check could be wrong. In particular, if the second check
2256 failed, it could be set empty. This bug probably also applied to
2259 41. The strip_trailing_dot option was not being applied to the address given
2260 with the -f command-line option.
2262 42. The code for reading a message's header from the spool was incrementing
2263 $received_count, but never initializing it. This meant that the value was
2264 incorrect (doubled) while delivering a message in the same process in which
2265 it was received. In the most common configuration of Exim, this never
2266 happens - a fresh exec is done - but it can happen when
2267 deliver_drop_privilege is set.
2269 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2270 it now includes up to 150 characters of the unexpected data in the log
2273 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2274 and building data strings. The size of both of these buffers was 10 000
2275 bytes - far larger than anybody would *ever* want, thought I. Needless to
2276 say, somebody hit the limit. I have increased the maximum line length to
2277 20 000 and the maximum data length of concatenated lines to 100 000. I have
2278 also fixed two bugs, because there was no checking on these buffers. Tsk,
2279 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2280 buffer is too small.
2282 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2283 lsearch lookups. Now it does.
2285 46. When parsing a route_list item in a manualroute router, a fixed-length
2286 buffer was used for the list of hosts. I made this 1024 bytes long,
2287 thinking that nobody would ever have a list of hosts that long. Wrong.
2288 Somebody had a whole pile of complicated expansion conditions, and the
2289 string was silently truncated, leading to an expansion error. It turns out
2290 that it is easier to change to an unlimited length (owing to other changes
2291 that have happened since this code was originally written) than to build
2292 structure for giving a limitation error. The length of the item that
2293 expands into the list of hosts is now unlimited.
2295 47. The lsearch lookup could not handle data where the length of text line was
2296 more than 4095 characters. Such lines were truncated, leading to shortened
2297 data being returned. It should now handle lines of any length.
2299 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2300 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2303 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2304 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2305 no longer happens in comment lines. A list of replacements is now placed
2306 at the head of all of the source files, except those whose only change is
2307 to replace PERL_COMMAND in the very first #! line.
2309 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2310 messages on the queue, with a bottom-up merge sort, using code contributed
2311 by Michael Haardt. This should make operations like -bp somewhat faster on
2312 large queues. It won't affect queue runners, except when queue_run_in_order
2315 51. Installed eximstats 1.31 in the distribution.
2317 52. Added support for SRV lookups to the dnslookup router.
2319 53. If an ACL referred to $message_body or $message_body_end, the value was not
2320 reset for any messages that followed in the same SMTP session.
2322 54. The store-handling optimization for building very long strings was not
2323 differentiating between the different store pools. I don't think this
2324 actually made any difference in practice, but I've tidied it.
2326 55. While running the routers to verify a sender address, $sender_address
2327 was still set to the sender address. This is wrong, because when routing to
2328 send a bounce to the sender, it would be empty. Therefore, I have changed
2329 it so that, while verifying a sender address, $sender_address is set to <>.
2330 (There is no change to what happens when verifying a recipient address.)
2332 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2333 target A or AAAA records (if not already returned) without resetting the
2334 qualify_single or search_parents options of the DNS resolver. These are
2335 inappropriate in this case because the targets of MX and SRV records must
2336 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2337 target that, when qualified, matched something in the local domain. These
2338 two options are now turned off when doing these lookups.
2340 57. It seems that at least some releases of Reiserfs (which does not have the
2341 concept of a fixed number of inodes) returns zero and not -1 for the
2342 number of available inodes. This interacted badly with check_spool_inodes,
2343 which assumed that -1 was the "no such thing" setting. What I have done is
2344 to check that the total number of inodes is greater than zero before doing
2345 the test of how many are available.
2347 58. When a "warn" ACL statement has a log_message modifier, the message is
2348 remembered, and not repeated. This is to avoid a lot of repetition when a
2349 message has many recipients that cause the same warning to be written.
2350 Howewer, Exim was preserving the list of already written lines for an
2351 entire SMTP session, which doesn't seem right. The memory is now reset if a
2352 new message is started.
2354 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2355 debugging output unless log_rewrite was also set.
2357 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2358 of (int)(handle) when we know that handle contains (void *)(-1).
2360 61. The Exim daemon panic-logs an error return when it closes the incoming
2361 connection. However "connection reset by peer" seems to be common, and
2362 isn't really an error worthy of noting specially, so that particular error
2365 62. When Exim is trying to find all the local interfaces, it used to panic and
2366 die if the ioctl to get the interface flags failed. However, it seems that
2367 on at least one OS (Solaris 9) it is possible to have an interface that is
2368 included in the list of interfaces, but for which you get a failure error
2369 for this call. This happens when the interface is not "plumbed" into a
2370 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2371 failure of the "get flags" call assumes that the interface is down.
2373 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2374 makes life easier for people who are doing arithmetic on fields extracted
2375 from dates, where you often get leading zeros that should not be
2376 interpreted as octal.
2378 64. Added qualify_domain to the redirect router, to override the global
2381 65. If a pathologically long header line contained very many addresses (the
2382 report of this problem mentioned 10 000) and each of them was rewritten,
2383 Exim could use up a very large amount of memory. (It kept on making new
2384 copies of the header line as it rewrote, and never released the old ones.)
2385 At the expense of a bit more processing, the header rewriting function has
2386 been changed so that it no longer eats memory in this way.
2388 66. The generation of the Received: header has been moved from the time that a
2389 message starts to be received, to the time that it finishes. The timestamp
2390 in the Received: header should now be very close to that of the <= log
2391 line. There are two side-effects of this change:
2393 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2394 logged header lines no longer include the local Received: line, because
2395 it has not yet been created. The same applies to a copy of the message
2396 that is returned to a non-SMTP sender when a message is rejected.
2398 (b) When a filter file is tested using -bf, no additional Received: header
2399 is added to the test message. After some thought, I decided that this
2402 This change does not affect the value of $received_for. It is still set
2403 after address rewriting, but before local_scan() is called.
2405 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2407 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2408 gave an unhelpful panic error message, and a defer error. I have managed to
2409 change this behaviour so that it now rejects any supplied certificate,
2410 which seems right, as the list of acceptable certificates is empty.
2412 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2413 gave an unhelpful defer error. I have not managed to make this reject any
2414 supplied certificates, but the error message it gives is "no certificate
2415 supplied", which is not helpful.
2417 70. exigrep's output now also includes lines that are not associated with any
2418 message, but which match the given pattern. Implemented by a patch from
2419 Martin Sluka, which also tidied up the Perl a bit.
2421 71. Recipient callout verification, like sender verification, was using <> in
2422 the MAIL FROM command. This isn't really the right thing, since the actual
2423 sender may affect whether the remote host accepts the recipient or not. I
2424 have changed it to use the actual sender in the callout; this means that
2425 the cache record is now keyed on a recipient/sender pair, not just the
2426 recipient address. There doesn't seem to be a real danger of callout loops,
2427 since a callout by the remote host to check the sender would use <>.
2428 [SEE ABOVE: changed after hitting problems.]
2430 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2431 temporary errors. However, in the case of such a code being given after
2432 the end of a data transmission (i.e. after ".") Exim was failing to write
2433 a retry record for the message. (Yes, there was some broken host that was
2434 actually sending 8xx at this point.)
2436 73. An unknown lookup type in a host list could cause Exim to panic-die when
2437 the list was checked. (An example that provoked this was putting <; in the
2438 middle of a list instead of at the start.) If this happened during a DATA
2439 ACL check, a -D file could be left lying around. This kind of configuration
2440 error no longer causes Exim to die; instead it causes a defer errror. The
2441 incident is still logged to the main and panic logs.
2443 74. Buglet left over from Exim 3 conversion. The message "too many messages
2444 in one connection" was written to the rejectlog but not the mainlog, except
2445 when address rewriting (yes!) was being logged.
2447 75. Added write_rejectlog option.
2449 76. When a system filter was run not as root (that is, when system_filter_user
2450 was set), the values of the $n variables were not being returned to the
2451 main process; thus, they were not subsequently available in the $sn
2454 77. Added +return_path_on_delivery log selector.
2456 78. A connection timeout was being treated differently from recipients deferred
2457 when testing hosts_max_try with a message that was older than the host's
2458 retry timeout. (The host should not be counted, thus allowing all hosts to
2459 be tried at least once before bouncing.) This may have been the cause of an
2460 occasionally reported bug whereby a message would remain on the queue
2461 longer than the retry timeout, but would be bounced if a delivery was
2462 forced. I say "may" because I never totally pinned down the problem;
2463 setting up timeout/retry tests is difficult. See also the next item.
2465 79. The ultimate address timeout was not being applied to errors that involved
2466 a combination of host plus message (for example, a timeout on a MAIL
2467 command). When an address resolved to a number of possible hosts, and they
2468 were not all tried for each delivery (e.g. because of hosts_max_try), a
2469 message could remain on the queue longer than the retry timeout.
2471 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2474 81. Fixed an obscure SMTP outgoing bug which required at least the following
2475 conditions: (a) there was another message waiting for the same server;
2476 (b) the server returned 5xx to all RCPT commands in the first message so
2477 that the message was not completed; (c) the server dropped the connection
2478 or gave a negative response to the RSET that Exim sends to abort the
2479 transaction. The observed case was a dropped connection after DATA that had
2480 been sent in pipelining mode. That is, the server had advertised PIPELINING
2481 but was not implementing it correctly. The effect of the bug was incorrect
2482 behaviour, such as trying another host, and this could lead to a crash.
2488 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2489 and daemon.c were passed as pointers to ints; they should have been
2490 pointers to socklen_t variables (which are typically unsigned ints).
2492 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2495 3. Fixed a really odd bug that affected only the testing scheme; patching a
2496 certain fixed string in the binary changed the value of another string that
2497 happened to be identical to the end of the original first string.
2499 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2500 name", it returns that address as the IP address. On some operating
2501 systems (e.g. Solaris), it also passes back the IP address string as the
2502 "host name". However, on others (e.g. Linux), it passes back an empty
2503 string. Exim wasn't checking for this, and was changing the host name to an
2504 empty string, assuming it had been canonicized.
2506 5. Although rare, it is permitted to have more than one PTR record for a given
2507 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2508 all the names associated with an address, because they do in Solaris.
2509 However, it seems that they do not in Linux for data that comes from the
2510 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2511 I found this out when I moved to a new Linux workstation and tried to run
2512 the Exim test suite.
2514 To get round this problem I have changed the code so that it now does its
2515 own call to the DNS to look up PTR records when searching for a host name.
2516 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2517 addresses that are only in /etc/hosts are still found.
2519 This behaviour is, however, controlled by an option called host_lookup_
2520 order, which defaults to "bydns:byaddr". If people want to use the other
2521 order, or indeed, just use one or the other means of lookup, they can
2522 specify it in this variable.
2524 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2525 some operating systems, this comes back from gethostbyaddr() as an empty
2526 string, and this is what Exim used to test for. However, it seems that in
2527 other systems, "." is yielded. Exim now tests for this case too.
2529 7. The values of check_spool_space and check_log_space are now held internally
2530 as a number of kilobytes instead of an absolute number of bytes. If a
2531 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2532 kilobyte. This means that much larger values can be stored.
2534 8. Exim monitor: an attempt to get the action menu when not actually pointing
2535 at a message produces an empty menu entitled "No message selected". This
2536 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2537 no entries in it ("Shell widget menu has zero width and/or height"). So I
2538 have added a single, blank menu entry in this case.
2540 9. Added ${quote_local_part.
2542 10. MIME decoding is now applied to the contents of Subject: header lines when
2545 11. Now that a reference to $sender_host_address automatically causes a reverse
2546 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2547 host lookup before query-style lookups in lists that might use this
2548 variable. This has therefore been abolished, and the "net-" prefix is no
2549 longer necessary for query-style lookups.
2551 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2552 have been a typo for LFLAGS, so it has been changed.
2554 13. The install script calls Exim with "-C /dev/null" in order to find the
2555 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2556 to be output. Howeve, since Exim outputs its version number before the
2557 error, it didn't break the script. It just looked ugly. I fixed this by
2558 always allowing "-C /dev/null" if the caller is root.
2560 14. Ignore overlarge ACL variable number when reading spool file - insurance
2561 against a later release with more variables having written the file.
2563 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2564 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2565 this, as well as the form without the "IPv6" on the front.
2567 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2568 OS/Makefile-Darwin file.
2570 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2571 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2573 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2574 variable. However, some OS use ints for inodes. I've added cast to long int
2575 to get rid of the compiler warning.
2577 19. I had forgotten to lock out "/../" in configuration file names when
2578 ALT_CONFIG_PREFIX was set.
2580 20. Routers used for verification do not need to specify transports. However,
2581 if such a router generated a host list, and callout was configured, Exim
2582 crashed, because it could not find a port number from the (non-existent)
2583 transport. It now assumes port 25 in this circumstance.
2585 21. Added the -t option to exigrep.
2587 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2588 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2589 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2590 them is set, LOOKUP_LSEARCH is forced.
2592 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2594 24. Added sender and host information to the "rejected by local_scan()" log
2595 line; previously there was no indication of these.
2597 25. Added .include_if_exists.
2599 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2600 Linux. It turns out that not all file systems support this. Apparently some
2601 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2602 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2603 not supported on the file descriptor, is now ignored when Exim is trying to
2604 sync a directory. This applies only to Linux.
2606 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2608 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2609 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2612 29. Added support for maildirsize files from supplied patch (modified a bit).
2614 30. The use of :fail: followed by an empty string could lead Exim to respond to
2615 sender verification failures with (e.g.):
2617 550 Verification failed for <xxx>
2618 550 Sender verify failed
2620 where the first response line was missing the '-' that indicates it is not
2621 the final line of the response.
2623 31. The loop for finding the name of the user that called Exim had a hardwired
2624 limit of 10; it now uses the value of finduser_retries, which is used for
2625 all other user lookups.
2627 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2630 33. Exim was neglecting to zero errno before one call of strtol() when
2631 expanding a string and expecting an integer value. On some systems this
2632 resulted in spurious "integer overflow" errors. Also, it was casting the
2633 result into an int without checking.
2635 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2636 did not work. The code looks as if it has *never* worked, though it appears
2637 to have been documented since at least releast 1.62. I have made it work.
2639 35. The "timeout_DNS" error in retry rules, also documented since at least
2640 1.62, also never worked. As it isn't clear exactly what this means, and
2641 clearly it isn't a major issue, I have abolished the feature by treating it
2642 as "timeout", and writing a warning to the main and panic logs.
2644 36. The display of retry rules for -brt wasn't always showing the error code
2647 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2648 timeout_connect_A, timeout_connect_MX.
2650 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2651 to the empty sender.
2653 39. The daemon was not analysing the content of -oX till after it had closed
2654 stderr and disconnected from the controlling terminal. This meant that any
2655 syntax errors were only noted on the panic log, and the return code from
2656 the command was 0. By re-arranging the code a little, I've made the
2657 decoding happen first, so such errors now appear on stderr, and the return
2658 code is 1. However, the actual setting up of the sockets still happens in
2659 the disconnected process, so errors there are still only recorded on the
2662 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2663 connections (as happens on some IP stacks) was logged at start up time as
2664 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2665 it from "IPv6 and IPv4", which means that two separate sockets are being
2668 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2669 says whether AF_INET or AF_INET6 was passed as an argument.
2671 42. Exiwhat output was messed up when time zones were included in log
2674 43. Exiwhat now gives more information about the daemon's listening ports,
2675 and whether -tls-on-connect was used.
2677 44. The "port" option of the smtp transport is now expanded.
2679 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2680 silently ignored. Now an error message is written to the main and panic
2683 46. There's a new ACL modifier called "logwrite" which writes to a log file
2684 as soon as it is encountered.
2686 47. Added $local_user_uid and $local_user_gid at routing time.
2688 48. Exim crashed when trying to verify a sender address that was being
2691 49. Exim was recognizing only a space character after ".include". It now also
2692 recognizes a tab character.
2694 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2695 extracting the relevant information from the specification. The man page no
2696 longer contains scrambled data for the -d option, and I've added a section
2697 at the front about calling Exim under different names.
2699 51. Added "extra_headers" argument to the "mail" command in filter files.
2701 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2704 53. Installed eximstats 1.29.
2706 54. Added transport_filter_timeout as a generic transport option.
2708 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2709 Cc: header lines. This was required by RFC 822, but it not required by RFC
2712 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2713 incoming messages that did not have them. Now it does so only if the
2714 message originates locally, that is, if there is no associated remote host
2715 address. When Resent- header lines are present, this applies to the Resent-
2716 lines rather than the non-Resent- lines.
2718 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2719 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2721 58. Messages for configuration errors now include the name of the main
2722 configuration file - useful now that there may be more than one file in a
2723 list (.included file names were always shown).
2725 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2726 for those rare installations that do not start the daemon as root or run it
2727 setuid root. I've cut out the call to initgroups() if the daemon is not
2730 60. The Exim user and group can now be bound into the binary as text strings
2731 that are looked up at the start of Exim's processing.
2733 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2735 62. Added $mailstore_basename variable.
2737 63. Installed patch to sieve.c from Michael Haardt.
2739 64. When Exim failed to open the panic log after failing to open the main log,
2740 the original message it was trying to log was written to stderr and debug
2741 output, but if they were not available (the usual case in production), it
2742 was lost. Now it is written to syslog before the two lines that record the
2743 failures to open the logs.
2745 65. Users' Exim filters run in subprocesses under the user's uid. It is
2746 possible for a "deliver" command or an alias in a "personal" command to
2747 provoke an address rewrite. If logging of address rewriting is configured,
2748 this fails because the process is not running as root or exim. There may be
2749 a better way of dealing with this, but for the moment (because 4.30 needs
2750 to be released), I have disabled address rewrite logging when running a
2751 filter in a non-root, non-exim process.
2757 1. The buildconfig auxiliary program wasn't quoting the value set for
2758 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2759 not defined. This bug was masked in 4.22 by the effect that was fixed in
2762 2. Some messages that were rejected after a message id was allocated were
2763 shown as "incomplete" by exigrep. It no longer does this for messages that
2764 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2766 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2767 have allow_domain_literals set, the ID did not get logged in the <= line.
2768 Domain literals are now always recognized in Message-ID: header lines.
2770 4. The first argument for a ${extract expansion item is the key name or field
2771 number. Leading and trailing spaces in this item were not being ignored,
2772 causing some misleading effects.
2774 5. When deliver_drop_privilege was set, single queue runner processes started
2775 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2776 same command in the process it spins off) were not dropping privilege.
2778 6. When the daemon running as "exim" started a queue runner, it always
2779 re-executed Exim in the spun-off process. This is a waste of effort when
2780 deliver_drop_privilege is set. The new process now just calls the
2781 queue-runner function directly.
2787 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2790 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2791 ignored. Though the use of -oP was forcing the writing of a pid file, it
2792 was always written to the default place.
2794 3. If the message "no IP address found for host xxxx" is generated during
2795 incoming verification, it is now followed by identification of the incoming
2796 connection (so you can more easily find what provoked it).
2798 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2800 5. Added some features to "harden" Exim a bit more against certain attacks:
2802 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2803 be put in Local/Makefile. This is like the never_users runtime option,
2804 but it cannot be overridden. The default setting is "root".
2806 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2807 prefix string with which any file named in a -C command line option
2810 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2811 is retained for -C and -D only if the caller of Exim is root. Without
2812 it, the exim user may also use -C and -D and retain privilege.
2814 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2815 command line option is disabled.
2817 6. Macro names set by the -D option must start with an upper case letter, just
2818 like macro names defined in the configuration file.
2820 7. Added "dereference=" facility to LDAP.
2822 8. Two instances of the typo "uknown" in the source files are fixed.
2824 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2825 the Configure-Makefile script screwed up while processing it.
2827 10. Incorporated PCRE 4.4.
2829 11. The SMTP synchronization check was not operating right at the start of an
2830 SMTP session. For example, it could not catch a HELO sent before the client
2831 waited for the greeting. There is now a check for outstanding input at the
2832 point when the greeting is written. Because of the duplex, asynchronous
2833 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2834 way, but not yet received, when the check is performed.
2836 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2837 on TCP/IP sockets, because this apparently causes some broken clients to
2840 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2841 unchanged) from the Cygwin maintainer.
2843 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2844 was supported instead of testing for mbx. Effectively a typo.
2846 15. The spa authenticator server code was not checking that the input it
2847 received was valid base64.
2849 16. The debug output line for the "set" modifier in ACLs was not showing the
2850 name of the variable that was being set.
2852 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2854 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2855 lookup on its own. Something else was needed to trigger the lookup. For
2856 example, a match in host_lookup or the need for a host name in a host list.
2857 Now, if $sender_host_name is referenced and the host name has not yet been
2858 looked up, a lookup is performed. If the lookup fails, the variable remains
2859 empty, and $host_lookup_failed is set to "1".
2861 19. Added "eqi" as a case-independent comparison operator.
2863 20. The saslauthd authentication condition could segfault if neither service
2864 nor realm was specified.
2866 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2867 error message that was logged was misleading, and incoming SMTP
2868 connections were dropped. The message is now more accurate, and temporary
2869 errors are given to SMTP connections.
2871 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2872 (except RSET) with a 421 error, until QUIT is received. However, it was
2873 failing to send a response to QUIT.
2875 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2876 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2877 ACL is now run after the helo_try_verify_hosts code.
2879 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2880 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2881 case-independent, so other case variants are also recognized). Apparently
2882 some systems use these upper case variants.
2884 25. If more than two messages were waiting for the same host, and a transport
2885 filter was specified for the transport, Exim sent two messages over the
2886 same TCP/IP connection, and then failed with "socket operation on non-
2887 socket" when it tried to send the third.
2889 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2891 27. The extern definition of crypt16() in expand.c was not being excluded when
2892 the OS had its own crypt16() function.
2894 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2895 as a preferred synonym for return_size_limit, both as an option and as an
2898 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2900 30. Changed the default configuration ACL to relax the local part checking rule
2901 for addresses that are not in any local domains. For these addresses,
2902 slashes and pipe symbols are allowed within local parts, but the sequence
2903 /../ is explicitly forbidden.
2905 31. SPA server authentication was not clearing the challenge buffer before
2908 32. log_message in a "warn" ACL statement was writing to the reject log as
2909 well as to the main log, which contradicts the documentation and doesn't
2910 seem right (because no rejection is happening). So I have stopped it.
2912 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2913 However, I am unable to do any testing of this.
2915 34. Fixed an infelicity in the appendfile transport. When checking directories
2916 for a mailbox, to see if any needed to be created, it was accidentally
2917 using path names with one or more superfluous leading slashes; tracing
2918 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2920 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2921 contents are added to the log line that is written for every discarded
2922 recipient. (Previously a log_message setting was ignored.)
2924 36. The ${quote: operator now quotes the string if it is empty.
2926 37. The install script runs exim in order to find its version number. If for
2927 some reason other than non-existence or emptiness, which it checks, it
2928 could not run './exim', it was installing it with an empty version number,
2929 i.e. as "exim-". This error state is now caught, and the installation is
2932 38. An argument was missing from the function that creates an error message
2933 when Exim fails to connect to the socket for saslauthd authentication.
2934 This could cause Exim to crash, or give a corrupted message.
2936 39. Added isip, isip4, and isip6 to ${if conditions.
2938 40. The ACL variables $acl_xx are now saved with the message, and can be
2939 accessed later in routers, transports, and filters.
2941 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2942 strings in the file are not string-expanded.
2944 42. If a MAIL command specified a SIZE value that was too large to fit into an
2945 int variable, the check against message_size_limit failed. Such values are
2946 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2947 day this will have to be increased, but I don't think I want to be around
2948 when emails are that large.
2955 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2956 iconv() is not standard in FreeBSD.
2958 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2959 IPv6 enabled. The observed symptom was a segmentation fault on return from
2960 the function os_common_find_running_interfaces() in src/os.c.
2962 3. In the check_special_case() function in daemon.c I had used "errno" as an
2963 argument name, which causes warnings on some systems. This was basically a
2964 typo, since it was named "eno" in the comments!
2966 4. The code that waits for the clock to tick (at a resolution of some fraction
2967 of a second) so as to ensure message-id uniqueness was always waiting for
2968 at least one whole tick, when it could have waited for less. [This is
2969 almost certainly not relevant at current processor speeds, where it is
2970 unlikely to ever wait at all. But we try to future-proof.]
2972 5. The function that sleeps for a time interval that includes fractions of a
2973 second contained a race. It did not block SIGALRM between setting the
2974 timer, and suspending (a couple of lines later). If the interval was short
2975 and the sigsuspend() was delayed until after it had expired, the suspension
2976 never ended. On busy systems this could lead to processes getting stuck for
2979 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2980 process, before it forks any delivery processes. The open lookup caching
2981 mechanism meant that the open file or database connection was passed into
2982 the delivery process. The problem was that delivery processes always tidy
2983 up cached lookup data. This could cause a problem for the next delivery
2984 process started by the queue runner, because the external queue runner
2985 process does not know about the closure. So the next delivery process
2986 still has data in the lookup cache. In the case of a file lookup, there was
2987 no problem because closing a file descriptor in a subprocess doesn't affect
2988 the parent. However, if the lookup was caching a connection to a database,
2989 the connection was closed, and the second delivery process was likely to
2990 see errors such as "PGSQL: query failed: server closed the connection
2991 unexpectedly". The problem has been fixed by closing all cached lookups
2992 in a queue runner before running a delivery process.
2994 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2995 seem to have the "const" qualifier which it has on other OS. I've
2998 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2999 *of the same type* (client or server) with the same public name that an
3000 error should be diagnosed.
3002 9. When Exim looked up a host name for an IP address, but failed to find the
3003 original IP address when looking up the host name (a safety check), it
3004 output the message "<ip address> does not match any IP for NULL", which was
3005 confusing, to say the least. The bug was that the host name should have
3006 appeared instead of "NULL".
3008 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
3009 user that is built into the binary, and the -C or -D options is used, root
3010 privilege is dropped before the configuration file is read. In addition,
3011 logging is switched to stderr instead of the normal log files. If the
3012 configuration then re-defines the Exim user, the unprivileged environment
3013 is probably not what is expected, so Exim logs a panic warning message (but
3016 However, if deliver_drop_privilege is set, the unprivileged state may well
3017 be exactly what is intended, so the warning has been cut out in that case,
3018 and Exim is allowed to try to write to its normal log files.
3024 1. smtp_return_error_details was not giving details for temporary sender
3025 or receiver verification errors.
3027 2. Diagnose a configuration error if two authenticators have the same public
3030 3. Exim used not to create the message log file for a message until the first
3031 delivery attempt. This could be confusing when incoming messages were held
3032 for policy or load reasons. The message log file is now created at the time
3033 the message is received, and an initial "Received" line is written to it.
3035 4. The automatically generated man page for command line options had a minor
3036 bug that caused no ill effects; however, a more serious problem was that
3037 the procedure for building the man page automatically didn't always
3038 operate. Consequently, release 4.20 contains an out-of-date version. This
3039 shouldn't happen again.
3041 5. When building Exim with embedded Perl support, the script that builds the
3042 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3043 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3045 6. The freeze_tell option was not being used for messages that were frozen on
3046 arrival, either by an ACL or by local_scan().
3048 7. Added the smtp_incomplete_transaction log selector.
3050 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3051 was accepting AUTH without a new EHLO.
3053 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3054 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3057 10. Logging of TCP/IP connections (when configured) now happens in the main
3058 daemon process instead of the child process, so that the TCP/IP connection
3059 count is more accurate (but it can never be perfect).
3061 11. The use of "drop" in a nested ACL was not being handled correctly in the
3062 outer ACL. Now, if condition failure induced by the nested "drop" causes
3063 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3064 or "require"), the connection is dropped.
3066 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3067 that yield "discard" can now be used with an "accept" or a "discard" verb,
3068 but an error is generated for any others (because I can't see a useful way
3069 to define what should happen).
3071 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3072 are now processed in the same way as lines in the Exim configuration file.
3073 In particular, continuation lines are supported.
3075 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3077 15. Added -ti meaning -t -i.
3079 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3080 domains, and warn by logging if others are found.
3082 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3083 returned from GIFCONF when Exim is trying to find the list of interfaces on
3084 a host. The code in os.c has been modified to copy each ifreq to an aligned
3085 structure in all cases.
3087 Also, in some cases, the returned ifreq's were being copied to a 'struct
3088 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3089 means the last couple of bytes of an IPv6 address could be chopped if the
3090 ifreq contained only a normal sockaddr (14 bytes storage).
3092 18. Named domain lists were not supported in the hosts_treat_as_local option.
3093 An entry such as +xxxx was not recognized, and was treated as a literal
3096 19. Ensure that header lines added by a DATA ACL are included in the reject log
3097 if the ACL subsequently rejects the message.
3099 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3100 MD5 (which is deprecated).
3102 21. When testing a filter file using -bf, Exim was writing a message when it
3103 took the sender from a "From " line in the message, but it was not doing so
3104 when it took $return_path from a Return-Path: header line. It now does.
3106 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3107 with a valid header line field name (a series of printing characters
3108 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3110 23. Changed "disc" in the source to "disk" to conform to the documentation and
3111 the book and for uniformity.
3113 24. Ignore Sendmail's -Ooption=value command line item.
3115 25. When execve() failed while trying to run a command in a pipe transport,
3116 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3117 could be confused with a return value of 69 from the command itself. This
3118 has been changed to 127, the value the shell returns if it is asked to run
3119 a non-existent command. The wording for the related log line suggests a
3120 non-existent command as the problem.
3122 26. If received_header_text expands to an empty string, do not add a Received:
3123 header line to the message. (Well, it adds a token one on the spool, but
3124 marks it "old" so that it doesn't get used or transmitted.)
3126 27. Installed eximstats 1.28 (addition of -nt option).
3128 28. There was no check for failure on the call to getsockname() in the daemon
3129 code. This can fail if there is a shortage of resources on the system, with
3130 ENOMEM, for example. A temporary error is now given on failure.
3132 29. Contrary to the C standard, it seems that in some environments, the
3133 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3134 program. Exim now does this explicitly; it affects the formatting of
3135 timestamps using strftime().
3137 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3138 complaints. I've now initialized all the variables, to avoid this.
3140 32. Header lines added by a system filter were not being "seen" during
3141 transport-time rewrites.
3143 33. The info_callback() function passed to OpenSSL is set up with type void
3144 (*)(SSL *, int, int), as described somewhere. However, when calling the
3145 function (actually a macro) that sets it up, the type void(*)() is
3146 expected. I've put in a cast to prevent warnings from picky compilers.
3148 34. If a DNS black list lookup found a CNAME record, but there were no A
3149 records associated with the domain it pointed at, Exim crashed.
3151 35. If a DNS black list lookup returned more than one A record, Exim ignored
3152 all but the first. It now scans all returned addresses if a particular IP
3153 value is being sought. In this situation, the contents of the
3154 $dnslist_value variable are a list of all the addresses, separated by a
3157 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3158 to accept a host name and all its aliases if the forward lookup for any of
3159 them yielded the IP address of the incoming connection. Now it accepts only
3160 those names whose forward lookup yields the correct IP address. Any other
3161 names are discarded. This closes a loophole whereby a rogue DNS
3162 administrator could create reverse DNS records to break through a
3163 wildcarded host restriction in an ACL.
3165 37. If a user filter or a system filter that ran in a subprocess used any of
3166 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3167 the wrong values were passed to the pipe command ($thisaddress had the
3168 value of $0, $0 had the value of $1, etc). This bug was introduced by
3169 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3171 38. Improved the line breaking for long SMTP error messages from ACLs.
3172 Previously, if there was no break point between 40 and 75 characters, Exim
3173 left the rest of the message alone. Two changes have been made: (a) I've
3174 reduced the minimum length to 35 characters; (b) if it can't find a break
3175 point between 35 and 75 characters, it looks ahead and uses the first one
3176 that it finds. This may give the occasional overlong line, but at least the
3177 remaining text gets split now.
3179 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3180 file descriptors might be low, and that setting 1000 would always raise it.
3181 It turns out that in some environments, the limit is already over 1000 and
3182 that lowering it causes trouble. So now Exim takes care not to decrease it.
3184 40. When delivering a message, the value of $return_path is set to $sender_
3185 address at the start of routing (routers may change the value). By an
3186 oversight, this default was not being set up when an address was tested by
3187 -bt or -bv, which affected the outcome if any router or filter referred to
3190 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3191 log only when "message" or "log_message" are set. However, if one of the
3192 conditions was an address verification, or a call to a nested ACL, the
3193 messages generated by the underlying test were being passed through. This
3194 no longer happens. The underlying message is available in $acl_verify_
3195 message for both "message" and "log_message" expansions, so it can be
3196 passed through if needed.
3198 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3199 new expansion $bh_ to give the encoded byte string without charset
3200 translation. Translation happens only if iconv() is available; HAVE_ICONV
3201 indicates this at build time. HEADERS_CHARSET gives the charset to
3202 translate to; headers_charset can change it in the configuration, and
3203 "headers charset" can change it in an individual filter file.
3205 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3206 that creates RFC 2047 encoded "words" labels them as that charset instead
3207 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3208 expansion operator; (ii) when Exim creates a From: line for a local
3209 message; (iii) when a header line is rewritten to include a "phrase" part.
3211 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3212 buggy, causing it to skip the first lines of messages whose message ID
3213 ended in 'D'. This would not have bitten before Exim release 4.14, because
3214 message IDs were unlikely to end in 'D' before then. The effect was to have
3215 incorrect size information for certain domains.
3217 45. #include "config.h" was missing at the start of the crypt16.c module. This
3218 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3221 46. If there was a timeout during a "random" callout check, Exim treated it as
3222 a failure of the random address, and carried on sending RSET and the real
3223 address. If the delay was just some slowness somewhere, the response to the
3224 original RCPT would be taken as a response to RSET and so on, causing
3225 mayhem of various kinds.
3227 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3228 when I implemented it. It didn't allow for the fact that some option values
3229 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3230 the right test for positive values.
3232 48. Domain names in DNS records are case-independent. Exim always looks them up
3233 in lower case. Some resolvers return domain names in exactly the case they
3234 appear in the zone file, that is, they may contain uppercase letters. Not
3235 all resolvers do this - some return always lower case. Exim was treating a
3236 change of case by a resolver as a change of domain, similar to a widening
3237 of a domain abbreviation. This triggered its re-routing code and so it was
3238 trying to route what was effectively the same domain again. This normally
3239 caused routing to fail (because the router wouldn't handle the domain
3240 twice). Now Exim checks for this case specially, and just changes the
3241 casing of the domain that it ultimately uses when it transmits the message
3244 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3247 50. If a filter generated a file delivery with a non-absolute name (possible if
3248 no home directory exists for the router), the forbid_file option was not
3251 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3252 the existing equality matching.
3254 52. Exim was using ints instead of ino_t variables in some places where it was
3255 dealing with inode numbers.
3257 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3258 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3259 called TMPDIR, and if it finds it is different, it changes its value.
3261 54. The smtp_printf() function is now made available to local_scan() so
3262 additional output lines can be written before returning. There is also an
3263 smtp_fflush() function to enable the detection of a dropped connection.
3264 The variables smtp_input and smtp_batched_input are exported to
3267 55. Changed the default runtime configuration: the message "Unknown user"
3268 has been removed from the ACL, and instead placed on the localuser router,
3269 using the cannot_route_message feature. This means that any verification
3270 failures that generate their own messages won't get overridden. Similarly,
3271 the "Unrouteable address" message that was in the ACL for unverifiable
3272 relay addresses has also been removed.
3274 56. Added hosts_avoid_esmtp to the smtp transport.
3276 57. The exicyclog script was not checking for the esoteric option
3277 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3278 will work only if exicyclog is run under the appropriate euid.
3280 58. Following a discussion on the list, the rules by which Exim recognises line
3281 endings on incoming messages have been changed. The -dropcr and drop_cr
3282 options are now no-ops, retained only for backwards compatibility. The
3283 following line terminators are recognized: LF CRLF CR. However, special
3284 processing applies to CR:
3286 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3287 nor a local message in the state where . is a terminator.
3289 (ii) If a bare CR is encountered in a header line, an extra space is added
3290 after the line terminator so as not to end the header. The reasoning
3291 behind this is that bare CRs in header lines are most likely either
3292 to be mistakes, or people trying to play silly games.
3294 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3295 was being incorrectly given as 18 bytes larger than it should have been.
3296 This is a VOB (very old bug).
3298 60. This may never have affected anything current, but just in case it has:
3299 When the local host is found other than at the start of a list of hosts,
3300 the local host, those with the same MX, and any that follow, are discarded.
3301 When the list in question was part of a longer list of hosts, the following
3302 hosts (not currently being processed) were also being discarded. This no
3303 longer happens. I'm not sure if this situation could ever has previously
3306 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3309 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3310 header. This is something that is recommended in a new Internet Draft, and
3311 is something that is documented as being done by Sendmail. There are two
3312 possible values. For messages generated by the autoreply transport, Exim
3315 Auto-Submitted: auto-replied
3317 whereas for all other generated messages (e.g. bounces) it adds
3319 Auto-Submitted: auto-generated
3321 63. The "personal" condition in filters now includes a test for the
3322 Auto-Submitted: header. If it contains the string "auto-" the message it
3323 not considered personal.
3325 64. Added rcpt_include_affixes as a generic transport option.
3327 65. Added queue_only_override (default true).
3329 66. Added the syslog_duplication option.
3331 67. If what should have been the first header line of a message consisted of
3332 a space followed by a colon, Exim was mis-interpreting it as a header line.
3333 It isn't of course - it is syntactically invalid and should therefore be
3334 treated as the start of the message body. The misbehaviour could have
3335 caused a number of strange effects, including loss of data in subsequent
3336 header lines, and spool format errors.
3338 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3339 client host had authenticated. This control can now be exercised by an ACL
3340 for more flexibility.
3342 69. By default, callouts do not happen when testing with -bh. There is now a
3343 variant, -bhc, which does actually run the callout code, including
3344 consulting and updating the callout cache.
3346 70. Added support for saslauthd authentication, courtesy of Alexander
3349 71. If statvfs() failed on the spool or log directories while checking their
3350 size for availability, Exim confusingly gave the error "space shortage".
3351 Furthermore, in debugging mode it crashed with a floating point exception.
3352 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3353 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3354 serious problem, Exim now writes to the main and panic logs when this
3355 happens, with details of the failure. It then refuses to accept the
3356 incoming message, giving the message "spool directory problem" or "log
3357 directory problem" with a 421 code for SMTP messages.
3359 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3360 0, 1, and 2 exist, because some OS complain for execs without them (see
3361 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3362 descriptors. However, the code omitted to check that the open succeeded,
3363 causing mysterious errors if for some reason the permissions on /dev/null
3364 got screwed. Now Exim writes a message to the main and panic logs, and
3365 bombs out if it can't open /dev/null.
3367 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3368 interact so that it is all more flexible. It is supposed to remain
3369 backwards compatible. Also added extra_local_interfaces.
3371 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3372 to bomb out with an assertion failure - to the client this appears as a
3373 connection drop. This problem occurs in the part of the code that was taken
3374 from the Samba project. Fortunately, the assertion is in a very simple
3375 function, so I have fixed this by reproducing the function inline in the
3376 one place where it is called, and arranging for authentication to fail
3377 instead of killing the process with assert().
3379 75. The SPA client code was not working when the server requested OEM rather
3380 than Unicode encoding.
3382 76. Added code to make require_files with a specific uid setting more usable in
3383 the case where statting the file as root fails - usually a non-root-mounted
3384 NFS file system. When this happens and the failure is EACCES, Exim now
3385 forks a subprocess and does the per-uid checking as the relevant uid.
3387 77. Added process_log_path.
3389 78. If log_file_path was not explicitly set, a setting of check_log_space or
3390 check_log_inodes was ignored.
3392 79. If a space check for the spool or log partitions fails, the incident is now
3393 logged. Of course, in the latter case the data may get lost...
3395 80. Added the %p formatting code to string_format() so that it can be used to
3396 print addresses in debug_print(). Adjusted all the address printing in the
3397 debugging in store.c to use %p rather than %d.
3399 81. There was a concern that a line of code in smtp_in.c could overflow a
3400 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3401 initially expressed, the concern was not well-founded, because trailing
3402 spaces are removed early. However, if the trailing spaces were followed by
3403 a NULL, they did not get removed, so the overflow was possible. Two fixes
3406 (a) I re-wrote the offending code in a cleaner fashion.
3407 (b) If an incoming SMTP command contains a NULL character, it is rejected
3410 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3411 runs initgroups(), so that if the Exim user is in any additional groups,
3412 they will be used during message reception.
3418 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / exim.git / blob