1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.241 2005/09/28 10:46:48 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
10 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
17 TK/01 Added the "success_on_redirect" address verification option. See
18 NewStuff for rationale and an example.
20 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
22 PH/02 Patch to exigrep to allow it to work on syslog lines.
24 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
25 fread() to skip over the body file's header line, because in Cygwin the
26 header line is locked and is inaccessible.
28 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
29 co-exist for some time) to make it clear that it is the Exim ID that is
30 referenced, not the Message-ID: header line.
32 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
33 string_format() function, because snprintf() does not exist on all
36 PH/06 The use of forbid_filter_existstest now also locks out the use of the
37 ${stat: expansion item.
39 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
40 protocol synchronization error", to keep the pedants happy.
42 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
43 well as for IRIX systems, when gcc is being used. See the host.c source
46 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
48 PH/10 Named domain lists were not working if used in a queue_smtp_domains
51 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
52 transport and to the smtp transport in LMTP mode.
54 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
56 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
57 run a filter in a subprocess. This could lead to confusion in subsequent
58 lookups in the parent process. There should also be a search_tidyup() at
59 the end of the subprocess.
61 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
62 only if the host matched helo_try_verify_hosts, which caused the
63 verification to occur when the EHLO/HELO command was issued. The ACL just
64 tested the remembered result. Now, if a previous verification attempt has
65 not happened, "verify = helo" does it there and then.
67 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
69 TK/03 Fix log output including CR from clamd.
71 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
72 exist provoked a memory error which could cause a segfault.
74 PH/15 Installed PCRE 6.2
76 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
78 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
79 of the problem. Specifically, suggested +O2 rather than +O1 for the
82 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
84 PH/20 If a delivery was routed to a non-standard port by means of an SRV
85 record, the port was not correctly logged when the outgoing_port log
86 selector was set (it logged the transort's default port).
88 PH/21 Added support for host-specific ports to manualroute, queryprogram,
89 fallback_hosts, and "hosts" in the smtp transport.
91 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
92 host errors such as "Connection refused".
94 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
95 authentication with radiusclient 0.4.9:
97 - Error returned from rc_read_config was caught wrongly
98 - Username/password not passed on to radius server due to wrong length.
100 The presumption is that some radiusclient API changes for 4.51/PH/17
101 were not taken care of correctly. The code is still untested by me (my
102 Linux distribution still has 0.3.2 of radiusclient), but it was
103 contributed by a Radius user.
105 PH/24 When doing a callout, the value of $domain wasn't set correctly when
106 expanding the "port" option of the smtp transport.
108 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
109 while reading a MIME header. Thanks to Tom Hughes for a patch.
111 PH/24 Include config.h inside local_scan.h so that configuration settings are
114 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
115 that in an ACL for RCPT (for example), you can examine exactly what was
118 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
119 commands, but it was not correctly comparing the address with the actual
120 client host address. Thus, it would show the EHLO address in Received:
121 header lines when this was not necessary.
123 PH/27 Added the % operator to ${eval:}.
125 PH/28 Exim tries to create and chdir to its spool directory when it starts;
126 it should be ignoring failures (because with -C, for example, it has lost
127 privilege). It wasn't ignoring creation failures other than "already
130 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
133 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
134 because an input line was too long, either on its own, or by virtue of
135 too many continuations, the temporary file was not being removed, and the
136 return code was incorrect.
138 PH/31 Missing "BOOL" in function definition in filtertest.c.
140 PH/32 Applied Sieve patches from the maintainer.
142 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
144 PH/33 Added "verify = not_blind".
146 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
147 Local/Makefile (with some defaults set). These are used in built scripts
148 such as exicyclog, but they have never been used in the exim_install
149 script (though there are many overriding facilities there). I have
150 arranged that the exim_install script now takes note of these two
153 PH/35 Installed configuration files for Dragonfly.
155 PH/36 When a locally submitted message by a trusted user did not contain a
156 From: header, and the sender address was obtained from -f or from an SMTP
157 MAIL command, and the trusted user did not use -F to supply a sender
158 name, $originator_name was incorrectly used when constructing a From:
159 header. Furthermore, $originator_name was used for submission mode
160 messages from external hosts without From: headers in a similar way,
161 which is clearly wrong.
163 PH/37 Added control=suppress_local_fixups.
165 PH/38 When log_selector = +received_sender was set, and the addition of the
166 sender made the log line's construction buffer exactly full, or one byte
167 less than full, an overflow happened when the terminating "\n" was
170 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
171 when the result of a list match is failure because a DNS lookup failed.
173 PH/40 RM_COMMAND is now used in the building process.
175 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
176 the "build-* directories that it finds.
178 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
179 address in a domain literal was a prefix of an interface address.
181 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
182 when verifying a sender address, unless rewrite_headers is false.
184 PH/44 Wrote a long comment about why errors_to addresses are verified as
185 recipients, not senders.
187 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
188 the ratelimit ACL was added.
190 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
192 PH/46 Added extra information about PostgreSQL errors to the error string.
194 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
195 avoiding the need to install special zones in a real server. This is
196 backwards compatible; if it can't find the fake resolver, it drops back.
197 Thus, both old and new test suites can be run.
199 TF/02 Added util/ratelimit.pl
201 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
202 clock is set back in time.
204 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
205 Candler <B.Candler@pobox.com>.
207 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
208 OK for addresses that are the result of redirections.
210 PH/48 A number of further additions for the benefit of the new test suite,
211 including a fake gethostbyname() that interfaces to the fake DNS resolver
214 TF/06 The fix for widen_domains has also been applied to qualify_single and
215 search_parents which are the other dnslookup options that can cause
218 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
221 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
223 TF/07 Exim produced the error message "an SRV record indicated no SMTP
224 service" if it encountered an MX record with an empty target hostname.
225 The message is now "an MX or SRV record indicated no SMTP service".
227 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
228 if the DNS of the sending site is misconfigured. This is quite a
229 common situation. This change restores the behaviour of treating a
230 helo verification defer as a failure.
232 PH/51 If self=fail was set on a router, the bounce message did not include the
233 actual error message.
239 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
241 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
242 command itself ended in error, the underlying message about the transport
243 filter timeout was being overwritten with the pipe command error. Now the
244 underlying error message should be appended to the second error message.
246 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
248 PH/02 Reduce the amount of output that "make" produces by default. Full output
249 can still be requested.
251 PH/03 The warning log line about a condition test deferring for a "warn" verb
252 was being output only once per connection, rather than after each
253 occurrence (because it was using the same function as for successful
254 "warn" verbs). This seems wrong, so I have changed it.
256 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
257 it should not have, which might have caused a crash in the right
258 circumstances, but probably never did.
260 PH/04 Installed a modified version of Tony Finch's patch to make submission
261 mode fix the return path as well as the Sender: header line, and to
262 add a /name= option so that you can make the user's friendly name appear
265 TF/03 Added the control = fakedefer ACL modifier.
267 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
268 Mark Lowes for thorough testing.
270 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
272 TK/03 Merged latest SRS patch from Miles Wilton.
274 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
275 with the definition in sysexits.h (which is #included earlier).
276 Fortunately, Exim does not actually use EX_OK. The code used to try to
277 preserve the sysexits.h value, by assumimg that macro definitions were
278 scanned for macro replacements. I have been disabused of this notion,
279 so now the code just undefines EX_OK before #including unistd.h.
281 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
282 in the smtp transport. When a block could not be written in a single
283 write() function, the timeout was being re-applied to each part-write.
284 This seems wrong - if the receiver was accepting one byte at a time it
285 would take for ever. The timeout is now adjusted when this happens. It
286 doesn't have to be particularly precise.
288 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
289 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
291 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
292 without a domain if the check to <postmaster@domain> fails.
294 SC/01 Eximstats: added -xls and the ability to specify output files
295 (patch written by Frank Heydlauf).
297 SC/02 Eximstats: use FileHandles for outputing results.
299 SC/03 Eximstats: allow any combination of xls, txt, and html output.
301 SC/04 Eximstats: fixed display of large numbers with -nvr option
303 SC/05 Eximstats: fixed merging of reports with empty tables.
305 SC/06 Eximstats: added the -include_original_destination flag
307 SC/07 Eximstats: removed tabs and trailing whitespace.
309 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
311 TK/06 MBOX spool code: Add real "From " MBOX separator line
312 so the .eml file is really in mbox format (even though
313 most programs do not really care). Patch from Alex Miller.
315 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
316 The latter is generated from $received_to and is only set if the
317 message has one envelope recipient. SA can use these headers,
318 obviously out-of-the-box. Patch from Alex Miller.
320 PH/08 The ${def test on a variable was returning false if the variable's
321 value was "0", contrary to what the specification has always said!
322 The result should be true unless the variable is empty.
324 PH/09 The syntax error of a character other than { following "${if
325 def:variable_name" (after optional whitespace) was not being diagnosed.
326 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
327 accidental colon was present, for example, could give incorrect results.
329 PH/10 Tidied the code in a number of places where the st_size field of a stat()
330 result is used (not including appendfile, where other changes are about
333 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
334 This involved changing a lot of size variables from int to off_t. It
335 should work with maildirs and everything.
337 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
338 spamd dying while we are connected to it.
340 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
343 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
344 function for BDB 4.3.
346 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
348 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
349 the sources for PCRE have been re-arranged and more files are now
352 PH/15 The code I had for printing potentially long long variables in PH/11
353 above was not the best (it lost precision). The length of off_t variables
354 is now inspected at build time, and an appropriate printing format (%ld
355 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
356 to be "long long int" or "long int". This is needed for the internal
357 formatting function string_vformat().
359 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
360 the configuration file to be ":syslog", then the script "guesses" where
361 the logs files are, rather than using the compiled in default. In our
362 case the guess is not the same as the compiled default, so the script
363 suddenly stopped working when I started to use syslog. The patch checks
364 to see if log_file_path is "". If so, it attempts to read it from exim
365 with no configuration file to get the compiled in version, before it
366 falls back to the previous guessing code."
368 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
369 implementing BATV in an Exim configuration. See NewStuff for the gory
372 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
373 Makefile that are specific to HP-UX.
375 PH/18 If the "use_postmaster" option was set for a recipient callout together
376 with the "random" option, the postmaster address was used as the MAIL
377 FROM address for the random test, but not for the subsequent recipient
378 test. It is now used for both.
380 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
381 patch removes a few documentation additions to RFC 3028, because the
382 latest draft now contains them. It adds the new en;ascii-case comparator
383 and a new error check for 8bit text in MIME parts. Comparator and
384 require names are now matched exactly. I enabled the subaddress
385 extension, but it is not well tested yet (read: it works for me)."
387 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
388 rework some of the code of TK/09 above to avoid the hardwired use of
389 "%lld" and "long long". Replaced the call to snprintf() with a call to
392 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
393 records point to non-existent hosts", "retry timeout exceeded", and
394 "retry time not reached for any host after a long failure period".
396 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
397 experimental DomainKeys support:
399 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
400 (2) On an error such as an illegally used "control", the wrong name for
401 the control was given.
403 These problems did NOT occur unless DomainKeys support was compiled.
405 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
407 PH/24 Added ${if match_ip condition.
409 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
410 databases so that it will be absolutely obvious if a crash occurs in the
411 DB library. This is a regular occurrence (often caused by mis-matched
414 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
415 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
416 picked up on a user's system that detects such things. There doesn't seem
417 to be a gcc warning option for this - only an attribute that has to be
418 put on the function's prototype. It seems that in Fedora Core 4 they have
419 set this on a number of new functions. No doubt there will be more in due
422 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
423 specify a transport. However, if an address that was verified by such a
424 router was the subject of a callout, Exim crashed because it tried to
425 read the rcpt_include_affixes from the non-existent transport. Now it
426 just assumes that the setting of that option is false. This bug was
427 introduced by 4.51/PH/31.
429 PH/28 Changed -d+all to exclude +memory, because that information is very
430 rarely of interest, but it makes the output a lot bigger. People tend to
431 do -d+all out of habit.
433 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
434 code in os-type was giving problems when libc.so lives in lib64, like on
437 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
438 aren't the modern standard, and it seems that some systems' include files
439 don't always have them. Exim was already checking for some of the newer
440 ones like T_AAAA, and defining it itself. I've added checks for all the
441 record types that Exim uses.
443 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
444 not automatically generating a new one, as it is supposed to. This
445 prevented TLS from working. If the file did exist, but contained invalid
446 data, a new version was generated, as expected. It was only the case of a
447 non-existent file that was broken.
449 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
450 with a change in libdomainkeys > 0.64.
452 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
453 from DNS. If the selector record carries the flag, it now has
454 precedence over the domain-wide flag.
456 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
458 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
459 the use of a port name that isn't defined in /etc/services) occurred, the
460 message was deferred as in a normal delivery, and thus remained on the
461 spool, instead of being failed because of the mua_wrapper setting. This
462 is now fixed, and I tidied up some of the mua_wrapper messages at the
465 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
466 the messages in a hash of arrays rather than using individual hashes.
467 This is a bit cleaner and results in dramatic memory savings, albeit
468 at a slight CPU cost.
470 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
471 as requested by Marc Sherman.
473 SC/10 Eximstats: added histograms for user specified patterns as requested
476 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
478 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
479 fopen() in the content-scanning modules that did not already have it.
485 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
486 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
488 TK/02 Fix ACL "control" statement not being available in MIME ACL.
490 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
492 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
493 to test Sieve filters that use "vacation".
495 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
496 that changes the way the GnuTLS parameters are stored in the cache file.
497 The new format can be generated externally. For backward compatibility,
498 if the data in the cache doesn't make sense, Exim assumes it has read an
499 old-format file, and it generates new data and writes a new file. This
500 means that you can't go back to an older release without removing the
503 PH/03 A redirect router that has both "unseen" and "one_time" set does not
504 work if there are any delivery delays because "one_time" forces the
505 parent to be marked "delivered", so its unseen clone is never tried
506 again. For this reason, Exim now forbids the simultaneous setting of
509 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
510 redirected to themselves ("homonym" addresses). Read the long ChangeLog
511 entry if you want to know the details. The fix, however, neglected to
512 consider the case when local delivery batching is involved. The test for
513 "previously delivered" was not happening when checking to see if an
514 address could be batched with a previous (undelivered) one; under
515 certain circumstances this could lead to multiple deliveries to the same
518 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
519 in its include files, and this causes problems building Exim.
521 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
522 header_syntax) but Exim was just ignoring anything given after a slash.
523 In particular, this caused confusion with an attempt to use "verify =
524 reverse_host_lookup/defer_ok". An error is now given when options are
525 supplied for verify items that do not have them. (Maybe reverse_host_
526 lookup should have a defer_ok option, but that's a different point.)
528 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
529 defined by RFC 821) to 2048, because there were problems with some AUTH
530 commands, and RFC 1869 says the size should be increased for extended
531 SMTP commands that take arguments.
533 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
536 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
537 "unknown" error; now it says that the functionality isn't in the binary.
539 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
540 an address' error message when a string expansion fails (syntax or
541 whatever). Otherwise the password may appear in the log. Following change
542 PH/42 below, there is no longer a chance of it appearing in a bounce
545 PH/11 Installed exipick version 20050225.0 from John Jetmore.
547 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
548 of its addresses was ever tried. (Bugzilla bug #2.)
550 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
551 the result incorrectly in the debug output. (It correctly added a newline
552 to what was transported.)
554 TF/01 Added $received_time.
556 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
557 commented out examples of how to interface to a virus scanner and to
558 SpamAssassin. Also added commented examples of av_scanner and
559 spamd_address settings.
561 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
562 and controls are allowed in which ACLs. There were a couple of minor
563 errors. Some of the entries in the conditions table (which is a table of
564 where they are NOT allowed) were getting very unwieldy; rewrote them as a
565 negation of where the condition IS allowed.
567 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
569 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
570 header file does not have a version number, so I've had to invent a new
571 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
572 API. The code is untested by me (my Linux distribution still has 0.3.2 of
573 radiusclient), but it was contributed by a Radius user.
575 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
576 files or directories, for OpenSSL.
578 PH/19 When an Exim process that is running as root has to create an Exim log
579 file, it does so in a subprocess that runs as exim:exim so as to get the
580 ownership right at creation (otherwise, other Exim processes might see
581 the file with the wrong ownership). There was no test for failure of this
582 fork() call, which would lead to the process getting stuck as it waited
583 for a non-existent subprocess. Forks do occasionally fail when resources
584 run out. I reviewed all the other calls to fork(); they all seem to check
587 PH/20 When checking for unexpected SMTP input at connect time (before writing
588 the banner), Exim was not dealing correctly with a non-positive return
589 from the read() function. If the client had disconnected by this time,
590 the result was a log entry for a synchronization error with an empty
591 string after "input=" when read() returned zero. If read() returned -1
592 (an event I could not check), uninitialized data bytes were printed.
593 There were reports of junk text (parts of files, etc) appearing after
596 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
598 PH/22 Added support for macro redefinition, and (re)definition in between
599 driver and ACL definitions.
601 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
602 forgetting to use the resulting value; it was using the unexpanded value.
604 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
605 hadn't been configured. The fix is from Juergen Kreileder, who
606 understands it better than I do:
608 "Here's what I see happening with three configured cyrus_sasl
609 authenticators configured (plain, login, cram-md5):
611 On startup auth_cyrus_sasl_init() gets called for each of these.
612 This means three calls to sasl_listmech() without a specified mech_list.
613 => SASL tests which mechs of all available mechs actually work
614 => three warnings about OTP not working
615 => the returned list contains: plain, login, cram-md5, digest-md5, ...
617 With the patch, sasl_listmech() also gets called three times. But now
618 SASL's mech_list option is set to the server_mech specified in the the
619 authenticator. Or in other words, the answer from sasl_listmech()
620 gets limited to just the mech you're testing for (which is different
622 => the return list contains just 'plain' or 'login', 'cram-md5' or
623 nothing depending on the value of ob->server_mech.
625 I've just tested the patch: Authentication still works fine,
626 unavailable mechs specified in the exim configuration are still
627 caught, and the auth.log warnings about OTP are gone."
629 PH/25 When debugging is enabled, the contents of the command line are added
630 to the debugging output, even when log_selector=+arguments is not
633 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
634 answer is "GNU", and only if the return is "GNU/something" is the answer
637 PH/27 $acl_verify_message is now set immediately after the failure of a
638 verification in an ACL, and so is available in subsequent modifiers. In
639 particular, the message can be preserved by coding like this:
641 warn !verify = sender
642 set acl_m0 = $acl_verify_message
644 Previously, $acl_verify_message was set only while expanding "message"
645 and "log_message" when a very denied access.
647 PH/28 Modified OS/os.c-Linux with
649 -#ifndef OS_LOAD_AVERAGE
650 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
652 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
653 nomenclature these days.)
655 PH/29 Installed patch from the Sieve maintainer that adds the options
656 sieve_useraddress and sieve_subaddress to the redirect router.
658 PH/30 In these circumstances:
659 . Two addresses routed to the same list of hosts;
660 . First host does not offer TLS;
661 . First host accepts first address;
662 . First host gives temporary error to second address;
663 . Second host offers TLS and a TLS session is established;
664 . Second host accepts second address.
665 Exim incorrectly logged both deliveries with the TLS parameters (cipher
666 and peerdn, if requested) that were in fact used only for the second
669 PH/31 When doing a callout as part of verifying an address, Exim was not paying
670 attention to any local part prefix or suffix that was matched by the
671 router that accepted the address. It now behaves in the same way as it
672 does for delivery: the affixes are removed from the local part unless
673 rcpt_include_affixes is set on the transport.
675 PH/32 Add the sender address, as F=<...>, to the log line when logging a
676 timeout during the DATA phase of an incoming message.
678 PH/33 Sieve envelope tests were broken for match types other than :is. I have
679 applied a patch sanctioned by the Sieve maintainer.
681 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
682 the uid or gid is negative. A case of a negative gid caused this to be
683 noticed. The fix allows for either to be negative.
685 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
686 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
689 PH/36 The change PH/12 above was broken. Fixed it.
691 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
692 the grounds that routing the same address twice would always produce the
693 same answer. This might have been true once, but it is certainly no
694 longer true now. Routing a child address may depend on the previous
695 routing that produced that child. Some complicated redirection strategies
696 went wrong when messages had multiple recipients, and made Exim's
697 behaviour dependent on the order in which the addresses were given.
699 I have moved the duplicate checking until after the routing is complete.
700 Exim scans the addresses that are assigned to local and remote
701 transports, and removes any duplicates. This means that more work will be
702 done, as duplicates will always all be routed, but duplicates are
703 presumably rare, so I don't expect this is of any significance.
705 For deliveries to pipes, files, and autoreplies, the duplicate checking
706 still happens during the routing process, since they are not going to be
709 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
710 It corrects a timeout issue with spamd. This is Ian's comment: "The
711 background is that sometimes spamd either never reads data from a
712 connection it has accepted, or it never writes response data. The exiscan
713 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
714 blindly assumes that writes won't block so it may never time out."
716 PH/39 Allow G after quota size as well as K and M.
718 PH/40 The value set for $authenticated_id in an authenticator may not contain
719 binary zeroes or newlines because the value is written to log lines and
720 to spool files. There was no check on this. Now the value is run through
721 the string_printing() function so that such characters are converted to
722 printable escape sequences.
724 PH/41 $message_linecount is a new variable that contains the total number of
725 lines in the message. Compare $body_linecount, which is the count for the
728 PH/42 Exim no longer gives details of delivery errors for specific addresses in
729 bounce and delay warning messages, except in certain special cases, which
732 (a) An SMTP error message from a remote host;
733 (b) A message specified in a :fail: redirection;
734 (c) A message specified in a "fail" command in a system filter;
735 (d) A message specified in a FAIL return from the queryprogram router;
736 (e) A message specified by the cannot_route_message router option.
738 In these cases only, Exim does include the error details in bounce and
739 warning messages. There are also a few cases where bland messages such
740 as "unrouteable address" or "local delivery error" are given.
742 PH/43 $value is now also set for the "else" part of a ${run expansion.
744 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
745 being worked on, but at least Exim now implements the latest version to
748 PH/45 In a pipe transport, although a timeout while waiting for the pipe
749 process to complete was treated as a delivery failure, a timeout while
750 writing the message to the pipe was logged, but erroneously treated as a
751 successful delivery. Such timeouts include transport filter timeouts. For
752 consistency with the overall process timeout, these timeouts are now
753 treated as errors, giving rise to delivery failures by default. However,
754 there is now a new Boolean option for the pipe transport called
755 timeout_defer, which, if set TRUE, converts the failures into defers for
756 both kinds of timeout. A transport filter timeout is now identified in
759 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
760 systems where "make" and "gmake" are different, calling "gmake" at top
761 level broke things. I've arranged for the value of $(MAKE) to be passed
762 from the Makefile to this script so that it can call the same version of
766 A note about Exim versions 4.44 and 4.50
767 ----------------------------------------
769 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
770 changes of various kinds. As a consequence, a big documentation update was
771 needed. This delayed the release for rather longer than seemed good, especially
772 in the light of a couple of (minor) security issues. Therefore, the changes
773 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
774 release. So 4.44 and 4.50 are in effect two different branches that both start
777 I have left the 4.50 change log unchanged; it contains all the changes since
778 4.43. The change log for 4.44 is below; many of its items are identical to
779 those for 4.50. This seems to be the most sensible way to preserve the
780 historical information.
786 1. Minor wording change to the doc/README.SIEVE file.
788 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
789 computation of the current number of files was incorrect.
791 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
792 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
795 4. Give more explanation in the error message when the command for a transport
796 filter fails to execute.
798 5. There are several places where Exim runs a non-Exim command in a
799 subprocess. The SIGUSR1 signal should be disabled for these processes. This
800 was being done only for the command run by the queryprogram router. It is
801 now done for all such subprocesses. The other cases are: ${run, transport
802 filters, and the commands run by the lmtp and pipe transports.
804 6. Added CONFIGURE_GROUP build-time option.
806 7. Some older OS have a limit of 256 on the maximum number of file
807 descriptors. Exim was using setrlimit() to set 1000 as a large value
808 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
809 systems. I've change it so that if it can't get 1000, it tries for 256.
811 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
812 was an oversight, and furthermore, ever since the addition of extra
813 controls (e.g. 4.43/32), the checks on when to allow different forms of
814 "control" were broken. There should now be diagnostics for all cases when a
815 control that does not make sense is encountered.
817 9. Added the /retain_sender option to "control=submission".
819 10. $recipients is now available in the predata ACL (oversight).
821 11. Tidy the search cache before the fork to do a delivery from a message
822 received from the command line. Otherwise the child will trigger a lookup
823 failure and thereby defer the delivery if it tries to use (for example) a
824 cached ldap connection that the parent has called unbind on.
826 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
827 of $address_data from the recipient verification was clobbered by the
830 13. The value of address_data from a sender verification is now available in
831 $sender_address_data in subsequent conditions in the ACL statement.
833 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
835 15. Added a new option "connect=<time>" to callout options, to set a different
838 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
839 was its contents. (It was OK if the option was not defined at all.)
841 17. A "Completed" log line is now written for messages that are removed from
842 the spool by the -Mrm option.
844 18. New variables $sender_verify_failure and $recipient_verify_failure contain
845 information about exactly what failed.
847 19. Added -dd to debug only the daemon process.
849 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
850 handles timeouts, both on the server side and network timeouts. Renamed the
851 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
853 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
854 It is now set to "smtps".
856 22. $host_address is now set to the target address during the checking of
859 23. When checking ignore_target_hosts for an ipliteral router, no host name was
860 being passed; this would have caused $sender_host_name to have been used if
861 matching the list had actually called for a host name (not very likely,
862 since this list is usually IP addresses). A host name is now passed as
865 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
866 code that specifies a non-restarting handler (typically sigaction() in
867 modern systems) in an attempt to fix a rare and obscure crash bug.
869 25. Narrowed the window for a race in the daemon that could cause it to ignore
870 SIGCHLD signals. This is not a major problem, because they are used only to
871 wake it up if nothing else does.
873 26. A malformed maildirsize file could cause Exim to calculate negative values
874 for the mailbox size or file count. Odd effects could occur as a result.
875 The maildirsize information is now recalculated if the size or filecount
878 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
879 support for a long time. Removed HAVE_SYS_VFS_H.
881 28. Installed the latest version of exipick from John Jetmore.
883 29. In an address list, if the pattern was not a regular expression, an empty
884 subject address (from a bounce message) matched only if the pattern was an
885 empty string. Non-empty patterns were not even tested. This was the wrong
886 because it is perfectly reasonable to use an empty address as part of a
887 database query. An empty address is now tested by patterns that are
888 lookups. However, all the other forms of pattern expect the subject to
889 contain a local part and a domain, and therefore, for them, an empty
890 address still always fails if the pattern is not itself empty.
892 30. Exim went into a mad DNS loop when attempting to do a callout where the
893 host was specified on an smtp transport, and looking it up yielded more
896 31. Re-factored the code for checking spool and log partition space into a
897 function that finds that data and another that does the check. The former
898 is then used to implement four new variables: $spool_space, $log_space,
899 $spool_inodes, and $log_inodes.
901 32. The RFC2047 encoding function was originally intended for short strings
902 such as real names; it was not keeping to the 75-character limit for
903 encoded words that the RFC imposes. It now respects the limit, and
904 generates multiple encoded words if necessary. To be on the safe side, I
905 have increased the buffer size for the ${rfc2047: expansion operator from
908 33. It is now permitted to omit both strings after an "if" condition; if the
909 condition is true, the result is "true". As before, when the second string
910 is omitted, a false condition yields an empty string. This makes it less
911 cumbersome to write custom ACL and router conditions.
913 34. Failure to deliver a bounce message always caused it to be frozen, even if
914 there was an errors_to setting on the router. The errors_to setting is now
917 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
918 canonical form (fully expanded) before being placed in
919 $sender_host_address.
921 36. The table in the code that translates DNS record types into text (T_A to
922 "A" for instance) was missing entries for NS and CNAME. It is just possible
923 that this could have caused confusion if both these types were looked up
924 for the same domain, because the text type is used as part of Exim's
925 per-process caching. But the chance of anyone hitting this buglet seems
928 37. The dnsdb lookup has been extended in a number of ways.
930 (1) There is a new type, "zns", which walks up the domain tree until it
931 finds some nameserver records. It should be used with care.
933 (2) There is a new type, "mxh", which is like "mx" except that it returns
934 just the host names, not the priorities.
936 (3) It is now possible to give a list of domains (or IP addresses) to be
937 looked up. The behaviour when one of the lookups defers can be
938 controlled by a keyword.
940 (4) It is now possible to specify the separator character for use when
941 multiple records are returned.
943 38. The dnslists ACL condition has been extended: it is now possible to supply
944 a list of IP addresses and/or domains to be looked up in a particular DNS
947 39. Added log_selector=+queue_time_overall.
949 40. When running the queue in the test harness, wait just a tad after forking a
950 delivery process, to get repeatability of debugging output.
952 41. Include certificate and key file names in error message when GnuTLS fails
953 to set them up, because the GnuTLS error message doesn't include the name
954 of the failing file when there is a problem reading it.
956 42. Allow both -bf and -bF in the same test run.
958 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
960 44. The "Exiscan patch" is now merged into the mainline Exim source.
962 45. Sometimes the final signoff response after QUIT could fail to get
963 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
964 < 0 before doing a fflush(). This bug looks as though it goes back to the
965 introduction of TLS in release 3.20, but "sometimes" must have been rare
966 because the tests only now provoked it.
968 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
969 (this can affect the format of dates).
971 47. exim_tidydb, when checking for the continued existence of a message for
972 which it has found a message-specific retry record, was not finding
973 messages that were in split spool directories. Consequently, it was
974 deleting retry records that should have stayed in existence.
976 48. Steve fixed some bugs in eximstats.
978 49. The SPA authentication driver was not abandoning authentication and moving
979 on to the next authenticator when an expansion was forced to fail,
980 contradicting the general specification for all authenticators. Instead it
981 was generating a temporary error. It now behaves as specified.
983 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
984 (the order specifies the preference for clients). The order is now AES256,
985 AES128, 3DES, ARCFOUR128.
987 51. Small patch to Sieve code - explicitly set From: when generating an
990 52. Exim crashed if a remote delivery caused a very long error message to be
991 recorded - for instance if somebody sent an entire SpamAssassin report back
992 as a large number of 550 error lines. This bug was coincidentally fixed by
993 increasing the size of one of Exim's internal buffers (big_buffer) that
994 happened as part of the Exiscan merge. However, to be on the safe side, I
995 have made the code more robust (and fixed the comments that describe what
998 53. Now that there can be additional text after "Completed" in log lines (if
999 the queue_time_overall log selector is set), a one-byte patch to exigrep
1000 was needed to allow it to recognize "Completed" as not the last thing in
1003 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1004 patch that reportedly fixes this has been added. I am not expert enough to
1005 create a test for it. This is what the patch creator wrote:
1007 "I found a little strange behaviour of ldap code when working with
1008 Windows 2003 AD Domain, where users was placed in more than one
1009 Organization Units. When I tried to give exim partial DN, the exit code
1010 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1011 But simultaneously result of request was absolutely normal ldap result,
1012 so I produce this patch..."
1014 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1015 REFERENCE, so I have modified the code to exclude the patch when that macro
1018 55. Some experimental protocols are using DNS PTR records for new purposes. The
1019 keys for these records are domain names, not reversed IP addresses. The
1020 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1021 leaves it alone. Component reversal etc. now happens only for IP addresses.
1024 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1026 57. Double the size of the debug message buffer (to 2048) so that more of very
1027 long debug lines gets shown.
1029 58. The exicyclog utility now does better if the number of log files to keep
1030 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1032 59. Two changes related to the smtp_active_hostname option:
1034 (1) $smtp_active_hostname is now available as a variable.
1035 (2) The default for smtp_banner uses $smtp_active_hostname instead
1036 of $primary_hostname.
1038 60. The host_aton() function is supposed to be passed a string that is known
1039 to be a valid IP address. However, in the case of IPv6 addresses, it was
1040 not checking this. This is a hostage to fortune. Exim now panics and dies
1041 if the condition is not met. A case was found where this could be provoked
1042 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1043 components; fortuitously, this particular loophole had already been fixed
1044 by change 4.50/55 above.
1046 If there are any other similar loopholes, the new check in host_aton()
1047 itself should stop them being exploited. The report I received stated that
1048 data on the command line could provoke the exploit when Exim was running as
1049 exim, but did not say which command line option was involved. All I could
1050 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1051 running as the user.
1054 61. There was a buffer overflow vulnerability in the SPA authentication code
1055 (which came originally from the Samba project). I have added a test to the
1056 spa_base64_to_bits() function which I hope fixes it.
1059 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1060 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1062 63. The daemon start-up calls getloadavg() while still root for those OS that
1063 need the first call to be done as root, but it missed one case: when
1064 deliver_queue_load_max is set with deliver_drop_privilege. This is
1065 necessary for the benefit of the queue runner, because there is no re-exec
1066 when deliver_drop_privilege is set.
1068 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1069 This has been fixed.
1071 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1072 was in use, was not putting the data itself into the right store pool;
1073 consequently, it could be overwritten for a subsequent message in the same
1074 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1077 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1079 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1080 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1081 the function were treating the return as a boolean value, which happened to
1082 work because 0=false and not-0=true, but is not correct code.
1084 68. The host_aton() function was not handling scoped IPv6 addresses (those
1085 with, for example, "%eth0" on the end) correctly.
1087 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1088 negated items (that is, ~something) in unsigned ints. Some compilers
1089 apparently mutter when there is no cast.
1091 70. If an address verification called from an ACL failed, and did not produce a
1092 user-specific message (i.e. there was only a "system" message), nothing was
1093 put in $acl_verify_message. In this situation, it now puts the system
1096 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1097 session; change 4.31/43 added the unwanted input to the log line - except
1098 that it did not do this in the start of session case. It now does.
1100 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1101 This is wrong and can cause the other end to generate a synchronization
1102 error if it is another Exim or anything else that does the synchronization
1103 check. A QUIT command is no longer sent after a timeout.
1105 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1106 during host lookups.
1108 74. The defer_ok option of callout verification was not working if it was used
1109 when verifying addresses in header lines, that is, for this case:
1111 verify = header_sender/callout=defer_ok
1113 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1114 those file descriptors could be used for SMTP connections. If anything
1115 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1116 could be sent to the SMTP client, causing chaos. The daemon now opens
1117 stdin, stdout, and stderr to /dev/null when it puts itself into the
1120 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1121 log by default. The user can override this with suitable Perl magic.
1123 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1124 the log message when discard triggers, was not working for the DATA ACL or
1125 for the non-SMTP ACL.
1127 78. Error message wording change in sieve.c.
1129 79. If smtp_accept_max_per_host was set, the number of connections could be
1130 restricted to fewer than expected, because the daemon was trying to set up
1131 a new connection before checking whether the processes handling previous
1132 connections had finished. The check for completed processes is now done
1133 earlier. On busy systems, this bug wouldn't be noticed because something
1134 else would have woken the daemon, and it would have reaped the completed
1137 80. If a message was submitted locally by a user whose login name contained one
1138 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1139 It caused a spool format error. I have fixed the spool reading code. A
1140 related problem was that the "from" clause in the Received: line became
1141 illegal because of the space(s). It is now covered by ${quote_local_part.
1143 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1146 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1148 83. Patch from Sieve maintainer to fix unterminated string problem in
1149 "vacation" handling.
1151 84. Some minor changes to the Linux configuration files to help with other
1152 OS variants using glibc.
1154 85. One more patch for Sieve to update vacation handling to latest spec.
1157 ----------------------------------------------------
1158 See the note above about the 4.44 and 4.50 releases.
1159 ----------------------------------------------------
1165 1. Change 4.43/35 introduced a bug that caused file counts to be
1166 incorrectly computed when quota_filecount was set in an appendfile
1169 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1170 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1173 3. Give more explanation in the error message when the command for a transport
1174 filter fails to execute.
1176 4. There are several places where Exim runs a non-Exim command in a
1177 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1178 was being done only for the command run by the queryprogram router. It is
1179 now done for all such subprocesses. The other cases are: ${run, transport
1180 filters, and the commands run by the lmtp and pipe transports.
1182 5. Some older OS have a limit of 256 on the maximum number of file
1183 descriptors. Exim was using setrlimit() to set 1000 as a large value
1184 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1185 systems. I've change it so that if it can't get 1000, it tries for 256.
1187 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1188 was an oversight, and furthermore, ever since the addition of extra
1189 controls (e.g. 4.43/32), the checks on when to allow different forms of
1190 "control" were broken. There should now be diagnostics for all cases when a
1191 control that does not make sense is encountered.
1193 7. $recipients is now available in the predata ACL (oversight).
1195 8. Tidy the search cache before the fork to do a delivery from a message
1196 received from the command line. Otherwise the child will trigger a lookup
1197 failure and thereby defer the delivery if it tries to use (for example) a
1198 cached ldap connection that the parent has called unbind on.
1200 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1201 of $address_data from the recipient verification was clobbered by the
1202 sender verification.
1204 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1205 was its contents. (It was OK if the option was not defined at all.)
1207 11. A "Completed" log line is now written for messages that are removed from
1208 the spool by the -Mrm option.
1210 12. $host_address is now set to the target address during the checking of
1211 ignore_target_hosts.
1213 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1214 being passed; this would have caused $sender_host_name to have been used if
1215 matching the list had actually called for a host name (not very likely,
1216 since this list is usually IP addresses). A host name is now passed as
1219 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1220 code that specifies a non-restarting handler (typically sigaction() in
1221 modern systems) in an attempt to fix a rare and obscure crash bug.
1223 15. Narrowed the window for a race in the daemon that could cause it to ignore
1224 SIGCHLD signals. This is not a major problem, because they are used only to
1225 wake it up if nothing else does.
1227 16. A malformed maildirsize file could cause Exim to calculate negative values
1228 for the mailbox size or file count. Odd effects could occur as a result.
1229 The maildirsize information is now recalculated if the size or filecount
1232 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1233 support for a long time. Removed HAVE_SYS_VFS_H.
1235 18. Updated exipick to current release from John Jetmore.
1237 19. Allow an empty sender to be matched against a lookup in an address list.
1238 Previously the only cases considered were a regular expression, or an
1241 20. Exim went into a mad DNS lookup loop when doing a callout where the
1242 host was specified on the transport, if the DNS lookup yielded more than
1245 21. The RFC2047 encoding function was originally intended for short strings
1246 such as real names; it was not keeping to the 75-character limit for
1247 encoded words that the RFC imposes. It now respects the limit, and
1248 generates multiple encoded words if necessary. To be on the safe side, I
1249 have increased the buffer size for the ${rfc2047: expansion operator from
1252 22. Failure to deliver a bounce message always caused it to be frozen, even if
1253 there was an errors_to setting on the router. The errors_to setting is now
1256 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1257 canonical form (fully expanded) before being placed in
1258 $sender_host_address.
1260 24. Updated eximstats to version 1.33
1262 25. Include certificate and key file names in error message when GnuTLS fails
1263 to set them up, because the GnuTLS error message doesn't include the name
1264 of the failing file when there is a problem reading it.
1266 26. Expand error message when OpenSSL has problems setting up cert/key files.
1269 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1270 (this can affect the format of dates).
1272 28. exim_tidydb, when checking for the continued existence of a message for
1273 which it has found a message-specific retry record, was not finding
1274 messages that were in split spool directories. Consequently, it was
1275 deleting retry records that should have stayed in existence.
1277 29. eximstats updated to version 1.35
1278 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1279 1.35 - bugfix such that pie charts by volume are generated correctly
1281 30. The SPA authentication driver was not abandoning authentication and moving
1282 on to the next authenticator when an expansion was forced to fail,
1283 contradicting the general specification for all authenticators. Instead it
1284 was generating a temporary error. It now behaves as specified.
1286 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1287 (the order specifies the preference for clients). The order is now AES256,
1288 AES128, 3DES, ARCFOUR128.
1290 31. Small patch to Sieve code - explicitly set From: when generating an
1293 32. Exim crashed if a remote delivery caused a very long error message to be
1294 recorded - for instance if somebody sent an entire SpamAssassin report back
1295 as a large number of 550 error lines. This bug was coincidentally fixed by
1296 increasing the size of one of Exim's internal buffers (big_buffer) that
1297 happened as part of the Exiscan merge. However, to be on the safe side, I
1298 have made the code more robust (and fixed the comments that describe what
1301 33. Some experimental protocols are using DNS PTR records for new purposes. The
1302 keys for these records are domain names, not reversed IP addresses. The
1303 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1304 leaves it alone. Component reversal etc. now happens only for IP addresses.
1307 34. The host_aton() function is supposed to be passed a string that is known
1308 to be a valid IP address. However, in the case of IPv6 addresses, it was
1309 not checking this. This is a hostage to fortune. Exim now panics and dies
1310 if the condition is not met. A case was found where this could be provoked
1311 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1312 components; fortuitously, this particular loophole had already been fixed
1313 by change 4.50/55 or 4.44/33 above.
1315 If there are any other similar loopholes, the new check in host_aton()
1316 itself should stop them being exploited. The report I received stated that
1317 data on the command line could provoke the exploit when Exim was running as
1318 exim, but did not say which command line option was involved. All I could
1319 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1320 running as the user.
1323 35. There was a buffer overflow vulnerability in the SPA authentication code
1324 (which came originally from the Samba project). I have added a test to the
1325 spa_base64_to_bits() function which I hope fixes it.
1328 36. The daemon start-up calls getloadavg() while still root for those OS that
1329 need the first call to be done as root, but it missed one case: when
1330 deliver_queue_load_max is set with deliver_drop_privilege. This is
1331 necessary for the benefit of the queue runner, because there is no re-exec
1332 when deliver_drop_privilege is set.
1334 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1335 was in use, was not putting the data itself into the right store pool;
1336 consequently, it could be overwritten for a subsequent message in the same
1337 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1340 38. Sometimes the final signoff response after QUIT could fail to get
1341 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1342 < 0 before doing a fflush(). This bug looks as though it goes back to the
1343 introduction of TLS in release 3.20, but "sometimes" must have been rare
1344 because the tests only now provoked it.
1350 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1351 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1352 Now it returns an integer. A number of calls were still expecting a T/F
1353 return. Fortuitously, in all cases, the tests worked in OK situations,
1354 which is the norm. However, things would have gone wrong on any write
1355 failures on the smtp file descriptor. This function is used when sending
1356 messages over SMTP and also when doing verify callouts.
1358 2. When Exim is called to do synchronous delivery of a locally submitted
1359 message (the -odf or -odi options), it no longer closes stderr before doing
1362 3. Implemented the mua_wrapper option.
1364 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1366 5. Implemented the functions header_remove(), header_testname(),
1367 header_add_at_position(), and receive_remove_recipient(), and exported them
1370 6. If an ACL "warn" statement specified the addition of headers, Exim already
1371 inserted X-ACL-Warn: at the start if there was no header name. However, it
1372 was not making this test for the second and subsequent header lines if
1373 there were newlines in the string. This meant that an invalid header could
1374 be inserted if Exim was badly configured.
1376 7. Allow an ACL "warn" statement to add header lines at the start or after all
1377 the Received: headers, as well as at the end.
1379 8. Added the rcpt_4xx retry error code.
1381 9. Added postmaster_mailfrom=xxx to callout verification option.
1383 10. Added mailfrom=xxxx to the callout verification option, for verify=
1386 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1387 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1389 12. Inserted some casts to stop certain compilers warning when using pointer
1390 differences as field lengths or precisions in printf-type calls (mostly
1391 affecting debugging statements).
1393 13. Added optional readline() support for -be (dynamically loaded).
1395 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1396 same clock tick as a message's arrival, so that its received time was the
1397 same as the "first fail" time on the retry record, and that message
1398 remained on the queue past the ultimate address timeout, every queue runner
1399 would try a delivery (because it was past the ultimate address timeout) but
1400 after another failure, the ultimate address timeout, which should have then
1401 bounced the address, did not kick in. This was a "< instead of <=" error;
1402 in most cases the first failure would have been in the next clock tick
1403 after the received time, and all would be well.
1405 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1406 being recognized when the domain list was tested by the match_domain
1407 condition in an expansion string.
1409 16. Added the ${str2b64: operator.
1411 17. Exim was always calling setrlimit() to set a large limit for the number of
1412 processes, without checking whether the existing limit was already
1413 adequate. (It did check for the limit on file descriptors.) Furthermore,
1414 errors from getrlimit() and setrlimit() were being ignored. Now they are
1415 logged to the main and panic logs, but Exim does carry on, to try to do its
1416 job under whatever limits there are.
1418 18. Imported PCRE 5.0.
1420 19. Trivial typo in log message " temporarily refused connection" (the leading
1423 20. If the log selector return_path_on_delivery was set and an address was
1424 redirected to /dev/null, the delivery process crashed because it assumed
1425 that a return path would always be set for a "successful" delivery. In this
1426 case, the whole delivery is bypassed as an optimization, and therefore no
1429 21. Internal re-arrangement: the function for sending a challenge and reading
1430 a response while authentication was assuming a zero-terminated challenge
1431 string. It's now changed to take a pointer and a length, to allow for
1432 binary data in such strings.
1434 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1436 23. Exim was not respecting finduser_retries when seeking the login of the
1437 uid under which it was called; it was always trying 10 times. (The default
1438 setting of finduser_retries is zero.) Also, it was sleeping after the final
1439 failure, which is pointless.
1441 24. Implemented tls_on_connect_ports.
1443 25. Implemented acl_smtp_predata.
1445 26. If the domain in control=submission is set empty, Exim assumes that the
1446 authenticated id is a complete email address when it generates From: or
1447 Sender: header lines.
1449 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1450 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1451 chown and chgrp in /bin and hostname in /usr/bin.
1453 28. Exim was keeping the "process log" file open after each use, just as it
1454 does for the main log. This opens the possibility of it remaining open for
1455 long periods when the USR1 signal hits a daemon. Occasional processlog
1456 errors were reported, that could have been caused by this. Anyway, it seems
1457 much more sensible not to leave this file open at all, so that is what now
1460 29. The long-running daemon process does not normally write to the log once it
1461 has entered its main loop, and it closes the log before doing so. This is
1462 so that log files can straightforwardly be renamed and moved. However,
1463 there are a couple of unusual error situations where the daemon does write
1464 log entries, and I had neglected to close the log afterwards.
1466 30. The text of an SMTP error response that was received during a remote
1467 delivery was being truncated at 512 bytes. This is too short for some of
1468 the long messages that one sometimes sees. I've increased the limit to
1471 31. It is now possible to make retry rules that apply only when a message has a
1472 specific sender, in particular, an empty sender.
1474 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1475 it possible to be selective about when SMTP synchronization is enforced.
1477 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1479 32. Implemented hosts_connection_nolog.
1481 33. Added an ACL for QUIT.
1483 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1486 35. Added mailbox_size and mailbox_filecount to appendfile.
1488 36. Added control = no_multiline_responses to ACLs.
1490 37. There was a bug in the logic of the code that waits for the clock to tick
1491 in the case where the clock went backwards by a substantial amount such
1492 that the microsecond fraction of "now" was more than the microsecond
1493 fraction of "then" (but the whole seconds number was less).
1495 38. Added support for the libradius Radius client library this is found on
1496 FreeBSD (previously only the radiusclient library was supported).
1502 1. When certain lookups returned multiple values in the form name=value, the
1503 quoting of the values was not always being done properly. Specifically:
1504 (a) If the value started with a double quote, but contained no whitespace,
1506 (b) If the value contained whitespace other than a space character (i.e.
1507 tabs or newlines or carriage returns) it was not quoted.
1508 This fix has been applied to the mysql and pgsql lookups by writing a
1509 separate quoting function and calling it from the lookup code. The fix
1510 should probably also be applied to nisplus, ibase and oracle lookups, but
1511 since I cannot test any of those, I have not disturbed their existing code.
1513 2. A hit in the callout cache for a specific address caused a log line with no
1514 reason for rejecting RCPT. Now it says "Previous (cached) callout
1515 verification failure".
1517 3. There was an off-by-one bug in the queryprogram router. An over-long
1518 return line was truncated at 256 instead of 255 characters, thereby
1519 overflowing its buffer with the terminating zero. As well as fixing this, I
1520 have increased the buffer size to 1024 (and made a note to document this).
1522 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1523 when Exim is waiting for an SMTP response from a remote server, Exim
1524 restarts its select() call on the socket, thereby resetting its timeout.
1525 This is not a problem when such interrupts are rare. Somebody set up a cron
1526 job to run exiwhat every 2 minutes, which is less than the normal select()
1527 timeout (5 or 10 minutes). This meant that the select() timeout never
1528 kicked in because it was always reset. I have fixed this by comparing the
1529 time when an interrupt arrives with the time at the start of the first call
1530 to select(). If more time than the timeout has elapsed, the interrupt is
1531 treated as a timeout.
1533 5. Some internal re-factoring in preparation for the addition of Sieve
1534 extensions (by MH). In particular, the "personal" test is moved to a
1535 separate function, and given an option for scanning Cc: and Bcc: (which is
1536 not set for Exim filters).
1538 6. When Exim created an email address using the login of the caller as the
1539 local part (e.g. when creating a From: or Sender: header line), it was not
1540 quoting the local part when it contained special characters such as @.
1542 7. Installed new OpenBSD configuration files.
1544 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1545 try to make them clearer.
1547 9. Callout options, other than the timeout value, were being ignored when
1548 verifying sender addresses in header lines. For example, when using
1550 verify = header_sender/callout=no_cache
1552 the cache was (incorrectly) being used.
1554 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1555 only the Cygwin environment.
1557 11. When return_path_on_delivery was set as a log selector, if different remote
1558 addresses in the same message used different return paths and parallel
1559 remote delivery occurred, the wrong values would sometimes be logged.
1560 (Whenever a remote delivery process finished, the return path value from
1561 the most recently started remote delivery process was logged.)
1563 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1564 lines when AUTH and/or TLS are in use. This is the "received protocol"
1565 field. Exim used to use "asmtp" for authenticated SMTP, without any
1566 indication (in the protocol name) for TLS use. Now it follows the RFC and
1567 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1568 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1569 names appear in log lines as well as in Received: header lines.
1571 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1572 extensions, and comparison tests, and to fix some bugs.
1574 14. Changes to the "personal" filter test:
1576 (1) The test was buggy in that it was just doing the equivalent of
1577 "contains" tests on header lines. For example, if a user's address was
1578 anne@some.where, the "personal" test would incorrectly be true for
1580 To: susanne@some.where
1582 This test is now done by extracting each address from the header in turn,
1583 and checking the entire address. Other tests that are part of "personal"
1584 are now done using regular expressions (for example, to check local parts
1585 of addresses in From: header lines).
1587 (2) The list of non-personal local parts in From: addresses has been
1588 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1589 taken from the Sieve specification recommendations.
1591 (3) If the message contains any header line starting with "List-" it is
1592 treated as non-personal.
1594 (4) The test for "circular" in the Subject: header line has been removed
1595 because it now seems ill-conceived.
1597 15. Minor typos in src/EDITME comments corrected.
1599 16. Installed latest exipick from John Jetmore.
1601 17. If headers_add on a router specified a text string that was too long for
1602 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1603 of string_sprintf() is now avoided.
1605 18. $message_body_size was not set (it was always zero) when running the DATA
1606 ACL and the local_scan() function.
1608 19. For the "mail" command in an Exim filter, no default was being set for
1609 the once_repeat time, causing a random time value to be used if "once" was
1610 specified. (If the value happened to be <= 0, no repeat happened.) The
1611 default is now 0s, meaning "never repeat". The "vacation" command was OK
1612 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1613 (I found it when inspecting the code).
1615 20. There is now an overall timeout for performing a callout verification. It
1616 defaults to 4 times the callout timeout, which applies to individual SMTP
1617 commands during the callout. The overall timeout applies when there is more
1618 than one host that can be tried. The timeout is checked before trying the
1619 next host. This prevents very long delays if there are a large number of
1620 hosts and all are timing out (e.g. when the network connections are timing
1621 out). The value of the overall timeout can be changed by specifying an
1622 additional sub-option for "callout", called "maxwait". For example:
1624 verify = sender/callout=5s,maxwait=20s
1626 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1627 to the end before writing, but this should make it even safer).
1629 22. Exim was forgetting that it had advertised PIPELINING for the second and
1630 subsequent messages on an SMTP connection. It was also not resetting its
1631 memory on STARTTLS and an internal HELO.
1633 23. When Exim logs an SMTP synchronization error within a session, it now
1634 records whether PIPELINING has been advertised or not.
1636 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1637 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1638 rather than long int.
1640 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1642 26. Added the never_mail option to autoreply.
1648 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1649 crash if the getsockname() call failed; this can happen if a connection is
1650 closed very soon after it is established. The problem was simply in the
1651 order in which certain operations were done, causing Exim to try to write
1652 to the SMTP stream before it had set up the file descriptor. The bug has
1653 been fixed by making things happen in the correct order.
1659 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1660 before the connection was closed, thus losing the rejection response.
1662 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1663 some early Solaris releases, but causes trouble in current releases where
1664 socklen_t is defined.
1666 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1669 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1672 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1673 that releasing the top of it at the end releases what was used for sub-
1674 expansions (unless the block got too big). However, discard this block if
1675 the first thing is a variable or header, so that we can use its block when
1676 it is dynamic (useful for very large $message_headers, for example).
1678 6. Lookups now cache *every* query, not just the most recent. A new, separate
1679 store pool is used for this. It can be recovered when all lookup caches are
1680 flushed. Lookups now release memory at the end of their result strings.
1681 This has involved some general refactoring of the lookup sources.
1683 7. Some code has been added to the store_xxx() functions to reduce the amount
1684 of flapping under certain conditions.
1686 8. log_incoming_interface used to affect only the <= reception log lines. Now
1687 it causes the local interface and port to be added to several more SMTP log
1688 lines, for example "SMTP connection from", and rejection lines.
1690 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1692 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1694 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1695 could be overwritten at the end of the current message (or the start of a
1696 new message if it was set in a HELO ACL). The value is now preserved for
1697 the duration of the SMTP connection.
1699 12. If a transport had a headers_rewrite setting, and a matching header line
1700 contained an unqualified address, that address was qualified, even if it
1701 did not match any rewriting rules. The underlying bug was that the values
1702 of the flags that permit the existence of unqualified sender and recipient
1703 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1704 non-local messages, and by -bnq for local messages) were not being
1705 preserved with the message after it was received.
1707 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1708 "next input=" as part of the text comprising the host identity instead of
1709 the correct text. The code was using the same buffer for two different
1710 strings. However, depending on which order the printing function evaluated
1711 its arguments, the bug did not always show up. Under Linux, for example, my
1712 test suite worked just fine.
1714 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1715 doesn't work with some older versions of Perl. It has been changed to "my",
1716 which in any case is probably the better facility to use.
1718 15. A really picky compiler found some instances of statements for creating
1719 error messages that either had too many or two few arguments for the format
1722 16. The size of the buffer for calls to the DNS resolver has been increased
1723 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1724 for addresses that have a lot of PTR records. This alleviates a problem; it
1725 does not fully solve it.
1727 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1728 buffer now truncates the list and logs the incident, which is the same
1729 action as happens when Exim is looking up a host name and its aliases.
1730 Previously in this situation something unpredictable would happen;
1731 sometimes it was "internal error: store_reset failed".
1733 18. If a server dropped the connection unexpectedly when an Exim client was
1734 using GnuTLS and trying to read a response, the client delivery process
1735 crashed while trying to generate an error log message.
1737 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1738 string, for example:
1740 warn message = H1: something\nH2: something
1742 the text was added as a single header line from Exim's point of view
1743 though it ended up OK in the delivered message. However, searching for the
1744 second and subsequent header lines using $h_h2: did not work. This has been
1745 fixed. Similarly, if a system filter added multiple headers in this way,
1746 the routers could not see them.
1748 20. Expanded the error message when iplsearch is called with an invalid key to
1749 suggest using net-iplsearch in a host list.
1751 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1752 ACLs are no longer actually imposed (and a message to that effect is
1755 22. If a "gecos" field in a passwd entry contained escaped characters, in
1756 particular, if it contained a \" sequence, Exim got it wrong when building
1757 a From: or a Sender: header from that name. A second bug also caused
1758 incorrect handling when an unquoted " was present following a character
1759 that needed quoting.
1761 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1762 was not being matched caselessly.
1764 24. Arranged for all hyphens in the exim.8 source to be escaped with
1767 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1768 properly. Recipient callout cache records were still being keyed to include
1769 the sender, even when use_sender was set false. This led to far more
1770 callouts that were necessary. The sender is no longer included in the key
1771 when use_sender is false.
1773 26. Added "control = submission" modifier to ACLs.
1775 27. Added the ${base62d: operator to decode base 62 numbers.
1777 28. dnsdb lookups can now access SRV records.
1779 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1780 the configuration file.
1782 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1783 (-v) mode. This makes the output for a verbose queue run more intelligible.
1785 31. Added a use_postmaster feature to recipient callouts.
1787 32. Added the $body_zerocount variable, containing the number of binary zero
1788 bytes in the message body.
1790 33. The time of last modification of the "new" subdirectory is now used as the
1791 "mailbox time last read" when there is a quota error for a maildir
1794 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1796 35. Added +ignore_unknown as a special item in host lists.
1798 36. Code for decoding IPv6 addresses in host lists is now included, even if
1799 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1800 address was recognized as an IP address, but was then not correctly decoded
1801 into binary, causing unexpected and incorrect effects when compared with
1808 1. Very minor rewording of debugging text in manualroute to say "list of
1809 hosts" instead of "hostlist".
1811 2. If verify=header_syntax was set, and a header line with an unqualified
1812 address (no domain) and a large number of spaces between the end of the
1813 name and the colon was received, the reception process suffered a buffer
1814 overflow, and (when I tested it) crashed. This was caused by some obsolete
1815 code that should have been removed. The fix is to remove it!
1817 3. When running in the test harness, delay a bit after writing a bounce
1818 message to get a bit more predictability in the log output.
1820 4. Added a call to search_tidyup() just before forking a reception process. In
1821 theory, someone could use a lookup in the expansion of smtp_accept_max_
1822 per_host which, without the tidyup, could leave open a database connection.
1824 5. Added the variables $recipient_data and $sender_data which get set from a
1825 lookup success in an ACL "recipients" or "senders" condition, or a router
1826 "senders" option, similar to $domain_data and $local_part_data.
1828 6. Moved the writing of debug_print from before to after the "senders" test
1831 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1832 problems for message scanning, either using a data ACL, or using
1833 local_scan() because the Received: header was not generated till after they
1834 were called (in order to set the time as the time of reception completion).
1835 I have revised the way this works. The header is now generated after the
1836 body is received, but before the ACL or local_scan() are called. After they
1837 are run, the timestamp in the header is updated.
1843 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1844 before starting a queue runner without re-exec. This happened only when
1845 deliver_drop_privilege was set or when the Exim user was set to root. The
1846 effect of the bug was that timeouts during subsequent deliveries caused
1847 crashes instead of being properly handled. The handler is now left at its
1848 default (and expected) setting.
1850 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1851 message, again when deliver_drop_privilege is set or Exim is run as root.
1852 The bug described in (1) was not present in this case, but the tidying up
1853 of the other signals was missing. I have made the two cases consistent.
1855 3. The ignore_target_hosts setting on a manualroute router was being ignored
1856 for hosts that were looked up using the /MX notation.
1858 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1861 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1862 operated on the sender address. After changing the $sender_address to <>
1863 for the sender address verify, Exim was re-instated it as the original
1864 (before rewriting) address, but remembering that it had rewritten it, so it
1865 wasn't rewriting it again. This bug also had the effect of breaking the
1866 sender address verification caching when the sender address was rewritten.
1868 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1869 This has been changed so that if the ip literal address matches
1870 ignore_target_hosts, the router declines.
1872 7. Added expansion conditions match_domain, match_address, and match_local_
1873 part (NOT match_host).
1875 8. The placeholder for the Received: header didn't have a length field set.
1877 9. Added code to Exim itself and to exim_lock to test for a specific race
1878 condition that could lead to file corruption when using MBX delivery. The
1879 issue is with the lockfile that is created in /tmp. If this file is removed
1880 after a process has opened it but before that process has acquired a lock,
1881 there is the potential for a second process to recreate the file and also
1882 acquire a lock. This could lead to two Exim processes writing to the file
1883 at the same time. The added code performs the same test as UW imapd; it
1884 checks after acquiring the lock that its file descriptor still refers to
1885 the same named file.
1887 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1888 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1891 11. Added the smtp_active_hostname option. If used, this will typically be made
1892 to depend on the incoming interface address. Because $interface_address is
1893 not set up until the daemon has forked a reception process, error responses
1894 that can happen earlier (such as "too many connections") no longer contain
1897 12. If an expansion in a condition on a "warn" statement fails because a lookup
1898 defers, the "warn" statement is abandoned, and the next ACL statement is
1899 processed. Previously this caused the whole ACL to be aborted.
1901 13. Added the iplsearch lookup type.
1903 14. Added ident_timeout as a log selector.
1905 15. Added tls_certificate_verified as a log selector.
1907 16. Added a global option tls_require_ciphers (compare the smtp transport
1908 option of the same name). This controls incoming TLS connections.
1910 17. I finally figured out how to make tls_require_ciphers do a similar thing
1911 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1912 before starting the TLS session.
1914 18. Tabs are now shown as \t in -bP output.
1916 19. If the log selector return_path_on_delivery was set, Exim crashed when
1917 bouncing a message because it had too many Received: header lines.
1919 20. If two routers both had headers_remove settings, and the first one included
1920 a superfluous trailing colon, the final name in the first list and the
1921 first name in the second list were incorrectly joined into one item (with a
1922 colon in the middle).
1928 1. Added -C and -D options to the exinext utility, mainly to make it easier
1929 to include in the automated testing, but these could be helpful when
1930 multiple configurations are in use.
1932 2. The exinext utility was not formatting the output nicely when there was
1933 an alternate port involved in the retry record key, nor when there was a
1934 message id as well (for retries that were specific to a specific message
1935 and a specific host). It was also confused by IPv6 addresses, because of
1936 the additional colons they contain. I have fixed the IPv4 problem, and
1937 patched it up to do a reasonable job for IPv6.
1939 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1940 delivery, the log line now contains "pipelined" if PIPELINING was used.
1942 4. An SMTP transport process used to panic and die if the bind() call to set
1943 an explicit outgoing interface failed. This has been changed; it is now
1944 treated in the same way as a connect() failure.
1946 5. A reference to $sender_host_name in the part of a conditional expansion
1947 that was being skipped was still causing a DNS lookup. This no longer
1950 6. The def: expansion condition was not recognizing references to header lines
1951 that used bh_ and bheader_.
1953 7. Added the _cache feature to named lists.
1955 8. The code for checking quota_filecount in the appendfile transport was
1956 allowing one more file than it should have been.
1958 9. For compatibility with Sendmail, the command line option
1966 and sets the incoming protocol and host name (for trusted callers). The
1967 host name and its colon can be omitted when only the protocol is to be set.
1968 Note the Exim already has two private options, -pd and -ps, that refer to
1969 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1970 "s", but I don't think that's a major issue.
1972 10. A number of refactoring changes to the code, none of which should affect
1975 (a) The number of logging options was getting close to filling up the
1976 32-bit word that was used as a bit map. I have split them into two classes:
1977 those that are passed in the argument to log_write(), and those that are
1978 only ever tested independently outside of that function. These are now in
1979 separate 32-bit words, so there is plenty of room for expansion again.
1980 There is no change in the user interface or the logging behaviour.
1982 (b) When building, for example, log lines, the code previously used a
1983 macro that called string_cat() twice, in order to add two strings. This is
1984 not really sufficiently general. Furthermore, there was one instance where
1985 it was actually wrong because one of the argument was used twice, and in
1986 one call a function was used. (As it happened, calling the function twice
1987 did not affect the overall behaviour.) The macro has been replaced by a
1988 function that can join an arbitrary number of extra strings onto a growing
1991 (c) The code for expansion conditions now uses a table and a binary chop
1992 instead of a serial search (which was left over from when there were very
1993 few conditions). Also, it now recognizes conditions like "pam" even when
1994 the relevant support is not compiled in: a suitably worded error message is
1995 given if an attempt is made to use such a condition.
1997 11. Added ${time_interval:xxxxx}.
1999 12. A bug was causing one of the ddress fields not to be passed back correctly
2000 from remote delivery subprocesses. The field in question was not being
2001 subsequently used, so this caused to problems in practice.
2003 13. Added new log selectors queue_time and deliver_time.
2005 14. Might have fixed a bug in maildirsizefile handling that threw up
2006 "unexpected character" debug warnings, and recalculated the data
2007 unnecessarily. In any case, I expanded the warning message to give more
2010 15. Added the message "Restricted characters in address" to the statements in
2011 the default ACL that block characters like @ and % in local parts.
2013 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2014 Three changes have been made:
2016 (a) There was a serious bug; a negative response to MAIL caused the whole
2017 recipient domain to be cached as invalid, thereby blocking all messages
2018 to all local parts at the same domain, from all senders. This bug has
2019 been fixed. The domain is no longer cached after a negative response to
2020 MAIL if the sender used is not empty.
2022 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2025 (c) A new callout option, "use_sender" has been added for people who want
2026 the modified behaviour.
2032 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2035 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2036 indeed breaks things for older releases.
2038 3. Added additional logging to the case where there is a problem reading data
2039 from a filter that is running in a subprocess using a pipe, in order to
2040 try to track down a specific problem.
2042 4. Testing facility fudge: when running in the test harness and attempting
2043 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2044 getting "No route to host". Convert this to a timeout.
2046 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2049 6. Some OS don't have socklen_t but use size_t instead. This affects the
2050 fifth argument of getsockopt() amongst other things. This is now
2051 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2052 can be set for individual OS. I have set it for SunOS5, OSF1, and
2053 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2054 some earlier ones do not.
2056 7. Change 4.30/15 was not doing the test caselessly.
2058 8. The standard form for an IPv6 address literal was being rejected by address
2059 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2060 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2061 this, as well as the form without the "IPv6" on the front (but only when
2062 address literals are enabled, of course).
2064 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2066 10. Exim crashed if a message with an empty sender address specified by -f
2067 encountered a router with an errors_to setting. This could be provoked only
2068 by a command such as
2072 where an empty string was supplied; "<>" did not hit this bug.
2074 11. Installed PCRE release 4.5.
2076 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2077 remained set. It is now erased.
2079 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2080 times from message ids (which are base 36 rather than the normal 62).
2082 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2083 were being counted as actual protocol errors, and logged if the log
2084 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2085 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2086 and DATA following a set of rejected RCPTs do not count as protocol errors.
2087 In other words, Exim assumes they were pipelined, though this may not
2088 actually be the case. Of course, in all cases the client gets an
2089 appropriate error code.
2091 15. If a lookup fails in an ACL condition, a message about the failure may
2092 be available; it is used if testing the ACL cannot continue, because most
2093 such messages specify what the cause of the deferral is. However, some
2094 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2095 that caused an old message to be retained and used if a later statement
2096 caused a defer, replacing the real cause of the deferral.
2098 16. If an IP address had so many PTR records that the DNS lookup buffer
2099 was not large enough to hold them, Exim could crash while trying to process
2100 the truncated data. It now detects and logs this case.
2102 17. Further to 4.21/58, another change has been made: if (and only if) the
2103 first line of a message (the first header line) ends with CRLF, a bare LF
2104 in a subsequent header line has a space inserted after it, so as not to
2105 terminate the header.
2107 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2108 unnecessarily, used atoi() instead of strtol(), and didn't check the
2109 termination when getting file sizes from file names by regex.
2111 19. Completely re-implemented the support for maildirsize files, in the light
2112 of a number of problems with the previous contributed implementation
2113 (4.30/29). In particular:
2115 . If the quota is zero, the maildirsize file is maintained, but no quota is
2118 . If the maildir directory does not exist, it is created before any attempt
2119 to write a maildirsize file.
2121 . The quota value in the file is just a cache; if the quota is changed in
2122 the transport, the new value overrides.
2124 . A regular expression is available for excluding directories from the
2127 20. The autoreply transport checks the characters in options that define the
2128 message's headers; it allows continued headers, but it was checking with
2129 isspace() after an embedded newline instead of explicitly looking for a
2132 21. If all the "regular" hosts to which an address was routed had passed their
2133 expiry times, and had not reached their retry times, the address was
2134 bounced, even if fallback hosts were defined. Now Exim should go on to try
2137 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2138 equivalent code in the SMTP transport. Some hosts send humungous responses
2139 to HELO/EHLO, more than 1024 it seems.
2141 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2142 gives compiler warnings in some environments. I've now done it "properly",
2145 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2146 (because of problems with the built-in one) was declared to return uschar *
2147 instead of char *, causing compiler failure.
2149 25. Fixed a file descriptor leak when processing alias/forward files.
2151 26. Fixed a minor format string issue in dbfn.c.
2153 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2155 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2156 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2159 29. When a sender address is verified, it is cached, to save repeating the test
2160 when there is more than one recipient in a message. However, when the
2161 verification involves a callout, it is possible for different callout
2162 options to be set for different recipients. It is too complicated to keep
2163 track of this in the cache, so now Exim always runs a verification when a
2164 callout is required, relying on the callout cache for the optimization.
2165 The overhead is duplication of the address routing, but this should not be
2168 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2169 to be verified with callout=postmaster, and the main callout worked but the
2170 postmaster check failed, the verification correctly failed. However, if a
2171 subsequent RCPT command asked for sender verification *without* the
2172 postmaster check, incorrect caching caused this verification also to fail,
2175 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2176 it was not caching the DNS options (qualify_single, search_parents) that
2177 were used when the lookup failed. A subsequent lookup with different
2178 options therefore always gave the same answer, though there were cases
2179 where it should not have. (Example: a "domains = !$mx_any" option on a
2180 dnslookup router: the "domains" option is always processed without any
2181 widening, but the router might have qualify_single set.) Now Exim uses the
2182 cached value only when the same options are set.
2184 32. Added John Jetmore's "exipick" utility to the distribution.
2186 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2187 than a timeout (e.g. a certificate is required, and is not provided), an
2188 Exim server now closes the connection immediately. Previously it waited for
2189 the client to close - but if the client is SSL, it seems that they each
2190 wait for each other, leading to a delay before one of them times out.
2192 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2193 maintained 0.8.x compatibility because I don't think many are using it, and
2194 it is clearly obsolete.
2196 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2199 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2200 client certificate was expired. A simple patch fixes this, though I don't
2201 understand the full logic of why the verify callback is called multiple
2204 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2205 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2206 which causes problems with some clients (such as the Certicom SSL Plus
2207 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2208 disables the coutermeasure allowing Eudora to connect."
2210 38. Exim was not checking that a write() to a log file succeeded. This could
2211 lead to Bad Things if a log got too big, in particular if it hit a file
2212 size limit. Exim now panics and dies if it cannot write to a log file, just
2213 as it does if it cannot open a log file.
2215 39. Modified OS/Makefile-Linux so that it now contains
2217 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2219 The two -D definitions ensure that Exim is compiled with large file
2220 support, which makes it possible to handle log files that are bigger than
2223 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2224 instance) a domain was checked against a named list that involved a lookup,
2225 causing $domain_data to be set, then another domain was checked against the
2226 same list, then the first domain was re-checked, the value of $domain_data
2227 after the final check could be wrong. In particular, if the second check
2228 failed, it could be set empty. This bug probably also applied to
2231 41. The strip_trailing_dot option was not being applied to the address given
2232 with the -f command-line option.
2234 42. The code for reading a message's header from the spool was incrementing
2235 $received_count, but never initializing it. This meant that the value was
2236 incorrect (doubled) while delivering a message in the same process in which
2237 it was received. In the most common configuration of Exim, this never
2238 happens - a fresh exec is done - but it can happen when
2239 deliver_drop_privilege is set.
2241 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2242 it now includes up to 150 characters of the unexpected data in the log
2245 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2246 and building data strings. The size of both of these buffers was 10 000
2247 bytes - far larger than anybody would *ever* want, thought I. Needless to
2248 say, somebody hit the limit. I have increased the maximum line length to
2249 20 000 and the maximum data length of concatenated lines to 100 000. I have
2250 also fixed two bugs, because there was no checking on these buffers. Tsk,
2251 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2252 buffer is too small.
2254 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2255 lsearch lookups. Now it does.
2257 46. When parsing a route_list item in a manualroute router, a fixed-length
2258 buffer was used for the list of hosts. I made this 1024 bytes long,
2259 thinking that nobody would ever have a list of hosts that long. Wrong.
2260 Somebody had a whole pile of complicated expansion conditions, and the
2261 string was silently truncated, leading to an expansion error. It turns out
2262 that it is easier to change to an unlimited length (owing to other changes
2263 that have happened since this code was originally written) than to build
2264 structure for giving a limitation error. The length of the item that
2265 expands into the list of hosts is now unlimited.
2267 47. The lsearch lookup could not handle data where the length of text line was
2268 more than 4095 characters. Such lines were truncated, leading to shortened
2269 data being returned. It should now handle lines of any length.
2271 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2272 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2275 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2276 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2277 no longer happens in comment lines. A list of replacements is now placed
2278 at the head of all of the source files, except those whose only change is
2279 to replace PERL_COMMAND in the very first #! line.
2281 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2282 messages on the queue, with a bottom-up merge sort, using code contributed
2283 by Michael Haardt. This should make operations like -bp somewhat faster on
2284 large queues. It won't affect queue runners, except when queue_run_in_order
2287 51. Installed eximstats 1.31 in the distribution.
2289 52. Added support for SRV lookups to the dnslookup router.
2291 53. If an ACL referred to $message_body or $message_body_end, the value was not
2292 reset for any messages that followed in the same SMTP session.
2294 54. The store-handling optimization for building very long strings was not
2295 differentiating between the different store pools. I don't think this
2296 actually made any difference in practice, but I've tidied it.
2298 55. While running the routers to verify a sender address, $sender_address
2299 was still set to the sender address. This is wrong, because when routing to
2300 send a bounce to the sender, it would be empty. Therefore, I have changed
2301 it so that, while verifying a sender address, $sender_address is set to <>.
2302 (There is no change to what happens when verifying a recipient address.)
2304 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2305 target A or AAAA records (if not already returned) without resetting the
2306 qualify_single or search_parents options of the DNS resolver. These are
2307 inappropriate in this case because the targets of MX and SRV records must
2308 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2309 target that, when qualified, matched something in the local domain. These
2310 two options are now turned off when doing these lookups.
2312 57. It seems that at least some releases of Reiserfs (which does not have the
2313 concept of a fixed number of inodes) returns zero and not -1 for the
2314 number of available inodes. This interacted badly with check_spool_inodes,
2315 which assumed that -1 was the "no such thing" setting. What I have done is
2316 to check that the total number of inodes is greater than zero before doing
2317 the test of how many are available.
2319 58. When a "warn" ACL statement has a log_message modifier, the message is
2320 remembered, and not repeated. This is to avoid a lot of repetition when a
2321 message has many recipients that cause the same warning to be written.
2322 Howewer, Exim was preserving the list of already written lines for an
2323 entire SMTP session, which doesn't seem right. The memory is now reset if a
2324 new message is started.
2326 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2327 debugging output unless log_rewrite was also set.
2329 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2330 of (int)(handle) when we know that handle contains (void *)(-1).
2332 61. The Exim daemon panic-logs an error return when it closes the incoming
2333 connection. However "connection reset by peer" seems to be common, and
2334 isn't really an error worthy of noting specially, so that particular error
2337 62. When Exim is trying to find all the local interfaces, it used to panic and
2338 die if the ioctl to get the interface flags failed. However, it seems that
2339 on at least one OS (Solaris 9) it is possible to have an interface that is
2340 included in the list of interfaces, but for which you get a failure error
2341 for this call. This happens when the interface is not "plumbed" into a
2342 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2343 failure of the "get flags" call assumes that the interface is down.
2345 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2346 makes life easier for people who are doing arithmetic on fields extracted
2347 from dates, where you often get leading zeros that should not be
2348 interpreted as octal.
2350 64. Added qualify_domain to the redirect router, to override the global
2353 65. If a pathologically long header line contained very many addresses (the
2354 report of this problem mentioned 10 000) and each of them was rewritten,
2355 Exim could use up a very large amount of memory. (It kept on making new
2356 copies of the header line as it rewrote, and never released the old ones.)
2357 At the expense of a bit more processing, the header rewriting function has
2358 been changed so that it no longer eats memory in this way.
2360 66. The generation of the Received: header has been moved from the time that a
2361 message starts to be received, to the time that it finishes. The timestamp
2362 in the Received: header should now be very close to that of the <= log
2363 line. There are two side-effects of this change:
2365 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2366 logged header lines no longer include the local Received: line, because
2367 it has not yet been created. The same applies to a copy of the message
2368 that is returned to a non-SMTP sender when a message is rejected.
2370 (b) When a filter file is tested using -bf, no additional Received: header
2371 is added to the test message. After some thought, I decided that this
2374 This change does not affect the value of $received_for. It is still set
2375 after address rewriting, but before local_scan() is called.
2377 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2379 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2380 gave an unhelpful panic error message, and a defer error. I have managed to
2381 change this behaviour so that it now rejects any supplied certificate,
2382 which seems right, as the list of acceptable certificates is empty.
2384 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2385 gave an unhelpful defer error. I have not managed to make this reject any
2386 supplied certificates, but the error message it gives is "no certificate
2387 supplied", which is not helpful.
2389 70. exigrep's output now also includes lines that are not associated with any
2390 message, but which match the given pattern. Implemented by a patch from
2391 Martin Sluka, which also tidied up the Perl a bit.
2393 71. Recipient callout verification, like sender verification, was using <> in
2394 the MAIL FROM command. This isn't really the right thing, since the actual
2395 sender may affect whether the remote host accepts the recipient or not. I
2396 have changed it to use the actual sender in the callout; this means that
2397 the cache record is now keyed on a recipient/sender pair, not just the
2398 recipient address. There doesn't seem to be a real danger of callout loops,
2399 since a callout by the remote host to check the sender would use <>.
2400 [SEE ABOVE: changed after hitting problems.]
2402 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2403 temporary errors. However, in the case of such a code being given after
2404 the end of a data transmission (i.e. after ".") Exim was failing to write
2405 a retry record for the message. (Yes, there was some broken host that was
2406 actually sending 8xx at this point.)
2408 73. An unknown lookup type in a host list could cause Exim to panic-die when
2409 the list was checked. (An example that provoked this was putting <; in the
2410 middle of a list instead of at the start.) If this happened during a DATA
2411 ACL check, a -D file could be left lying around. This kind of configuration
2412 error no longer causes Exim to die; instead it causes a defer errror. The
2413 incident is still logged to the main and panic logs.
2415 74. Buglet left over from Exim 3 conversion. The message "too many messages
2416 in one connection" was written to the rejectlog but not the mainlog, except
2417 when address rewriting (yes!) was being logged.
2419 75. Added write_rejectlog option.
2421 76. When a system filter was run not as root (that is, when system_filter_user
2422 was set), the values of the $n variables were not being returned to the
2423 main process; thus, they were not subsequently available in the $sn
2426 77. Added +return_path_on_delivery log selector.
2428 78. A connection timeout was being treated differently from recipients deferred
2429 when testing hosts_max_try with a message that was older than the host's
2430 retry timeout. (The host should not be counted, thus allowing all hosts to
2431 be tried at least once before bouncing.) This may have been the cause of an
2432 occasionally reported bug whereby a message would remain on the queue
2433 longer than the retry timeout, but would be bounced if a delivery was
2434 forced. I say "may" because I never totally pinned down the problem;
2435 setting up timeout/retry tests is difficult. See also the next item.
2437 79. The ultimate address timeout was not being applied to errors that involved
2438 a combination of host plus message (for example, a timeout on a MAIL
2439 command). When an address resolved to a number of possible hosts, and they
2440 were not all tried for each delivery (e.g. because of hosts_max_try), a
2441 message could remain on the queue longer than the retry timeout.
2443 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2446 81. Fixed an obscure SMTP outgoing bug which required at least the following
2447 conditions: (a) there was another message waiting for the same server;
2448 (b) the server returned 5xx to all RCPT commands in the first message so
2449 that the message was not completed; (c) the server dropped the connection
2450 or gave a negative response to the RSET that Exim sends to abort the
2451 transaction. The observed case was a dropped connection after DATA that had
2452 been sent in pipelining mode. That is, the server had advertised PIPELINING
2453 but was not implementing it correctly. The effect of the bug was incorrect
2454 behaviour, such as trying another host, and this could lead to a crash.
2460 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2461 and daemon.c were passed as pointers to ints; they should have been
2462 pointers to socklen_t variables (which are typically unsigned ints).
2464 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2467 3. Fixed a really odd bug that affected only the testing scheme; patching a
2468 certain fixed string in the binary changed the value of another string that
2469 happened to be identical to the end of the original first string.
2471 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2472 name", it returns that address as the IP address. On some operating
2473 systems (e.g. Solaris), it also passes back the IP address string as the
2474 "host name". However, on others (e.g. Linux), it passes back an empty
2475 string. Exim wasn't checking for this, and was changing the host name to an
2476 empty string, assuming it had been canonicized.
2478 5. Although rare, it is permitted to have more than one PTR record for a given
2479 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2480 all the names associated with an address, because they do in Solaris.
2481 However, it seems that they do not in Linux for data that comes from the
2482 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2483 I found this out when I moved to a new Linux workstation and tried to run
2484 the Exim test suite.
2486 To get round this problem I have changed the code so that it now does its
2487 own call to the DNS to look up PTR records when searching for a host name.
2488 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2489 addresses that are only in /etc/hosts are still found.
2491 This behaviour is, however, controlled by an option called host_lookup_
2492 order, which defaults to "bydns:byaddr". If people want to use the other
2493 order, or indeed, just use one or the other means of lookup, they can
2494 specify it in this variable.
2496 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2497 some operating systems, this comes back from gethostbyaddr() as an empty
2498 string, and this is what Exim used to test for. However, it seems that in
2499 other systems, "." is yielded. Exim now tests for this case too.
2501 7. The values of check_spool_space and check_log_space are now held internally
2502 as a number of kilobytes instead of an absolute number of bytes. If a
2503 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2504 kilobyte. This means that much larger values can be stored.
2506 8. Exim monitor: an attempt to get the action menu when not actually pointing
2507 at a message produces an empty menu entitled "No message selected". This
2508 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2509 no entries in it ("Shell widget menu has zero width and/or height"). So I
2510 have added a single, blank menu entry in this case.
2512 9. Added ${quote_local_part.
2514 10. MIME decoding is now applied to the contents of Subject: header lines when
2517 11. Now that a reference to $sender_host_address automatically causes a reverse
2518 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2519 host lookup before query-style lookups in lists that might use this
2520 variable. This has therefore been abolished, and the "net-" prefix is no
2521 longer necessary for query-style lookups.
2523 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2524 have been a typo for LFLAGS, so it has been changed.
2526 13. The install script calls Exim with "-C /dev/null" in order to find the
2527 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2528 to be output. Howeve, since Exim outputs its version number before the
2529 error, it didn't break the script. It just looked ugly. I fixed this by
2530 always allowing "-C /dev/null" if the caller is root.
2532 14. Ignore overlarge ACL variable number when reading spool file - insurance
2533 against a later release with more variables having written the file.
2535 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2536 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2537 this, as well as the form without the "IPv6" on the front.
2539 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2540 OS/Makefile-Darwin file.
2542 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2543 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2545 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2546 variable. However, some OS use ints for inodes. I've added cast to long int
2547 to get rid of the compiler warning.
2549 19. I had forgotten to lock out "/../" in configuration file names when
2550 ALT_CONFIG_PREFIX was set.
2552 20. Routers used for verification do not need to specify transports. However,
2553 if such a router generated a host list, and callout was configured, Exim
2554 crashed, because it could not find a port number from the (non-existent)
2555 transport. It now assumes port 25 in this circumstance.
2557 21. Added the -t option to exigrep.
2559 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2560 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2561 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2562 them is set, LOOKUP_LSEARCH is forced.
2564 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2566 24. Added sender and host information to the "rejected by local_scan()" log
2567 line; previously there was no indication of these.
2569 25. Added .include_if_exists.
2571 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2572 Linux. It turns out that not all file systems support this. Apparently some
2573 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2574 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2575 not supported on the file descriptor, is now ignored when Exim is trying to
2576 sync a directory. This applies only to Linux.
2578 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2580 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2581 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2584 29. Added support for maildirsize files from supplied patch (modified a bit).
2586 30. The use of :fail: followed by an empty string could lead Exim to respond to
2587 sender verification failures with (e.g.):
2589 550 Verification failed for <xxx>
2590 550 Sender verify failed
2592 where the first response line was missing the '-' that indicates it is not
2593 the final line of the response.
2595 31. The loop for finding the name of the user that called Exim had a hardwired
2596 limit of 10; it now uses the value of finduser_retries, which is used for
2597 all other user lookups.
2599 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2602 33. Exim was neglecting to zero errno before one call of strtol() when
2603 expanding a string and expecting an integer value. On some systems this
2604 resulted in spurious "integer overflow" errors. Also, it was casting the
2605 result into an int without checking.
2607 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2608 did not work. The code looks as if it has *never* worked, though it appears
2609 to have been documented since at least releast 1.62. I have made it work.
2611 35. The "timeout_DNS" error in retry rules, also documented since at least
2612 1.62, also never worked. As it isn't clear exactly what this means, and
2613 clearly it isn't a major issue, I have abolished the feature by treating it
2614 as "timeout", and writing a warning to the main and panic logs.
2616 36. The display of retry rules for -brt wasn't always showing the error code
2619 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2620 timeout_connect_A, timeout_connect_MX.
2622 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2623 to the empty sender.
2625 39. The daemon was not analysing the content of -oX till after it had closed
2626 stderr and disconnected from the controlling terminal. This meant that any
2627 syntax errors were only noted on the panic log, and the return code from
2628 the command was 0. By re-arranging the code a little, I've made the
2629 decoding happen first, so such errors now appear on stderr, and the return
2630 code is 1. However, the actual setting up of the sockets still happens in
2631 the disconnected process, so errors there are still only recorded on the
2634 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2635 connections (as happens on some IP stacks) was logged at start up time as
2636 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2637 it from "IPv6 and IPv4", which means that two separate sockets are being
2640 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2641 says whether AF_INET or AF_INET6 was passed as an argument.
2643 42. Exiwhat output was messed up when time zones were included in log
2646 43. Exiwhat now gives more information about the daemon's listening ports,
2647 and whether -tls-on-connect was used.
2649 44. The "port" option of the smtp transport is now expanded.
2651 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2652 silently ignored. Now an error message is written to the main and panic
2655 46. There's a new ACL modifier called "logwrite" which writes to a log file
2656 as soon as it is encountered.
2658 47. Added $local_user_uid and $local_user_gid at routing time.
2660 48. Exim crashed when trying to verify a sender address that was being
2663 49. Exim was recognizing only a space character after ".include". It now also
2664 recognizes a tab character.
2666 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2667 extracting the relevant information from the specification. The man page no
2668 longer contains scrambled data for the -d option, and I've added a section
2669 at the front about calling Exim under different names.
2671 51. Added "extra_headers" argument to the "mail" command in filter files.
2673 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2676 53. Installed eximstats 1.29.
2678 54. Added transport_filter_timeout as a generic transport option.
2680 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2681 Cc: header lines. This was required by RFC 822, but it not required by RFC
2684 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2685 incoming messages that did not have them. Now it does so only if the
2686 message originates locally, that is, if there is no associated remote host
2687 address. When Resent- header lines are present, this applies to the Resent-
2688 lines rather than the non-Resent- lines.
2690 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2691 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2693 58. Messages for configuration errors now include the name of the main
2694 configuration file - useful now that there may be more than one file in a
2695 list (.included file names were always shown).
2697 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2698 for those rare installations that do not start the daemon as root or run it
2699 setuid root. I've cut out the call to initgroups() if the daemon is not
2702 60. The Exim user and group can now be bound into the binary as text strings
2703 that are looked up at the start of Exim's processing.
2705 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2707 62. Added $mailstore_basename variable.
2709 63. Installed patch to sieve.c from Michael Haardt.
2711 64. When Exim failed to open the panic log after failing to open the main log,
2712 the original message it was trying to log was written to stderr and debug
2713 output, but if they were not available (the usual case in production), it
2714 was lost. Now it is written to syslog before the two lines that record the
2715 failures to open the logs.
2717 65. Users' Exim filters run in subprocesses under the user's uid. It is
2718 possible for a "deliver" command or an alias in a "personal" command to
2719 provoke an address rewrite. If logging of address rewriting is configured,
2720 this fails because the process is not running as root or exim. There may be
2721 a better way of dealing with this, but for the moment (because 4.30 needs
2722 to be released), I have disabled address rewrite logging when running a
2723 filter in a non-root, non-exim process.
2729 1. The buildconfig auxiliary program wasn't quoting the value set for
2730 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2731 not defined. This bug was masked in 4.22 by the effect that was fixed in
2734 2. Some messages that were rejected after a message id was allocated were
2735 shown as "incomplete" by exigrep. It no longer does this for messages that
2736 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2738 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2739 have allow_domain_literals set, the ID did not get logged in the <= line.
2740 Domain literals are now always recognized in Message-ID: header lines.
2742 4. The first argument for a ${extract expansion item is the key name or field
2743 number. Leading and trailing spaces in this item were not being ignored,
2744 causing some misleading effects.
2746 5. When deliver_drop_privilege was set, single queue runner processes started
2747 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2748 same command in the process it spins off) were not dropping privilege.
2750 6. When the daemon running as "exim" started a queue runner, it always
2751 re-executed Exim in the spun-off process. This is a waste of effort when
2752 deliver_drop_privilege is set. The new process now just calls the
2753 queue-runner function directly.
2759 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2762 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2763 ignored. Though the use of -oP was forcing the writing of a pid file, it
2764 was always written to the default place.
2766 3. If the message "no IP address found for host xxxx" is generated during
2767 incoming verification, it is now followed by identification of the incoming
2768 connection (so you can more easily find what provoked it).
2770 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2772 5. Added some features to "harden" Exim a bit more against certain attacks:
2774 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2775 be put in Local/Makefile. This is like the never_users runtime option,
2776 but it cannot be overridden. The default setting is "root".
2778 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2779 prefix string with which any file named in a -C command line option
2782 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2783 is retained for -C and -D only if the caller of Exim is root. Without
2784 it, the exim user may also use -C and -D and retain privilege.
2786 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2787 command line option is disabled.
2789 6. Macro names set by the -D option must start with an upper case letter, just
2790 like macro names defined in the configuration file.
2792 7. Added "dereference=" facility to LDAP.
2794 8. Two instances of the typo "uknown" in the source files are fixed.
2796 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2797 the Configure-Makefile script screwed up while processing it.
2799 10. Incorporated PCRE 4.4.
2801 11. The SMTP synchronization check was not operating right at the start of an
2802 SMTP session. For example, it could not catch a HELO sent before the client
2803 waited for the greeting. There is now a check for outstanding input at the
2804 point when the greeting is written. Because of the duplex, asynchronous
2805 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2806 way, but not yet received, when the check is performed.
2808 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2809 on TCP/IP sockets, because this apparently causes some broken clients to
2812 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2813 unchanged) from the Cygwin maintainer.
2815 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2816 was supported instead of testing for mbx. Effectively a typo.
2818 15. The spa authenticator server code was not checking that the input it
2819 received was valid base64.
2821 16. The debug output line for the "set" modifier in ACLs was not showing the
2822 name of the variable that was being set.
2824 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2826 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2827 lookup on its own. Something else was needed to trigger the lookup. For
2828 example, a match in host_lookup or the need for a host name in a host list.
2829 Now, if $sender_host_name is referenced and the host name has not yet been
2830 looked up, a lookup is performed. If the lookup fails, the variable remains
2831 empty, and $host_lookup_failed is set to "1".
2833 19. Added "eqi" as a case-independent comparison operator.
2835 20. The saslauthd authentication condition could segfault if neither service
2836 nor realm was specified.
2838 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2839 error message that was logged was misleading, and incoming SMTP
2840 connections were dropped. The message is now more accurate, and temporary
2841 errors are given to SMTP connections.
2843 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2844 (except RSET) with a 421 error, until QUIT is received. However, it was
2845 failing to send a response to QUIT.
2847 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2848 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2849 ACL is now run after the helo_try_verify_hosts code.
2851 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2852 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2853 case-independent, so other case variants are also recognized). Apparently
2854 some systems use these upper case variants.
2856 25. If more than two messages were waiting for the same host, and a transport
2857 filter was specified for the transport, Exim sent two messages over the
2858 same TCP/IP connection, and then failed with "socket operation on non-
2859 socket" when it tried to send the third.
2861 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2863 27. The extern definition of crypt16() in expand.c was not being excluded when
2864 the OS had its own crypt16() function.
2866 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2867 as a preferred synonym for return_size_limit, both as an option and as an
2870 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2872 30. Changed the default configuration ACL to relax the local part checking rule
2873 for addresses that are not in any local domains. For these addresses,
2874 slashes and pipe symbols are allowed within local parts, but the sequence
2875 /../ is explicitly forbidden.
2877 31. SPA server authentication was not clearing the challenge buffer before
2880 32. log_message in a "warn" ACL statement was writing to the reject log as
2881 well as to the main log, which contradicts the documentation and doesn't
2882 seem right (because no rejection is happening). So I have stopped it.
2884 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2885 However, I am unable to do any testing of this.
2887 34. Fixed an infelicity in the appendfile transport. When checking directories
2888 for a mailbox, to see if any needed to be created, it was accidentally
2889 using path names with one or more superfluous leading slashes; tracing
2890 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2892 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2893 contents are added to the log line that is written for every discarded
2894 recipient. (Previously a log_message setting was ignored.)
2896 36. The ${quote: operator now quotes the string if it is empty.
2898 37. The install script runs exim in order to find its version number. If for
2899 some reason other than non-existence or emptiness, which it checks, it
2900 could not run './exim', it was installing it with an empty version number,
2901 i.e. as "exim-". This error state is now caught, and the installation is
2904 38. An argument was missing from the function that creates an error message
2905 when Exim fails to connect to the socket for saslauthd authentication.
2906 This could cause Exim to crash, or give a corrupted message.
2908 39. Added isip, isip4, and isip6 to ${if conditions.
2910 40. The ACL variables $acl_xx are now saved with the message, and can be
2911 accessed later in routers, transports, and filters.
2913 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2914 strings in the file are not string-expanded.
2916 42. If a MAIL command specified a SIZE value that was too large to fit into an
2917 int variable, the check against message_size_limit failed. Such values are
2918 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2919 day this will have to be increased, but I don't think I want to be around
2920 when emails are that large.
2927 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2928 iconv() is not standard in FreeBSD.
2930 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2931 IPv6 enabled. The observed symptom was a segmentation fault on return from
2932 the function os_common_find_running_interfaces() in src/os.c.
2934 3. In the check_special_case() function in daemon.c I had used "errno" as an
2935 argument name, which causes warnings on some systems. This was basically a
2936 typo, since it was named "eno" in the comments!
2938 4. The code that waits for the clock to tick (at a resolution of some fraction
2939 of a second) so as to ensure message-id uniqueness was always waiting for
2940 at least one whole tick, when it could have waited for less. [This is
2941 almost certainly not relevant at current processor speeds, where it is
2942 unlikely to ever wait at all. But we try to future-proof.]
2944 5. The function that sleeps for a time interval that includes fractions of a
2945 second contained a race. It did not block SIGALRM between setting the
2946 timer, and suspending (a couple of lines later). If the interval was short
2947 and the sigsuspend() was delayed until after it had expired, the suspension
2948 never ended. On busy systems this could lead to processes getting stuck for
2951 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2952 process, before it forks any delivery processes. The open lookup caching
2953 mechanism meant that the open file or database connection was passed into
2954 the delivery process. The problem was that delivery processes always tidy
2955 up cached lookup data. This could cause a problem for the next delivery
2956 process started by the queue runner, because the external queue runner
2957 process does not know about the closure. So the next delivery process
2958 still has data in the lookup cache. In the case of a file lookup, there was
2959 no problem because closing a file descriptor in a subprocess doesn't affect
2960 the parent. However, if the lookup was caching a connection to a database,
2961 the connection was closed, and the second delivery process was likely to
2962 see errors such as "PGSQL: query failed: server closed the connection
2963 unexpectedly". The problem has been fixed by closing all cached lookups
2964 in a queue runner before running a delivery process.
2966 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2967 seem to have the "const" qualifier which it has on other OS. I've
2970 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2971 *of the same type* (client or server) with the same public name that an
2972 error should be diagnosed.
2974 9. When Exim looked up a host name for an IP address, but failed to find the
2975 original IP address when looking up the host name (a safety check), it
2976 output the message "<ip address> does not match any IP for NULL", which was
2977 confusing, to say the least. The bug was that the host name should have
2978 appeared instead of "NULL".
2980 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2981 user that is built into the binary, and the -C or -D options is used, root
2982 privilege is dropped before the configuration file is read. In addition,
2983 logging is switched to stderr instead of the normal log files. If the
2984 configuration then re-defines the Exim user, the unprivileged environment
2985 is probably not what is expected, so Exim logs a panic warning message (but
2988 However, if deliver_drop_privilege is set, the unprivileged state may well
2989 be exactly what is intended, so the warning has been cut out in that case,
2990 and Exim is allowed to try to write to its normal log files.
2996 1. smtp_return_error_details was not giving details for temporary sender
2997 or receiver verification errors.
2999 2. Diagnose a configuration error if two authenticators have the same public
3002 3. Exim used not to create the message log file for a message until the first
3003 delivery attempt. This could be confusing when incoming messages were held
3004 for policy or load reasons. The message log file is now created at the time
3005 the message is received, and an initial "Received" line is written to it.
3007 4. The automatically generated man page for command line options had a minor
3008 bug that caused no ill effects; however, a more serious problem was that
3009 the procedure for building the man page automatically didn't always
3010 operate. Consequently, release 4.20 contains an out-of-date version. This
3011 shouldn't happen again.
3013 5. When building Exim with embedded Perl support, the script that builds the
3014 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3015 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3017 6. The freeze_tell option was not being used for messages that were frozen on
3018 arrival, either by an ACL or by local_scan().
3020 7. Added the smtp_incomplete_transaction log selector.
3022 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3023 was accepting AUTH without a new EHLO.
3025 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3026 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3029 10. Logging of TCP/IP connections (when configured) now happens in the main
3030 daemon process instead of the child process, so that the TCP/IP connection
3031 count is more accurate (but it can never be perfect).
3033 11. The use of "drop" in a nested ACL was not being handled correctly in the
3034 outer ACL. Now, if condition failure induced by the nested "drop" causes
3035 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3036 or "require"), the connection is dropped.
3038 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3039 that yield "discard" can now be used with an "accept" or a "discard" verb,
3040 but an error is generated for any others (because I can't see a useful way
3041 to define what should happen).
3043 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3044 are now processed in the same way as lines in the Exim configuration file.
3045 In particular, continuation lines are supported.
3047 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3049 15. Added -ti meaning -t -i.
3051 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3052 domains, and warn by logging if others are found.
3054 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3055 returned from GIFCONF when Exim is trying to find the list of interfaces on
3056 a host. The code in os.c has been modified to copy each ifreq to an aligned
3057 structure in all cases.
3059 Also, in some cases, the returned ifreq's were being copied to a 'struct
3060 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3061 means the last couple of bytes of an IPv6 address could be chopped if the
3062 ifreq contained only a normal sockaddr (14 bytes storage).
3064 18. Named domain lists were not supported in the hosts_treat_as_local option.
3065 An entry such as +xxxx was not recognized, and was treated as a literal
3068 19. Ensure that header lines added by a DATA ACL are included in the reject log
3069 if the ACL subsequently rejects the message.
3071 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3072 MD5 (which is deprecated).
3074 21. When testing a filter file using -bf, Exim was writing a message when it
3075 took the sender from a "From " line in the message, but it was not doing so
3076 when it took $return_path from a Return-Path: header line. It now does.
3078 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3079 with a valid header line field name (a series of printing characters
3080 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3082 23. Changed "disc" in the source to "disk" to conform to the documentation and
3083 the book and for uniformity.
3085 24. Ignore Sendmail's -Ooption=value command line item.
3087 25. When execve() failed while trying to run a command in a pipe transport,
3088 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3089 could be confused with a return value of 69 from the command itself. This
3090 has been changed to 127, the value the shell returns if it is asked to run
3091 a non-existent command. The wording for the related log line suggests a
3092 non-existent command as the problem.
3094 26. If received_header_text expands to an empty string, do not add a Received:
3095 header line to the message. (Well, it adds a token one on the spool, but
3096 marks it "old" so that it doesn't get used or transmitted.)
3098 27. Installed eximstats 1.28 (addition of -nt option).
3100 28. There was no check for failure on the call to getsockname() in the daemon
3101 code. This can fail if there is a shortage of resources on the system, with
3102 ENOMEM, for example. A temporary error is now given on failure.
3104 29. Contrary to the C standard, it seems that in some environments, the
3105 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3106 program. Exim now does this explicitly; it affects the formatting of
3107 timestamps using strftime().
3109 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3110 complaints. I've now initialized all the variables, to avoid this.
3112 32. Header lines added by a system filter were not being "seen" during
3113 transport-time rewrites.
3115 33. The info_callback() function passed to OpenSSL is set up with type void
3116 (*)(SSL *, int, int), as described somewhere. However, when calling the
3117 function (actually a macro) that sets it up, the type void(*)() is
3118 expected. I've put in a cast to prevent warnings from picky compilers.
3120 34. If a DNS black list lookup found a CNAME record, but there were no A
3121 records associated with the domain it pointed at, Exim crashed.
3123 35. If a DNS black list lookup returned more than one A record, Exim ignored
3124 all but the first. It now scans all returned addresses if a particular IP
3125 value is being sought. In this situation, the contents of the
3126 $dnslist_value variable are a list of all the addresses, separated by a
3129 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3130 to accept a host name and all its aliases if the forward lookup for any of
3131 them yielded the IP address of the incoming connection. Now it accepts only
3132 those names whose forward lookup yields the correct IP address. Any other
3133 names are discarded. This closes a loophole whereby a rogue DNS
3134 administrator could create reverse DNS records to break through a
3135 wildcarded host restriction in an ACL.
3137 37. If a user filter or a system filter that ran in a subprocess used any of
3138 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3139 the wrong values were passed to the pipe command ($thisaddress had the
3140 value of $0, $0 had the value of $1, etc). This bug was introduced by
3141 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3143 38. Improved the line breaking for long SMTP error messages from ACLs.
3144 Previously, if there was no break point between 40 and 75 characters, Exim
3145 left the rest of the message alone. Two changes have been made: (a) I've
3146 reduced the minimum length to 35 characters; (b) if it can't find a break
3147 point between 35 and 75 characters, it looks ahead and uses the first one
3148 that it finds. This may give the occasional overlong line, but at least the
3149 remaining text gets split now.
3151 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3152 file descriptors might be low, and that setting 1000 would always raise it.
3153 It turns out that in some environments, the limit is already over 1000 and
3154 that lowering it causes trouble. So now Exim takes care not to decrease it.
3156 40. When delivering a message, the value of $return_path is set to $sender_
3157 address at the start of routing (routers may change the value). By an
3158 oversight, this default was not being set up when an address was tested by
3159 -bt or -bv, which affected the outcome if any router or filter referred to
3162 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3163 log only when "message" or "log_message" are set. However, if one of the
3164 conditions was an address verification, or a call to a nested ACL, the
3165 messages generated by the underlying test were being passed through. This
3166 no longer happens. The underlying message is available in $acl_verify_
3167 message for both "message" and "log_message" expansions, so it can be
3168 passed through if needed.
3170 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3171 new expansion $bh_ to give the encoded byte string without charset
3172 translation. Translation happens only if iconv() is available; HAVE_ICONV
3173 indicates this at build time. HEADERS_CHARSET gives the charset to
3174 translate to; headers_charset can change it in the configuration, and
3175 "headers charset" can change it in an individual filter file.
3177 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3178 that creates RFC 2047 encoded "words" labels them as that charset instead
3179 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3180 expansion operator; (ii) when Exim creates a From: line for a local
3181 message; (iii) when a header line is rewritten to include a "phrase" part.
3183 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3184 buggy, causing it to skip the first lines of messages whose message ID
3185 ended in 'D'. This would not have bitten before Exim release 4.14, because
3186 message IDs were unlikely to end in 'D' before then. The effect was to have
3187 incorrect size information for certain domains.
3189 45. #include "config.h" was missing at the start of the crypt16.c module. This
3190 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3193 46. If there was a timeout during a "random" callout check, Exim treated it as
3194 a failure of the random address, and carried on sending RSET and the real
3195 address. If the delay was just some slowness somewhere, the response to the
3196 original RCPT would be taken as a response to RSET and so on, causing
3197 mayhem of various kinds.
3199 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3200 when I implemented it. It didn't allow for the fact that some option values
3201 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3202 the right test for positive values.
3204 48. Domain names in DNS records are case-independent. Exim always looks them up
3205 in lower case. Some resolvers return domain names in exactly the case they
3206 appear in the zone file, that is, they may contain uppercase letters. Not
3207 all resolvers do this - some return always lower case. Exim was treating a
3208 change of case by a resolver as a change of domain, similar to a widening
3209 of a domain abbreviation. This triggered its re-routing code and so it was
3210 trying to route what was effectively the same domain again. This normally
3211 caused routing to fail (because the router wouldn't handle the domain
3212 twice). Now Exim checks for this case specially, and just changes the
3213 casing of the domain that it ultimately uses when it transmits the message
3216 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3219 50. If a filter generated a file delivery with a non-absolute name (possible if
3220 no home directory exists for the router), the forbid_file option was not
3223 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3224 the existing equality matching.
3226 52. Exim was using ints instead of ino_t variables in some places where it was
3227 dealing with inode numbers.
3229 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3230 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3231 called TMPDIR, and if it finds it is different, it changes its value.
3233 54. The smtp_printf() function is now made available to local_scan() so
3234 additional output lines can be written before returning. There is also an
3235 smtp_fflush() function to enable the detection of a dropped connection.
3236 The variables smtp_input and smtp_batched_input are exported to
3239 55. Changed the default runtime configuration: the message "Unknown user"
3240 has been removed from the ACL, and instead placed on the localuser router,
3241 using the cannot_route_message feature. This means that any verification
3242 failures that generate their own messages won't get overridden. Similarly,
3243 the "Unrouteable address" message that was in the ACL for unverifiable
3244 relay addresses has also been removed.
3246 56. Added hosts_avoid_esmtp to the smtp transport.
3248 57. The exicyclog script was not checking for the esoteric option
3249 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3250 will work only if exicyclog is run under the appropriate euid.
3252 58. Following a discussion on the list, the rules by which Exim recognises line
3253 endings on incoming messages have been changed. The -dropcr and drop_cr
3254 options are now no-ops, retained only for backwards compatibility. The
3255 following line terminators are recognized: LF CRLF CR. However, special
3256 processing applies to CR:
3258 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3259 nor a local message in the state where . is a terminator.
3261 (ii) If a bare CR is encountered in a header line, an extra space is added
3262 after the line terminator so as not to end the header. The reasoning
3263 behind this is that bare CRs in header lines are most likely either
3264 to be mistakes, or people trying to play silly games.
3266 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3267 was being incorrectly given as 18 bytes larger than it should have been.
3268 This is a VOB (very old bug).
3270 60. This may never have affected anything current, but just in case it has:
3271 When the local host is found other than at the start of a list of hosts,
3272 the local host, those with the same MX, and any that follow, are discarded.
3273 When the list in question was part of a longer list of hosts, the following
3274 hosts (not currently being processed) were also being discarded. This no
3275 longer happens. I'm not sure if this situation could ever has previously
3278 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3281 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3282 header. This is something that is recommended in a new Internet Draft, and
3283 is something that is documented as being done by Sendmail. There are two
3284 possible values. For messages generated by the autoreply transport, Exim
3287 Auto-Submitted: auto-replied
3289 whereas for all other generated messages (e.g. bounces) it adds
3291 Auto-Submitted: auto-generated
3293 63. The "personal" condition in filters now includes a test for the
3294 Auto-Submitted: header. If it contains the string "auto-" the message it
3295 not considered personal.
3297 64. Added rcpt_include_affixes as a generic transport option.
3299 65. Added queue_only_override (default true).
3301 66. Added the syslog_duplication option.
3303 67. If what should have been the first header line of a message consisted of
3304 a space followed by a colon, Exim was mis-interpreting it as a header line.
3305 It isn't of course - it is syntactically invalid and should therefore be
3306 treated as the start of the message body. The misbehaviour could have
3307 caused a number of strange effects, including loss of data in subsequent
3308 header lines, and spool format errors.
3310 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3311 client host had authenticated. This control can now be exercised by an ACL
3312 for more flexibility.
3314 69. By default, callouts do not happen when testing with -bh. There is now a
3315 variant, -bhc, which does actually run the callout code, including
3316 consulting and updating the callout cache.
3318 70. Added support for saslauthd authentication, courtesy of Alexander
3321 71. If statvfs() failed on the spool or log directories while checking their
3322 size for availability, Exim confusingly gave the error "space shortage".
3323 Furthermore, in debugging mode it crashed with a floating point exception.
3324 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3325 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3326 serious problem, Exim now writes to the main and panic logs when this
3327 happens, with details of the failure. It then refuses to accept the
3328 incoming message, giving the message "spool directory problem" or "log
3329 directory problem" with a 421 code for SMTP messages.
3331 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3332 0, 1, and 2 exist, because some OS complain for execs without them (see
3333 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3334 descriptors. However, the code omitted to check that the open succeeded,
3335 causing mysterious errors if for some reason the permissions on /dev/null
3336 got screwed. Now Exim writes a message to the main and panic logs, and
3337 bombs out if it can't open /dev/null.
3339 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3340 interact so that it is all more flexible. It is supposed to remain
3341 backwards compatible. Also added extra_local_interfaces.
3343 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3344 to bomb out with an assertion failure - to the client this appears as a
3345 connection drop. This problem occurs in the part of the code that was taken
3346 from the Samba project. Fortunately, the assertion is in a very simple
3347 function, so I have fixed this by reproducing the function inline in the
3348 one place where it is called, and arranging for authentication to fail
3349 instead of killing the process with assert().
3351 75. The SPA client code was not working when the server requested OEM rather
3352 than Unicode encoding.
3354 76. Added code to make require_files with a specific uid setting more usable in
3355 the case where statting the file as root fails - usually a non-root-mounted
3356 NFS file system. When this happens and the failure is EACCES, Exim now
3357 forks a subprocess and does the per-uid checking as the relevant uid.
3359 77. Added process_log_path.
3361 78. If log_file_path was not explicitly set, a setting of check_log_space or
3362 check_log_inodes was ignored.
3364 79. If a space check for the spool or log partitions fails, the incident is now
3365 logged. Of course, in the latter case the data may get lost...
3367 80. Added the %p formatting code to string_format() so that it can be used to
3368 print addresses in debug_print(). Adjusted all the address printing in the
3369 debugging in store.c to use %p rather than %d.
3371 81. There was a concern that a line of code in smtp_in.c could overflow a
3372 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3373 initially expressed, the concern was not well-founded, because trailing
3374 spaces are removed early. However, if the trailing spaces were followed by
3375 a NULL, they did not get removed, so the overflow was possible. Two fixes
3378 (a) I re-wrote the offending code in a cleaner fashion.
3379 (b) If an incoming SMTP command contains a NULL character, it is rejected
3382 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3383 runs initgroups(), so that if the Exim user is in any additional groups,
3384 they will be used during message reception.
3390 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob