1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.250 2005/10/13 13:21:00 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 PH/01 Two changes to the default runtime configuration:
11 (1) Move the checks for relay_from_hosts and authenticated clients from
12 after to before the (commented out) DNS black list checks.
14 (2) Add control=submission to the relay_from_hosts and authenticated
15 clients checks, on the grounds that messages accepted by these
16 statements are most likely to be submissions.
18 PH/02 Several tidies to the handling of ${prvs and ${prvscheck:
20 (1) Generate an error if the third argument for the ${prvs expansion is
23 (2) Treat a missing third argument of ${prvscheck as if it were an empty
26 (3) Reset the variables that are obtained from the first argument of
27 ${prvscheck and used in the second argument before leaving the code,
28 because their memory is reclaimed, so using them afterwards may do
31 (4) Tidy up the code for expanding the arguments of ${prvscheck one by
32 one (it's much easier than Tom thought :-).
34 (5) Because of (4), we can now allow for the use of $prvscheck_result
35 inside the third argument.
37 PH/03 For some reason, the default setting of PATH when running a command from
38 a pipe transport was just "/usr/bin". I have changed it to
45 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
46 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
49 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
50 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
52 PH/03 Typo: missing ".o" in src/pcre/Makefile.
54 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
55 header line, restrict the check to what is listed in RFCs 2369 and 2929.
56 Also, for "Auto-Submitted", treat anything other than "no" as
57 non-personal, in accordance with RFC 3834. (Previously it treated
58 anything starting "auto-" as non-personal.)
60 TF/01 The control=submission/name=... option had a problem with syntax
61 errors if the name included a slash character. The /name= option
62 now slurps the rest of the string, so it can include any characters
63 but it must come last in the list of options (after /sender_retain
66 PH/05 Some modifications to the interface to the fake nameserver for the new
73 TK/01 Added the "success_on_redirect" address verification option. See
74 NewStuff for rationale and an example.
76 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
78 PH/02 Patch to exigrep to allow it to work on syslog lines.
80 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
81 fread() to skip over the body file's header line, because in Cygwin the
82 header line is locked and is inaccessible.
84 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
85 co-exist for some time) to make it clear that it is the Exim ID that is
86 referenced, not the Message-ID: header line.
88 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
89 string_format() function, because snprintf() does not exist on all
92 PH/06 The use of forbid_filter_existstest now also locks out the use of the
93 ${stat: expansion item.
95 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
96 protocol synchronization error", to keep the pedants happy.
98 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
99 well as for IRIX systems, when gcc is being used. See the host.c source
102 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
104 PH/10 Named domain lists were not working if used in a queue_smtp_domains
107 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
108 transport and to the smtp transport in LMTP mode.
110 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
112 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
113 run a filter in a subprocess. This could lead to confusion in subsequent
114 lookups in the parent process. There should also be a search_tidyup() at
115 the end of the subprocess.
117 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
118 only if the host matched helo_try_verify_hosts, which caused the
119 verification to occur when the EHLO/HELO command was issued. The ACL just
120 tested the remembered result. Now, if a previous verification attempt has
121 not happened, "verify = helo" does it there and then.
123 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
125 TK/03 Fix log output including CR from clamd.
127 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
128 exist provoked a memory error which could cause a segfault.
130 PH/15 Installed PCRE 6.2
132 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
134 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
135 of the problem. Specifically, suggested +O2 rather than +O1 for the
138 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
140 PH/20 If a delivery was routed to a non-standard port by means of an SRV
141 record, the port was not correctly logged when the outgoing_port log
142 selector was set (it logged the transort's default port).
144 PH/21 Added support for host-specific ports to manualroute, queryprogram,
145 fallback_hosts, and "hosts" in the smtp transport.
147 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
148 host errors such as "Connection refused".
150 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
151 authentication with radiusclient 0.4.9:
153 - Error returned from rc_read_config was caught wrongly
154 - Username/password not passed on to radius server due to wrong length.
156 The presumption is that some radiusclient API changes for 4.51/PH/17
157 were not taken care of correctly. The code is still untested by me (my
158 Linux distribution still has 0.3.2 of radiusclient), but it was
159 contributed by a Radius user.
161 PH/24 When doing a callout, the value of $domain wasn't set correctly when
162 expanding the "port" option of the smtp transport.
164 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
165 while reading a MIME header. Thanks to Tom Hughes for a patch.
167 PH/24 Include config.h inside local_scan.h so that configuration settings are
170 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
171 that in an ACL for RCPT (for example), you can examine exactly what was
174 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
175 commands, but it was not correctly comparing the address with the actual
176 client host address. Thus, it would show the EHLO address in Received:
177 header lines when this was not necessary.
179 PH/27 Added the % operator to ${eval:}.
181 PH/28 Exim tries to create and chdir to its spool directory when it starts;
182 it should be ignoring failures (because with -C, for example, it has lost
183 privilege). It wasn't ignoring creation failures other than "already
186 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
189 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
190 because an input line was too long, either on its own, or by virtue of
191 too many continuations, the temporary file was not being removed, and the
192 return code was incorrect.
194 PH/31 Missing "BOOL" in function definition in filtertest.c.
196 PH/32 Applied Sieve patches from the maintainer.
198 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
200 PH/33 Added "verify = not_blind".
202 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
203 Local/Makefile (with some defaults set). These are used in built scripts
204 such as exicyclog, but they have never been used in the exim_install
205 script (though there are many overriding facilities there). I have
206 arranged that the exim_install script now takes note of these two
209 PH/35 Installed configuration files for Dragonfly.
211 PH/36 When a locally submitted message by a trusted user did not contain a
212 From: header, and the sender address was obtained from -f or from an SMTP
213 MAIL command, and the trusted user did not use -F to supply a sender
214 name, $originator_name was incorrectly used when constructing a From:
215 header. Furthermore, $originator_name was used for submission mode
216 messages from external hosts without From: headers in a similar way,
217 which is clearly wrong.
219 PH/37 Added control=suppress_local_fixups.
221 PH/38 When log_selector = +received_sender was set, and the addition of the
222 sender made the log line's construction buffer exactly full, or one byte
223 less than full, an overflow happened when the terminating "\n" was
226 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
227 when the result of a list match is failure because a DNS lookup failed.
229 PH/40 RM_COMMAND is now used in the building process.
231 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
232 the "build-* directories that it finds.
234 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
235 address in a domain literal was a prefix of an interface address.
237 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
238 when verifying a sender address, unless rewrite_headers is false.
240 PH/44 Wrote a long comment about why errors_to addresses are verified as
241 recipients, not senders.
243 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
244 the ratelimit ACL was added.
246 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
248 PH/46 Added extra information about PostgreSQL errors to the error string.
250 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
251 avoiding the need to install special zones in a real server. This is
252 backwards compatible; if it can't find the fake resolver, it drops back.
253 Thus, both old and new test suites can be run.
255 TF/02 Added util/ratelimit.pl
257 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
258 clock is set back in time.
260 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
261 Candler <B.Candler@pobox.com>.
263 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
264 OK for addresses that are the result of redirections.
266 PH/48 A number of further additions for the benefit of the new test suite,
267 including a fake gethostbyname() that interfaces to the fake DNS resolver
270 TF/06 The fix for widen_domains has also been applied to qualify_single and
271 search_parents which are the other dnslookup options that can cause
274 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
277 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
279 TF/07 Exim produced the error message "an SRV record indicated no SMTP
280 service" if it encountered an MX record with an empty target hostname.
281 The message is now "an MX or SRV record indicated no SMTP service".
283 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
284 if the DNS of the sending site is misconfigured. This is quite a
285 common situation. This change restores the behaviour of treating a
286 helo verification defer as a failure.
288 PH/51 If self=fail was set on a router, the bounce message did not include the
289 actual error message.
295 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
297 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
298 command itself ended in error, the underlying message about the transport
299 filter timeout was being overwritten with the pipe command error. Now the
300 underlying error message should be appended to the second error message.
302 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
304 PH/02 Reduce the amount of output that "make" produces by default. Full output
305 can still be requested.
307 PH/03 The warning log line about a condition test deferring for a "warn" verb
308 was being output only once per connection, rather than after each
309 occurrence (because it was using the same function as for successful
310 "warn" verbs). This seems wrong, so I have changed it.
312 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
313 it should not have, which might have caused a crash in the right
314 circumstances, but probably never did.
316 PH/04 Installed a modified version of Tony Finch's patch to make submission
317 mode fix the return path as well as the Sender: header line, and to
318 add a /name= option so that you can make the user's friendly name appear
321 TF/03 Added the control = fakedefer ACL modifier.
323 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
324 Mark Lowes for thorough testing.
326 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
328 TK/03 Merged latest SRS patch from Miles Wilton.
330 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
331 with the definition in sysexits.h (which is #included earlier).
332 Fortunately, Exim does not actually use EX_OK. The code used to try to
333 preserve the sysexits.h value, by assumimg that macro definitions were
334 scanned for macro replacements. I have been disabused of this notion,
335 so now the code just undefines EX_OK before #including unistd.h.
337 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
338 in the smtp transport. When a block could not be written in a single
339 write() function, the timeout was being re-applied to each part-write.
340 This seems wrong - if the receiver was accepting one byte at a time it
341 would take for ever. The timeout is now adjusted when this happens. It
342 doesn't have to be particularly precise.
344 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
345 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
347 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
348 without a domain if the check to <postmaster@domain> fails.
350 SC/01 Eximstats: added -xls and the ability to specify output files
351 (patch written by Frank Heydlauf).
353 SC/02 Eximstats: use FileHandles for outputing results.
355 SC/03 Eximstats: allow any combination of xls, txt, and html output.
357 SC/04 Eximstats: fixed display of large numbers with -nvr option
359 SC/05 Eximstats: fixed merging of reports with empty tables.
361 SC/06 Eximstats: added the -include_original_destination flag
363 SC/07 Eximstats: removed tabs and trailing whitespace.
365 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
367 TK/06 MBOX spool code: Add real "From " MBOX separator line
368 so the .eml file is really in mbox format (even though
369 most programs do not really care). Patch from Alex Miller.
371 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
372 The latter is generated from $received_to and is only set if the
373 message has one envelope recipient. SA can use these headers,
374 obviously out-of-the-box. Patch from Alex Miller.
376 PH/08 The ${def test on a variable was returning false if the variable's
377 value was "0", contrary to what the specification has always said!
378 The result should be true unless the variable is empty.
380 PH/09 The syntax error of a character other than { following "${if
381 def:variable_name" (after optional whitespace) was not being diagnosed.
382 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
383 accidental colon was present, for example, could give incorrect results.
385 PH/10 Tidied the code in a number of places where the st_size field of a stat()
386 result is used (not including appendfile, where other changes are about
389 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
390 This involved changing a lot of size variables from int to off_t. It
391 should work with maildirs and everything.
393 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
394 spamd dying while we are connected to it.
396 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
399 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
400 function for BDB 4.3.
402 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
404 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
405 the sources for PCRE have been re-arranged and more files are now
408 PH/15 The code I had for printing potentially long long variables in PH/11
409 above was not the best (it lost precision). The length of off_t variables
410 is now inspected at build time, and an appropriate printing format (%ld
411 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
412 to be "long long int" or "long int". This is needed for the internal
413 formatting function string_vformat().
415 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
416 the configuration file to be ":syslog", then the script "guesses" where
417 the logs files are, rather than using the compiled in default. In our
418 case the guess is not the same as the compiled default, so the script
419 suddenly stopped working when I started to use syslog. The patch checks
420 to see if log_file_path is "". If so, it attempts to read it from exim
421 with no configuration file to get the compiled in version, before it
422 falls back to the previous guessing code."
424 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
425 implementing BATV in an Exim configuration. See NewStuff for the gory
428 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
429 Makefile that are specific to HP-UX.
431 PH/18 If the "use_postmaster" option was set for a recipient callout together
432 with the "random" option, the postmaster address was used as the MAIL
433 FROM address for the random test, but not for the subsequent recipient
434 test. It is now used for both.
436 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
437 patch removes a few documentation additions to RFC 3028, because the
438 latest draft now contains them. It adds the new en;ascii-case comparator
439 and a new error check for 8bit text in MIME parts. Comparator and
440 require names are now matched exactly. I enabled the subaddress
441 extension, but it is not well tested yet (read: it works for me)."
443 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
444 rework some of the code of TK/09 above to avoid the hardwired use of
445 "%lld" and "long long". Replaced the call to snprintf() with a call to
448 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
449 records point to non-existent hosts", "retry timeout exceeded", and
450 "retry time not reached for any host after a long failure period".
452 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
453 experimental DomainKeys support:
455 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
456 (2) On an error such as an illegally used "control", the wrong name for
457 the control was given.
459 These problems did NOT occur unless DomainKeys support was compiled.
461 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
463 PH/24 Added ${if match_ip condition.
465 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
466 databases so that it will be absolutely obvious if a crash occurs in the
467 DB library. This is a regular occurrence (often caused by mis-matched
470 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
471 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
472 picked up on a user's system that detects such things. There doesn't seem
473 to be a gcc warning option for this - only an attribute that has to be
474 put on the function's prototype. It seems that in Fedora Core 4 they have
475 set this on a number of new functions. No doubt there will be more in due
478 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
479 specify a transport. However, if an address that was verified by such a
480 router was the subject of a callout, Exim crashed because it tried to
481 read the rcpt_include_affixes from the non-existent transport. Now it
482 just assumes that the setting of that option is false. This bug was
483 introduced by 4.51/PH/31.
485 PH/28 Changed -d+all to exclude +memory, because that information is very
486 rarely of interest, but it makes the output a lot bigger. People tend to
487 do -d+all out of habit.
489 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
490 code in os-type was giving problems when libc.so lives in lib64, like on
493 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
494 aren't the modern standard, and it seems that some systems' include files
495 don't always have them. Exim was already checking for some of the newer
496 ones like T_AAAA, and defining it itself. I've added checks for all the
497 record types that Exim uses.
499 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
500 not automatically generating a new one, as it is supposed to. This
501 prevented TLS from working. If the file did exist, but contained invalid
502 data, a new version was generated, as expected. It was only the case of a
503 non-existent file that was broken.
505 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
506 with a change in libdomainkeys > 0.64.
508 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
509 from DNS. If the selector record carries the flag, it now has
510 precedence over the domain-wide flag.
512 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
514 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
515 the use of a port name that isn't defined in /etc/services) occurred, the
516 message was deferred as in a normal delivery, and thus remained on the
517 spool, instead of being failed because of the mua_wrapper setting. This
518 is now fixed, and I tidied up some of the mua_wrapper messages at the
521 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
522 the messages in a hash of arrays rather than using individual hashes.
523 This is a bit cleaner and results in dramatic memory savings, albeit
524 at a slight CPU cost.
526 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
527 as requested by Marc Sherman.
529 SC/10 Eximstats: added histograms for user specified patterns as requested
532 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
534 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
535 fopen() in the content-scanning modules that did not already have it.
541 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
542 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
544 TK/02 Fix ACL "control" statement not being available in MIME ACL.
546 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
548 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
549 to test Sieve filters that use "vacation".
551 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
552 that changes the way the GnuTLS parameters are stored in the cache file.
553 The new format can be generated externally. For backward compatibility,
554 if the data in the cache doesn't make sense, Exim assumes it has read an
555 old-format file, and it generates new data and writes a new file. This
556 means that you can't go back to an older release without removing the
559 PH/03 A redirect router that has both "unseen" and "one_time" set does not
560 work if there are any delivery delays because "one_time" forces the
561 parent to be marked "delivered", so its unseen clone is never tried
562 again. For this reason, Exim now forbids the simultaneous setting of
565 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
566 redirected to themselves ("homonym" addresses). Read the long ChangeLog
567 entry if you want to know the details. The fix, however, neglected to
568 consider the case when local delivery batching is involved. The test for
569 "previously delivered" was not happening when checking to see if an
570 address could be batched with a previous (undelivered) one; under
571 certain circumstances this could lead to multiple deliveries to the same
574 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
575 in its include files, and this causes problems building Exim.
577 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
578 header_syntax) but Exim was just ignoring anything given after a slash.
579 In particular, this caused confusion with an attempt to use "verify =
580 reverse_host_lookup/defer_ok". An error is now given when options are
581 supplied for verify items that do not have them. (Maybe reverse_host_
582 lookup should have a defer_ok option, but that's a different point.)
584 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
585 defined by RFC 821) to 2048, because there were problems with some AUTH
586 commands, and RFC 1869 says the size should be increased for extended
587 SMTP commands that take arguments.
589 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
592 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
593 "unknown" error; now it says that the functionality isn't in the binary.
595 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
596 an address' error message when a string expansion fails (syntax or
597 whatever). Otherwise the password may appear in the log. Following change
598 PH/42 below, there is no longer a chance of it appearing in a bounce
601 PH/11 Installed exipick version 20050225.0 from John Jetmore.
603 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
604 of its addresses was ever tried. (Bugzilla bug #2.)
606 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
607 the result incorrectly in the debug output. (It correctly added a newline
608 to what was transported.)
610 TF/01 Added $received_time.
612 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
613 commented out examples of how to interface to a virus scanner and to
614 SpamAssassin. Also added commented examples of av_scanner and
615 spamd_address settings.
617 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
618 and controls are allowed in which ACLs. There were a couple of minor
619 errors. Some of the entries in the conditions table (which is a table of
620 where they are NOT allowed) were getting very unwieldy; rewrote them as a
621 negation of where the condition IS allowed.
623 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
625 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
626 header file does not have a version number, so I've had to invent a new
627 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
628 API. The code is untested by me (my Linux distribution still has 0.3.2 of
629 radiusclient), but it was contributed by a Radius user.
631 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
632 files or directories, for OpenSSL.
634 PH/19 When an Exim process that is running as root has to create an Exim log
635 file, it does so in a subprocess that runs as exim:exim so as to get the
636 ownership right at creation (otherwise, other Exim processes might see
637 the file with the wrong ownership). There was no test for failure of this
638 fork() call, which would lead to the process getting stuck as it waited
639 for a non-existent subprocess. Forks do occasionally fail when resources
640 run out. I reviewed all the other calls to fork(); they all seem to check
643 PH/20 When checking for unexpected SMTP input at connect time (before writing
644 the banner), Exim was not dealing correctly with a non-positive return
645 from the read() function. If the client had disconnected by this time,
646 the result was a log entry for a synchronization error with an empty
647 string after "input=" when read() returned zero. If read() returned -1
648 (an event I could not check), uninitialized data bytes were printed.
649 There were reports of junk text (parts of files, etc) appearing after
652 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
654 PH/22 Added support for macro redefinition, and (re)definition in between
655 driver and ACL definitions.
657 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
658 forgetting to use the resulting value; it was using the unexpanded value.
660 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
661 hadn't been configured. The fix is from Juergen Kreileder, who
662 understands it better than I do:
664 "Here's what I see happening with three configured cyrus_sasl
665 authenticators configured (plain, login, cram-md5):
667 On startup auth_cyrus_sasl_init() gets called for each of these.
668 This means three calls to sasl_listmech() without a specified mech_list.
669 => SASL tests which mechs of all available mechs actually work
670 => three warnings about OTP not working
671 => the returned list contains: plain, login, cram-md5, digest-md5, ...
673 With the patch, sasl_listmech() also gets called three times. But now
674 SASL's mech_list option is set to the server_mech specified in the the
675 authenticator. Or in other words, the answer from sasl_listmech()
676 gets limited to just the mech you're testing for (which is different
678 => the return list contains just 'plain' or 'login', 'cram-md5' or
679 nothing depending on the value of ob->server_mech.
681 I've just tested the patch: Authentication still works fine,
682 unavailable mechs specified in the exim configuration are still
683 caught, and the auth.log warnings about OTP are gone."
685 PH/25 When debugging is enabled, the contents of the command line are added
686 to the debugging output, even when log_selector=+arguments is not
689 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
690 answer is "GNU", and only if the return is "GNU/something" is the answer
693 PH/27 $acl_verify_message is now set immediately after the failure of a
694 verification in an ACL, and so is available in subsequent modifiers. In
695 particular, the message can be preserved by coding like this:
697 warn !verify = sender
698 set acl_m0 = $acl_verify_message
700 Previously, $acl_verify_message was set only while expanding "message"
701 and "log_message" when a very denied access.
703 PH/28 Modified OS/os.c-Linux with
705 -#ifndef OS_LOAD_AVERAGE
706 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
708 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
709 nomenclature these days.)
711 PH/29 Installed patch from the Sieve maintainer that adds the options
712 sieve_useraddress and sieve_subaddress to the redirect router.
714 PH/30 In these circumstances:
715 . Two addresses routed to the same list of hosts;
716 . First host does not offer TLS;
717 . First host accepts first address;
718 . First host gives temporary error to second address;
719 . Second host offers TLS and a TLS session is established;
720 . Second host accepts second address.
721 Exim incorrectly logged both deliveries with the TLS parameters (cipher
722 and peerdn, if requested) that were in fact used only for the second
725 PH/31 When doing a callout as part of verifying an address, Exim was not paying
726 attention to any local part prefix or suffix that was matched by the
727 router that accepted the address. It now behaves in the same way as it
728 does for delivery: the affixes are removed from the local part unless
729 rcpt_include_affixes is set on the transport.
731 PH/32 Add the sender address, as F=<...>, to the log line when logging a
732 timeout during the DATA phase of an incoming message.
734 PH/33 Sieve envelope tests were broken for match types other than :is. I have
735 applied a patch sanctioned by the Sieve maintainer.
737 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
738 the uid or gid is negative. A case of a negative gid caused this to be
739 noticed. The fix allows for either to be negative.
741 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
742 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
745 PH/36 The change PH/12 above was broken. Fixed it.
747 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
748 the grounds that routing the same address twice would always produce the
749 same answer. This might have been true once, but it is certainly no
750 longer true now. Routing a child address may depend on the previous
751 routing that produced that child. Some complicated redirection strategies
752 went wrong when messages had multiple recipients, and made Exim's
753 behaviour dependent on the order in which the addresses were given.
755 I have moved the duplicate checking until after the routing is complete.
756 Exim scans the addresses that are assigned to local and remote
757 transports, and removes any duplicates. This means that more work will be
758 done, as duplicates will always all be routed, but duplicates are
759 presumably rare, so I don't expect this is of any significance.
761 For deliveries to pipes, files, and autoreplies, the duplicate checking
762 still happens during the routing process, since they are not going to be
765 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
766 It corrects a timeout issue with spamd. This is Ian's comment: "The
767 background is that sometimes spamd either never reads data from a
768 connection it has accepted, or it never writes response data. The exiscan
769 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
770 blindly assumes that writes won't block so it may never time out."
772 PH/39 Allow G after quota size as well as K and M.
774 PH/40 The value set for $authenticated_id in an authenticator may not contain
775 binary zeroes or newlines because the value is written to log lines and
776 to spool files. There was no check on this. Now the value is run through
777 the string_printing() function so that such characters are converted to
778 printable escape sequences.
780 PH/41 $message_linecount is a new variable that contains the total number of
781 lines in the message. Compare $body_linecount, which is the count for the
784 PH/42 Exim no longer gives details of delivery errors for specific addresses in
785 bounce and delay warning messages, except in certain special cases, which
788 (a) An SMTP error message from a remote host;
789 (b) A message specified in a :fail: redirection;
790 (c) A message specified in a "fail" command in a system filter;
791 (d) A message specified in a FAIL return from the queryprogram router;
792 (e) A message specified by the cannot_route_message router option.
794 In these cases only, Exim does include the error details in bounce and
795 warning messages. There are also a few cases where bland messages such
796 as "unrouteable address" or "local delivery error" are given.
798 PH/43 $value is now also set for the "else" part of a ${run expansion.
800 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
801 being worked on, but at least Exim now implements the latest version to
804 PH/45 In a pipe transport, although a timeout while waiting for the pipe
805 process to complete was treated as a delivery failure, a timeout while
806 writing the message to the pipe was logged, but erroneously treated as a
807 successful delivery. Such timeouts include transport filter timeouts. For
808 consistency with the overall process timeout, these timeouts are now
809 treated as errors, giving rise to delivery failures by default. However,
810 there is now a new Boolean option for the pipe transport called
811 timeout_defer, which, if set TRUE, converts the failures into defers for
812 both kinds of timeout. A transport filter timeout is now identified in
815 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
816 systems where "make" and "gmake" are different, calling "gmake" at top
817 level broke things. I've arranged for the value of $(MAKE) to be passed
818 from the Makefile to this script so that it can call the same version of
822 A note about Exim versions 4.44 and 4.50
823 ----------------------------------------
825 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
826 changes of various kinds. As a consequence, a big documentation update was
827 needed. This delayed the release for rather longer than seemed good, especially
828 in the light of a couple of (minor) security issues. Therefore, the changes
829 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
830 release. So 4.44 and 4.50 are in effect two different branches that both start
833 I have left the 4.50 change log unchanged; it contains all the changes since
834 4.43. The change log for 4.44 is below; many of its items are identical to
835 those for 4.50. This seems to be the most sensible way to preserve the
836 historical information.
842 1. Minor wording change to the doc/README.SIEVE file.
844 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
845 computation of the current number of files was incorrect.
847 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
848 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
851 4. Give more explanation in the error message when the command for a transport
852 filter fails to execute.
854 5. There are several places where Exim runs a non-Exim command in a
855 subprocess. The SIGUSR1 signal should be disabled for these processes. This
856 was being done only for the command run by the queryprogram router. It is
857 now done for all such subprocesses. The other cases are: ${run, transport
858 filters, and the commands run by the lmtp and pipe transports.
860 6. Added CONFIGURE_GROUP build-time option.
862 7. Some older OS have a limit of 256 on the maximum number of file
863 descriptors. Exim was using setrlimit() to set 1000 as a large value
864 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
865 systems. I've change it so that if it can't get 1000, it tries for 256.
867 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
868 was an oversight, and furthermore, ever since the addition of extra
869 controls (e.g. 4.43/32), the checks on when to allow different forms of
870 "control" were broken. There should now be diagnostics for all cases when a
871 control that does not make sense is encountered.
873 9. Added the /retain_sender option to "control=submission".
875 10. $recipients is now available in the predata ACL (oversight).
877 11. Tidy the search cache before the fork to do a delivery from a message
878 received from the command line. Otherwise the child will trigger a lookup
879 failure and thereby defer the delivery if it tries to use (for example) a
880 cached ldap connection that the parent has called unbind on.
882 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
883 of $address_data from the recipient verification was clobbered by the
886 13. The value of address_data from a sender verification is now available in
887 $sender_address_data in subsequent conditions in the ACL statement.
889 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
891 15. Added a new option "connect=<time>" to callout options, to set a different
894 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
895 was its contents. (It was OK if the option was not defined at all.)
897 17. A "Completed" log line is now written for messages that are removed from
898 the spool by the -Mrm option.
900 18. New variables $sender_verify_failure and $recipient_verify_failure contain
901 information about exactly what failed.
903 19. Added -dd to debug only the daemon process.
905 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
906 handles timeouts, both on the server side and network timeouts. Renamed the
907 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
909 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
910 It is now set to "smtps".
912 22. $host_address is now set to the target address during the checking of
915 23. When checking ignore_target_hosts for an ipliteral router, no host name was
916 being passed; this would have caused $sender_host_name to have been used if
917 matching the list had actually called for a host name (not very likely,
918 since this list is usually IP addresses). A host name is now passed as
921 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
922 code that specifies a non-restarting handler (typically sigaction() in
923 modern systems) in an attempt to fix a rare and obscure crash bug.
925 25. Narrowed the window for a race in the daemon that could cause it to ignore
926 SIGCHLD signals. This is not a major problem, because they are used only to
927 wake it up if nothing else does.
929 26. A malformed maildirsize file could cause Exim to calculate negative values
930 for the mailbox size or file count. Odd effects could occur as a result.
931 The maildirsize information is now recalculated if the size or filecount
934 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
935 support for a long time. Removed HAVE_SYS_VFS_H.
937 28. Installed the latest version of exipick from John Jetmore.
939 29. In an address list, if the pattern was not a regular expression, an empty
940 subject address (from a bounce message) matched only if the pattern was an
941 empty string. Non-empty patterns were not even tested. This was the wrong
942 because it is perfectly reasonable to use an empty address as part of a
943 database query. An empty address is now tested by patterns that are
944 lookups. However, all the other forms of pattern expect the subject to
945 contain a local part and a domain, and therefore, for them, an empty
946 address still always fails if the pattern is not itself empty.
948 30. Exim went into a mad DNS loop when attempting to do a callout where the
949 host was specified on an smtp transport, and looking it up yielded more
952 31. Re-factored the code for checking spool and log partition space into a
953 function that finds that data and another that does the check. The former
954 is then used to implement four new variables: $spool_space, $log_space,
955 $spool_inodes, and $log_inodes.
957 32. The RFC2047 encoding function was originally intended for short strings
958 such as real names; it was not keeping to the 75-character limit for
959 encoded words that the RFC imposes. It now respects the limit, and
960 generates multiple encoded words if necessary. To be on the safe side, I
961 have increased the buffer size for the ${rfc2047: expansion operator from
964 33. It is now permitted to omit both strings after an "if" condition; if the
965 condition is true, the result is "true". As before, when the second string
966 is omitted, a false condition yields an empty string. This makes it less
967 cumbersome to write custom ACL and router conditions.
969 34. Failure to deliver a bounce message always caused it to be frozen, even if
970 there was an errors_to setting on the router. The errors_to setting is now
973 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
974 canonical form (fully expanded) before being placed in
975 $sender_host_address.
977 36. The table in the code that translates DNS record types into text (T_A to
978 "A" for instance) was missing entries for NS and CNAME. It is just possible
979 that this could have caused confusion if both these types were looked up
980 for the same domain, because the text type is used as part of Exim's
981 per-process caching. But the chance of anyone hitting this buglet seems
984 37. The dnsdb lookup has been extended in a number of ways.
986 (1) There is a new type, "zns", which walks up the domain tree until it
987 finds some nameserver records. It should be used with care.
989 (2) There is a new type, "mxh", which is like "mx" except that it returns
990 just the host names, not the priorities.
992 (3) It is now possible to give a list of domains (or IP addresses) to be
993 looked up. The behaviour when one of the lookups defers can be
994 controlled by a keyword.
996 (4) It is now possible to specify the separator character for use when
997 multiple records are returned.
999 38. The dnslists ACL condition has been extended: it is now possible to supply
1000 a list of IP addresses and/or domains to be looked up in a particular DNS
1003 39. Added log_selector=+queue_time_overall.
1005 40. When running the queue in the test harness, wait just a tad after forking a
1006 delivery process, to get repeatability of debugging output.
1008 41. Include certificate and key file names in error message when GnuTLS fails
1009 to set them up, because the GnuTLS error message doesn't include the name
1010 of the failing file when there is a problem reading it.
1012 42. Allow both -bf and -bF in the same test run.
1014 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
1016 44. The "Exiscan patch" is now merged into the mainline Exim source.
1018 45. Sometimes the final signoff response after QUIT could fail to get
1019 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1020 < 0 before doing a fflush(). This bug looks as though it goes back to the
1021 introduction of TLS in release 3.20, but "sometimes" must have been rare
1022 because the tests only now provoked it.
1024 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
1025 (this can affect the format of dates).
1027 47. exim_tidydb, when checking for the continued existence of a message for
1028 which it has found a message-specific retry record, was not finding
1029 messages that were in split spool directories. Consequently, it was
1030 deleting retry records that should have stayed in existence.
1032 48. Steve fixed some bugs in eximstats.
1034 49. The SPA authentication driver was not abandoning authentication and moving
1035 on to the next authenticator when an expansion was forced to fail,
1036 contradicting the general specification for all authenticators. Instead it
1037 was generating a temporary error. It now behaves as specified.
1039 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
1040 (the order specifies the preference for clients). The order is now AES256,
1041 AES128, 3DES, ARCFOUR128.
1043 51. Small patch to Sieve code - explicitly set From: when generating an
1046 52. Exim crashed if a remote delivery caused a very long error message to be
1047 recorded - for instance if somebody sent an entire SpamAssassin report back
1048 as a large number of 550 error lines. This bug was coincidentally fixed by
1049 increasing the size of one of Exim's internal buffers (big_buffer) that
1050 happened as part of the Exiscan merge. However, to be on the safe side, I
1051 have made the code more robust (and fixed the comments that describe what
1054 53. Now that there can be additional text after "Completed" in log lines (if
1055 the queue_time_overall log selector is set), a one-byte patch to exigrep
1056 was needed to allow it to recognize "Completed" as not the last thing in
1059 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1060 patch that reportedly fixes this has been added. I am not expert enough to
1061 create a test for it. This is what the patch creator wrote:
1063 "I found a little strange behaviour of ldap code when working with
1064 Windows 2003 AD Domain, where users was placed in more than one
1065 Organization Units. When I tried to give exim partial DN, the exit code
1066 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1067 But simultaneously result of request was absolutely normal ldap result,
1068 so I produce this patch..."
1070 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1071 REFERENCE, so I have modified the code to exclude the patch when that macro
1074 55. Some experimental protocols are using DNS PTR records for new purposes. The
1075 keys for these records are domain names, not reversed IP addresses. The
1076 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1077 leaves it alone. Component reversal etc. now happens only for IP addresses.
1080 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1082 57. Double the size of the debug message buffer (to 2048) so that more of very
1083 long debug lines gets shown.
1085 58. The exicyclog utility now does better if the number of log files to keep
1086 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1088 59. Two changes related to the smtp_active_hostname option:
1090 (1) $smtp_active_hostname is now available as a variable.
1091 (2) The default for smtp_banner uses $smtp_active_hostname instead
1092 of $primary_hostname.
1094 60. The host_aton() function is supposed to be passed a string that is known
1095 to be a valid IP address. However, in the case of IPv6 addresses, it was
1096 not checking this. This is a hostage to fortune. Exim now panics and dies
1097 if the condition is not met. A case was found where this could be provoked
1098 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1099 components; fortuitously, this particular loophole had already been fixed
1100 by change 4.50/55 above.
1102 If there are any other similar loopholes, the new check in host_aton()
1103 itself should stop them being exploited. The report I received stated that
1104 data on the command line could provoke the exploit when Exim was running as
1105 exim, but did not say which command line option was involved. All I could
1106 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1107 running as the user.
1110 61. There was a buffer overflow vulnerability in the SPA authentication code
1111 (which came originally from the Samba project). I have added a test to the
1112 spa_base64_to_bits() function which I hope fixes it.
1115 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1116 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1118 63. The daemon start-up calls getloadavg() while still root for those OS that
1119 need the first call to be done as root, but it missed one case: when
1120 deliver_queue_load_max is set with deliver_drop_privilege. This is
1121 necessary for the benefit of the queue runner, because there is no re-exec
1122 when deliver_drop_privilege is set.
1124 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1125 This has been fixed.
1127 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1128 was in use, was not putting the data itself into the right store pool;
1129 consequently, it could be overwritten for a subsequent message in the same
1130 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1133 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1135 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1136 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1137 the function were treating the return as a boolean value, which happened to
1138 work because 0=false and not-0=true, but is not correct code.
1140 68. The host_aton() function was not handling scoped IPv6 addresses (those
1141 with, for example, "%eth0" on the end) correctly.
1143 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1144 negated items (that is, ~something) in unsigned ints. Some compilers
1145 apparently mutter when there is no cast.
1147 70. If an address verification called from an ACL failed, and did not produce a
1148 user-specific message (i.e. there was only a "system" message), nothing was
1149 put in $acl_verify_message. In this situation, it now puts the system
1152 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1153 session; change 4.31/43 added the unwanted input to the log line - except
1154 that it did not do this in the start of session case. It now does.
1156 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1157 This is wrong and can cause the other end to generate a synchronization
1158 error if it is another Exim or anything else that does the synchronization
1159 check. A QUIT command is no longer sent after a timeout.
1161 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1162 during host lookups.
1164 74. The defer_ok option of callout verification was not working if it was used
1165 when verifying addresses in header lines, that is, for this case:
1167 verify = header_sender/callout=defer_ok
1169 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1170 those file descriptors could be used for SMTP connections. If anything
1171 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1172 could be sent to the SMTP client, causing chaos. The daemon now opens
1173 stdin, stdout, and stderr to /dev/null when it puts itself into the
1176 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1177 log by default. The user can override this with suitable Perl magic.
1179 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1180 the log message when discard triggers, was not working for the DATA ACL or
1181 for the non-SMTP ACL.
1183 78. Error message wording change in sieve.c.
1185 79. If smtp_accept_max_per_host was set, the number of connections could be
1186 restricted to fewer than expected, because the daemon was trying to set up
1187 a new connection before checking whether the processes handling previous
1188 connections had finished. The check for completed processes is now done
1189 earlier. On busy systems, this bug wouldn't be noticed because something
1190 else would have woken the daemon, and it would have reaped the completed
1193 80. If a message was submitted locally by a user whose login name contained one
1194 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1195 It caused a spool format error. I have fixed the spool reading code. A
1196 related problem was that the "from" clause in the Received: line became
1197 illegal because of the space(s). It is now covered by ${quote_local_part.
1199 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1202 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1204 83. Patch from Sieve maintainer to fix unterminated string problem in
1205 "vacation" handling.
1207 84. Some minor changes to the Linux configuration files to help with other
1208 OS variants using glibc.
1210 85. One more patch for Sieve to update vacation handling to latest spec.
1213 ----------------------------------------------------
1214 See the note above about the 4.44 and 4.50 releases.
1215 ----------------------------------------------------
1221 1. Change 4.43/35 introduced a bug that caused file counts to be
1222 incorrectly computed when quota_filecount was set in an appendfile
1225 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1226 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1229 3. Give more explanation in the error message when the command for a transport
1230 filter fails to execute.
1232 4. There are several places where Exim runs a non-Exim command in a
1233 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1234 was being done only for the command run by the queryprogram router. It is
1235 now done for all such subprocesses. The other cases are: ${run, transport
1236 filters, and the commands run by the lmtp and pipe transports.
1238 5. Some older OS have a limit of 256 on the maximum number of file
1239 descriptors. Exim was using setrlimit() to set 1000 as a large value
1240 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1241 systems. I've change it so that if it can't get 1000, it tries for 256.
1243 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1244 was an oversight, and furthermore, ever since the addition of extra
1245 controls (e.g. 4.43/32), the checks on when to allow different forms of
1246 "control" were broken. There should now be diagnostics for all cases when a
1247 control that does not make sense is encountered.
1249 7. $recipients is now available in the predata ACL (oversight).
1251 8. Tidy the search cache before the fork to do a delivery from a message
1252 received from the command line. Otherwise the child will trigger a lookup
1253 failure and thereby defer the delivery if it tries to use (for example) a
1254 cached ldap connection that the parent has called unbind on.
1256 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1257 of $address_data from the recipient verification was clobbered by the
1258 sender verification.
1260 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1261 was its contents. (It was OK if the option was not defined at all.)
1263 11. A "Completed" log line is now written for messages that are removed from
1264 the spool by the -Mrm option.
1266 12. $host_address is now set to the target address during the checking of
1267 ignore_target_hosts.
1269 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1270 being passed; this would have caused $sender_host_name to have been used if
1271 matching the list had actually called for a host name (not very likely,
1272 since this list is usually IP addresses). A host name is now passed as
1275 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1276 code that specifies a non-restarting handler (typically sigaction() in
1277 modern systems) in an attempt to fix a rare and obscure crash bug.
1279 15. Narrowed the window for a race in the daemon that could cause it to ignore
1280 SIGCHLD signals. This is not a major problem, because they are used only to
1281 wake it up if nothing else does.
1283 16. A malformed maildirsize file could cause Exim to calculate negative values
1284 for the mailbox size or file count. Odd effects could occur as a result.
1285 The maildirsize information is now recalculated if the size or filecount
1288 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1289 support for a long time. Removed HAVE_SYS_VFS_H.
1291 18. Updated exipick to current release from John Jetmore.
1293 19. Allow an empty sender to be matched against a lookup in an address list.
1294 Previously the only cases considered were a regular expression, or an
1297 20. Exim went into a mad DNS lookup loop when doing a callout where the
1298 host was specified on the transport, if the DNS lookup yielded more than
1301 21. The RFC2047 encoding function was originally intended for short strings
1302 such as real names; it was not keeping to the 75-character limit for
1303 encoded words that the RFC imposes. It now respects the limit, and
1304 generates multiple encoded words if necessary. To be on the safe side, I
1305 have increased the buffer size for the ${rfc2047: expansion operator from
1308 22. Failure to deliver a bounce message always caused it to be frozen, even if
1309 there was an errors_to setting on the router. The errors_to setting is now
1312 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1313 canonical form (fully expanded) before being placed in
1314 $sender_host_address.
1316 24. Updated eximstats to version 1.33
1318 25. Include certificate and key file names in error message when GnuTLS fails
1319 to set them up, because the GnuTLS error message doesn't include the name
1320 of the failing file when there is a problem reading it.
1322 26. Expand error message when OpenSSL has problems setting up cert/key files.
1325 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1326 (this can affect the format of dates).
1328 28. exim_tidydb, when checking for the continued existence of a message for
1329 which it has found a message-specific retry record, was not finding
1330 messages that were in split spool directories. Consequently, it was
1331 deleting retry records that should have stayed in existence.
1333 29. eximstats updated to version 1.35
1334 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1335 1.35 - bugfix such that pie charts by volume are generated correctly
1337 30. The SPA authentication driver was not abandoning authentication and moving
1338 on to the next authenticator when an expansion was forced to fail,
1339 contradicting the general specification for all authenticators. Instead it
1340 was generating a temporary error. It now behaves as specified.
1342 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1343 (the order specifies the preference for clients). The order is now AES256,
1344 AES128, 3DES, ARCFOUR128.
1346 31. Small patch to Sieve code - explicitly set From: when generating an
1349 32. Exim crashed if a remote delivery caused a very long error message to be
1350 recorded - for instance if somebody sent an entire SpamAssassin report back
1351 as a large number of 550 error lines. This bug was coincidentally fixed by
1352 increasing the size of one of Exim's internal buffers (big_buffer) that
1353 happened as part of the Exiscan merge. However, to be on the safe side, I
1354 have made the code more robust (and fixed the comments that describe what
1357 33. Some experimental protocols are using DNS PTR records for new purposes. The
1358 keys for these records are domain names, not reversed IP addresses. The
1359 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1360 leaves it alone. Component reversal etc. now happens only for IP addresses.
1363 34. The host_aton() function is supposed to be passed a string that is known
1364 to be a valid IP address. However, in the case of IPv6 addresses, it was
1365 not checking this. This is a hostage to fortune. Exim now panics and dies
1366 if the condition is not met. A case was found where this could be provoked
1367 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1368 components; fortuitously, this particular loophole had already been fixed
1369 by change 4.50/55 or 4.44/33 above.
1371 If there are any other similar loopholes, the new check in host_aton()
1372 itself should stop them being exploited. The report I received stated that
1373 data on the command line could provoke the exploit when Exim was running as
1374 exim, but did not say which command line option was involved. All I could
1375 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1376 running as the user.
1379 35. There was a buffer overflow vulnerability in the SPA authentication code
1380 (which came originally from the Samba project). I have added a test to the
1381 spa_base64_to_bits() function which I hope fixes it.
1384 36. The daemon start-up calls getloadavg() while still root for those OS that
1385 need the first call to be done as root, but it missed one case: when
1386 deliver_queue_load_max is set with deliver_drop_privilege. This is
1387 necessary for the benefit of the queue runner, because there is no re-exec
1388 when deliver_drop_privilege is set.
1390 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1391 was in use, was not putting the data itself into the right store pool;
1392 consequently, it could be overwritten for a subsequent message in the same
1393 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1396 38. Sometimes the final signoff response after QUIT could fail to get
1397 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1398 < 0 before doing a fflush(). This bug looks as though it goes back to the
1399 introduction of TLS in release 3.20, but "sometimes" must have been rare
1400 because the tests only now provoked it.
1406 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1407 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1408 Now it returns an integer. A number of calls were still expecting a T/F
1409 return. Fortuitously, in all cases, the tests worked in OK situations,
1410 which is the norm. However, things would have gone wrong on any write
1411 failures on the smtp file descriptor. This function is used when sending
1412 messages over SMTP and also when doing verify callouts.
1414 2. When Exim is called to do synchronous delivery of a locally submitted
1415 message (the -odf or -odi options), it no longer closes stderr before doing
1418 3. Implemented the mua_wrapper option.
1420 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1422 5. Implemented the functions header_remove(), header_testname(),
1423 header_add_at_position(), and receive_remove_recipient(), and exported them
1426 6. If an ACL "warn" statement specified the addition of headers, Exim already
1427 inserted X-ACL-Warn: at the start if there was no header name. However, it
1428 was not making this test for the second and subsequent header lines if
1429 there were newlines in the string. This meant that an invalid header could
1430 be inserted if Exim was badly configured.
1432 7. Allow an ACL "warn" statement to add header lines at the start or after all
1433 the Received: headers, as well as at the end.
1435 8. Added the rcpt_4xx retry error code.
1437 9. Added postmaster_mailfrom=xxx to callout verification option.
1439 10. Added mailfrom=xxxx to the callout verification option, for verify=
1442 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1443 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1445 12. Inserted some casts to stop certain compilers warning when using pointer
1446 differences as field lengths or precisions in printf-type calls (mostly
1447 affecting debugging statements).
1449 13. Added optional readline() support for -be (dynamically loaded).
1451 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1452 same clock tick as a message's arrival, so that its received time was the
1453 same as the "first fail" time on the retry record, and that message
1454 remained on the queue past the ultimate address timeout, every queue runner
1455 would try a delivery (because it was past the ultimate address timeout) but
1456 after another failure, the ultimate address timeout, which should have then
1457 bounced the address, did not kick in. This was a "< instead of <=" error;
1458 in most cases the first failure would have been in the next clock tick
1459 after the received time, and all would be well.
1461 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1462 being recognized when the domain list was tested by the match_domain
1463 condition in an expansion string.
1465 16. Added the ${str2b64: operator.
1467 17. Exim was always calling setrlimit() to set a large limit for the number of
1468 processes, without checking whether the existing limit was already
1469 adequate. (It did check for the limit on file descriptors.) Furthermore,
1470 errors from getrlimit() and setrlimit() were being ignored. Now they are
1471 logged to the main and panic logs, but Exim does carry on, to try to do its
1472 job under whatever limits there are.
1474 18. Imported PCRE 5.0.
1476 19. Trivial typo in log message " temporarily refused connection" (the leading
1479 20. If the log selector return_path_on_delivery was set and an address was
1480 redirected to /dev/null, the delivery process crashed because it assumed
1481 that a return path would always be set for a "successful" delivery. In this
1482 case, the whole delivery is bypassed as an optimization, and therefore no
1485 21. Internal re-arrangement: the function for sending a challenge and reading
1486 a response while authentication was assuming a zero-terminated challenge
1487 string. It's now changed to take a pointer and a length, to allow for
1488 binary data in such strings.
1490 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1492 23. Exim was not respecting finduser_retries when seeking the login of the
1493 uid under which it was called; it was always trying 10 times. (The default
1494 setting of finduser_retries is zero.) Also, it was sleeping after the final
1495 failure, which is pointless.
1497 24. Implemented tls_on_connect_ports.
1499 25. Implemented acl_smtp_predata.
1501 26. If the domain in control=submission is set empty, Exim assumes that the
1502 authenticated id is a complete email address when it generates From: or
1503 Sender: header lines.
1505 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1506 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1507 chown and chgrp in /bin and hostname in /usr/bin.
1509 28. Exim was keeping the "process log" file open after each use, just as it
1510 does for the main log. This opens the possibility of it remaining open for
1511 long periods when the USR1 signal hits a daemon. Occasional processlog
1512 errors were reported, that could have been caused by this. Anyway, it seems
1513 much more sensible not to leave this file open at all, so that is what now
1516 29. The long-running daemon process does not normally write to the log once it
1517 has entered its main loop, and it closes the log before doing so. This is
1518 so that log files can straightforwardly be renamed and moved. However,
1519 there are a couple of unusual error situations where the daemon does write
1520 log entries, and I had neglected to close the log afterwards.
1522 30. The text of an SMTP error response that was received during a remote
1523 delivery was being truncated at 512 bytes. This is too short for some of
1524 the long messages that one sometimes sees. I've increased the limit to
1527 31. It is now possible to make retry rules that apply only when a message has a
1528 specific sender, in particular, an empty sender.
1530 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1531 it possible to be selective about when SMTP synchronization is enforced.
1533 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1535 32. Implemented hosts_connection_nolog.
1537 33. Added an ACL for QUIT.
1539 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1542 35. Added mailbox_size and mailbox_filecount to appendfile.
1544 36. Added control = no_multiline_responses to ACLs.
1546 37. There was a bug in the logic of the code that waits for the clock to tick
1547 in the case where the clock went backwards by a substantial amount such
1548 that the microsecond fraction of "now" was more than the microsecond
1549 fraction of "then" (but the whole seconds number was less).
1551 38. Added support for the libradius Radius client library this is found on
1552 FreeBSD (previously only the radiusclient library was supported).
1558 1. When certain lookups returned multiple values in the form name=value, the
1559 quoting of the values was not always being done properly. Specifically:
1560 (a) If the value started with a double quote, but contained no whitespace,
1562 (b) If the value contained whitespace other than a space character (i.e.
1563 tabs or newlines or carriage returns) it was not quoted.
1564 This fix has been applied to the mysql and pgsql lookups by writing a
1565 separate quoting function and calling it from the lookup code. The fix
1566 should probably also be applied to nisplus, ibase and oracle lookups, but
1567 since I cannot test any of those, I have not disturbed their existing code.
1569 2. A hit in the callout cache for a specific address caused a log line with no
1570 reason for rejecting RCPT. Now it says "Previous (cached) callout
1571 verification failure".
1573 3. There was an off-by-one bug in the queryprogram router. An over-long
1574 return line was truncated at 256 instead of 255 characters, thereby
1575 overflowing its buffer with the terminating zero. As well as fixing this, I
1576 have increased the buffer size to 1024 (and made a note to document this).
1578 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1579 when Exim is waiting for an SMTP response from a remote server, Exim
1580 restarts its select() call on the socket, thereby resetting its timeout.
1581 This is not a problem when such interrupts are rare. Somebody set up a cron
1582 job to run exiwhat every 2 minutes, which is less than the normal select()
1583 timeout (5 or 10 minutes). This meant that the select() timeout never
1584 kicked in because it was always reset. I have fixed this by comparing the
1585 time when an interrupt arrives with the time at the start of the first call
1586 to select(). If more time than the timeout has elapsed, the interrupt is
1587 treated as a timeout.
1589 5. Some internal re-factoring in preparation for the addition of Sieve
1590 extensions (by MH). In particular, the "personal" test is moved to a
1591 separate function, and given an option for scanning Cc: and Bcc: (which is
1592 not set for Exim filters).
1594 6. When Exim created an email address using the login of the caller as the
1595 local part (e.g. when creating a From: or Sender: header line), it was not
1596 quoting the local part when it contained special characters such as @.
1598 7. Installed new OpenBSD configuration files.
1600 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1601 try to make them clearer.
1603 9. Callout options, other than the timeout value, were being ignored when
1604 verifying sender addresses in header lines. For example, when using
1606 verify = header_sender/callout=no_cache
1608 the cache was (incorrectly) being used.
1610 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1611 only the Cygwin environment.
1613 11. When return_path_on_delivery was set as a log selector, if different remote
1614 addresses in the same message used different return paths and parallel
1615 remote delivery occurred, the wrong values would sometimes be logged.
1616 (Whenever a remote delivery process finished, the return path value from
1617 the most recently started remote delivery process was logged.)
1619 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1620 lines when AUTH and/or TLS are in use. This is the "received protocol"
1621 field. Exim used to use "asmtp" for authenticated SMTP, without any
1622 indication (in the protocol name) for TLS use. Now it follows the RFC and
1623 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1624 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1625 names appear in log lines as well as in Received: header lines.
1627 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1628 extensions, and comparison tests, and to fix some bugs.
1630 14. Changes to the "personal" filter test:
1632 (1) The test was buggy in that it was just doing the equivalent of
1633 "contains" tests on header lines. For example, if a user's address was
1634 anne@some.where, the "personal" test would incorrectly be true for
1636 To: susanne@some.where
1638 This test is now done by extracting each address from the header in turn,
1639 and checking the entire address. Other tests that are part of "personal"
1640 are now done using regular expressions (for example, to check local parts
1641 of addresses in From: header lines).
1643 (2) The list of non-personal local parts in From: addresses has been
1644 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1645 taken from the Sieve specification recommendations.
1647 (3) If the message contains any header line starting with "List-" it is
1648 treated as non-personal.
1650 (4) The test for "circular" in the Subject: header line has been removed
1651 because it now seems ill-conceived.
1653 15. Minor typos in src/EDITME comments corrected.
1655 16. Installed latest exipick from John Jetmore.
1657 17. If headers_add on a router specified a text string that was too long for
1658 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1659 of string_sprintf() is now avoided.
1661 18. $message_body_size was not set (it was always zero) when running the DATA
1662 ACL and the local_scan() function.
1664 19. For the "mail" command in an Exim filter, no default was being set for
1665 the once_repeat time, causing a random time value to be used if "once" was
1666 specified. (If the value happened to be <= 0, no repeat happened.) The
1667 default is now 0s, meaning "never repeat". The "vacation" command was OK
1668 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1669 (I found it when inspecting the code).
1671 20. There is now an overall timeout for performing a callout verification. It
1672 defaults to 4 times the callout timeout, which applies to individual SMTP
1673 commands during the callout. The overall timeout applies when there is more
1674 than one host that can be tried. The timeout is checked before trying the
1675 next host. This prevents very long delays if there are a large number of
1676 hosts and all are timing out (e.g. when the network connections are timing
1677 out). The value of the overall timeout can be changed by specifying an
1678 additional sub-option for "callout", called "maxwait". For example:
1680 verify = sender/callout=5s,maxwait=20s
1682 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1683 to the end before writing, but this should make it even safer).
1685 22. Exim was forgetting that it had advertised PIPELINING for the second and
1686 subsequent messages on an SMTP connection. It was also not resetting its
1687 memory on STARTTLS and an internal HELO.
1689 23. When Exim logs an SMTP synchronization error within a session, it now
1690 records whether PIPELINING has been advertised or not.
1692 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1693 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1694 rather than long int.
1696 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1698 26. Added the never_mail option to autoreply.
1704 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1705 crash if the getsockname() call failed; this can happen if a connection is
1706 closed very soon after it is established. The problem was simply in the
1707 order in which certain operations were done, causing Exim to try to write
1708 to the SMTP stream before it had set up the file descriptor. The bug has
1709 been fixed by making things happen in the correct order.
1715 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1716 before the connection was closed, thus losing the rejection response.
1718 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1719 some early Solaris releases, but causes trouble in current releases where
1720 socklen_t is defined.
1722 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1725 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1728 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1729 that releasing the top of it at the end releases what was used for sub-
1730 expansions (unless the block got too big). However, discard this block if
1731 the first thing is a variable or header, so that we can use its block when
1732 it is dynamic (useful for very large $message_headers, for example).
1734 6. Lookups now cache *every* query, not just the most recent. A new, separate
1735 store pool is used for this. It can be recovered when all lookup caches are
1736 flushed. Lookups now release memory at the end of their result strings.
1737 This has involved some general refactoring of the lookup sources.
1739 7. Some code has been added to the store_xxx() functions to reduce the amount
1740 of flapping under certain conditions.
1742 8. log_incoming_interface used to affect only the <= reception log lines. Now
1743 it causes the local interface and port to be added to several more SMTP log
1744 lines, for example "SMTP connection from", and rejection lines.
1746 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1748 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1750 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1751 could be overwritten at the end of the current message (or the start of a
1752 new message if it was set in a HELO ACL). The value is now preserved for
1753 the duration of the SMTP connection.
1755 12. If a transport had a headers_rewrite setting, and a matching header line
1756 contained an unqualified address, that address was qualified, even if it
1757 did not match any rewriting rules. The underlying bug was that the values
1758 of the flags that permit the existence of unqualified sender and recipient
1759 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1760 non-local messages, and by -bnq for local messages) were not being
1761 preserved with the message after it was received.
1763 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1764 "next input=" as part of the text comprising the host identity instead of
1765 the correct text. The code was using the same buffer for two different
1766 strings. However, depending on which order the printing function evaluated
1767 its arguments, the bug did not always show up. Under Linux, for example, my
1768 test suite worked just fine.
1770 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1771 doesn't work with some older versions of Perl. It has been changed to "my",
1772 which in any case is probably the better facility to use.
1774 15. A really picky compiler found some instances of statements for creating
1775 error messages that either had too many or two few arguments for the format
1778 16. The size of the buffer for calls to the DNS resolver has been increased
1779 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1780 for addresses that have a lot of PTR records. This alleviates a problem; it
1781 does not fully solve it.
1783 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1784 buffer now truncates the list and logs the incident, which is the same
1785 action as happens when Exim is looking up a host name and its aliases.
1786 Previously in this situation something unpredictable would happen;
1787 sometimes it was "internal error: store_reset failed".
1789 18. If a server dropped the connection unexpectedly when an Exim client was
1790 using GnuTLS and trying to read a response, the client delivery process
1791 crashed while trying to generate an error log message.
1793 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1794 string, for example:
1796 warn message = H1: something\nH2: something
1798 the text was added as a single header line from Exim's point of view
1799 though it ended up OK in the delivered message. However, searching for the
1800 second and subsequent header lines using $h_h2: did not work. This has been
1801 fixed. Similarly, if a system filter added multiple headers in this way,
1802 the routers could not see them.
1804 20. Expanded the error message when iplsearch is called with an invalid key to
1805 suggest using net-iplsearch in a host list.
1807 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1808 ACLs are no longer actually imposed (and a message to that effect is
1811 22. If a "gecos" field in a passwd entry contained escaped characters, in
1812 particular, if it contained a \" sequence, Exim got it wrong when building
1813 a From: or a Sender: header from that name. A second bug also caused
1814 incorrect handling when an unquoted " was present following a character
1815 that needed quoting.
1817 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1818 was not being matched caselessly.
1820 24. Arranged for all hyphens in the exim.8 source to be escaped with
1823 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1824 properly. Recipient callout cache records were still being keyed to include
1825 the sender, even when use_sender was set false. This led to far more
1826 callouts that were necessary. The sender is no longer included in the key
1827 when use_sender is false.
1829 26. Added "control = submission" modifier to ACLs.
1831 27. Added the ${base62d: operator to decode base 62 numbers.
1833 28. dnsdb lookups can now access SRV records.
1835 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1836 the configuration file.
1838 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1839 (-v) mode. This makes the output for a verbose queue run more intelligible.
1841 31. Added a use_postmaster feature to recipient callouts.
1843 32. Added the $body_zerocount variable, containing the number of binary zero
1844 bytes in the message body.
1846 33. The time of last modification of the "new" subdirectory is now used as the
1847 "mailbox time last read" when there is a quota error for a maildir
1850 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1852 35. Added +ignore_unknown as a special item in host lists.
1854 36. Code for decoding IPv6 addresses in host lists is now included, even if
1855 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1856 address was recognized as an IP address, but was then not correctly decoded
1857 into binary, causing unexpected and incorrect effects when compared with
1864 1. Very minor rewording of debugging text in manualroute to say "list of
1865 hosts" instead of "hostlist".
1867 2. If verify=header_syntax was set, and a header line with an unqualified
1868 address (no domain) and a large number of spaces between the end of the
1869 name and the colon was received, the reception process suffered a buffer
1870 overflow, and (when I tested it) crashed. This was caused by some obsolete
1871 code that should have been removed. The fix is to remove it!
1873 3. When running in the test harness, delay a bit after writing a bounce
1874 message to get a bit more predictability in the log output.
1876 4. Added a call to search_tidyup() just before forking a reception process. In
1877 theory, someone could use a lookup in the expansion of smtp_accept_max_
1878 per_host which, without the tidyup, could leave open a database connection.
1880 5. Added the variables $recipient_data and $sender_data which get set from a
1881 lookup success in an ACL "recipients" or "senders" condition, or a router
1882 "senders" option, similar to $domain_data and $local_part_data.
1884 6. Moved the writing of debug_print from before to after the "senders" test
1887 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1888 problems for message scanning, either using a data ACL, or using
1889 local_scan() because the Received: header was not generated till after they
1890 were called (in order to set the time as the time of reception completion).
1891 I have revised the way this works. The header is now generated after the
1892 body is received, but before the ACL or local_scan() are called. After they
1893 are run, the timestamp in the header is updated.
1899 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1900 before starting a queue runner without re-exec. This happened only when
1901 deliver_drop_privilege was set or when the Exim user was set to root. The
1902 effect of the bug was that timeouts during subsequent deliveries caused
1903 crashes instead of being properly handled. The handler is now left at its
1904 default (and expected) setting.
1906 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1907 message, again when deliver_drop_privilege is set or Exim is run as root.
1908 The bug described in (1) was not present in this case, but the tidying up
1909 of the other signals was missing. I have made the two cases consistent.
1911 3. The ignore_target_hosts setting on a manualroute router was being ignored
1912 for hosts that were looked up using the /MX notation.
1914 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1917 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1918 operated on the sender address. After changing the $sender_address to <>
1919 for the sender address verify, Exim was re-instated it as the original
1920 (before rewriting) address, but remembering that it had rewritten it, so it
1921 wasn't rewriting it again. This bug also had the effect of breaking the
1922 sender address verification caching when the sender address was rewritten.
1924 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1925 This has been changed so that if the ip literal address matches
1926 ignore_target_hosts, the router declines.
1928 7. Added expansion conditions match_domain, match_address, and match_local_
1929 part (NOT match_host).
1931 8. The placeholder for the Received: header didn't have a length field set.
1933 9. Added code to Exim itself and to exim_lock to test for a specific race
1934 condition that could lead to file corruption when using MBX delivery. The
1935 issue is with the lockfile that is created in /tmp. If this file is removed
1936 after a process has opened it but before that process has acquired a lock,
1937 there is the potential for a second process to recreate the file and also
1938 acquire a lock. This could lead to two Exim processes writing to the file
1939 at the same time. The added code performs the same test as UW imapd; it
1940 checks after acquiring the lock that its file descriptor still refers to
1941 the same named file.
1943 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1944 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1947 11. Added the smtp_active_hostname option. If used, this will typically be made
1948 to depend on the incoming interface address. Because $interface_address is
1949 not set up until the daemon has forked a reception process, error responses
1950 that can happen earlier (such as "too many connections") no longer contain
1953 12. If an expansion in a condition on a "warn" statement fails because a lookup
1954 defers, the "warn" statement is abandoned, and the next ACL statement is
1955 processed. Previously this caused the whole ACL to be aborted.
1957 13. Added the iplsearch lookup type.
1959 14. Added ident_timeout as a log selector.
1961 15. Added tls_certificate_verified as a log selector.
1963 16. Added a global option tls_require_ciphers (compare the smtp transport
1964 option of the same name). This controls incoming TLS connections.
1966 17. I finally figured out how to make tls_require_ciphers do a similar thing
1967 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1968 before starting the TLS session.
1970 18. Tabs are now shown as \t in -bP output.
1972 19. If the log selector return_path_on_delivery was set, Exim crashed when
1973 bouncing a message because it had too many Received: header lines.
1975 20. If two routers both had headers_remove settings, and the first one included
1976 a superfluous trailing colon, the final name in the first list and the
1977 first name in the second list were incorrectly joined into one item (with a
1978 colon in the middle).
1984 1. Added -C and -D options to the exinext utility, mainly to make it easier
1985 to include in the automated testing, but these could be helpful when
1986 multiple configurations are in use.
1988 2. The exinext utility was not formatting the output nicely when there was
1989 an alternate port involved in the retry record key, nor when there was a
1990 message id as well (for retries that were specific to a specific message
1991 and a specific host). It was also confused by IPv6 addresses, because of
1992 the additional colons they contain. I have fixed the IPv4 problem, and
1993 patched it up to do a reasonable job for IPv6.
1995 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1996 delivery, the log line now contains "pipelined" if PIPELINING was used.
1998 4. An SMTP transport process used to panic and die if the bind() call to set
1999 an explicit outgoing interface failed. This has been changed; it is now
2000 treated in the same way as a connect() failure.
2002 5. A reference to $sender_host_name in the part of a conditional expansion
2003 that was being skipped was still causing a DNS lookup. This no longer
2006 6. The def: expansion condition was not recognizing references to header lines
2007 that used bh_ and bheader_.
2009 7. Added the _cache feature to named lists.
2011 8. The code for checking quota_filecount in the appendfile transport was
2012 allowing one more file than it should have been.
2014 9. For compatibility with Sendmail, the command line option
2022 and sets the incoming protocol and host name (for trusted callers). The
2023 host name and its colon can be omitted when only the protocol is to be set.
2024 Note the Exim already has two private options, -pd and -ps, that refer to
2025 embedded Perl. It is therefore impossible to set a protocol value of "d" or
2026 "s", but I don't think that's a major issue.
2028 10. A number of refactoring changes to the code, none of which should affect
2031 (a) The number of logging options was getting close to filling up the
2032 32-bit word that was used as a bit map. I have split them into two classes:
2033 those that are passed in the argument to log_write(), and those that are
2034 only ever tested independently outside of that function. These are now in
2035 separate 32-bit words, so there is plenty of room for expansion again.
2036 There is no change in the user interface or the logging behaviour.
2038 (b) When building, for example, log lines, the code previously used a
2039 macro that called string_cat() twice, in order to add two strings. This is
2040 not really sufficiently general. Furthermore, there was one instance where
2041 it was actually wrong because one of the argument was used twice, and in
2042 one call a function was used. (As it happened, calling the function twice
2043 did not affect the overall behaviour.) The macro has been replaced by a
2044 function that can join an arbitrary number of extra strings onto a growing
2047 (c) The code for expansion conditions now uses a table and a binary chop
2048 instead of a serial search (which was left over from when there were very
2049 few conditions). Also, it now recognizes conditions like "pam" even when
2050 the relevant support is not compiled in: a suitably worded error message is
2051 given if an attempt is made to use such a condition.
2053 11. Added ${time_interval:xxxxx}.
2055 12. A bug was causing one of the ddress fields not to be passed back correctly
2056 from remote delivery subprocesses. The field in question was not being
2057 subsequently used, so this caused to problems in practice.
2059 13. Added new log selectors queue_time and deliver_time.
2061 14. Might have fixed a bug in maildirsizefile handling that threw up
2062 "unexpected character" debug warnings, and recalculated the data
2063 unnecessarily. In any case, I expanded the warning message to give more
2066 15. Added the message "Restricted characters in address" to the statements in
2067 the default ACL that block characters like @ and % in local parts.
2069 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2070 Three changes have been made:
2072 (a) There was a serious bug; a negative response to MAIL caused the whole
2073 recipient domain to be cached as invalid, thereby blocking all messages
2074 to all local parts at the same domain, from all senders. This bug has
2075 been fixed. The domain is no longer cached after a negative response to
2076 MAIL if the sender used is not empty.
2078 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2081 (c) A new callout option, "use_sender" has been added for people who want
2082 the modified behaviour.
2088 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2091 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2092 indeed breaks things for older releases.
2094 3. Added additional logging to the case where there is a problem reading data
2095 from a filter that is running in a subprocess using a pipe, in order to
2096 try to track down a specific problem.
2098 4. Testing facility fudge: when running in the test harness and attempting
2099 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2100 getting "No route to host". Convert this to a timeout.
2102 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2105 6. Some OS don't have socklen_t but use size_t instead. This affects the
2106 fifth argument of getsockopt() amongst other things. This is now
2107 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2108 can be set for individual OS. I have set it for SunOS5, OSF1, and
2109 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2110 some earlier ones do not.
2112 7. Change 4.30/15 was not doing the test caselessly.
2114 8. The standard form for an IPv6 address literal was being rejected by address
2115 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2116 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2117 this, as well as the form without the "IPv6" on the front (but only when
2118 address literals are enabled, of course).
2120 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2122 10. Exim crashed if a message with an empty sender address specified by -f
2123 encountered a router with an errors_to setting. This could be provoked only
2124 by a command such as
2128 where an empty string was supplied; "<>" did not hit this bug.
2130 11. Installed PCRE release 4.5.
2132 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2133 remained set. It is now erased.
2135 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2136 times from message ids (which are base 36 rather than the normal 62).
2138 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2139 were being counted as actual protocol errors, and logged if the log
2140 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2141 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2142 and DATA following a set of rejected RCPTs do not count as protocol errors.
2143 In other words, Exim assumes they were pipelined, though this may not
2144 actually be the case. Of course, in all cases the client gets an
2145 appropriate error code.
2147 15. If a lookup fails in an ACL condition, a message about the failure may
2148 be available; it is used if testing the ACL cannot continue, because most
2149 such messages specify what the cause of the deferral is. However, some
2150 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2151 that caused an old message to be retained and used if a later statement
2152 caused a defer, replacing the real cause of the deferral.
2154 16. If an IP address had so many PTR records that the DNS lookup buffer
2155 was not large enough to hold them, Exim could crash while trying to process
2156 the truncated data. It now detects and logs this case.
2158 17. Further to 4.21/58, another change has been made: if (and only if) the
2159 first line of a message (the first header line) ends with CRLF, a bare LF
2160 in a subsequent header line has a space inserted after it, so as not to
2161 terminate the header.
2163 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2164 unnecessarily, used atoi() instead of strtol(), and didn't check the
2165 termination when getting file sizes from file names by regex.
2167 19. Completely re-implemented the support for maildirsize files, in the light
2168 of a number of problems with the previous contributed implementation
2169 (4.30/29). In particular:
2171 . If the quota is zero, the maildirsize file is maintained, but no quota is
2174 . If the maildir directory does not exist, it is created before any attempt
2175 to write a maildirsize file.
2177 . The quota value in the file is just a cache; if the quota is changed in
2178 the transport, the new value overrides.
2180 . A regular expression is available for excluding directories from the
2183 20. The autoreply transport checks the characters in options that define the
2184 message's headers; it allows continued headers, but it was checking with
2185 isspace() after an embedded newline instead of explicitly looking for a
2188 21. If all the "regular" hosts to which an address was routed had passed their
2189 expiry times, and had not reached their retry times, the address was
2190 bounced, even if fallback hosts were defined. Now Exim should go on to try
2193 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2194 equivalent code in the SMTP transport. Some hosts send humungous responses
2195 to HELO/EHLO, more than 1024 it seems.
2197 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2198 gives compiler warnings in some environments. I've now done it "properly",
2201 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2202 (because of problems with the built-in one) was declared to return uschar *
2203 instead of char *, causing compiler failure.
2205 25. Fixed a file descriptor leak when processing alias/forward files.
2207 26. Fixed a minor format string issue in dbfn.c.
2209 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2211 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2212 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2215 29. When a sender address is verified, it is cached, to save repeating the test
2216 when there is more than one recipient in a message. However, when the
2217 verification involves a callout, it is possible for different callout
2218 options to be set for different recipients. It is too complicated to keep
2219 track of this in the cache, so now Exim always runs a verification when a
2220 callout is required, relying on the callout cache for the optimization.
2221 The overhead is duplication of the address routing, but this should not be
2224 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2225 to be verified with callout=postmaster, and the main callout worked but the
2226 postmaster check failed, the verification correctly failed. However, if a
2227 subsequent RCPT command asked for sender verification *without* the
2228 postmaster check, incorrect caching caused this verification also to fail,
2231 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2232 it was not caching the DNS options (qualify_single, search_parents) that
2233 were used when the lookup failed. A subsequent lookup with different
2234 options therefore always gave the same answer, though there were cases
2235 where it should not have. (Example: a "domains = !$mx_any" option on a
2236 dnslookup router: the "domains" option is always processed without any
2237 widening, but the router might have qualify_single set.) Now Exim uses the
2238 cached value only when the same options are set.
2240 32. Added John Jetmore's "exipick" utility to the distribution.
2242 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2243 than a timeout (e.g. a certificate is required, and is not provided), an
2244 Exim server now closes the connection immediately. Previously it waited for
2245 the client to close - but if the client is SSL, it seems that they each
2246 wait for each other, leading to a delay before one of them times out.
2248 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2249 maintained 0.8.x compatibility because I don't think many are using it, and
2250 it is clearly obsolete.
2252 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2255 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2256 client certificate was expired. A simple patch fixes this, though I don't
2257 understand the full logic of why the verify callback is called multiple
2260 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2261 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2262 which causes problems with some clients (such as the Certicom SSL Plus
2263 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2264 disables the coutermeasure allowing Eudora to connect."
2266 38. Exim was not checking that a write() to a log file succeeded. This could
2267 lead to Bad Things if a log got too big, in particular if it hit a file
2268 size limit. Exim now panics and dies if it cannot write to a log file, just
2269 as it does if it cannot open a log file.
2271 39. Modified OS/Makefile-Linux so that it now contains
2273 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2275 The two -D definitions ensure that Exim is compiled with large file
2276 support, which makes it possible to handle log files that are bigger than
2279 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2280 instance) a domain was checked against a named list that involved a lookup,
2281 causing $domain_data to be set, then another domain was checked against the
2282 same list, then the first domain was re-checked, the value of $domain_data
2283 after the final check could be wrong. In particular, if the second check
2284 failed, it could be set empty. This bug probably also applied to
2287 41. The strip_trailing_dot option was not being applied to the address given
2288 with the -f command-line option.
2290 42. The code for reading a message's header from the spool was incrementing
2291 $received_count, but never initializing it. This meant that the value was
2292 incorrect (doubled) while delivering a message in the same process in which
2293 it was received. In the most common configuration of Exim, this never
2294 happens - a fresh exec is done - but it can happen when
2295 deliver_drop_privilege is set.
2297 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2298 it now includes up to 150 characters of the unexpected data in the log
2301 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2302 and building data strings. The size of both of these buffers was 10 000
2303 bytes - far larger than anybody would *ever* want, thought I. Needless to
2304 say, somebody hit the limit. I have increased the maximum line length to
2305 20 000 and the maximum data length of concatenated lines to 100 000. I have
2306 also fixed two bugs, because there was no checking on these buffers. Tsk,
2307 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2308 buffer is too small.
2310 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2311 lsearch lookups. Now it does.
2313 46. When parsing a route_list item in a manualroute router, a fixed-length
2314 buffer was used for the list of hosts. I made this 1024 bytes long,
2315 thinking that nobody would ever have a list of hosts that long. Wrong.
2316 Somebody had a whole pile of complicated expansion conditions, and the
2317 string was silently truncated, leading to an expansion error. It turns out
2318 that it is easier to change to an unlimited length (owing to other changes
2319 that have happened since this code was originally written) than to build
2320 structure for giving a limitation error. The length of the item that
2321 expands into the list of hosts is now unlimited.
2323 47. The lsearch lookup could not handle data where the length of text line was
2324 more than 4095 characters. Such lines were truncated, leading to shortened
2325 data being returned. It should now handle lines of any length.
2327 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2328 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2331 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2332 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2333 no longer happens in comment lines. A list of replacements is now placed
2334 at the head of all of the source files, except those whose only change is
2335 to replace PERL_COMMAND in the very first #! line.
2337 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2338 messages on the queue, with a bottom-up merge sort, using code contributed
2339 by Michael Haardt. This should make operations like -bp somewhat faster on
2340 large queues. It won't affect queue runners, except when queue_run_in_order
2343 51. Installed eximstats 1.31 in the distribution.
2345 52. Added support for SRV lookups to the dnslookup router.
2347 53. If an ACL referred to $message_body or $message_body_end, the value was not
2348 reset for any messages that followed in the same SMTP session.
2350 54. The store-handling optimization for building very long strings was not
2351 differentiating between the different store pools. I don't think this
2352 actually made any difference in practice, but I've tidied it.
2354 55. While running the routers to verify a sender address, $sender_address
2355 was still set to the sender address. This is wrong, because when routing to
2356 send a bounce to the sender, it would be empty. Therefore, I have changed
2357 it so that, while verifying a sender address, $sender_address is set to <>.
2358 (There is no change to what happens when verifying a recipient address.)
2360 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2361 target A or AAAA records (if not already returned) without resetting the
2362 qualify_single or search_parents options of the DNS resolver. These are
2363 inappropriate in this case because the targets of MX and SRV records must
2364 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2365 target that, when qualified, matched something in the local domain. These
2366 two options are now turned off when doing these lookups.
2368 57. It seems that at least some releases of Reiserfs (which does not have the
2369 concept of a fixed number of inodes) returns zero and not -1 for the
2370 number of available inodes. This interacted badly with check_spool_inodes,
2371 which assumed that -1 was the "no such thing" setting. What I have done is
2372 to check that the total number of inodes is greater than zero before doing
2373 the test of how many are available.
2375 58. When a "warn" ACL statement has a log_message modifier, the message is
2376 remembered, and not repeated. This is to avoid a lot of repetition when a
2377 message has many recipients that cause the same warning to be written.
2378 Howewer, Exim was preserving the list of already written lines for an
2379 entire SMTP session, which doesn't seem right. The memory is now reset if a
2380 new message is started.
2382 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2383 debugging output unless log_rewrite was also set.
2385 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2386 of (int)(handle) when we know that handle contains (void *)(-1).
2388 61. The Exim daemon panic-logs an error return when it closes the incoming
2389 connection. However "connection reset by peer" seems to be common, and
2390 isn't really an error worthy of noting specially, so that particular error
2393 62. When Exim is trying to find all the local interfaces, it used to panic and
2394 die if the ioctl to get the interface flags failed. However, it seems that
2395 on at least one OS (Solaris 9) it is possible to have an interface that is
2396 included in the list of interfaces, but for which you get a failure error
2397 for this call. This happens when the interface is not "plumbed" into a
2398 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2399 failure of the "get flags" call assumes that the interface is down.
2401 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2402 makes life easier for people who are doing arithmetic on fields extracted
2403 from dates, where you often get leading zeros that should not be
2404 interpreted as octal.
2406 64. Added qualify_domain to the redirect router, to override the global
2409 65. If a pathologically long header line contained very many addresses (the
2410 report of this problem mentioned 10 000) and each of them was rewritten,
2411 Exim could use up a very large amount of memory. (It kept on making new
2412 copies of the header line as it rewrote, and never released the old ones.)
2413 At the expense of a bit more processing, the header rewriting function has
2414 been changed so that it no longer eats memory in this way.
2416 66. The generation of the Received: header has been moved from the time that a
2417 message starts to be received, to the time that it finishes. The timestamp
2418 in the Received: header should now be very close to that of the <= log
2419 line. There are two side-effects of this change:
2421 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2422 logged header lines no longer include the local Received: line, because
2423 it has not yet been created. The same applies to a copy of the message
2424 that is returned to a non-SMTP sender when a message is rejected.
2426 (b) When a filter file is tested using -bf, no additional Received: header
2427 is added to the test message. After some thought, I decided that this
2430 This change does not affect the value of $received_for. It is still set
2431 after address rewriting, but before local_scan() is called.
2433 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2435 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2436 gave an unhelpful panic error message, and a defer error. I have managed to
2437 change this behaviour so that it now rejects any supplied certificate,
2438 which seems right, as the list of acceptable certificates is empty.
2440 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2441 gave an unhelpful defer error. I have not managed to make this reject any
2442 supplied certificates, but the error message it gives is "no certificate
2443 supplied", which is not helpful.
2445 70. exigrep's output now also includes lines that are not associated with any
2446 message, but which match the given pattern. Implemented by a patch from
2447 Martin Sluka, which also tidied up the Perl a bit.
2449 71. Recipient callout verification, like sender verification, was using <> in
2450 the MAIL FROM command. This isn't really the right thing, since the actual
2451 sender may affect whether the remote host accepts the recipient or not. I
2452 have changed it to use the actual sender in the callout; this means that
2453 the cache record is now keyed on a recipient/sender pair, not just the
2454 recipient address. There doesn't seem to be a real danger of callout loops,
2455 since a callout by the remote host to check the sender would use <>.
2456 [SEE ABOVE: changed after hitting problems.]
2458 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2459 temporary errors. However, in the case of such a code being given after
2460 the end of a data transmission (i.e. after ".") Exim was failing to write
2461 a retry record for the message. (Yes, there was some broken host that was
2462 actually sending 8xx at this point.)
2464 73. An unknown lookup type in a host list could cause Exim to panic-die when
2465 the list was checked. (An example that provoked this was putting <; in the
2466 middle of a list instead of at the start.) If this happened during a DATA
2467 ACL check, a -D file could be left lying around. This kind of configuration
2468 error no longer causes Exim to die; instead it causes a defer errror. The
2469 incident is still logged to the main and panic logs.
2471 74. Buglet left over from Exim 3 conversion. The message "too many messages
2472 in one connection" was written to the rejectlog but not the mainlog, except
2473 when address rewriting (yes!) was being logged.
2475 75. Added write_rejectlog option.
2477 76. When a system filter was run not as root (that is, when system_filter_user
2478 was set), the values of the $n variables were not being returned to the
2479 main process; thus, they were not subsequently available in the $sn
2482 77. Added +return_path_on_delivery log selector.
2484 78. A connection timeout was being treated differently from recipients deferred
2485 when testing hosts_max_try with a message that was older than the host's
2486 retry timeout. (The host should not be counted, thus allowing all hosts to
2487 be tried at least once before bouncing.) This may have been the cause of an
2488 occasionally reported bug whereby a message would remain on the queue
2489 longer than the retry timeout, but would be bounced if a delivery was
2490 forced. I say "may" because I never totally pinned down the problem;
2491 setting up timeout/retry tests is difficult. See also the next item.
2493 79. The ultimate address timeout was not being applied to errors that involved
2494 a combination of host plus message (for example, a timeout on a MAIL
2495 command). When an address resolved to a number of possible hosts, and they
2496 were not all tried for each delivery (e.g. because of hosts_max_try), a
2497 message could remain on the queue longer than the retry timeout.
2499 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2502 81. Fixed an obscure SMTP outgoing bug which required at least the following
2503 conditions: (a) there was another message waiting for the same server;
2504 (b) the server returned 5xx to all RCPT commands in the first message so
2505 that the message was not completed; (c) the server dropped the connection
2506 or gave a negative response to the RSET that Exim sends to abort the
2507 transaction. The observed case was a dropped connection after DATA that had
2508 been sent in pipelining mode. That is, the server had advertised PIPELINING
2509 but was not implementing it correctly. The effect of the bug was incorrect
2510 behaviour, such as trying another host, and this could lead to a crash.
2516 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2517 and daemon.c were passed as pointers to ints; they should have been
2518 pointers to socklen_t variables (which are typically unsigned ints).
2520 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2523 3. Fixed a really odd bug that affected only the testing scheme; patching a
2524 certain fixed string in the binary changed the value of another string that
2525 happened to be identical to the end of the original first string.
2527 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2528 name", it returns that address as the IP address. On some operating
2529 systems (e.g. Solaris), it also passes back the IP address string as the
2530 "host name". However, on others (e.g. Linux), it passes back an empty
2531 string. Exim wasn't checking for this, and was changing the host name to an
2532 empty string, assuming it had been canonicized.
2534 5. Although rare, it is permitted to have more than one PTR record for a given
2535 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2536 all the names associated with an address, because they do in Solaris.
2537 However, it seems that they do not in Linux for data that comes from the
2538 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2539 I found this out when I moved to a new Linux workstation and tried to run
2540 the Exim test suite.
2542 To get round this problem I have changed the code so that it now does its
2543 own call to the DNS to look up PTR records when searching for a host name.
2544 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2545 addresses that are only in /etc/hosts are still found.
2547 This behaviour is, however, controlled by an option called host_lookup_
2548 order, which defaults to "bydns:byaddr". If people want to use the other
2549 order, or indeed, just use one or the other means of lookup, they can
2550 specify it in this variable.
2552 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2553 some operating systems, this comes back from gethostbyaddr() as an empty
2554 string, and this is what Exim used to test for. However, it seems that in
2555 other systems, "." is yielded. Exim now tests for this case too.
2557 7. The values of check_spool_space and check_log_space are now held internally
2558 as a number of kilobytes instead of an absolute number of bytes. If a
2559 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2560 kilobyte. This means that much larger values can be stored.
2562 8. Exim monitor: an attempt to get the action menu when not actually pointing
2563 at a message produces an empty menu entitled "No message selected". This
2564 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2565 no entries in it ("Shell widget menu has zero width and/or height"). So I
2566 have added a single, blank menu entry in this case.
2568 9. Added ${quote_local_part.
2570 10. MIME decoding is now applied to the contents of Subject: header lines when
2573 11. Now that a reference to $sender_host_address automatically causes a reverse
2574 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2575 host lookup before query-style lookups in lists that might use this
2576 variable. This has therefore been abolished, and the "net-" prefix is no
2577 longer necessary for query-style lookups.
2579 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2580 have been a typo for LFLAGS, so it has been changed.
2582 13. The install script calls Exim with "-C /dev/null" in order to find the
2583 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2584 to be output. Howeve, since Exim outputs its version number before the
2585 error, it didn't break the script. It just looked ugly. I fixed this by
2586 always allowing "-C /dev/null" if the caller is root.
2588 14. Ignore overlarge ACL variable number when reading spool file - insurance
2589 against a later release with more variables having written the file.
2591 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2592 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2593 this, as well as the form without the "IPv6" on the front.
2595 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2596 OS/Makefile-Darwin file.
2598 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2599 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2601 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2602 variable. However, some OS use ints for inodes. I've added cast to long int
2603 to get rid of the compiler warning.
2605 19. I had forgotten to lock out "/../" in configuration file names when
2606 ALT_CONFIG_PREFIX was set.
2608 20. Routers used for verification do not need to specify transports. However,
2609 if such a router generated a host list, and callout was configured, Exim
2610 crashed, because it could not find a port number from the (non-existent)
2611 transport. It now assumes port 25 in this circumstance.
2613 21. Added the -t option to exigrep.
2615 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2616 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2617 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2618 them is set, LOOKUP_LSEARCH is forced.
2620 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2622 24. Added sender and host information to the "rejected by local_scan()" log
2623 line; previously there was no indication of these.
2625 25. Added .include_if_exists.
2627 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2628 Linux. It turns out that not all file systems support this. Apparently some
2629 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2630 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2631 not supported on the file descriptor, is now ignored when Exim is trying to
2632 sync a directory. This applies only to Linux.
2634 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2636 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2637 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2640 29. Added support for maildirsize files from supplied patch (modified a bit).
2642 30. The use of :fail: followed by an empty string could lead Exim to respond to
2643 sender verification failures with (e.g.):
2645 550 Verification failed for <xxx>
2646 550 Sender verify failed
2648 where the first response line was missing the '-' that indicates it is not
2649 the final line of the response.
2651 31. The loop for finding the name of the user that called Exim had a hardwired
2652 limit of 10; it now uses the value of finduser_retries, which is used for
2653 all other user lookups.
2655 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2658 33. Exim was neglecting to zero errno before one call of strtol() when
2659 expanding a string and expecting an integer value. On some systems this
2660 resulted in spurious "integer overflow" errors. Also, it was casting the
2661 result into an int without checking.
2663 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2664 did not work. The code looks as if it has *never* worked, though it appears
2665 to have been documented since at least releast 1.62. I have made it work.
2667 35. The "timeout_DNS" error in retry rules, also documented since at least
2668 1.62, also never worked. As it isn't clear exactly what this means, and
2669 clearly it isn't a major issue, I have abolished the feature by treating it
2670 as "timeout", and writing a warning to the main and panic logs.
2672 36. The display of retry rules for -brt wasn't always showing the error code
2675 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2676 timeout_connect_A, timeout_connect_MX.
2678 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2679 to the empty sender.
2681 39. The daemon was not analysing the content of -oX till after it had closed
2682 stderr and disconnected from the controlling terminal. This meant that any
2683 syntax errors were only noted on the panic log, and the return code from
2684 the command was 0. By re-arranging the code a little, I've made the
2685 decoding happen first, so such errors now appear on stderr, and the return
2686 code is 1. However, the actual setting up of the sockets still happens in
2687 the disconnected process, so errors there are still only recorded on the
2690 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2691 connections (as happens on some IP stacks) was logged at start up time as
2692 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2693 it from "IPv6 and IPv4", which means that two separate sockets are being
2696 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2697 says whether AF_INET or AF_INET6 was passed as an argument.
2699 42. Exiwhat output was messed up when time zones were included in log
2702 43. Exiwhat now gives more information about the daemon's listening ports,
2703 and whether -tls-on-connect was used.
2705 44. The "port" option of the smtp transport is now expanded.
2707 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2708 silently ignored. Now an error message is written to the main and panic
2711 46. There's a new ACL modifier called "logwrite" which writes to a log file
2712 as soon as it is encountered.
2714 47. Added $local_user_uid and $local_user_gid at routing time.
2716 48. Exim crashed when trying to verify a sender address that was being
2719 49. Exim was recognizing only a space character after ".include". It now also
2720 recognizes a tab character.
2722 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2723 extracting the relevant information from the specification. The man page no
2724 longer contains scrambled data for the -d option, and I've added a section
2725 at the front about calling Exim under different names.
2727 51. Added "extra_headers" argument to the "mail" command in filter files.
2729 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2732 53. Installed eximstats 1.29.
2734 54. Added transport_filter_timeout as a generic transport option.
2736 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2737 Cc: header lines. This was required by RFC 822, but it not required by RFC
2740 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2741 incoming messages that did not have them. Now it does so only if the
2742 message originates locally, that is, if there is no associated remote host
2743 address. When Resent- header lines are present, this applies to the Resent-
2744 lines rather than the non-Resent- lines.
2746 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2747 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2749 58. Messages for configuration errors now include the name of the main
2750 configuration file - useful now that there may be more than one file in a
2751 list (.included file names were always shown).
2753 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2754 for those rare installations that do not start the daemon as root or run it
2755 setuid root. I've cut out the call to initgroups() if the daemon is not
2758 60. The Exim user and group can now be bound into the binary as text strings
2759 that are looked up at the start of Exim's processing.
2761 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2763 62. Added $mailstore_basename variable.
2765 63. Installed patch to sieve.c from Michael Haardt.
2767 64. When Exim failed to open the panic log after failing to open the main log,
2768 the original message it was trying to log was written to stderr and debug
2769 output, but if they were not available (the usual case in production), it
2770 was lost. Now it is written to syslog before the two lines that record the
2771 failures to open the logs.
2773 65. Users' Exim filters run in subprocesses under the user's uid. It is
2774 possible for a "deliver" command or an alias in a "personal" command to
2775 provoke an address rewrite. If logging of address rewriting is configured,
2776 this fails because the process is not running as root or exim. There may be
2777 a better way of dealing with this, but for the moment (because 4.30 needs
2778 to be released), I have disabled address rewrite logging when running a
2779 filter in a non-root, non-exim process.
2785 1. The buildconfig auxiliary program wasn't quoting the value set for
2786 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2787 not defined. This bug was masked in 4.22 by the effect that was fixed in
2790 2. Some messages that were rejected after a message id was allocated were
2791 shown as "incomplete" by exigrep. It no longer does this for messages that
2792 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2794 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2795 have allow_domain_literals set, the ID did not get logged in the <= line.
2796 Domain literals are now always recognized in Message-ID: header lines.
2798 4. The first argument for a ${extract expansion item is the key name or field
2799 number. Leading and trailing spaces in this item were not being ignored,
2800 causing some misleading effects.
2802 5. When deliver_drop_privilege was set, single queue runner processes started
2803 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2804 same command in the process it spins off) were not dropping privilege.
2806 6. When the daemon running as "exim" started a queue runner, it always
2807 re-executed Exim in the spun-off process. This is a waste of effort when
2808 deliver_drop_privilege is set. The new process now just calls the
2809 queue-runner function directly.
2815 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2818 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2819 ignored. Though the use of -oP was forcing the writing of a pid file, it
2820 was always written to the default place.
2822 3. If the message "no IP address found for host xxxx" is generated during
2823 incoming verification, it is now followed by identification of the incoming
2824 connection (so you can more easily find what provoked it).
2826 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2828 5. Added some features to "harden" Exim a bit more against certain attacks:
2830 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2831 be put in Local/Makefile. This is like the never_users runtime option,
2832 but it cannot be overridden. The default setting is "root".
2834 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2835 prefix string with which any file named in a -C command line option
2838 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2839 is retained for -C and -D only if the caller of Exim is root. Without
2840 it, the exim user may also use -C and -D and retain privilege.
2842 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2843 command line option is disabled.
2845 6. Macro names set by the -D option must start with an upper case letter, just
2846 like macro names defined in the configuration file.
2848 7. Added "dereference=" facility to LDAP.
2850 8. Two instances of the typo "uknown" in the source files are fixed.
2852 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2853 the Configure-Makefile script screwed up while processing it.
2855 10. Incorporated PCRE 4.4.
2857 11. The SMTP synchronization check was not operating right at the start of an
2858 SMTP session. For example, it could not catch a HELO sent before the client
2859 waited for the greeting. There is now a check for outstanding input at the
2860 point when the greeting is written. Because of the duplex, asynchronous
2861 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2862 way, but not yet received, when the check is performed.
2864 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2865 on TCP/IP sockets, because this apparently causes some broken clients to
2868 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2869 unchanged) from the Cygwin maintainer.
2871 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2872 was supported instead of testing for mbx. Effectively a typo.
2874 15. The spa authenticator server code was not checking that the input it
2875 received was valid base64.
2877 16. The debug output line for the "set" modifier in ACLs was not showing the
2878 name of the variable that was being set.
2880 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2882 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2883 lookup on its own. Something else was needed to trigger the lookup. For
2884 example, a match in host_lookup or the need for a host name in a host list.
2885 Now, if $sender_host_name is referenced and the host name has not yet been
2886 looked up, a lookup is performed. If the lookup fails, the variable remains
2887 empty, and $host_lookup_failed is set to "1".
2889 19. Added "eqi" as a case-independent comparison operator.
2891 20. The saslauthd authentication condition could segfault if neither service
2892 nor realm was specified.
2894 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2895 error message that was logged was misleading, and incoming SMTP
2896 connections were dropped. The message is now more accurate, and temporary
2897 errors are given to SMTP connections.
2899 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2900 (except RSET) with a 421 error, until QUIT is received. However, it was
2901 failing to send a response to QUIT.
2903 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2904 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2905 ACL is now run after the helo_try_verify_hosts code.
2907 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2908 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2909 case-independent, so other case variants are also recognized). Apparently
2910 some systems use these upper case variants.
2912 25. If more than two messages were waiting for the same host, and a transport
2913 filter was specified for the transport, Exim sent two messages over the
2914 same TCP/IP connection, and then failed with "socket operation on non-
2915 socket" when it tried to send the third.
2917 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2919 27. The extern definition of crypt16() in expand.c was not being excluded when
2920 the OS had its own crypt16() function.
2922 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2923 as a preferred synonym for return_size_limit, both as an option and as an
2926 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2928 30. Changed the default configuration ACL to relax the local part checking rule
2929 for addresses that are not in any local domains. For these addresses,
2930 slashes and pipe symbols are allowed within local parts, but the sequence
2931 /../ is explicitly forbidden.
2933 31. SPA server authentication was not clearing the challenge buffer before
2936 32. log_message in a "warn" ACL statement was writing to the reject log as
2937 well as to the main log, which contradicts the documentation and doesn't
2938 seem right (because no rejection is happening). So I have stopped it.
2940 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2941 However, I am unable to do any testing of this.
2943 34. Fixed an infelicity in the appendfile transport. When checking directories
2944 for a mailbox, to see if any needed to be created, it was accidentally
2945 using path names with one or more superfluous leading slashes; tracing
2946 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2948 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2949 contents are added to the log line that is written for every discarded
2950 recipient. (Previously a log_message setting was ignored.)
2952 36. The ${quote: operator now quotes the string if it is empty.
2954 37. The install script runs exim in order to find its version number. If for
2955 some reason other than non-existence or emptiness, which it checks, it
2956 could not run './exim', it was installing it with an empty version number,
2957 i.e. as "exim-". This error state is now caught, and the installation is
2960 38. An argument was missing from the function that creates an error message
2961 when Exim fails to connect to the socket for saslauthd authentication.
2962 This could cause Exim to crash, or give a corrupted message.
2964 39. Added isip, isip4, and isip6 to ${if conditions.
2966 40. The ACL variables $acl_xx are now saved with the message, and can be
2967 accessed later in routers, transports, and filters.
2969 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2970 strings in the file are not string-expanded.
2972 42. If a MAIL command specified a SIZE value that was too large to fit into an
2973 int variable, the check against message_size_limit failed. Such values are
2974 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2975 day this will have to be increased, but I don't think I want to be around
2976 when emails are that large.
2983 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2984 iconv() is not standard in FreeBSD.
2986 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2987 IPv6 enabled. The observed symptom was a segmentation fault on return from
2988 the function os_common_find_running_interfaces() in src/os.c.
2990 3. In the check_special_case() function in daemon.c I had used "errno" as an
2991 argument name, which causes warnings on some systems. This was basically a
2992 typo, since it was named "eno" in the comments!
2994 4. The code that waits for the clock to tick (at a resolution of some fraction
2995 of a second) so as to ensure message-id uniqueness was always waiting for
2996 at least one whole tick, when it could have waited for less. [This is
2997 almost certainly not relevant at current processor speeds, where it is
2998 unlikely to ever wait at all. But we try to future-proof.]
3000 5. The function that sleeps for a time interval that includes fractions of a
3001 second contained a race. It did not block SIGALRM between setting the
3002 timer, and suspending (a couple of lines later). If the interval was short
3003 and the sigsuspend() was delayed until after it had expired, the suspension
3004 never ended. On busy systems this could lead to processes getting stuck for
3007 6. Some uncommon configurations may cause a lookup to happen in a queue runner
3008 process, before it forks any delivery processes. The open lookup caching
3009 mechanism meant that the open file or database connection was passed into
3010 the delivery process. The problem was that delivery processes always tidy
3011 up cached lookup data. This could cause a problem for the next delivery
3012 process started by the queue runner, because the external queue runner
3013 process does not know about the closure. So the next delivery process
3014 still has data in the lookup cache. In the case of a file lookup, there was
3015 no problem because closing a file descriptor in a subprocess doesn't affect
3016 the parent. However, if the lookup was caching a connection to a database,
3017 the connection was closed, and the second delivery process was likely to
3018 see errors such as "PGSQL: query failed: server closed the connection
3019 unexpectedly". The problem has been fixed by closing all cached lookups
3020 in a queue runner before running a delivery process.
3022 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
3023 seem to have the "const" qualifier which it has on other OS. I've
3026 8. Change 4.21/2 was too strict. It is only if there are two authenticators
3027 *of the same type* (client or server) with the same public name that an
3028 error should be diagnosed.
3030 9. When Exim looked up a host name for an IP address, but failed to find the
3031 original IP address when looking up the host name (a safety check), it
3032 output the message "<ip address> does not match any IP for NULL", which was
3033 confusing, to say the least. The bug was that the host name should have
3034 appeared instead of "NULL".
3036 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
3037 user that is built into the binary, and the -C or -D options is used, root
3038 privilege is dropped before the configuration file is read. In addition,
3039 logging is switched to stderr instead of the normal log files. If the
3040 configuration then re-defines the Exim user, the unprivileged environment
3041 is probably not what is expected, so Exim logs a panic warning message (but
3044 However, if deliver_drop_privilege is set, the unprivileged state may well
3045 be exactly what is intended, so the warning has been cut out in that case,
3046 and Exim is allowed to try to write to its normal log files.
3052 1. smtp_return_error_details was not giving details for temporary sender
3053 or receiver verification errors.
3055 2. Diagnose a configuration error if two authenticators have the same public
3058 3. Exim used not to create the message log file for a message until the first
3059 delivery attempt. This could be confusing when incoming messages were held
3060 for policy or load reasons. The message log file is now created at the time
3061 the message is received, and an initial "Received" line is written to it.
3063 4. The automatically generated man page for command line options had a minor
3064 bug that caused no ill effects; however, a more serious problem was that
3065 the procedure for building the man page automatically didn't always
3066 operate. Consequently, release 4.20 contains an out-of-date version. This
3067 shouldn't happen again.
3069 5. When building Exim with embedded Perl support, the script that builds the
3070 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3071 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3073 6. The freeze_tell option was not being used for messages that were frozen on
3074 arrival, either by an ACL or by local_scan().
3076 7. Added the smtp_incomplete_transaction log selector.
3078 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3079 was accepting AUTH without a new EHLO.
3081 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3082 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3085 10. Logging of TCP/IP connections (when configured) now happens in the main
3086 daemon process instead of the child process, so that the TCP/IP connection
3087 count is more accurate (but it can never be perfect).
3089 11. The use of "drop" in a nested ACL was not being handled correctly in the
3090 outer ACL. Now, if condition failure induced by the nested "drop" causes
3091 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3092 or "require"), the connection is dropped.
3094 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3095 that yield "discard" can now be used with an "accept" or a "discard" verb,
3096 but an error is generated for any others (because I can't see a useful way
3097 to define what should happen).
3099 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3100 are now processed in the same way as lines in the Exim configuration file.
3101 In particular, continuation lines are supported.
3103 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3105 15. Added -ti meaning -t -i.
3107 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3108 domains, and warn by logging if others are found.
3110 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3111 returned from GIFCONF when Exim is trying to find the list of interfaces on
3112 a host. The code in os.c has been modified to copy each ifreq to an aligned
3113 structure in all cases.
3115 Also, in some cases, the returned ifreq's were being copied to a 'struct
3116 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3117 means the last couple of bytes of an IPv6 address could be chopped if the
3118 ifreq contained only a normal sockaddr (14 bytes storage).
3120 18. Named domain lists were not supported in the hosts_treat_as_local option.
3121 An entry such as +xxxx was not recognized, and was treated as a literal
3124 19. Ensure that header lines added by a DATA ACL are included in the reject log
3125 if the ACL subsequently rejects the message.
3127 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3128 MD5 (which is deprecated).
3130 21. When testing a filter file using -bf, Exim was writing a message when it
3131 took the sender from a "From " line in the message, but it was not doing so
3132 when it took $return_path from a Return-Path: header line. It now does.
3134 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3135 with a valid header line field name (a series of printing characters
3136 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3138 23. Changed "disc" in the source to "disk" to conform to the documentation and
3139 the book and for uniformity.
3141 24. Ignore Sendmail's -Ooption=value command line item.
3143 25. When execve() failed while trying to run a command in a pipe transport,
3144 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3145 could be confused with a return value of 69 from the command itself. This
3146 has been changed to 127, the value the shell returns if it is asked to run
3147 a non-existent command. The wording for the related log line suggests a
3148 non-existent command as the problem.
3150 26. If received_header_text expands to an empty string, do not add a Received:
3151 header line to the message. (Well, it adds a token one on the spool, but
3152 marks it "old" so that it doesn't get used or transmitted.)
3154 27. Installed eximstats 1.28 (addition of -nt option).
3156 28. There was no check for failure on the call to getsockname() in the daemon
3157 code. This can fail if there is a shortage of resources on the system, with
3158 ENOMEM, for example. A temporary error is now given on failure.
3160 29. Contrary to the C standard, it seems that in some environments, the
3161 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3162 program. Exim now does this explicitly; it affects the formatting of
3163 timestamps using strftime().
3165 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3166 complaints. I've now initialized all the variables, to avoid this.
3168 32. Header lines added by a system filter were not being "seen" during
3169 transport-time rewrites.
3171 33. The info_callback() function passed to OpenSSL is set up with type void
3172 (*)(SSL *, int, int), as described somewhere. However, when calling the
3173 function (actually a macro) that sets it up, the type void(*)() is
3174 expected. I've put in a cast to prevent warnings from picky compilers.
3176 34. If a DNS black list lookup found a CNAME record, but there were no A
3177 records associated with the domain it pointed at, Exim crashed.
3179 35. If a DNS black list lookup returned more than one A record, Exim ignored
3180 all but the first. It now scans all returned addresses if a particular IP
3181 value is being sought. In this situation, the contents of the
3182 $dnslist_value variable are a list of all the addresses, separated by a
3185 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3186 to accept a host name and all its aliases if the forward lookup for any of
3187 them yielded the IP address of the incoming connection. Now it accepts only
3188 those names whose forward lookup yields the correct IP address. Any other
3189 names are discarded. This closes a loophole whereby a rogue DNS
3190 administrator could create reverse DNS records to break through a
3191 wildcarded host restriction in an ACL.
3193 37. If a user filter or a system filter that ran in a subprocess used any of
3194 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3195 the wrong values were passed to the pipe command ($thisaddress had the
3196 value of $0, $0 had the value of $1, etc). This bug was introduced by
3197 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3199 38. Improved the line breaking for long SMTP error messages from ACLs.
3200 Previously, if there was no break point between 40 and 75 characters, Exim
3201 left the rest of the message alone. Two changes have been made: (a) I've
3202 reduced the minimum length to 35 characters; (b) if it can't find a break
3203 point between 35 and 75 characters, it looks ahead and uses the first one
3204 that it finds. This may give the occasional overlong line, but at least the
3205 remaining text gets split now.
3207 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3208 file descriptors might be low, and that setting 1000 would always raise it.
3209 It turns out that in some environments, the limit is already over 1000 and
3210 that lowering it causes trouble. So now Exim takes care not to decrease it.
3212 40. When delivering a message, the value of $return_path is set to $sender_
3213 address at the start of routing (routers may change the value). By an
3214 oversight, this default was not being set up when an address was tested by
3215 -bt or -bv, which affected the outcome if any router or filter referred to
3218 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3219 log only when "message" or "log_message" are set. However, if one of the
3220 conditions was an address verification, or a call to a nested ACL, the
3221 messages generated by the underlying test were being passed through. This
3222 no longer happens. The underlying message is available in $acl_verify_
3223 message for both "message" and "log_message" expansions, so it can be
3224 passed through if needed.
3226 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3227 new expansion $bh_ to give the encoded byte string without charset
3228 translation. Translation happens only if iconv() is available; HAVE_ICONV
3229 indicates this at build time. HEADERS_CHARSET gives the charset to
3230 translate to; headers_charset can change it in the configuration, and
3231 "headers charset" can change it in an individual filter file.
3233 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3234 that creates RFC 2047 encoded "words" labels them as that charset instead
3235 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3236 expansion operator; (ii) when Exim creates a From: line for a local
3237 message; (iii) when a header line is rewritten to include a "phrase" part.
3239 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3240 buggy, causing it to skip the first lines of messages whose message ID
3241 ended in 'D'. This would not have bitten before Exim release 4.14, because
3242 message IDs were unlikely to end in 'D' before then. The effect was to have
3243 incorrect size information for certain domains.
3245 45. #include "config.h" was missing at the start of the crypt16.c module. This
3246 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3249 46. If there was a timeout during a "random" callout check, Exim treated it as
3250 a failure of the random address, and carried on sending RSET and the real
3251 address. If the delay was just some slowness somewhere, the response to the
3252 original RCPT would be taken as a response to RSET and so on, causing
3253 mayhem of various kinds.
3255 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3256 when I implemented it. It didn't allow for the fact that some option values
3257 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3258 the right test for positive values.
3260 48. Domain names in DNS records are case-independent. Exim always looks them up
3261 in lower case. Some resolvers return domain names in exactly the case they
3262 appear in the zone file, that is, they may contain uppercase letters. Not
3263 all resolvers do this - some return always lower case. Exim was treating a
3264 change of case by a resolver as a change of domain, similar to a widening
3265 of a domain abbreviation. This triggered its re-routing code and so it was
3266 trying to route what was effectively the same domain again. This normally
3267 caused routing to fail (because the router wouldn't handle the domain
3268 twice). Now Exim checks for this case specially, and just changes the
3269 casing of the domain that it ultimately uses when it transmits the message
3272 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3275 50. If a filter generated a file delivery with a non-absolute name (possible if
3276 no home directory exists for the router), the forbid_file option was not
3279 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3280 the existing equality matching.
3282 52. Exim was using ints instead of ino_t variables in some places where it was
3283 dealing with inode numbers.
3285 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3286 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3287 called TMPDIR, and if it finds it is different, it changes its value.
3289 54. The smtp_printf() function is now made available to local_scan() so
3290 additional output lines can be written before returning. There is also an
3291 smtp_fflush() function to enable the detection of a dropped connection.
3292 The variables smtp_input and smtp_batched_input are exported to
3295 55. Changed the default runtime configuration: the message "Unknown user"
3296 has been removed from the ACL, and instead placed on the localuser router,
3297 using the cannot_route_message feature. This means that any verification
3298 failures that generate their own messages won't get overridden. Similarly,
3299 the "Unrouteable address" message that was in the ACL for unverifiable
3300 relay addresses has also been removed.
3302 56. Added hosts_avoid_esmtp to the smtp transport.
3304 57. The exicyclog script was not checking for the esoteric option
3305 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3306 will work only if exicyclog is run under the appropriate euid.
3308 58. Following a discussion on the list, the rules by which Exim recognises line
3309 endings on incoming messages have been changed. The -dropcr and drop_cr
3310 options are now no-ops, retained only for backwards compatibility. The
3311 following line terminators are recognized: LF CRLF CR. However, special
3312 processing applies to CR:
3314 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3315 nor a local message in the state where . is a terminator.
3317 (ii) If a bare CR is encountered in a header line, an extra space is added
3318 after the line terminator so as not to end the header. The reasoning
3319 behind this is that bare CRs in header lines are most likely either
3320 to be mistakes, or people trying to play silly games.
3322 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3323 was being incorrectly given as 18 bytes larger than it should have been.
3324 This is a VOB (very old bug).
3326 60. This may never have affected anything current, but just in case it has:
3327 When the local host is found other than at the start of a list of hosts,
3328 the local host, those with the same MX, and any that follow, are discarded.
3329 When the list in question was part of a longer list of hosts, the following
3330 hosts (not currently being processed) were also being discarded. This no
3331 longer happens. I'm not sure if this situation could ever has previously
3334 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3337 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3338 header. This is something that is recommended in a new Internet Draft, and
3339 is something that is documented as being done by Sendmail. There are two
3340 possible values. For messages generated by the autoreply transport, Exim
3343 Auto-Submitted: auto-replied
3345 whereas for all other generated messages (e.g. bounces) it adds
3347 Auto-Submitted: auto-generated
3349 63. The "personal" condition in filters now includes a test for the
3350 Auto-Submitted: header. If it contains the string "auto-" the message it
3351 not considered personal.
3353 64. Added rcpt_include_affixes as a generic transport option.
3355 65. Added queue_only_override (default true).
3357 66. Added the syslog_duplication option.
3359 67. If what should have been the first header line of a message consisted of
3360 a space followed by a colon, Exim was mis-interpreting it as a header line.
3361 It isn't of course - it is syntactically invalid and should therefore be
3362 treated as the start of the message body. The misbehaviour could have
3363 caused a number of strange effects, including loss of data in subsequent
3364 header lines, and spool format errors.
3366 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3367 client host had authenticated. This control can now be exercised by an ACL
3368 for more flexibility.
3370 69. By default, callouts do not happen when testing with -bh. There is now a
3371 variant, -bhc, which does actually run the callout code, including
3372 consulting and updating the callout cache.
3374 70. Added support for saslauthd authentication, courtesy of Alexander
3377 71. If statvfs() failed on the spool or log directories while checking their
3378 size for availability, Exim confusingly gave the error "space shortage".
3379 Furthermore, in debugging mode it crashed with a floating point exception.
3380 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3381 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3382 serious problem, Exim now writes to the main and panic logs when this
3383 happens, with details of the failure. It then refuses to accept the
3384 incoming message, giving the message "spool directory problem" or "log
3385 directory problem" with a 421 code for SMTP messages.
3387 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3388 0, 1, and 2 exist, because some OS complain for execs without them (see
3389 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3390 descriptors. However, the code omitted to check that the open succeeded,
3391 causing mysterious errors if for some reason the permissions on /dev/null
3392 got screwed. Now Exim writes a message to the main and panic logs, and
3393 bombs out if it can't open /dev/null.
3395 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3396 interact so that it is all more flexible. It is supposed to remain
3397 backwards compatible. Also added extra_local_interfaces.
3399 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3400 to bomb out with an assertion failure - to the client this appears as a
3401 connection drop. This problem occurs in the part of the code that was taken
3402 from the Samba project. Fortunately, the assertion is in a very simple
3403 function, so I have fixed this by reproducing the function inline in the
3404 one place where it is called, and arranging for authentication to fail
3405 instead of killing the process with assert().
3407 75. The SPA client code was not working when the server requested OEM rather
3408 than Unicode encoding.
3410 76. Added code to make require_files with a specific uid setting more usable in
3411 the case where statting the file as root fails - usually a non-root-mounted
3412 NFS file system. When this happens and the failure is EACCES, Exim now
3413 forks a subprocess and does the per-uid checking as the relevant uid.
3415 77. Added process_log_path.
3417 78. If log_file_path was not explicitly set, a setting of check_log_space or
3418 check_log_inodes was ignored.
3420 79. If a space check for the spool or log partitions fails, the incident is now
3421 logged. Of course, in the latter case the data may get lost...
3423 80. Added the %p formatting code to string_format() so that it can be used to
3424 print addresses in debug_print(). Adjusted all the address printing in the
3425 debugging in store.c to use %p rather than %d.
3427 81. There was a concern that a line of code in smtp_in.c could overflow a
3428 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3429 initially expressed, the concern was not well-founded, because trailing
3430 spaces are removed early. However, if the trailing spaces were followed by
3431 a NULL, they did not get removed, so the overflow was possible. Two fixes
3434 (a) I re-wrote the offending code in a cleaner fashion.
3435 (b) If an incoming SMTP command contains a NULL character, it is rejected
3438 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3439 runs initgroups(), so that if the Exim user is in any additional groups,
3440 they will be used during message reception.
3446 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob