1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.251 2005/10/20 14:03:22 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
41 PH/04 SUPPORT_TRANSLATE_IP_ADDRESS did not cause anything to be listed in the
48 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
49 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
52 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
53 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
55 PH/03 Typo: missing ".o" in src/pcre/Makefile.
57 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
58 header line, restrict the check to what is listed in RFCs 2369 and 2929.
59 Also, for "Auto-Submitted", treat anything other than "no" as
60 non-personal, in accordance with RFC 3834. (Previously it treated
61 anything starting "auto-" as non-personal.)
63 TF/01 The control=submission/name=... option had a problem with syntax
64 errors if the name included a slash character. The /name= option
65 now slurps the rest of the string, so it can include any characters
66 but it must come last in the list of options (after /sender_retain
69 PH/05 Some modifications to the interface to the fake nameserver for the new
76 TK/01 Added the "success_on_redirect" address verification option. See
77 NewStuff for rationale and an example.
79 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
81 PH/02 Patch to exigrep to allow it to work on syslog lines.
83 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
84 fread() to skip over the body file's header line, because in Cygwin the
85 header line is locked and is inaccessible.
87 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
88 co-exist for some time) to make it clear that it is the Exim ID that is
89 referenced, not the Message-ID: header line.
91 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
92 string_format() function, because snprintf() does not exist on all
95 PH/06 The use of forbid_filter_existstest now also locks out the use of the
96 ${stat: expansion item.
98 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
99 protocol synchronization error", to keep the pedants happy.
101 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
102 well as for IRIX systems, when gcc is being used. See the host.c source
105 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
107 PH/10 Named domain lists were not working if used in a queue_smtp_domains
110 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
111 transport and to the smtp transport in LMTP mode.
113 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
115 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
116 run a filter in a subprocess. This could lead to confusion in subsequent
117 lookups in the parent process. There should also be a search_tidyup() at
118 the end of the subprocess.
120 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
121 only if the host matched helo_try_verify_hosts, which caused the
122 verification to occur when the EHLO/HELO command was issued. The ACL just
123 tested the remembered result. Now, if a previous verification attempt has
124 not happened, "verify = helo" does it there and then.
126 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
128 TK/03 Fix log output including CR from clamd.
130 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
131 exist provoked a memory error which could cause a segfault.
133 PH/15 Installed PCRE 6.2
135 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
137 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
138 of the problem. Specifically, suggested +O2 rather than +O1 for the
141 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
143 PH/20 If a delivery was routed to a non-standard port by means of an SRV
144 record, the port was not correctly logged when the outgoing_port log
145 selector was set (it logged the transort's default port).
147 PH/21 Added support for host-specific ports to manualroute, queryprogram,
148 fallback_hosts, and "hosts" in the smtp transport.
150 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
151 host errors such as "Connection refused".
153 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
154 authentication with radiusclient 0.4.9:
156 - Error returned from rc_read_config was caught wrongly
157 - Username/password not passed on to radius server due to wrong length.
159 The presumption is that some radiusclient API changes for 4.51/PH/17
160 were not taken care of correctly. The code is still untested by me (my
161 Linux distribution still has 0.3.2 of radiusclient), but it was
162 contributed by a Radius user.
164 PH/24 When doing a callout, the value of $domain wasn't set correctly when
165 expanding the "port" option of the smtp transport.
167 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
168 while reading a MIME header. Thanks to Tom Hughes for a patch.
170 PH/24 Include config.h inside local_scan.h so that configuration settings are
173 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
174 that in an ACL for RCPT (for example), you can examine exactly what was
177 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
178 commands, but it was not correctly comparing the address with the actual
179 client host address. Thus, it would show the EHLO address in Received:
180 header lines when this was not necessary.
182 PH/27 Added the % operator to ${eval:}.
184 PH/28 Exim tries to create and chdir to its spool directory when it starts;
185 it should be ignoring failures (because with -C, for example, it has lost
186 privilege). It wasn't ignoring creation failures other than "already
189 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
192 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
193 because an input line was too long, either on its own, or by virtue of
194 too many continuations, the temporary file was not being removed, and the
195 return code was incorrect.
197 PH/31 Missing "BOOL" in function definition in filtertest.c.
199 PH/32 Applied Sieve patches from the maintainer.
201 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
203 PH/33 Added "verify = not_blind".
205 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
206 Local/Makefile (with some defaults set). These are used in built scripts
207 such as exicyclog, but they have never been used in the exim_install
208 script (though there are many overriding facilities there). I have
209 arranged that the exim_install script now takes note of these two
212 PH/35 Installed configuration files for Dragonfly.
214 PH/36 When a locally submitted message by a trusted user did not contain a
215 From: header, and the sender address was obtained from -f or from an SMTP
216 MAIL command, and the trusted user did not use -F to supply a sender
217 name, $originator_name was incorrectly used when constructing a From:
218 header. Furthermore, $originator_name was used for submission mode
219 messages from external hosts without From: headers in a similar way,
220 which is clearly wrong.
222 PH/37 Added control=suppress_local_fixups.
224 PH/38 When log_selector = +received_sender was set, and the addition of the
225 sender made the log line's construction buffer exactly full, or one byte
226 less than full, an overflow happened when the terminating "\n" was
229 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
230 when the result of a list match is failure because a DNS lookup failed.
232 PH/40 RM_COMMAND is now used in the building process.
234 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
235 the "build-* directories that it finds.
237 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
238 address in a domain literal was a prefix of an interface address.
240 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
241 when verifying a sender address, unless rewrite_headers is false.
243 PH/44 Wrote a long comment about why errors_to addresses are verified as
244 recipients, not senders.
246 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
247 the ratelimit ACL was added.
249 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
251 PH/46 Added extra information about PostgreSQL errors to the error string.
253 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
254 avoiding the need to install special zones in a real server. This is
255 backwards compatible; if it can't find the fake resolver, it drops back.
256 Thus, both old and new test suites can be run.
258 TF/02 Added util/ratelimit.pl
260 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
261 clock is set back in time.
263 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
264 Candler <B.Candler@pobox.com>.
266 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
267 OK for addresses that are the result of redirections.
269 PH/48 A number of further additions for the benefit of the new test suite,
270 including a fake gethostbyname() that interfaces to the fake DNS resolver
273 TF/06 The fix for widen_domains has also been applied to qualify_single and
274 search_parents which are the other dnslookup options that can cause
277 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
280 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
282 TF/07 Exim produced the error message "an SRV record indicated no SMTP
283 service" if it encountered an MX record with an empty target hostname.
284 The message is now "an MX or SRV record indicated no SMTP service".
286 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
287 if the DNS of the sending site is misconfigured. This is quite a
288 common situation. This change restores the behaviour of treating a
289 helo verification defer as a failure.
291 PH/51 If self=fail was set on a router, the bounce message did not include the
292 actual error message.
298 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
300 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
301 command itself ended in error, the underlying message about the transport
302 filter timeout was being overwritten with the pipe command error. Now the
303 underlying error message should be appended to the second error message.
305 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
307 PH/02 Reduce the amount of output that "make" produces by default. Full output
308 can still be requested.
310 PH/03 The warning log line about a condition test deferring for a "warn" verb
311 was being output only once per connection, rather than after each
312 occurrence (because it was using the same function as for successful
313 "warn" verbs). This seems wrong, so I have changed it.
315 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
316 it should not have, which might have caused a crash in the right
317 circumstances, but probably never did.
319 PH/04 Installed a modified version of Tony Finch's patch to make submission
320 mode fix the return path as well as the Sender: header line, and to
321 add a /name= option so that you can make the user's friendly name appear
324 TF/03 Added the control = fakedefer ACL modifier.
326 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
327 Mark Lowes for thorough testing.
329 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
331 TK/03 Merged latest SRS patch from Miles Wilton.
333 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
334 with the definition in sysexits.h (which is #included earlier).
335 Fortunately, Exim does not actually use EX_OK. The code used to try to
336 preserve the sysexits.h value, by assumimg that macro definitions were
337 scanned for macro replacements. I have been disabused of this notion,
338 so now the code just undefines EX_OK before #including unistd.h.
340 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
341 in the smtp transport. When a block could not be written in a single
342 write() function, the timeout was being re-applied to each part-write.
343 This seems wrong - if the receiver was accepting one byte at a time it
344 would take for ever. The timeout is now adjusted when this happens. It
345 doesn't have to be particularly precise.
347 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
348 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
350 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
351 without a domain if the check to <postmaster@domain> fails.
353 SC/01 Eximstats: added -xls and the ability to specify output files
354 (patch written by Frank Heydlauf).
356 SC/02 Eximstats: use FileHandles for outputing results.
358 SC/03 Eximstats: allow any combination of xls, txt, and html output.
360 SC/04 Eximstats: fixed display of large numbers with -nvr option
362 SC/05 Eximstats: fixed merging of reports with empty tables.
364 SC/06 Eximstats: added the -include_original_destination flag
366 SC/07 Eximstats: removed tabs and trailing whitespace.
368 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
370 TK/06 MBOX spool code: Add real "From " MBOX separator line
371 so the .eml file is really in mbox format (even though
372 most programs do not really care). Patch from Alex Miller.
374 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
375 The latter is generated from $received_to and is only set if the
376 message has one envelope recipient. SA can use these headers,
377 obviously out-of-the-box. Patch from Alex Miller.
379 PH/08 The ${def test on a variable was returning false if the variable's
380 value was "0", contrary to what the specification has always said!
381 The result should be true unless the variable is empty.
383 PH/09 The syntax error of a character other than { following "${if
384 def:variable_name" (after optional whitespace) was not being diagnosed.
385 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
386 accidental colon was present, for example, could give incorrect results.
388 PH/10 Tidied the code in a number of places where the st_size field of a stat()
389 result is used (not including appendfile, where other changes are about
392 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
393 This involved changing a lot of size variables from int to off_t. It
394 should work with maildirs and everything.
396 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
397 spamd dying while we are connected to it.
399 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
402 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
403 function for BDB 4.3.
405 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
407 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
408 the sources for PCRE have been re-arranged and more files are now
411 PH/15 The code I had for printing potentially long long variables in PH/11
412 above was not the best (it lost precision). The length of off_t variables
413 is now inspected at build time, and an appropriate printing format (%ld
414 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
415 to be "long long int" or "long int". This is needed for the internal
416 formatting function string_vformat().
418 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
419 the configuration file to be ":syslog", then the script "guesses" where
420 the logs files are, rather than using the compiled in default. In our
421 case the guess is not the same as the compiled default, so the script
422 suddenly stopped working when I started to use syslog. The patch checks
423 to see if log_file_path is "". If so, it attempts to read it from exim
424 with no configuration file to get the compiled in version, before it
425 falls back to the previous guessing code."
427 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
428 implementing BATV in an Exim configuration. See NewStuff for the gory
431 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
432 Makefile that are specific to HP-UX.
434 PH/18 If the "use_postmaster" option was set for a recipient callout together
435 with the "random" option, the postmaster address was used as the MAIL
436 FROM address for the random test, but not for the subsequent recipient
437 test. It is now used for both.
439 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
440 patch removes a few documentation additions to RFC 3028, because the
441 latest draft now contains them. It adds the new en;ascii-case comparator
442 and a new error check for 8bit text in MIME parts. Comparator and
443 require names are now matched exactly. I enabled the subaddress
444 extension, but it is not well tested yet (read: it works for me)."
446 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
447 rework some of the code of TK/09 above to avoid the hardwired use of
448 "%lld" and "long long". Replaced the call to snprintf() with a call to
451 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
452 records point to non-existent hosts", "retry timeout exceeded", and
453 "retry time not reached for any host after a long failure period".
455 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
456 experimental DomainKeys support:
458 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
459 (2) On an error such as an illegally used "control", the wrong name for
460 the control was given.
462 These problems did NOT occur unless DomainKeys support was compiled.
464 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
466 PH/24 Added ${if match_ip condition.
468 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
469 databases so that it will be absolutely obvious if a crash occurs in the
470 DB library. This is a regular occurrence (often caused by mis-matched
473 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
474 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
475 picked up on a user's system that detects such things. There doesn't seem
476 to be a gcc warning option for this - only an attribute that has to be
477 put on the function's prototype. It seems that in Fedora Core 4 they have
478 set this on a number of new functions. No doubt there will be more in due
481 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
482 specify a transport. However, if an address that was verified by such a
483 router was the subject of a callout, Exim crashed because it tried to
484 read the rcpt_include_affixes from the non-existent transport. Now it
485 just assumes that the setting of that option is false. This bug was
486 introduced by 4.51/PH/31.
488 PH/28 Changed -d+all to exclude +memory, because that information is very
489 rarely of interest, but it makes the output a lot bigger. People tend to
490 do -d+all out of habit.
492 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
493 code in os-type was giving problems when libc.so lives in lib64, like on
496 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
497 aren't the modern standard, and it seems that some systems' include files
498 don't always have them. Exim was already checking for some of the newer
499 ones like T_AAAA, and defining it itself. I've added checks for all the
500 record types that Exim uses.
502 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
503 not automatically generating a new one, as it is supposed to. This
504 prevented TLS from working. If the file did exist, but contained invalid
505 data, a new version was generated, as expected. It was only the case of a
506 non-existent file that was broken.
508 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
509 with a change in libdomainkeys > 0.64.
511 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
512 from DNS. If the selector record carries the flag, it now has
513 precedence over the domain-wide flag.
515 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
517 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
518 the use of a port name that isn't defined in /etc/services) occurred, the
519 message was deferred as in a normal delivery, and thus remained on the
520 spool, instead of being failed because of the mua_wrapper setting. This
521 is now fixed, and I tidied up some of the mua_wrapper messages at the
524 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
525 the messages in a hash of arrays rather than using individual hashes.
526 This is a bit cleaner and results in dramatic memory savings, albeit
527 at a slight CPU cost.
529 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
530 as requested by Marc Sherman.
532 SC/10 Eximstats: added histograms for user specified patterns as requested
535 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
537 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
538 fopen() in the content-scanning modules that did not already have it.
544 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
545 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
547 TK/02 Fix ACL "control" statement not being available in MIME ACL.
549 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
551 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
552 to test Sieve filters that use "vacation".
554 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
555 that changes the way the GnuTLS parameters are stored in the cache file.
556 The new format can be generated externally. For backward compatibility,
557 if the data in the cache doesn't make sense, Exim assumes it has read an
558 old-format file, and it generates new data and writes a new file. This
559 means that you can't go back to an older release without removing the
562 PH/03 A redirect router that has both "unseen" and "one_time" set does not
563 work if there are any delivery delays because "one_time" forces the
564 parent to be marked "delivered", so its unseen clone is never tried
565 again. For this reason, Exim now forbids the simultaneous setting of
568 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
569 redirected to themselves ("homonym" addresses). Read the long ChangeLog
570 entry if you want to know the details. The fix, however, neglected to
571 consider the case when local delivery batching is involved. The test for
572 "previously delivered" was not happening when checking to see if an
573 address could be batched with a previous (undelivered) one; under
574 certain circumstances this could lead to multiple deliveries to the same
577 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
578 in its include files, and this causes problems building Exim.
580 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
581 header_syntax) but Exim was just ignoring anything given after a slash.
582 In particular, this caused confusion with an attempt to use "verify =
583 reverse_host_lookup/defer_ok". An error is now given when options are
584 supplied for verify items that do not have them. (Maybe reverse_host_
585 lookup should have a defer_ok option, but that's a different point.)
587 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
588 defined by RFC 821) to 2048, because there were problems with some AUTH
589 commands, and RFC 1869 says the size should be increased for extended
590 SMTP commands that take arguments.
592 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
595 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
596 "unknown" error; now it says that the functionality isn't in the binary.
598 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
599 an address' error message when a string expansion fails (syntax or
600 whatever). Otherwise the password may appear in the log. Following change
601 PH/42 below, there is no longer a chance of it appearing in a bounce
604 PH/11 Installed exipick version 20050225.0 from John Jetmore.
606 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
607 of its addresses was ever tried. (Bugzilla bug #2.)
609 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
610 the result incorrectly in the debug output. (It correctly added a newline
611 to what was transported.)
613 TF/01 Added $received_time.
615 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
616 commented out examples of how to interface to a virus scanner and to
617 SpamAssassin. Also added commented examples of av_scanner and
618 spamd_address settings.
620 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
621 and controls are allowed in which ACLs. There were a couple of minor
622 errors. Some of the entries in the conditions table (which is a table of
623 where they are NOT allowed) were getting very unwieldy; rewrote them as a
624 negation of where the condition IS allowed.
626 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
628 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
629 header file does not have a version number, so I've had to invent a new
630 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
631 API. The code is untested by me (my Linux distribution still has 0.3.2 of
632 radiusclient), but it was contributed by a Radius user.
634 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
635 files or directories, for OpenSSL.
637 PH/19 When an Exim process that is running as root has to create an Exim log
638 file, it does so in a subprocess that runs as exim:exim so as to get the
639 ownership right at creation (otherwise, other Exim processes might see
640 the file with the wrong ownership). There was no test for failure of this
641 fork() call, which would lead to the process getting stuck as it waited
642 for a non-existent subprocess. Forks do occasionally fail when resources
643 run out. I reviewed all the other calls to fork(); they all seem to check
646 PH/20 When checking for unexpected SMTP input at connect time (before writing
647 the banner), Exim was not dealing correctly with a non-positive return
648 from the read() function. If the client had disconnected by this time,
649 the result was a log entry for a synchronization error with an empty
650 string after "input=" when read() returned zero. If read() returned -1
651 (an event I could not check), uninitialized data bytes were printed.
652 There were reports of junk text (parts of files, etc) appearing after
655 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
657 PH/22 Added support for macro redefinition, and (re)definition in between
658 driver and ACL definitions.
660 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
661 forgetting to use the resulting value; it was using the unexpanded value.
663 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
664 hadn't been configured. The fix is from Juergen Kreileder, who
665 understands it better than I do:
667 "Here's what I see happening with three configured cyrus_sasl
668 authenticators configured (plain, login, cram-md5):
670 On startup auth_cyrus_sasl_init() gets called for each of these.
671 This means three calls to sasl_listmech() without a specified mech_list.
672 => SASL tests which mechs of all available mechs actually work
673 => three warnings about OTP not working
674 => the returned list contains: plain, login, cram-md5, digest-md5, ...
676 With the patch, sasl_listmech() also gets called three times. But now
677 SASL's mech_list option is set to the server_mech specified in the the
678 authenticator. Or in other words, the answer from sasl_listmech()
679 gets limited to just the mech you're testing for (which is different
681 => the return list contains just 'plain' or 'login', 'cram-md5' or
682 nothing depending on the value of ob->server_mech.
684 I've just tested the patch: Authentication still works fine,
685 unavailable mechs specified in the exim configuration are still
686 caught, and the auth.log warnings about OTP are gone."
688 PH/25 When debugging is enabled, the contents of the command line are added
689 to the debugging output, even when log_selector=+arguments is not
692 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
693 answer is "GNU", and only if the return is "GNU/something" is the answer
696 PH/27 $acl_verify_message is now set immediately after the failure of a
697 verification in an ACL, and so is available in subsequent modifiers. In
698 particular, the message can be preserved by coding like this:
700 warn !verify = sender
701 set acl_m0 = $acl_verify_message
703 Previously, $acl_verify_message was set only while expanding "message"
704 and "log_message" when a very denied access.
706 PH/28 Modified OS/os.c-Linux with
708 -#ifndef OS_LOAD_AVERAGE
709 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
711 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
712 nomenclature these days.)
714 PH/29 Installed patch from the Sieve maintainer that adds the options
715 sieve_useraddress and sieve_subaddress to the redirect router.
717 PH/30 In these circumstances:
718 . Two addresses routed to the same list of hosts;
719 . First host does not offer TLS;
720 . First host accepts first address;
721 . First host gives temporary error to second address;
722 . Second host offers TLS and a TLS session is established;
723 . Second host accepts second address.
724 Exim incorrectly logged both deliveries with the TLS parameters (cipher
725 and peerdn, if requested) that were in fact used only for the second
728 PH/31 When doing a callout as part of verifying an address, Exim was not paying
729 attention to any local part prefix or suffix that was matched by the
730 router that accepted the address. It now behaves in the same way as it
731 does for delivery: the affixes are removed from the local part unless
732 rcpt_include_affixes is set on the transport.
734 PH/32 Add the sender address, as F=<...>, to the log line when logging a
735 timeout during the DATA phase of an incoming message.
737 PH/33 Sieve envelope tests were broken for match types other than :is. I have
738 applied a patch sanctioned by the Sieve maintainer.
740 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
741 the uid or gid is negative. A case of a negative gid caused this to be
742 noticed. The fix allows for either to be negative.
744 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
745 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
748 PH/36 The change PH/12 above was broken. Fixed it.
750 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
751 the grounds that routing the same address twice would always produce the
752 same answer. This might have been true once, but it is certainly no
753 longer true now. Routing a child address may depend on the previous
754 routing that produced that child. Some complicated redirection strategies
755 went wrong when messages had multiple recipients, and made Exim's
756 behaviour dependent on the order in which the addresses were given.
758 I have moved the duplicate checking until after the routing is complete.
759 Exim scans the addresses that are assigned to local and remote
760 transports, and removes any duplicates. This means that more work will be
761 done, as duplicates will always all be routed, but duplicates are
762 presumably rare, so I don't expect this is of any significance.
764 For deliveries to pipes, files, and autoreplies, the duplicate checking
765 still happens during the routing process, since they are not going to be
768 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
769 It corrects a timeout issue with spamd. This is Ian's comment: "The
770 background is that sometimes spamd either never reads data from a
771 connection it has accepted, or it never writes response data. The exiscan
772 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
773 blindly assumes that writes won't block so it may never time out."
775 PH/39 Allow G after quota size as well as K and M.
777 PH/40 The value set for $authenticated_id in an authenticator may not contain
778 binary zeroes or newlines because the value is written to log lines and
779 to spool files. There was no check on this. Now the value is run through
780 the string_printing() function so that such characters are converted to
781 printable escape sequences.
783 PH/41 $message_linecount is a new variable that contains the total number of
784 lines in the message. Compare $body_linecount, which is the count for the
787 PH/42 Exim no longer gives details of delivery errors for specific addresses in
788 bounce and delay warning messages, except in certain special cases, which
791 (a) An SMTP error message from a remote host;
792 (b) A message specified in a :fail: redirection;
793 (c) A message specified in a "fail" command in a system filter;
794 (d) A message specified in a FAIL return from the queryprogram router;
795 (e) A message specified by the cannot_route_message router option.
797 In these cases only, Exim does include the error details in bounce and
798 warning messages. There are also a few cases where bland messages such
799 as "unrouteable address" or "local delivery error" are given.
801 PH/43 $value is now also set for the "else" part of a ${run expansion.
803 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
804 being worked on, but at least Exim now implements the latest version to
807 PH/45 In a pipe transport, although a timeout while waiting for the pipe
808 process to complete was treated as a delivery failure, a timeout while
809 writing the message to the pipe was logged, but erroneously treated as a
810 successful delivery. Such timeouts include transport filter timeouts. For
811 consistency with the overall process timeout, these timeouts are now
812 treated as errors, giving rise to delivery failures by default. However,
813 there is now a new Boolean option for the pipe transport called
814 timeout_defer, which, if set TRUE, converts the failures into defers for
815 both kinds of timeout. A transport filter timeout is now identified in
818 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
819 systems where "make" and "gmake" are different, calling "gmake" at top
820 level broke things. I've arranged for the value of $(MAKE) to be passed
821 from the Makefile to this script so that it can call the same version of
825 A note about Exim versions 4.44 and 4.50
826 ----------------------------------------
828 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
829 changes of various kinds. As a consequence, a big documentation update was
830 needed. This delayed the release for rather longer than seemed good, especially
831 in the light of a couple of (minor) security issues. Therefore, the changes
832 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
833 release. So 4.44 and 4.50 are in effect two different branches that both start
836 I have left the 4.50 change log unchanged; it contains all the changes since
837 4.43. The change log for 4.44 is below; many of its items are identical to
838 those for 4.50. This seems to be the most sensible way to preserve the
839 historical information.
845 1. Minor wording change to the doc/README.SIEVE file.
847 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
848 computation of the current number of files was incorrect.
850 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
851 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
854 4. Give more explanation in the error message when the command for a transport
855 filter fails to execute.
857 5. There are several places where Exim runs a non-Exim command in a
858 subprocess. The SIGUSR1 signal should be disabled for these processes. This
859 was being done only for the command run by the queryprogram router. It is
860 now done for all such subprocesses. The other cases are: ${run, transport
861 filters, and the commands run by the lmtp and pipe transports.
863 6. Added CONFIGURE_GROUP build-time option.
865 7. Some older OS have a limit of 256 on the maximum number of file
866 descriptors. Exim was using setrlimit() to set 1000 as a large value
867 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
868 systems. I've change it so that if it can't get 1000, it tries for 256.
870 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
871 was an oversight, and furthermore, ever since the addition of extra
872 controls (e.g. 4.43/32), the checks on when to allow different forms of
873 "control" were broken. There should now be diagnostics for all cases when a
874 control that does not make sense is encountered.
876 9. Added the /retain_sender option to "control=submission".
878 10. $recipients is now available in the predata ACL (oversight).
880 11. Tidy the search cache before the fork to do a delivery from a message
881 received from the command line. Otherwise the child will trigger a lookup
882 failure and thereby defer the delivery if it tries to use (for example) a
883 cached ldap connection that the parent has called unbind on.
885 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
886 of $address_data from the recipient verification was clobbered by the
889 13. The value of address_data from a sender verification is now available in
890 $sender_address_data in subsequent conditions in the ACL statement.
892 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
894 15. Added a new option "connect=<time>" to callout options, to set a different
897 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
898 was its contents. (It was OK if the option was not defined at all.)
900 17. A "Completed" log line is now written for messages that are removed from
901 the spool by the -Mrm option.
903 18. New variables $sender_verify_failure and $recipient_verify_failure contain
904 information about exactly what failed.
906 19. Added -dd to debug only the daemon process.
908 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
909 handles timeouts, both on the server side and network timeouts. Renamed the
910 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
912 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
913 It is now set to "smtps".
915 22. $host_address is now set to the target address during the checking of
918 23. When checking ignore_target_hosts for an ipliteral router, no host name was
919 being passed; this would have caused $sender_host_name to have been used if
920 matching the list had actually called for a host name (not very likely,
921 since this list is usually IP addresses). A host name is now passed as
924 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
925 code that specifies a non-restarting handler (typically sigaction() in
926 modern systems) in an attempt to fix a rare and obscure crash bug.
928 25. Narrowed the window for a race in the daemon that could cause it to ignore
929 SIGCHLD signals. This is not a major problem, because they are used only to
930 wake it up if nothing else does.
932 26. A malformed maildirsize file could cause Exim to calculate negative values
933 for the mailbox size or file count. Odd effects could occur as a result.
934 The maildirsize information is now recalculated if the size or filecount
937 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
938 support for a long time. Removed HAVE_SYS_VFS_H.
940 28. Installed the latest version of exipick from John Jetmore.
942 29. In an address list, if the pattern was not a regular expression, an empty
943 subject address (from a bounce message) matched only if the pattern was an
944 empty string. Non-empty patterns were not even tested. This was the wrong
945 because it is perfectly reasonable to use an empty address as part of a
946 database query. An empty address is now tested by patterns that are
947 lookups. However, all the other forms of pattern expect the subject to
948 contain a local part and a domain, and therefore, for them, an empty
949 address still always fails if the pattern is not itself empty.
951 30. Exim went into a mad DNS loop when attempting to do a callout where the
952 host was specified on an smtp transport, and looking it up yielded more
955 31. Re-factored the code for checking spool and log partition space into a
956 function that finds that data and another that does the check. The former
957 is then used to implement four new variables: $spool_space, $log_space,
958 $spool_inodes, and $log_inodes.
960 32. The RFC2047 encoding function was originally intended for short strings
961 such as real names; it was not keeping to the 75-character limit for
962 encoded words that the RFC imposes. It now respects the limit, and
963 generates multiple encoded words if necessary. To be on the safe side, I
964 have increased the buffer size for the ${rfc2047: expansion operator from
967 33. It is now permitted to omit both strings after an "if" condition; if the
968 condition is true, the result is "true". As before, when the second string
969 is omitted, a false condition yields an empty string. This makes it less
970 cumbersome to write custom ACL and router conditions.
972 34. Failure to deliver a bounce message always caused it to be frozen, even if
973 there was an errors_to setting on the router. The errors_to setting is now
976 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
977 canonical form (fully expanded) before being placed in
978 $sender_host_address.
980 36. The table in the code that translates DNS record types into text (T_A to
981 "A" for instance) was missing entries for NS and CNAME. It is just possible
982 that this could have caused confusion if both these types were looked up
983 for the same domain, because the text type is used as part of Exim's
984 per-process caching. But the chance of anyone hitting this buglet seems
987 37. The dnsdb lookup has been extended in a number of ways.
989 (1) There is a new type, "zns", which walks up the domain tree until it
990 finds some nameserver records. It should be used with care.
992 (2) There is a new type, "mxh", which is like "mx" except that it returns
993 just the host names, not the priorities.
995 (3) It is now possible to give a list of domains (or IP addresses) to be
996 looked up. The behaviour when one of the lookups defers can be
997 controlled by a keyword.
999 (4) It is now possible to specify the separator character for use when
1000 multiple records are returned.
1002 38. The dnslists ACL condition has been extended: it is now possible to supply
1003 a list of IP addresses and/or domains to be looked up in a particular DNS
1006 39. Added log_selector=+queue_time_overall.
1008 40. When running the queue in the test harness, wait just a tad after forking a
1009 delivery process, to get repeatability of debugging output.
1011 41. Include certificate and key file names in error message when GnuTLS fails
1012 to set them up, because the GnuTLS error message doesn't include the name
1013 of the failing file when there is a problem reading it.
1015 42. Allow both -bf and -bF in the same test run.
1017 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
1019 44. The "Exiscan patch" is now merged into the mainline Exim source.
1021 45. Sometimes the final signoff response after QUIT could fail to get
1022 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1023 < 0 before doing a fflush(). This bug looks as though it goes back to the
1024 introduction of TLS in release 3.20, but "sometimes" must have been rare
1025 because the tests only now provoked it.
1027 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
1028 (this can affect the format of dates).
1030 47. exim_tidydb, when checking for the continued existence of a message for
1031 which it has found a message-specific retry record, was not finding
1032 messages that were in split spool directories. Consequently, it was
1033 deleting retry records that should have stayed in existence.
1035 48. Steve fixed some bugs in eximstats.
1037 49. The SPA authentication driver was not abandoning authentication and moving
1038 on to the next authenticator when an expansion was forced to fail,
1039 contradicting the general specification for all authenticators. Instead it
1040 was generating a temporary error. It now behaves as specified.
1042 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
1043 (the order specifies the preference for clients). The order is now AES256,
1044 AES128, 3DES, ARCFOUR128.
1046 51. Small patch to Sieve code - explicitly set From: when generating an
1049 52. Exim crashed if a remote delivery caused a very long error message to be
1050 recorded - for instance if somebody sent an entire SpamAssassin report back
1051 as a large number of 550 error lines. This bug was coincidentally fixed by
1052 increasing the size of one of Exim's internal buffers (big_buffer) that
1053 happened as part of the Exiscan merge. However, to be on the safe side, I
1054 have made the code more robust (and fixed the comments that describe what
1057 53. Now that there can be additional text after "Completed" in log lines (if
1058 the queue_time_overall log selector is set), a one-byte patch to exigrep
1059 was needed to allow it to recognize "Completed" as not the last thing in
1062 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1063 patch that reportedly fixes this has been added. I am not expert enough to
1064 create a test for it. This is what the patch creator wrote:
1066 "I found a little strange behaviour of ldap code when working with
1067 Windows 2003 AD Domain, where users was placed in more than one
1068 Organization Units. When I tried to give exim partial DN, the exit code
1069 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1070 But simultaneously result of request was absolutely normal ldap result,
1071 so I produce this patch..."
1073 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1074 REFERENCE, so I have modified the code to exclude the patch when that macro
1077 55. Some experimental protocols are using DNS PTR records for new purposes. The
1078 keys for these records are domain names, not reversed IP addresses. The
1079 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1080 leaves it alone. Component reversal etc. now happens only for IP addresses.
1083 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1085 57. Double the size of the debug message buffer (to 2048) so that more of very
1086 long debug lines gets shown.
1088 58. The exicyclog utility now does better if the number of log files to keep
1089 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1091 59. Two changes related to the smtp_active_hostname option:
1093 (1) $smtp_active_hostname is now available as a variable.
1094 (2) The default for smtp_banner uses $smtp_active_hostname instead
1095 of $primary_hostname.
1097 60. The host_aton() function is supposed to be passed a string that is known
1098 to be a valid IP address. However, in the case of IPv6 addresses, it was
1099 not checking this. This is a hostage to fortune. Exim now panics and dies
1100 if the condition is not met. A case was found where this could be provoked
1101 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1102 components; fortuitously, this particular loophole had already been fixed
1103 by change 4.50/55 above.
1105 If there are any other similar loopholes, the new check in host_aton()
1106 itself should stop them being exploited. The report I received stated that
1107 data on the command line could provoke the exploit when Exim was running as
1108 exim, but did not say which command line option was involved. All I could
1109 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1110 running as the user.
1113 61. There was a buffer overflow vulnerability in the SPA authentication code
1114 (which came originally from the Samba project). I have added a test to the
1115 spa_base64_to_bits() function which I hope fixes it.
1118 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1119 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1121 63. The daemon start-up calls getloadavg() while still root for those OS that
1122 need the first call to be done as root, but it missed one case: when
1123 deliver_queue_load_max is set with deliver_drop_privilege. This is
1124 necessary for the benefit of the queue runner, because there is no re-exec
1125 when deliver_drop_privilege is set.
1127 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1128 This has been fixed.
1130 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1131 was in use, was not putting the data itself into the right store pool;
1132 consequently, it could be overwritten for a subsequent message in the same
1133 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1136 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1138 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1139 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1140 the function were treating the return as a boolean value, which happened to
1141 work because 0=false and not-0=true, but is not correct code.
1143 68. The host_aton() function was not handling scoped IPv6 addresses (those
1144 with, for example, "%eth0" on the end) correctly.
1146 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1147 negated items (that is, ~something) in unsigned ints. Some compilers
1148 apparently mutter when there is no cast.
1150 70. If an address verification called from an ACL failed, and did not produce a
1151 user-specific message (i.e. there was only a "system" message), nothing was
1152 put in $acl_verify_message. In this situation, it now puts the system
1155 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1156 session; change 4.31/43 added the unwanted input to the log line - except
1157 that it did not do this in the start of session case. It now does.
1159 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1160 This is wrong and can cause the other end to generate a synchronization
1161 error if it is another Exim or anything else that does the synchronization
1162 check. A QUIT command is no longer sent after a timeout.
1164 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1165 during host lookups.
1167 74. The defer_ok option of callout verification was not working if it was used
1168 when verifying addresses in header lines, that is, for this case:
1170 verify = header_sender/callout=defer_ok
1172 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1173 those file descriptors could be used for SMTP connections. If anything
1174 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1175 could be sent to the SMTP client, causing chaos. The daemon now opens
1176 stdin, stdout, and stderr to /dev/null when it puts itself into the
1179 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1180 log by default. The user can override this with suitable Perl magic.
1182 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1183 the log message when discard triggers, was not working for the DATA ACL or
1184 for the non-SMTP ACL.
1186 78. Error message wording change in sieve.c.
1188 79. If smtp_accept_max_per_host was set, the number of connections could be
1189 restricted to fewer than expected, because the daemon was trying to set up
1190 a new connection before checking whether the processes handling previous
1191 connections had finished. The check for completed processes is now done
1192 earlier. On busy systems, this bug wouldn't be noticed because something
1193 else would have woken the daemon, and it would have reaped the completed
1196 80. If a message was submitted locally by a user whose login name contained one
1197 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1198 It caused a spool format error. I have fixed the spool reading code. A
1199 related problem was that the "from" clause in the Received: line became
1200 illegal because of the space(s). It is now covered by ${quote_local_part.
1202 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1205 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1207 83. Patch from Sieve maintainer to fix unterminated string problem in
1208 "vacation" handling.
1210 84. Some minor changes to the Linux configuration files to help with other
1211 OS variants using glibc.
1213 85. One more patch for Sieve to update vacation handling to latest spec.
1216 ----------------------------------------------------
1217 See the note above about the 4.44 and 4.50 releases.
1218 ----------------------------------------------------
1224 1. Change 4.43/35 introduced a bug that caused file counts to be
1225 incorrectly computed when quota_filecount was set in an appendfile
1228 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1229 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1232 3. Give more explanation in the error message when the command for a transport
1233 filter fails to execute.
1235 4. There are several places where Exim runs a non-Exim command in a
1236 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1237 was being done only for the command run by the queryprogram router. It is
1238 now done for all such subprocesses. The other cases are: ${run, transport
1239 filters, and the commands run by the lmtp and pipe transports.
1241 5. Some older OS have a limit of 256 on the maximum number of file
1242 descriptors. Exim was using setrlimit() to set 1000 as a large value
1243 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1244 systems. I've change it so that if it can't get 1000, it tries for 256.
1246 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1247 was an oversight, and furthermore, ever since the addition of extra
1248 controls (e.g. 4.43/32), the checks on when to allow different forms of
1249 "control" were broken. There should now be diagnostics for all cases when a
1250 control that does not make sense is encountered.
1252 7. $recipients is now available in the predata ACL (oversight).
1254 8. Tidy the search cache before the fork to do a delivery from a message
1255 received from the command line. Otherwise the child will trigger a lookup
1256 failure and thereby defer the delivery if it tries to use (for example) a
1257 cached ldap connection that the parent has called unbind on.
1259 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1260 of $address_data from the recipient verification was clobbered by the
1261 sender verification.
1263 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1264 was its contents. (It was OK if the option was not defined at all.)
1266 11. A "Completed" log line is now written for messages that are removed from
1267 the spool by the -Mrm option.
1269 12. $host_address is now set to the target address during the checking of
1270 ignore_target_hosts.
1272 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1273 being passed; this would have caused $sender_host_name to have been used if
1274 matching the list had actually called for a host name (not very likely,
1275 since this list is usually IP addresses). A host name is now passed as
1278 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1279 code that specifies a non-restarting handler (typically sigaction() in
1280 modern systems) in an attempt to fix a rare and obscure crash bug.
1282 15. Narrowed the window for a race in the daemon that could cause it to ignore
1283 SIGCHLD signals. This is not a major problem, because they are used only to
1284 wake it up if nothing else does.
1286 16. A malformed maildirsize file could cause Exim to calculate negative values
1287 for the mailbox size or file count. Odd effects could occur as a result.
1288 The maildirsize information is now recalculated if the size or filecount
1291 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1292 support for a long time. Removed HAVE_SYS_VFS_H.
1294 18. Updated exipick to current release from John Jetmore.
1296 19. Allow an empty sender to be matched against a lookup in an address list.
1297 Previously the only cases considered were a regular expression, or an
1300 20. Exim went into a mad DNS lookup loop when doing a callout where the
1301 host was specified on the transport, if the DNS lookup yielded more than
1304 21. The RFC2047 encoding function was originally intended for short strings
1305 such as real names; it was not keeping to the 75-character limit for
1306 encoded words that the RFC imposes. It now respects the limit, and
1307 generates multiple encoded words if necessary. To be on the safe side, I
1308 have increased the buffer size for the ${rfc2047: expansion operator from
1311 22. Failure to deliver a bounce message always caused it to be frozen, even if
1312 there was an errors_to setting on the router. The errors_to setting is now
1315 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1316 canonical form (fully expanded) before being placed in
1317 $sender_host_address.
1319 24. Updated eximstats to version 1.33
1321 25. Include certificate and key file names in error message when GnuTLS fails
1322 to set them up, because the GnuTLS error message doesn't include the name
1323 of the failing file when there is a problem reading it.
1325 26. Expand error message when OpenSSL has problems setting up cert/key files.
1328 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1329 (this can affect the format of dates).
1331 28. exim_tidydb, when checking for the continued existence of a message for
1332 which it has found a message-specific retry record, was not finding
1333 messages that were in split spool directories. Consequently, it was
1334 deleting retry records that should have stayed in existence.
1336 29. eximstats updated to version 1.35
1337 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1338 1.35 - bugfix such that pie charts by volume are generated correctly
1340 30. The SPA authentication driver was not abandoning authentication and moving
1341 on to the next authenticator when an expansion was forced to fail,
1342 contradicting the general specification for all authenticators. Instead it
1343 was generating a temporary error. It now behaves as specified.
1345 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1346 (the order specifies the preference for clients). The order is now AES256,
1347 AES128, 3DES, ARCFOUR128.
1349 31. Small patch to Sieve code - explicitly set From: when generating an
1352 32. Exim crashed if a remote delivery caused a very long error message to be
1353 recorded - for instance if somebody sent an entire SpamAssassin report back
1354 as a large number of 550 error lines. This bug was coincidentally fixed by
1355 increasing the size of one of Exim's internal buffers (big_buffer) that
1356 happened as part of the Exiscan merge. However, to be on the safe side, I
1357 have made the code more robust (and fixed the comments that describe what
1360 33. Some experimental protocols are using DNS PTR records for new purposes. The
1361 keys for these records are domain names, not reversed IP addresses. The
1362 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1363 leaves it alone. Component reversal etc. now happens only for IP addresses.
1366 34. The host_aton() function is supposed to be passed a string that is known
1367 to be a valid IP address. However, in the case of IPv6 addresses, it was
1368 not checking this. This is a hostage to fortune. Exim now panics and dies
1369 if the condition is not met. A case was found where this could be provoked
1370 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1371 components; fortuitously, this particular loophole had already been fixed
1372 by change 4.50/55 or 4.44/33 above.
1374 If there are any other similar loopholes, the new check in host_aton()
1375 itself should stop them being exploited. The report I received stated that
1376 data on the command line could provoke the exploit when Exim was running as
1377 exim, but did not say which command line option was involved. All I could
1378 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1379 running as the user.
1382 35. There was a buffer overflow vulnerability in the SPA authentication code
1383 (which came originally from the Samba project). I have added a test to the
1384 spa_base64_to_bits() function which I hope fixes it.
1387 36. The daemon start-up calls getloadavg() while still root for those OS that
1388 need the first call to be done as root, but it missed one case: when
1389 deliver_queue_load_max is set with deliver_drop_privilege. This is
1390 necessary for the benefit of the queue runner, because there is no re-exec
1391 when deliver_drop_privilege is set.
1393 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1394 was in use, was not putting the data itself into the right store pool;
1395 consequently, it could be overwritten for a subsequent message in the same
1396 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1399 38. Sometimes the final signoff response after QUIT could fail to get
1400 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1401 < 0 before doing a fflush(). This bug looks as though it goes back to the
1402 introduction of TLS in release 3.20, but "sometimes" must have been rare
1403 because the tests only now provoked it.
1409 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1410 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1411 Now it returns an integer. A number of calls were still expecting a T/F
1412 return. Fortuitously, in all cases, the tests worked in OK situations,
1413 which is the norm. However, things would have gone wrong on any write
1414 failures on the smtp file descriptor. This function is used when sending
1415 messages over SMTP and also when doing verify callouts.
1417 2. When Exim is called to do synchronous delivery of a locally submitted
1418 message (the -odf or -odi options), it no longer closes stderr before doing
1421 3. Implemented the mua_wrapper option.
1423 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1425 5. Implemented the functions header_remove(), header_testname(),
1426 header_add_at_position(), and receive_remove_recipient(), and exported them
1429 6. If an ACL "warn" statement specified the addition of headers, Exim already
1430 inserted X-ACL-Warn: at the start if there was no header name. However, it
1431 was not making this test for the second and subsequent header lines if
1432 there were newlines in the string. This meant that an invalid header could
1433 be inserted if Exim was badly configured.
1435 7. Allow an ACL "warn" statement to add header lines at the start or after all
1436 the Received: headers, as well as at the end.
1438 8. Added the rcpt_4xx retry error code.
1440 9. Added postmaster_mailfrom=xxx to callout verification option.
1442 10. Added mailfrom=xxxx to the callout verification option, for verify=
1445 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1446 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1448 12. Inserted some casts to stop certain compilers warning when using pointer
1449 differences as field lengths or precisions in printf-type calls (mostly
1450 affecting debugging statements).
1452 13. Added optional readline() support for -be (dynamically loaded).
1454 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1455 same clock tick as a message's arrival, so that its received time was the
1456 same as the "first fail" time on the retry record, and that message
1457 remained on the queue past the ultimate address timeout, every queue runner
1458 would try a delivery (because it was past the ultimate address timeout) but
1459 after another failure, the ultimate address timeout, which should have then
1460 bounced the address, did not kick in. This was a "< instead of <=" error;
1461 in most cases the first failure would have been in the next clock tick
1462 after the received time, and all would be well.
1464 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1465 being recognized when the domain list was tested by the match_domain
1466 condition in an expansion string.
1468 16. Added the ${str2b64: operator.
1470 17. Exim was always calling setrlimit() to set a large limit for the number of
1471 processes, without checking whether the existing limit was already
1472 adequate. (It did check for the limit on file descriptors.) Furthermore,
1473 errors from getrlimit() and setrlimit() were being ignored. Now they are
1474 logged to the main and panic logs, but Exim does carry on, to try to do its
1475 job under whatever limits there are.
1477 18. Imported PCRE 5.0.
1479 19. Trivial typo in log message " temporarily refused connection" (the leading
1482 20. If the log selector return_path_on_delivery was set and an address was
1483 redirected to /dev/null, the delivery process crashed because it assumed
1484 that a return path would always be set for a "successful" delivery. In this
1485 case, the whole delivery is bypassed as an optimization, and therefore no
1488 21. Internal re-arrangement: the function for sending a challenge and reading
1489 a response while authentication was assuming a zero-terminated challenge
1490 string. It's now changed to take a pointer and a length, to allow for
1491 binary data in such strings.
1493 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1495 23. Exim was not respecting finduser_retries when seeking the login of the
1496 uid under which it was called; it was always trying 10 times. (The default
1497 setting of finduser_retries is zero.) Also, it was sleeping after the final
1498 failure, which is pointless.
1500 24. Implemented tls_on_connect_ports.
1502 25. Implemented acl_smtp_predata.
1504 26. If the domain in control=submission is set empty, Exim assumes that the
1505 authenticated id is a complete email address when it generates From: or
1506 Sender: header lines.
1508 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1509 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1510 chown and chgrp in /bin and hostname in /usr/bin.
1512 28. Exim was keeping the "process log" file open after each use, just as it
1513 does for the main log. This opens the possibility of it remaining open for
1514 long periods when the USR1 signal hits a daemon. Occasional processlog
1515 errors were reported, that could have been caused by this. Anyway, it seems
1516 much more sensible not to leave this file open at all, so that is what now
1519 29. The long-running daemon process does not normally write to the log once it
1520 has entered its main loop, and it closes the log before doing so. This is
1521 so that log files can straightforwardly be renamed and moved. However,
1522 there are a couple of unusual error situations where the daemon does write
1523 log entries, and I had neglected to close the log afterwards.
1525 30. The text of an SMTP error response that was received during a remote
1526 delivery was being truncated at 512 bytes. This is too short for some of
1527 the long messages that one sometimes sees. I've increased the limit to
1530 31. It is now possible to make retry rules that apply only when a message has a
1531 specific sender, in particular, an empty sender.
1533 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1534 it possible to be selective about when SMTP synchronization is enforced.
1536 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1538 32. Implemented hosts_connection_nolog.
1540 33. Added an ACL for QUIT.
1542 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1545 35. Added mailbox_size and mailbox_filecount to appendfile.
1547 36. Added control = no_multiline_responses to ACLs.
1549 37. There was a bug in the logic of the code that waits for the clock to tick
1550 in the case where the clock went backwards by a substantial amount such
1551 that the microsecond fraction of "now" was more than the microsecond
1552 fraction of "then" (but the whole seconds number was less).
1554 38. Added support for the libradius Radius client library this is found on
1555 FreeBSD (previously only the radiusclient library was supported).
1561 1. When certain lookups returned multiple values in the form name=value, the
1562 quoting of the values was not always being done properly. Specifically:
1563 (a) If the value started with a double quote, but contained no whitespace,
1565 (b) If the value contained whitespace other than a space character (i.e.
1566 tabs or newlines or carriage returns) it was not quoted.
1567 This fix has been applied to the mysql and pgsql lookups by writing a
1568 separate quoting function and calling it from the lookup code. The fix
1569 should probably also be applied to nisplus, ibase and oracle lookups, but
1570 since I cannot test any of those, I have not disturbed their existing code.
1572 2. A hit in the callout cache for a specific address caused a log line with no
1573 reason for rejecting RCPT. Now it says "Previous (cached) callout
1574 verification failure".
1576 3. There was an off-by-one bug in the queryprogram router. An over-long
1577 return line was truncated at 256 instead of 255 characters, thereby
1578 overflowing its buffer with the terminating zero. As well as fixing this, I
1579 have increased the buffer size to 1024 (and made a note to document this).
1581 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1582 when Exim is waiting for an SMTP response from a remote server, Exim
1583 restarts its select() call on the socket, thereby resetting its timeout.
1584 This is not a problem when such interrupts are rare. Somebody set up a cron
1585 job to run exiwhat every 2 minutes, which is less than the normal select()
1586 timeout (5 or 10 minutes). This meant that the select() timeout never
1587 kicked in because it was always reset. I have fixed this by comparing the
1588 time when an interrupt arrives with the time at the start of the first call
1589 to select(). If more time than the timeout has elapsed, the interrupt is
1590 treated as a timeout.
1592 5. Some internal re-factoring in preparation for the addition of Sieve
1593 extensions (by MH). In particular, the "personal" test is moved to a
1594 separate function, and given an option for scanning Cc: and Bcc: (which is
1595 not set for Exim filters).
1597 6. When Exim created an email address using the login of the caller as the
1598 local part (e.g. when creating a From: or Sender: header line), it was not
1599 quoting the local part when it contained special characters such as @.
1601 7. Installed new OpenBSD configuration files.
1603 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1604 try to make them clearer.
1606 9. Callout options, other than the timeout value, were being ignored when
1607 verifying sender addresses in header lines. For example, when using
1609 verify = header_sender/callout=no_cache
1611 the cache was (incorrectly) being used.
1613 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1614 only the Cygwin environment.
1616 11. When return_path_on_delivery was set as a log selector, if different remote
1617 addresses in the same message used different return paths and parallel
1618 remote delivery occurred, the wrong values would sometimes be logged.
1619 (Whenever a remote delivery process finished, the return path value from
1620 the most recently started remote delivery process was logged.)
1622 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1623 lines when AUTH and/or TLS are in use. This is the "received protocol"
1624 field. Exim used to use "asmtp" for authenticated SMTP, without any
1625 indication (in the protocol name) for TLS use. Now it follows the RFC and
1626 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1627 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1628 names appear in log lines as well as in Received: header lines.
1630 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1631 extensions, and comparison tests, and to fix some bugs.
1633 14. Changes to the "personal" filter test:
1635 (1) The test was buggy in that it was just doing the equivalent of
1636 "contains" tests on header lines. For example, if a user's address was
1637 anne@some.where, the "personal" test would incorrectly be true for
1639 To: susanne@some.where
1641 This test is now done by extracting each address from the header in turn,
1642 and checking the entire address. Other tests that are part of "personal"
1643 are now done using regular expressions (for example, to check local parts
1644 of addresses in From: header lines).
1646 (2) The list of non-personal local parts in From: addresses has been
1647 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1648 taken from the Sieve specification recommendations.
1650 (3) If the message contains any header line starting with "List-" it is
1651 treated as non-personal.
1653 (4) The test for "circular" in the Subject: header line has been removed
1654 because it now seems ill-conceived.
1656 15. Minor typos in src/EDITME comments corrected.
1658 16. Installed latest exipick from John Jetmore.
1660 17. If headers_add on a router specified a text string that was too long for
1661 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1662 of string_sprintf() is now avoided.
1664 18. $message_body_size was not set (it was always zero) when running the DATA
1665 ACL and the local_scan() function.
1667 19. For the "mail" command in an Exim filter, no default was being set for
1668 the once_repeat time, causing a random time value to be used if "once" was
1669 specified. (If the value happened to be <= 0, no repeat happened.) The
1670 default is now 0s, meaning "never repeat". The "vacation" command was OK
1671 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1672 (I found it when inspecting the code).
1674 20. There is now an overall timeout for performing a callout verification. It
1675 defaults to 4 times the callout timeout, which applies to individual SMTP
1676 commands during the callout. The overall timeout applies when there is more
1677 than one host that can be tried. The timeout is checked before trying the
1678 next host. This prevents very long delays if there are a large number of
1679 hosts and all are timing out (e.g. when the network connections are timing
1680 out). The value of the overall timeout can be changed by specifying an
1681 additional sub-option for "callout", called "maxwait". For example:
1683 verify = sender/callout=5s,maxwait=20s
1685 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1686 to the end before writing, but this should make it even safer).
1688 22. Exim was forgetting that it had advertised PIPELINING for the second and
1689 subsequent messages on an SMTP connection. It was also not resetting its
1690 memory on STARTTLS and an internal HELO.
1692 23. When Exim logs an SMTP synchronization error within a session, it now
1693 records whether PIPELINING has been advertised or not.
1695 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1696 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1697 rather than long int.
1699 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1701 26. Added the never_mail option to autoreply.
1707 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1708 crash if the getsockname() call failed; this can happen if a connection is
1709 closed very soon after it is established. The problem was simply in the
1710 order in which certain operations were done, causing Exim to try to write
1711 to the SMTP stream before it had set up the file descriptor. The bug has
1712 been fixed by making things happen in the correct order.
1718 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1719 before the connection was closed, thus losing the rejection response.
1721 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1722 some early Solaris releases, but causes trouble in current releases where
1723 socklen_t is defined.
1725 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1728 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1731 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1732 that releasing the top of it at the end releases what was used for sub-
1733 expansions (unless the block got too big). However, discard this block if
1734 the first thing is a variable or header, so that we can use its block when
1735 it is dynamic (useful for very large $message_headers, for example).
1737 6. Lookups now cache *every* query, not just the most recent. A new, separate
1738 store pool is used for this. It can be recovered when all lookup caches are
1739 flushed. Lookups now release memory at the end of their result strings.
1740 This has involved some general refactoring of the lookup sources.
1742 7. Some code has been added to the store_xxx() functions to reduce the amount
1743 of flapping under certain conditions.
1745 8. log_incoming_interface used to affect only the <= reception log lines. Now
1746 it causes the local interface and port to be added to several more SMTP log
1747 lines, for example "SMTP connection from", and rejection lines.
1749 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1751 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1753 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1754 could be overwritten at the end of the current message (or the start of a
1755 new message if it was set in a HELO ACL). The value is now preserved for
1756 the duration of the SMTP connection.
1758 12. If a transport had a headers_rewrite setting, and a matching header line
1759 contained an unqualified address, that address was qualified, even if it
1760 did not match any rewriting rules. The underlying bug was that the values
1761 of the flags that permit the existence of unqualified sender and recipient
1762 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1763 non-local messages, and by -bnq for local messages) were not being
1764 preserved with the message after it was received.
1766 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1767 "next input=" as part of the text comprising the host identity instead of
1768 the correct text. The code was using the same buffer for two different
1769 strings. However, depending on which order the printing function evaluated
1770 its arguments, the bug did not always show up. Under Linux, for example, my
1771 test suite worked just fine.
1773 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1774 doesn't work with some older versions of Perl. It has been changed to "my",
1775 which in any case is probably the better facility to use.
1777 15. A really picky compiler found some instances of statements for creating
1778 error messages that either had too many or two few arguments for the format
1781 16. The size of the buffer for calls to the DNS resolver has been increased
1782 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1783 for addresses that have a lot of PTR records. This alleviates a problem; it
1784 does not fully solve it.
1786 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1787 buffer now truncates the list and logs the incident, which is the same
1788 action as happens when Exim is looking up a host name and its aliases.
1789 Previously in this situation something unpredictable would happen;
1790 sometimes it was "internal error: store_reset failed".
1792 18. If a server dropped the connection unexpectedly when an Exim client was
1793 using GnuTLS and trying to read a response, the client delivery process
1794 crashed while trying to generate an error log message.
1796 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1797 string, for example:
1799 warn message = H1: something\nH2: something
1801 the text was added as a single header line from Exim's point of view
1802 though it ended up OK in the delivered message. However, searching for the
1803 second and subsequent header lines using $h_h2: did not work. This has been
1804 fixed. Similarly, if a system filter added multiple headers in this way,
1805 the routers could not see them.
1807 20. Expanded the error message when iplsearch is called with an invalid key to
1808 suggest using net-iplsearch in a host list.
1810 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1811 ACLs are no longer actually imposed (and a message to that effect is
1814 22. If a "gecos" field in a passwd entry contained escaped characters, in
1815 particular, if it contained a \" sequence, Exim got it wrong when building
1816 a From: or a Sender: header from that name. A second bug also caused
1817 incorrect handling when an unquoted " was present following a character
1818 that needed quoting.
1820 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1821 was not being matched caselessly.
1823 24. Arranged for all hyphens in the exim.8 source to be escaped with
1826 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1827 properly. Recipient callout cache records were still being keyed to include
1828 the sender, even when use_sender was set false. This led to far more
1829 callouts that were necessary. The sender is no longer included in the key
1830 when use_sender is false.
1832 26. Added "control = submission" modifier to ACLs.
1834 27. Added the ${base62d: operator to decode base 62 numbers.
1836 28. dnsdb lookups can now access SRV records.
1838 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1839 the configuration file.
1841 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1842 (-v) mode. This makes the output for a verbose queue run more intelligible.
1844 31. Added a use_postmaster feature to recipient callouts.
1846 32. Added the $body_zerocount variable, containing the number of binary zero
1847 bytes in the message body.
1849 33. The time of last modification of the "new" subdirectory is now used as the
1850 "mailbox time last read" when there is a quota error for a maildir
1853 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1855 35. Added +ignore_unknown as a special item in host lists.
1857 36. Code for decoding IPv6 addresses in host lists is now included, even if
1858 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1859 address was recognized as an IP address, but was then not correctly decoded
1860 into binary, causing unexpected and incorrect effects when compared with
1867 1. Very minor rewording of debugging text in manualroute to say "list of
1868 hosts" instead of "hostlist".
1870 2. If verify=header_syntax was set, and a header line with an unqualified
1871 address (no domain) and a large number of spaces between the end of the
1872 name and the colon was received, the reception process suffered a buffer
1873 overflow, and (when I tested it) crashed. This was caused by some obsolete
1874 code that should have been removed. The fix is to remove it!
1876 3. When running in the test harness, delay a bit after writing a bounce
1877 message to get a bit more predictability in the log output.
1879 4. Added a call to search_tidyup() just before forking a reception process. In
1880 theory, someone could use a lookup in the expansion of smtp_accept_max_
1881 per_host which, without the tidyup, could leave open a database connection.
1883 5. Added the variables $recipient_data and $sender_data which get set from a
1884 lookup success in an ACL "recipients" or "senders" condition, or a router
1885 "senders" option, similar to $domain_data and $local_part_data.
1887 6. Moved the writing of debug_print from before to after the "senders" test
1890 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1891 problems for message scanning, either using a data ACL, or using
1892 local_scan() because the Received: header was not generated till after they
1893 were called (in order to set the time as the time of reception completion).
1894 I have revised the way this works. The header is now generated after the
1895 body is received, but before the ACL or local_scan() are called. After they
1896 are run, the timestamp in the header is updated.
1902 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1903 before starting a queue runner without re-exec. This happened only when
1904 deliver_drop_privilege was set or when the Exim user was set to root. The
1905 effect of the bug was that timeouts during subsequent deliveries caused
1906 crashes instead of being properly handled. The handler is now left at its
1907 default (and expected) setting.
1909 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1910 message, again when deliver_drop_privilege is set or Exim is run as root.
1911 The bug described in (1) was not present in this case, but the tidying up
1912 of the other signals was missing. I have made the two cases consistent.
1914 3. The ignore_target_hosts setting on a manualroute router was being ignored
1915 for hosts that were looked up using the /MX notation.
1917 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1920 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1921 operated on the sender address. After changing the $sender_address to <>
1922 for the sender address verify, Exim was re-instated it as the original
1923 (before rewriting) address, but remembering that it had rewritten it, so it
1924 wasn't rewriting it again. This bug also had the effect of breaking the
1925 sender address verification caching when the sender address was rewritten.
1927 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1928 This has been changed so that if the ip literal address matches
1929 ignore_target_hosts, the router declines.
1931 7. Added expansion conditions match_domain, match_address, and match_local_
1932 part (NOT match_host).
1934 8. The placeholder for the Received: header didn't have a length field set.
1936 9. Added code to Exim itself and to exim_lock to test for a specific race
1937 condition that could lead to file corruption when using MBX delivery. The
1938 issue is with the lockfile that is created in /tmp. If this file is removed
1939 after a process has opened it but before that process has acquired a lock,
1940 there is the potential for a second process to recreate the file and also
1941 acquire a lock. This could lead to two Exim processes writing to the file
1942 at the same time. The added code performs the same test as UW imapd; it
1943 checks after acquiring the lock that its file descriptor still refers to
1944 the same named file.
1946 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1947 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1950 11. Added the smtp_active_hostname option. If used, this will typically be made
1951 to depend on the incoming interface address. Because $interface_address is
1952 not set up until the daemon has forked a reception process, error responses
1953 that can happen earlier (such as "too many connections") no longer contain
1956 12. If an expansion in a condition on a "warn" statement fails because a lookup
1957 defers, the "warn" statement is abandoned, and the next ACL statement is
1958 processed. Previously this caused the whole ACL to be aborted.
1960 13. Added the iplsearch lookup type.
1962 14. Added ident_timeout as a log selector.
1964 15. Added tls_certificate_verified as a log selector.
1966 16. Added a global option tls_require_ciphers (compare the smtp transport
1967 option of the same name). This controls incoming TLS connections.
1969 17. I finally figured out how to make tls_require_ciphers do a similar thing
1970 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1971 before starting the TLS session.
1973 18. Tabs are now shown as \t in -bP output.
1975 19. If the log selector return_path_on_delivery was set, Exim crashed when
1976 bouncing a message because it had too many Received: header lines.
1978 20. If two routers both had headers_remove settings, and the first one included
1979 a superfluous trailing colon, the final name in the first list and the
1980 first name in the second list were incorrectly joined into one item (with a
1981 colon in the middle).
1987 1. Added -C and -D options to the exinext utility, mainly to make it easier
1988 to include in the automated testing, but these could be helpful when
1989 multiple configurations are in use.
1991 2. The exinext utility was not formatting the output nicely when there was
1992 an alternate port involved in the retry record key, nor when there was a
1993 message id as well (for retries that were specific to a specific message
1994 and a specific host). It was also confused by IPv6 addresses, because of
1995 the additional colons they contain. I have fixed the IPv4 problem, and
1996 patched it up to do a reasonable job for IPv6.
1998 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1999 delivery, the log line now contains "pipelined" if PIPELINING was used.
2001 4. An SMTP transport process used to panic and die if the bind() call to set
2002 an explicit outgoing interface failed. This has been changed; it is now
2003 treated in the same way as a connect() failure.
2005 5. A reference to $sender_host_name in the part of a conditional expansion
2006 that was being skipped was still causing a DNS lookup. This no longer
2009 6. The def: expansion condition was not recognizing references to header lines
2010 that used bh_ and bheader_.
2012 7. Added the _cache feature to named lists.
2014 8. The code for checking quota_filecount in the appendfile transport was
2015 allowing one more file than it should have been.
2017 9. For compatibility with Sendmail, the command line option
2025 and sets the incoming protocol and host name (for trusted callers). The
2026 host name and its colon can be omitted when only the protocol is to be set.
2027 Note the Exim already has two private options, -pd and -ps, that refer to
2028 embedded Perl. It is therefore impossible to set a protocol value of "d" or
2029 "s", but I don't think that's a major issue.
2031 10. A number of refactoring changes to the code, none of which should affect
2034 (a) The number of logging options was getting close to filling up the
2035 32-bit word that was used as a bit map. I have split them into two classes:
2036 those that are passed in the argument to log_write(), and those that are
2037 only ever tested independently outside of that function. These are now in
2038 separate 32-bit words, so there is plenty of room for expansion again.
2039 There is no change in the user interface or the logging behaviour.
2041 (b) When building, for example, log lines, the code previously used a
2042 macro that called string_cat() twice, in order to add two strings. This is
2043 not really sufficiently general. Furthermore, there was one instance where
2044 it was actually wrong because one of the argument was used twice, and in
2045 one call a function was used. (As it happened, calling the function twice
2046 did not affect the overall behaviour.) The macro has been replaced by a
2047 function that can join an arbitrary number of extra strings onto a growing
2050 (c) The code for expansion conditions now uses a table and a binary chop
2051 instead of a serial search (which was left over from when there were very
2052 few conditions). Also, it now recognizes conditions like "pam" even when
2053 the relevant support is not compiled in: a suitably worded error message is
2054 given if an attempt is made to use such a condition.
2056 11. Added ${time_interval:xxxxx}.
2058 12. A bug was causing one of the ddress fields not to be passed back correctly
2059 from remote delivery subprocesses. The field in question was not being
2060 subsequently used, so this caused to problems in practice.
2062 13. Added new log selectors queue_time and deliver_time.
2064 14. Might have fixed a bug in maildirsizefile handling that threw up
2065 "unexpected character" debug warnings, and recalculated the data
2066 unnecessarily. In any case, I expanded the warning message to give more
2069 15. Added the message "Restricted characters in address" to the statements in
2070 the default ACL that block characters like @ and % in local parts.
2072 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2073 Three changes have been made:
2075 (a) There was a serious bug; a negative response to MAIL caused the whole
2076 recipient domain to be cached as invalid, thereby blocking all messages
2077 to all local parts at the same domain, from all senders. This bug has
2078 been fixed. The domain is no longer cached after a negative response to
2079 MAIL if the sender used is not empty.
2081 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2084 (c) A new callout option, "use_sender" has been added for people who want
2085 the modified behaviour.
2091 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2094 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2095 indeed breaks things for older releases.
2097 3. Added additional logging to the case where there is a problem reading data
2098 from a filter that is running in a subprocess using a pipe, in order to
2099 try to track down a specific problem.
2101 4. Testing facility fudge: when running in the test harness and attempting
2102 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2103 getting "No route to host". Convert this to a timeout.
2105 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2108 6. Some OS don't have socklen_t but use size_t instead. This affects the
2109 fifth argument of getsockopt() amongst other things. This is now
2110 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2111 can be set for individual OS. I have set it for SunOS5, OSF1, and
2112 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2113 some earlier ones do not.
2115 7. Change 4.30/15 was not doing the test caselessly.
2117 8. The standard form for an IPv6 address literal was being rejected by address
2118 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2119 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2120 this, as well as the form without the "IPv6" on the front (but only when
2121 address literals are enabled, of course).
2123 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2125 10. Exim crashed if a message with an empty sender address specified by -f
2126 encountered a router with an errors_to setting. This could be provoked only
2127 by a command such as
2131 where an empty string was supplied; "<>" did not hit this bug.
2133 11. Installed PCRE release 4.5.
2135 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2136 remained set. It is now erased.
2138 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2139 times from message ids (which are base 36 rather than the normal 62).
2141 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2142 were being counted as actual protocol errors, and logged if the log
2143 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2144 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2145 and DATA following a set of rejected RCPTs do not count as protocol errors.
2146 In other words, Exim assumes they were pipelined, though this may not
2147 actually be the case. Of course, in all cases the client gets an
2148 appropriate error code.
2150 15. If a lookup fails in an ACL condition, a message about the failure may
2151 be available; it is used if testing the ACL cannot continue, because most
2152 such messages specify what the cause of the deferral is. However, some
2153 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2154 that caused an old message to be retained and used if a later statement
2155 caused a defer, replacing the real cause of the deferral.
2157 16. If an IP address had so many PTR records that the DNS lookup buffer
2158 was not large enough to hold them, Exim could crash while trying to process
2159 the truncated data. It now detects and logs this case.
2161 17. Further to 4.21/58, another change has been made: if (and only if) the
2162 first line of a message (the first header line) ends with CRLF, a bare LF
2163 in a subsequent header line has a space inserted after it, so as not to
2164 terminate the header.
2166 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2167 unnecessarily, used atoi() instead of strtol(), and didn't check the
2168 termination when getting file sizes from file names by regex.
2170 19. Completely re-implemented the support for maildirsize files, in the light
2171 of a number of problems with the previous contributed implementation
2172 (4.30/29). In particular:
2174 . If the quota is zero, the maildirsize file is maintained, but no quota is
2177 . If the maildir directory does not exist, it is created before any attempt
2178 to write a maildirsize file.
2180 . The quota value in the file is just a cache; if the quota is changed in
2181 the transport, the new value overrides.
2183 . A regular expression is available for excluding directories from the
2186 20. The autoreply transport checks the characters in options that define the
2187 message's headers; it allows continued headers, but it was checking with
2188 isspace() after an embedded newline instead of explicitly looking for a
2191 21. If all the "regular" hosts to which an address was routed had passed their
2192 expiry times, and had not reached their retry times, the address was
2193 bounced, even if fallback hosts were defined. Now Exim should go on to try
2196 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2197 equivalent code in the SMTP transport. Some hosts send humungous responses
2198 to HELO/EHLO, more than 1024 it seems.
2200 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2201 gives compiler warnings in some environments. I've now done it "properly",
2204 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2205 (because of problems with the built-in one) was declared to return uschar *
2206 instead of char *, causing compiler failure.
2208 25. Fixed a file descriptor leak when processing alias/forward files.
2210 26. Fixed a minor format string issue in dbfn.c.
2212 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2214 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2215 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2218 29. When a sender address is verified, it is cached, to save repeating the test
2219 when there is more than one recipient in a message. However, when the
2220 verification involves a callout, it is possible for different callout
2221 options to be set for different recipients. It is too complicated to keep
2222 track of this in the cache, so now Exim always runs a verification when a
2223 callout is required, relying on the callout cache for the optimization.
2224 The overhead is duplication of the address routing, but this should not be
2227 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2228 to be verified with callout=postmaster, and the main callout worked but the
2229 postmaster check failed, the verification correctly failed. However, if a
2230 subsequent RCPT command asked for sender verification *without* the
2231 postmaster check, incorrect caching caused this verification also to fail,
2234 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2235 it was not caching the DNS options (qualify_single, search_parents) that
2236 were used when the lookup failed. A subsequent lookup with different
2237 options therefore always gave the same answer, though there were cases
2238 where it should not have. (Example: a "domains = !$mx_any" option on a
2239 dnslookup router: the "domains" option is always processed without any
2240 widening, but the router might have qualify_single set.) Now Exim uses the
2241 cached value only when the same options are set.
2243 32. Added John Jetmore's "exipick" utility to the distribution.
2245 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2246 than a timeout (e.g. a certificate is required, and is not provided), an
2247 Exim server now closes the connection immediately. Previously it waited for
2248 the client to close - but if the client is SSL, it seems that they each
2249 wait for each other, leading to a delay before one of them times out.
2251 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2252 maintained 0.8.x compatibility because I don't think many are using it, and
2253 it is clearly obsolete.
2255 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2258 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2259 client certificate was expired. A simple patch fixes this, though I don't
2260 understand the full logic of why the verify callback is called multiple
2263 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2264 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2265 which causes problems with some clients (such as the Certicom SSL Plus
2266 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2267 disables the coutermeasure allowing Eudora to connect."
2269 38. Exim was not checking that a write() to a log file succeeded. This could
2270 lead to Bad Things if a log got too big, in particular if it hit a file
2271 size limit. Exim now panics and dies if it cannot write to a log file, just
2272 as it does if it cannot open a log file.
2274 39. Modified OS/Makefile-Linux so that it now contains
2276 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2278 The two -D definitions ensure that Exim is compiled with large file
2279 support, which makes it possible to handle log files that are bigger than
2282 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2283 instance) a domain was checked against a named list that involved a lookup,
2284 causing $domain_data to be set, then another domain was checked against the
2285 same list, then the first domain was re-checked, the value of $domain_data
2286 after the final check could be wrong. In particular, if the second check
2287 failed, it could be set empty. This bug probably also applied to
2290 41. The strip_trailing_dot option was not being applied to the address given
2291 with the -f command-line option.
2293 42. The code for reading a message's header from the spool was incrementing
2294 $received_count, but never initializing it. This meant that the value was
2295 incorrect (doubled) while delivering a message in the same process in which
2296 it was received. In the most common configuration of Exim, this never
2297 happens - a fresh exec is done - but it can happen when
2298 deliver_drop_privilege is set.
2300 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2301 it now includes up to 150 characters of the unexpected data in the log
2304 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2305 and building data strings. The size of both of these buffers was 10 000
2306 bytes - far larger than anybody would *ever* want, thought I. Needless to
2307 say, somebody hit the limit. I have increased the maximum line length to
2308 20 000 and the maximum data length of concatenated lines to 100 000. I have
2309 also fixed two bugs, because there was no checking on these buffers. Tsk,
2310 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2311 buffer is too small.
2313 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2314 lsearch lookups. Now it does.
2316 46. When parsing a route_list item in a manualroute router, a fixed-length
2317 buffer was used for the list of hosts. I made this 1024 bytes long,
2318 thinking that nobody would ever have a list of hosts that long. Wrong.
2319 Somebody had a whole pile of complicated expansion conditions, and the
2320 string was silently truncated, leading to an expansion error. It turns out
2321 that it is easier to change to an unlimited length (owing to other changes
2322 that have happened since this code was originally written) than to build
2323 structure for giving a limitation error. The length of the item that
2324 expands into the list of hosts is now unlimited.
2326 47. The lsearch lookup could not handle data where the length of text line was
2327 more than 4095 characters. Such lines were truncated, leading to shortened
2328 data being returned. It should now handle lines of any length.
2330 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2331 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2334 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2335 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2336 no longer happens in comment lines. A list of replacements is now placed
2337 at the head of all of the source files, except those whose only change is
2338 to replace PERL_COMMAND in the very first #! line.
2340 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2341 messages on the queue, with a bottom-up merge sort, using code contributed
2342 by Michael Haardt. This should make operations like -bp somewhat faster on
2343 large queues. It won't affect queue runners, except when queue_run_in_order
2346 51. Installed eximstats 1.31 in the distribution.
2348 52. Added support for SRV lookups to the dnslookup router.
2350 53. If an ACL referred to $message_body or $message_body_end, the value was not
2351 reset for any messages that followed in the same SMTP session.
2353 54. The store-handling optimization for building very long strings was not
2354 differentiating between the different store pools. I don't think this
2355 actually made any difference in practice, but I've tidied it.
2357 55. While running the routers to verify a sender address, $sender_address
2358 was still set to the sender address. This is wrong, because when routing to
2359 send a bounce to the sender, it would be empty. Therefore, I have changed
2360 it so that, while verifying a sender address, $sender_address is set to <>.
2361 (There is no change to what happens when verifying a recipient address.)
2363 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2364 target A or AAAA records (if not already returned) without resetting the
2365 qualify_single or search_parents options of the DNS resolver. These are
2366 inappropriate in this case because the targets of MX and SRV records must
2367 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2368 target that, when qualified, matched something in the local domain. These
2369 two options are now turned off when doing these lookups.
2371 57. It seems that at least some releases of Reiserfs (which does not have the
2372 concept of a fixed number of inodes) returns zero and not -1 for the
2373 number of available inodes. This interacted badly with check_spool_inodes,
2374 which assumed that -1 was the "no such thing" setting. What I have done is
2375 to check that the total number of inodes is greater than zero before doing
2376 the test of how many are available.
2378 58. When a "warn" ACL statement has a log_message modifier, the message is
2379 remembered, and not repeated. This is to avoid a lot of repetition when a
2380 message has many recipients that cause the same warning to be written.
2381 Howewer, Exim was preserving the list of already written lines for an
2382 entire SMTP session, which doesn't seem right. The memory is now reset if a
2383 new message is started.
2385 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2386 debugging output unless log_rewrite was also set.
2388 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2389 of (int)(handle) when we know that handle contains (void *)(-1).
2391 61. The Exim daemon panic-logs an error return when it closes the incoming
2392 connection. However "connection reset by peer" seems to be common, and
2393 isn't really an error worthy of noting specially, so that particular error
2396 62. When Exim is trying to find all the local interfaces, it used to panic and
2397 die if the ioctl to get the interface flags failed. However, it seems that
2398 on at least one OS (Solaris 9) it is possible to have an interface that is
2399 included in the list of interfaces, but for which you get a failure error
2400 for this call. This happens when the interface is not "plumbed" into a
2401 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2402 failure of the "get flags" call assumes that the interface is down.
2404 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2405 makes life easier for people who are doing arithmetic on fields extracted
2406 from dates, where you often get leading zeros that should not be
2407 interpreted as octal.
2409 64. Added qualify_domain to the redirect router, to override the global
2412 65. If a pathologically long header line contained very many addresses (the
2413 report of this problem mentioned 10 000) and each of them was rewritten,
2414 Exim could use up a very large amount of memory. (It kept on making new
2415 copies of the header line as it rewrote, and never released the old ones.)
2416 At the expense of a bit more processing, the header rewriting function has
2417 been changed so that it no longer eats memory in this way.
2419 66. The generation of the Received: header has been moved from the time that a
2420 message starts to be received, to the time that it finishes. The timestamp
2421 in the Received: header should now be very close to that of the <= log
2422 line. There are two side-effects of this change:
2424 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2425 logged header lines no longer include the local Received: line, because
2426 it has not yet been created. The same applies to a copy of the message
2427 that is returned to a non-SMTP sender when a message is rejected.
2429 (b) When a filter file is tested using -bf, no additional Received: header
2430 is added to the test message. After some thought, I decided that this
2433 This change does not affect the value of $received_for. It is still set
2434 after address rewriting, but before local_scan() is called.
2436 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2438 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2439 gave an unhelpful panic error message, and a defer error. I have managed to
2440 change this behaviour so that it now rejects any supplied certificate,
2441 which seems right, as the list of acceptable certificates is empty.
2443 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2444 gave an unhelpful defer error. I have not managed to make this reject any
2445 supplied certificates, but the error message it gives is "no certificate
2446 supplied", which is not helpful.
2448 70. exigrep's output now also includes lines that are not associated with any
2449 message, but which match the given pattern. Implemented by a patch from
2450 Martin Sluka, which also tidied up the Perl a bit.
2452 71. Recipient callout verification, like sender verification, was using <> in
2453 the MAIL FROM command. This isn't really the right thing, since the actual
2454 sender may affect whether the remote host accepts the recipient or not. I
2455 have changed it to use the actual sender in the callout; this means that
2456 the cache record is now keyed on a recipient/sender pair, not just the
2457 recipient address. There doesn't seem to be a real danger of callout loops,
2458 since a callout by the remote host to check the sender would use <>.
2459 [SEE ABOVE: changed after hitting problems.]
2461 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2462 temporary errors. However, in the case of such a code being given after
2463 the end of a data transmission (i.e. after ".") Exim was failing to write
2464 a retry record for the message. (Yes, there was some broken host that was
2465 actually sending 8xx at this point.)
2467 73. An unknown lookup type in a host list could cause Exim to panic-die when
2468 the list was checked. (An example that provoked this was putting <; in the
2469 middle of a list instead of at the start.) If this happened during a DATA
2470 ACL check, a -D file could be left lying around. This kind of configuration
2471 error no longer causes Exim to die; instead it causes a defer errror. The
2472 incident is still logged to the main and panic logs.
2474 74. Buglet left over from Exim 3 conversion. The message "too many messages
2475 in one connection" was written to the rejectlog but not the mainlog, except
2476 when address rewriting (yes!) was being logged.
2478 75. Added write_rejectlog option.
2480 76. When a system filter was run not as root (that is, when system_filter_user
2481 was set), the values of the $n variables were not being returned to the
2482 main process; thus, they were not subsequently available in the $sn
2485 77. Added +return_path_on_delivery log selector.
2487 78. A connection timeout was being treated differently from recipients deferred
2488 when testing hosts_max_try with a message that was older than the host's
2489 retry timeout. (The host should not be counted, thus allowing all hosts to
2490 be tried at least once before bouncing.) This may have been the cause of an
2491 occasionally reported bug whereby a message would remain on the queue
2492 longer than the retry timeout, but would be bounced if a delivery was
2493 forced. I say "may" because I never totally pinned down the problem;
2494 setting up timeout/retry tests is difficult. See also the next item.
2496 79. The ultimate address timeout was not being applied to errors that involved
2497 a combination of host plus message (for example, a timeout on a MAIL
2498 command). When an address resolved to a number of possible hosts, and they
2499 were not all tried for each delivery (e.g. because of hosts_max_try), a
2500 message could remain on the queue longer than the retry timeout.
2502 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2505 81. Fixed an obscure SMTP outgoing bug which required at least the following
2506 conditions: (a) there was another message waiting for the same server;
2507 (b) the server returned 5xx to all RCPT commands in the first message so
2508 that the message was not completed; (c) the server dropped the connection
2509 or gave a negative response to the RSET that Exim sends to abort the
2510 transaction. The observed case was a dropped connection after DATA that had
2511 been sent in pipelining mode. That is, the server had advertised PIPELINING
2512 but was not implementing it correctly. The effect of the bug was incorrect
2513 behaviour, such as trying another host, and this could lead to a crash.
2519 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2520 and daemon.c were passed as pointers to ints; they should have been
2521 pointers to socklen_t variables (which are typically unsigned ints).
2523 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2526 3. Fixed a really odd bug that affected only the testing scheme; patching a
2527 certain fixed string in the binary changed the value of another string that
2528 happened to be identical to the end of the original first string.
2530 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2531 name", it returns that address as the IP address. On some operating
2532 systems (e.g. Solaris), it also passes back the IP address string as the
2533 "host name". However, on others (e.g. Linux), it passes back an empty
2534 string. Exim wasn't checking for this, and was changing the host name to an
2535 empty string, assuming it had been canonicized.
2537 5. Although rare, it is permitted to have more than one PTR record for a given
2538 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2539 all the names associated with an address, because they do in Solaris.
2540 However, it seems that they do not in Linux for data that comes from the
2541 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2542 I found this out when I moved to a new Linux workstation and tried to run
2543 the Exim test suite.
2545 To get round this problem I have changed the code so that it now does its
2546 own call to the DNS to look up PTR records when searching for a host name.
2547 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2548 addresses that are only in /etc/hosts are still found.
2550 This behaviour is, however, controlled by an option called host_lookup_
2551 order, which defaults to "bydns:byaddr". If people want to use the other
2552 order, or indeed, just use one or the other means of lookup, they can
2553 specify it in this variable.
2555 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2556 some operating systems, this comes back from gethostbyaddr() as an empty
2557 string, and this is what Exim used to test for. However, it seems that in
2558 other systems, "." is yielded. Exim now tests for this case too.
2560 7. The values of check_spool_space and check_log_space are now held internally
2561 as a number of kilobytes instead of an absolute number of bytes. If a
2562 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2563 kilobyte. This means that much larger values can be stored.
2565 8. Exim monitor: an attempt to get the action menu when not actually pointing
2566 at a message produces an empty menu entitled "No message selected". This
2567 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2568 no entries in it ("Shell widget menu has zero width and/or height"). So I
2569 have added a single, blank menu entry in this case.
2571 9. Added ${quote_local_part.
2573 10. MIME decoding is now applied to the contents of Subject: header lines when
2576 11. Now that a reference to $sender_host_address automatically causes a reverse
2577 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2578 host lookup before query-style lookups in lists that might use this
2579 variable. This has therefore been abolished, and the "net-" prefix is no
2580 longer necessary for query-style lookups.
2582 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2583 have been a typo for LFLAGS, so it has been changed.
2585 13. The install script calls Exim with "-C /dev/null" in order to find the
2586 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2587 to be output. Howeve, since Exim outputs its version number before the
2588 error, it didn't break the script. It just looked ugly. I fixed this by
2589 always allowing "-C /dev/null" if the caller is root.
2591 14. Ignore overlarge ACL variable number when reading spool file - insurance
2592 against a later release with more variables having written the file.
2594 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2595 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2596 this, as well as the form without the "IPv6" on the front.
2598 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2599 OS/Makefile-Darwin file.
2601 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2602 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2604 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2605 variable. However, some OS use ints for inodes. I've added cast to long int
2606 to get rid of the compiler warning.
2608 19. I had forgotten to lock out "/../" in configuration file names when
2609 ALT_CONFIG_PREFIX was set.
2611 20. Routers used for verification do not need to specify transports. However,
2612 if such a router generated a host list, and callout was configured, Exim
2613 crashed, because it could not find a port number from the (non-existent)
2614 transport. It now assumes port 25 in this circumstance.
2616 21. Added the -t option to exigrep.
2618 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2619 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2620 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2621 them is set, LOOKUP_LSEARCH is forced.
2623 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2625 24. Added sender and host information to the "rejected by local_scan()" log
2626 line; previously there was no indication of these.
2628 25. Added .include_if_exists.
2630 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2631 Linux. It turns out that not all file systems support this. Apparently some
2632 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2633 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2634 not supported on the file descriptor, is now ignored when Exim is trying to
2635 sync a directory. This applies only to Linux.
2637 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2639 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2640 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2643 29. Added support for maildirsize files from supplied patch (modified a bit).
2645 30. The use of :fail: followed by an empty string could lead Exim to respond to
2646 sender verification failures with (e.g.):
2648 550 Verification failed for <xxx>
2649 550 Sender verify failed
2651 where the first response line was missing the '-' that indicates it is not
2652 the final line of the response.
2654 31. The loop for finding the name of the user that called Exim had a hardwired
2655 limit of 10; it now uses the value of finduser_retries, which is used for
2656 all other user lookups.
2658 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2661 33. Exim was neglecting to zero errno before one call of strtol() when
2662 expanding a string and expecting an integer value. On some systems this
2663 resulted in spurious "integer overflow" errors. Also, it was casting the
2664 result into an int without checking.
2666 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2667 did not work. The code looks as if it has *never* worked, though it appears
2668 to have been documented since at least releast 1.62. I have made it work.
2670 35. The "timeout_DNS" error in retry rules, also documented since at least
2671 1.62, also never worked. As it isn't clear exactly what this means, and
2672 clearly it isn't a major issue, I have abolished the feature by treating it
2673 as "timeout", and writing a warning to the main and panic logs.
2675 36. The display of retry rules for -brt wasn't always showing the error code
2678 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2679 timeout_connect_A, timeout_connect_MX.
2681 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2682 to the empty sender.
2684 39. The daemon was not analysing the content of -oX till after it had closed
2685 stderr and disconnected from the controlling terminal. This meant that any
2686 syntax errors were only noted on the panic log, and the return code from
2687 the command was 0. By re-arranging the code a little, I've made the
2688 decoding happen first, so such errors now appear on stderr, and the return
2689 code is 1. However, the actual setting up of the sockets still happens in
2690 the disconnected process, so errors there are still only recorded on the
2693 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2694 connections (as happens on some IP stacks) was logged at start up time as
2695 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2696 it from "IPv6 and IPv4", which means that two separate sockets are being
2699 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2700 says whether AF_INET or AF_INET6 was passed as an argument.
2702 42. Exiwhat output was messed up when time zones were included in log
2705 43. Exiwhat now gives more information about the daemon's listening ports,
2706 and whether -tls-on-connect was used.
2708 44. The "port" option of the smtp transport is now expanded.
2710 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2711 silently ignored. Now an error message is written to the main and panic
2714 46. There's a new ACL modifier called "logwrite" which writes to a log file
2715 as soon as it is encountered.
2717 47. Added $local_user_uid and $local_user_gid at routing time.
2719 48. Exim crashed when trying to verify a sender address that was being
2722 49. Exim was recognizing only a space character after ".include". It now also
2723 recognizes a tab character.
2725 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2726 extracting the relevant information from the specification. The man page no
2727 longer contains scrambled data for the -d option, and I've added a section
2728 at the front about calling Exim under different names.
2730 51. Added "extra_headers" argument to the "mail" command in filter files.
2732 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2735 53. Installed eximstats 1.29.
2737 54. Added transport_filter_timeout as a generic transport option.
2739 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2740 Cc: header lines. This was required by RFC 822, but it not required by RFC
2743 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2744 incoming messages that did not have them. Now it does so only if the
2745 message originates locally, that is, if there is no associated remote host
2746 address. When Resent- header lines are present, this applies to the Resent-
2747 lines rather than the non-Resent- lines.
2749 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2750 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2752 58. Messages for configuration errors now include the name of the main
2753 configuration file - useful now that there may be more than one file in a
2754 list (.included file names were always shown).
2756 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2757 for those rare installations that do not start the daemon as root or run it
2758 setuid root. I've cut out the call to initgroups() if the daemon is not
2761 60. The Exim user and group can now be bound into the binary as text strings
2762 that are looked up at the start of Exim's processing.
2764 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2766 62. Added $mailstore_basename variable.
2768 63. Installed patch to sieve.c from Michael Haardt.
2770 64. When Exim failed to open the panic log after failing to open the main log,
2771 the original message it was trying to log was written to stderr and debug
2772 output, but if they were not available (the usual case in production), it
2773 was lost. Now it is written to syslog before the two lines that record the
2774 failures to open the logs.
2776 65. Users' Exim filters run in subprocesses under the user's uid. It is
2777 possible for a "deliver" command or an alias in a "personal" command to
2778 provoke an address rewrite. If logging of address rewriting is configured,
2779 this fails because the process is not running as root or exim. There may be
2780 a better way of dealing with this, but for the moment (because 4.30 needs
2781 to be released), I have disabled address rewrite logging when running a
2782 filter in a non-root, non-exim process.
2788 1. The buildconfig auxiliary program wasn't quoting the value set for
2789 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2790 not defined. This bug was masked in 4.22 by the effect that was fixed in
2793 2. Some messages that were rejected after a message id was allocated were
2794 shown as "incomplete" by exigrep. It no longer does this for messages that
2795 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2797 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2798 have allow_domain_literals set, the ID did not get logged in the <= line.
2799 Domain literals are now always recognized in Message-ID: header lines.
2801 4. The first argument for a ${extract expansion item is the key name or field
2802 number. Leading and trailing spaces in this item were not being ignored,
2803 causing some misleading effects.
2805 5. When deliver_drop_privilege was set, single queue runner processes started
2806 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2807 same command in the process it spins off) were not dropping privilege.
2809 6. When the daemon running as "exim" started a queue runner, it always
2810 re-executed Exim in the spun-off process. This is a waste of effort when
2811 deliver_drop_privilege is set. The new process now just calls the
2812 queue-runner function directly.
2818 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2821 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2822 ignored. Though the use of -oP was forcing the writing of a pid file, it
2823 was always written to the default place.
2825 3. If the message "no IP address found for host xxxx" is generated during
2826 incoming verification, it is now followed by identification of the incoming
2827 connection (so you can more easily find what provoked it).
2829 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2831 5. Added some features to "harden" Exim a bit more against certain attacks:
2833 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2834 be put in Local/Makefile. This is like the never_users runtime option,
2835 but it cannot be overridden. The default setting is "root".
2837 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2838 prefix string with which any file named in a -C command line option
2841 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2842 is retained for -C and -D only if the caller of Exim is root. Without
2843 it, the exim user may also use -C and -D and retain privilege.
2845 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2846 command line option is disabled.
2848 6. Macro names set by the -D option must start with an upper case letter, just
2849 like macro names defined in the configuration file.
2851 7. Added "dereference=" facility to LDAP.
2853 8. Two instances of the typo "uknown" in the source files are fixed.
2855 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2856 the Configure-Makefile script screwed up while processing it.
2858 10. Incorporated PCRE 4.4.
2860 11. The SMTP synchronization check was not operating right at the start of an
2861 SMTP session. For example, it could not catch a HELO sent before the client
2862 waited for the greeting. There is now a check for outstanding input at the
2863 point when the greeting is written. Because of the duplex, asynchronous
2864 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2865 way, but not yet received, when the check is performed.
2867 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2868 on TCP/IP sockets, because this apparently causes some broken clients to
2871 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2872 unchanged) from the Cygwin maintainer.
2874 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2875 was supported instead of testing for mbx. Effectively a typo.
2877 15. The spa authenticator server code was not checking that the input it
2878 received was valid base64.
2880 16. The debug output line for the "set" modifier in ACLs was not showing the
2881 name of the variable that was being set.
2883 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2885 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2886 lookup on its own. Something else was needed to trigger the lookup. For
2887 example, a match in host_lookup or the need for a host name in a host list.
2888 Now, if $sender_host_name is referenced and the host name has not yet been
2889 looked up, a lookup is performed. If the lookup fails, the variable remains
2890 empty, and $host_lookup_failed is set to "1".
2892 19. Added "eqi" as a case-independent comparison operator.
2894 20. The saslauthd authentication condition could segfault if neither service
2895 nor realm was specified.
2897 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2898 error message that was logged was misleading, and incoming SMTP
2899 connections were dropped. The message is now more accurate, and temporary
2900 errors are given to SMTP connections.
2902 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2903 (except RSET) with a 421 error, until QUIT is received. However, it was
2904 failing to send a response to QUIT.
2906 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2907 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2908 ACL is now run after the helo_try_verify_hosts code.
2910 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2911 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2912 case-independent, so other case variants are also recognized). Apparently
2913 some systems use these upper case variants.
2915 25. If more than two messages were waiting for the same host, and a transport
2916 filter was specified for the transport, Exim sent two messages over the
2917 same TCP/IP connection, and then failed with "socket operation on non-
2918 socket" when it tried to send the third.
2920 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2922 27. The extern definition of crypt16() in expand.c was not being excluded when
2923 the OS had its own crypt16() function.
2925 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2926 as a preferred synonym for return_size_limit, both as an option and as an
2929 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2931 30. Changed the default configuration ACL to relax the local part checking rule
2932 for addresses that are not in any local domains. For these addresses,
2933 slashes and pipe symbols are allowed within local parts, but the sequence
2934 /../ is explicitly forbidden.
2936 31. SPA server authentication was not clearing the challenge buffer before
2939 32. log_message in a "warn" ACL statement was writing to the reject log as
2940 well as to the main log, which contradicts the documentation and doesn't
2941 seem right (because no rejection is happening). So I have stopped it.
2943 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2944 However, I am unable to do any testing of this.
2946 34. Fixed an infelicity in the appendfile transport. When checking directories
2947 for a mailbox, to see if any needed to be created, it was accidentally
2948 using path names with one or more superfluous leading slashes; tracing
2949 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2951 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2952 contents are added to the log line that is written for every discarded
2953 recipient. (Previously a log_message setting was ignored.)
2955 36. The ${quote: operator now quotes the string if it is empty.
2957 37. The install script runs exim in order to find its version number. If for
2958 some reason other than non-existence or emptiness, which it checks, it
2959 could not run './exim', it was installing it with an empty version number,
2960 i.e. as "exim-". This error state is now caught, and the installation is
2963 38. An argument was missing from the function that creates an error message
2964 when Exim fails to connect to the socket for saslauthd authentication.
2965 This could cause Exim to crash, or give a corrupted message.
2967 39. Added isip, isip4, and isip6 to ${if conditions.
2969 40. The ACL variables $acl_xx are now saved with the message, and can be
2970 accessed later in routers, transports, and filters.
2972 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2973 strings in the file are not string-expanded.
2975 42. If a MAIL command specified a SIZE value that was too large to fit into an
2976 int variable, the check against message_size_limit failed. Such values are
2977 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2978 day this will have to be increased, but I don't think I want to be around
2979 when emails are that large.
2986 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2987 iconv() is not standard in FreeBSD.
2989 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2990 IPv6 enabled. The observed symptom was a segmentation fault on return from
2991 the function os_common_find_running_interfaces() in src/os.c.
2993 3. In the check_special_case() function in daemon.c I had used "errno" as an
2994 argument name, which causes warnings on some systems. This was basically a
2995 typo, since it was named "eno" in the comments!
2997 4. The code that waits for the clock to tick (at a resolution of some fraction
2998 of a second) so as to ensure message-id uniqueness was always waiting for
2999 at least one whole tick, when it could have waited for less. [This is
3000 almost certainly not relevant at current processor speeds, where it is
3001 unlikely to ever wait at all. But we try to future-proof.]
3003 5. The function that sleeps for a time interval that includes fractions of a
3004 second contained a race. It did not block SIGALRM between setting the
3005 timer, and suspending (a couple of lines later). If the interval was short
3006 and the sigsuspend() was delayed until after it had expired, the suspension
3007 never ended. On busy systems this could lead to processes getting stuck for
3010 6. Some uncommon configurations may cause a lookup to happen in a queue runner
3011 process, before it forks any delivery processes. The open lookup caching
3012 mechanism meant that the open file or database connection was passed into
3013 the delivery process. The problem was that delivery processes always tidy
3014 up cached lookup data. This could cause a problem for the next delivery
3015 process started by the queue runner, because the external queue runner
3016 process does not know about the closure. So the next delivery process
3017 still has data in the lookup cache. In the case of a file lookup, there was
3018 no problem because closing a file descriptor in a subprocess doesn't affect
3019 the parent. However, if the lookup was caching a connection to a database,
3020 the connection was closed, and the second delivery process was likely to
3021 see errors such as "PGSQL: query failed: server closed the connection
3022 unexpectedly". The problem has been fixed by closing all cached lookups
3023 in a queue runner before running a delivery process.
3025 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
3026 seem to have the "const" qualifier which it has on other OS. I've
3029 8. Change 4.21/2 was too strict. It is only if there are two authenticators
3030 *of the same type* (client or server) with the same public name that an
3031 error should be diagnosed.
3033 9. When Exim looked up a host name for an IP address, but failed to find the
3034 original IP address when looking up the host name (a safety check), it
3035 output the message "<ip address> does not match any IP for NULL", which was
3036 confusing, to say the least. The bug was that the host name should have
3037 appeared instead of "NULL".
3039 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
3040 user that is built into the binary, and the -C or -D options is used, root
3041 privilege is dropped before the configuration file is read. In addition,
3042 logging is switched to stderr instead of the normal log files. If the
3043 configuration then re-defines the Exim user, the unprivileged environment
3044 is probably not what is expected, so Exim logs a panic warning message (but
3047 However, if deliver_drop_privilege is set, the unprivileged state may well
3048 be exactly what is intended, so the warning has been cut out in that case,
3049 and Exim is allowed to try to write to its normal log files.
3055 1. smtp_return_error_details was not giving details for temporary sender
3056 or receiver verification errors.
3058 2. Diagnose a configuration error if two authenticators have the same public
3061 3. Exim used not to create the message log file for a message until the first
3062 delivery attempt. This could be confusing when incoming messages were held
3063 for policy or load reasons. The message log file is now created at the time
3064 the message is received, and an initial "Received" line is written to it.
3066 4. The automatically generated man page for command line options had a minor
3067 bug that caused no ill effects; however, a more serious problem was that
3068 the procedure for building the man page automatically didn't always
3069 operate. Consequently, release 4.20 contains an out-of-date version. This
3070 shouldn't happen again.
3072 5. When building Exim with embedded Perl support, the script that builds the
3073 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3074 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3076 6. The freeze_tell option was not being used for messages that were frozen on
3077 arrival, either by an ACL or by local_scan().
3079 7. Added the smtp_incomplete_transaction log selector.
3081 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3082 was accepting AUTH without a new EHLO.
3084 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3085 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3088 10. Logging of TCP/IP connections (when configured) now happens in the main
3089 daemon process instead of the child process, so that the TCP/IP connection
3090 count is more accurate (but it can never be perfect).
3092 11. The use of "drop" in a nested ACL was not being handled correctly in the
3093 outer ACL. Now, if condition failure induced by the nested "drop" causes
3094 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3095 or "require"), the connection is dropped.
3097 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3098 that yield "discard" can now be used with an "accept" or a "discard" verb,
3099 but an error is generated for any others (because I can't see a useful way
3100 to define what should happen).
3102 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3103 are now processed in the same way as lines in the Exim configuration file.
3104 In particular, continuation lines are supported.
3106 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3108 15. Added -ti meaning -t -i.
3110 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3111 domains, and warn by logging if others are found.
3113 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3114 returned from GIFCONF when Exim is trying to find the list of interfaces on
3115 a host. The code in os.c has been modified to copy each ifreq to an aligned
3116 structure in all cases.
3118 Also, in some cases, the returned ifreq's were being copied to a 'struct
3119 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3120 means the last couple of bytes of an IPv6 address could be chopped if the
3121 ifreq contained only a normal sockaddr (14 bytes storage).
3123 18. Named domain lists were not supported in the hosts_treat_as_local option.
3124 An entry such as +xxxx was not recognized, and was treated as a literal
3127 19. Ensure that header lines added by a DATA ACL are included in the reject log
3128 if the ACL subsequently rejects the message.
3130 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3131 MD5 (which is deprecated).
3133 21. When testing a filter file using -bf, Exim was writing a message when it
3134 took the sender from a "From " line in the message, but it was not doing so
3135 when it took $return_path from a Return-Path: header line. It now does.
3137 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3138 with a valid header line field name (a series of printing characters
3139 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3141 23. Changed "disc" in the source to "disk" to conform to the documentation and
3142 the book and for uniformity.
3144 24. Ignore Sendmail's -Ooption=value command line item.
3146 25. When execve() failed while trying to run a command in a pipe transport,
3147 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3148 could be confused with a return value of 69 from the command itself. This
3149 has been changed to 127, the value the shell returns if it is asked to run
3150 a non-existent command. The wording for the related log line suggests a
3151 non-existent command as the problem.
3153 26. If received_header_text expands to an empty string, do not add a Received:
3154 header line to the message. (Well, it adds a token one on the spool, but
3155 marks it "old" so that it doesn't get used or transmitted.)
3157 27. Installed eximstats 1.28 (addition of -nt option).
3159 28. There was no check for failure on the call to getsockname() in the daemon
3160 code. This can fail if there is a shortage of resources on the system, with
3161 ENOMEM, for example. A temporary error is now given on failure.
3163 29. Contrary to the C standard, it seems that in some environments, the
3164 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3165 program. Exim now does this explicitly; it affects the formatting of
3166 timestamps using strftime().
3168 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3169 complaints. I've now initialized all the variables, to avoid this.
3171 32. Header lines added by a system filter were not being "seen" during
3172 transport-time rewrites.
3174 33. The info_callback() function passed to OpenSSL is set up with type void
3175 (*)(SSL *, int, int), as described somewhere. However, when calling the
3176 function (actually a macro) that sets it up, the type void(*)() is
3177 expected. I've put in a cast to prevent warnings from picky compilers.
3179 34. If a DNS black list lookup found a CNAME record, but there were no A
3180 records associated with the domain it pointed at, Exim crashed.
3182 35. If a DNS black list lookup returned more than one A record, Exim ignored
3183 all but the first. It now scans all returned addresses if a particular IP
3184 value is being sought. In this situation, the contents of the
3185 $dnslist_value variable are a list of all the addresses, separated by a
3188 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3189 to accept a host name and all its aliases if the forward lookup for any of
3190 them yielded the IP address of the incoming connection. Now it accepts only
3191 those names whose forward lookup yields the correct IP address. Any other
3192 names are discarded. This closes a loophole whereby a rogue DNS
3193 administrator could create reverse DNS records to break through a
3194 wildcarded host restriction in an ACL.
3196 37. If a user filter or a system filter that ran in a subprocess used any of
3197 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3198 the wrong values were passed to the pipe command ($thisaddress had the
3199 value of $0, $0 had the value of $1, etc). This bug was introduced by
3200 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3202 38. Improved the line breaking for long SMTP error messages from ACLs.
3203 Previously, if there was no break point between 40 and 75 characters, Exim
3204 left the rest of the message alone. Two changes have been made: (a) I've
3205 reduced the minimum length to 35 characters; (b) if it can't find a break
3206 point between 35 and 75 characters, it looks ahead and uses the first one
3207 that it finds. This may give the occasional overlong line, but at least the
3208 remaining text gets split now.
3210 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3211 file descriptors might be low, and that setting 1000 would always raise it.
3212 It turns out that in some environments, the limit is already over 1000 and
3213 that lowering it causes trouble. So now Exim takes care not to decrease it.
3215 40. When delivering a message, the value of $return_path is set to $sender_
3216 address at the start of routing (routers may change the value). By an
3217 oversight, this default was not being set up when an address was tested by
3218 -bt or -bv, which affected the outcome if any router or filter referred to
3221 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3222 log only when "message" or "log_message" are set. However, if one of the
3223 conditions was an address verification, or a call to a nested ACL, the
3224 messages generated by the underlying test were being passed through. This
3225 no longer happens. The underlying message is available in $acl_verify_
3226 message for both "message" and "log_message" expansions, so it can be
3227 passed through if needed.
3229 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3230 new expansion $bh_ to give the encoded byte string without charset
3231 translation. Translation happens only if iconv() is available; HAVE_ICONV
3232 indicates this at build time. HEADERS_CHARSET gives the charset to
3233 translate to; headers_charset can change it in the configuration, and
3234 "headers charset" can change it in an individual filter file.
3236 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3237 that creates RFC 2047 encoded "words" labels them as that charset instead
3238 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3239 expansion operator; (ii) when Exim creates a From: line for a local
3240 message; (iii) when a header line is rewritten to include a "phrase" part.
3242 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3243 buggy, causing it to skip the first lines of messages whose message ID
3244 ended in 'D'. This would not have bitten before Exim release 4.14, because
3245 message IDs were unlikely to end in 'D' before then. The effect was to have
3246 incorrect size information for certain domains.
3248 45. #include "config.h" was missing at the start of the crypt16.c module. This
3249 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3252 46. If there was a timeout during a "random" callout check, Exim treated it as
3253 a failure of the random address, and carried on sending RSET and the real
3254 address. If the delay was just some slowness somewhere, the response to the
3255 original RCPT would be taken as a response to RSET and so on, causing
3256 mayhem of various kinds.
3258 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3259 when I implemented it. It didn't allow for the fact that some option values
3260 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3261 the right test for positive values.
3263 48. Domain names in DNS records are case-independent. Exim always looks them up
3264 in lower case. Some resolvers return domain names in exactly the case they
3265 appear in the zone file, that is, they may contain uppercase letters. Not
3266 all resolvers do this - some return always lower case. Exim was treating a
3267 change of case by a resolver as a change of domain, similar to a widening
3268 of a domain abbreviation. This triggered its re-routing code and so it was
3269 trying to route what was effectively the same domain again. This normally
3270 caused routing to fail (because the router wouldn't handle the domain
3271 twice). Now Exim checks for this case specially, and just changes the
3272 casing of the domain that it ultimately uses when it transmits the message
3275 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3278 50. If a filter generated a file delivery with a non-absolute name (possible if
3279 no home directory exists for the router), the forbid_file option was not
3282 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3283 the existing equality matching.
3285 52. Exim was using ints instead of ino_t variables in some places where it was
3286 dealing with inode numbers.
3288 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3289 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3290 called TMPDIR, and if it finds it is different, it changes its value.
3292 54. The smtp_printf() function is now made available to local_scan() so
3293 additional output lines can be written before returning. There is also an
3294 smtp_fflush() function to enable the detection of a dropped connection.
3295 The variables smtp_input and smtp_batched_input are exported to
3298 55. Changed the default runtime configuration: the message "Unknown user"
3299 has been removed from the ACL, and instead placed on the localuser router,
3300 using the cannot_route_message feature. This means that any verification
3301 failures that generate their own messages won't get overridden. Similarly,
3302 the "Unrouteable address" message that was in the ACL for unverifiable
3303 relay addresses has also been removed.
3305 56. Added hosts_avoid_esmtp to the smtp transport.
3307 57. The exicyclog script was not checking for the esoteric option
3308 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3309 will work only if exicyclog is run under the appropriate euid.
3311 58. Following a discussion on the list, the rules by which Exim recognises line
3312 endings on incoming messages have been changed. The -dropcr and drop_cr
3313 options are now no-ops, retained only for backwards compatibility. The
3314 following line terminators are recognized: LF CRLF CR. However, special
3315 processing applies to CR:
3317 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3318 nor a local message in the state where . is a terminator.
3320 (ii) If a bare CR is encountered in a header line, an extra space is added
3321 after the line terminator so as not to end the header. The reasoning
3322 behind this is that bare CRs in header lines are most likely either
3323 to be mistakes, or people trying to play silly games.
3325 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3326 was being incorrectly given as 18 bytes larger than it should have been.
3327 This is a VOB (very old bug).
3329 60. This may never have affected anything current, but just in case it has:
3330 When the local host is found other than at the start of a list of hosts,
3331 the local host, those with the same MX, and any that follow, are discarded.
3332 When the list in question was part of a longer list of hosts, the following
3333 hosts (not currently being processed) were also being discarded. This no
3334 longer happens. I'm not sure if this situation could ever has previously
3337 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3340 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3341 header. This is something that is recommended in a new Internet Draft, and
3342 is something that is documented as being done by Sendmail. There are two
3343 possible values. For messages generated by the autoreply transport, Exim
3346 Auto-Submitted: auto-replied
3348 whereas for all other generated messages (e.g. bounces) it adds
3350 Auto-Submitted: auto-generated
3352 63. The "personal" condition in filters now includes a test for the
3353 Auto-Submitted: header. If it contains the string "auto-" the message it
3354 not considered personal.
3356 64. Added rcpt_include_affixes as a generic transport option.
3358 65. Added queue_only_override (default true).
3360 66. Added the syslog_duplication option.
3362 67. If what should have been the first header line of a message consisted of
3363 a space followed by a colon, Exim was mis-interpreting it as a header line.
3364 It isn't of course - it is syntactically invalid and should therefore be
3365 treated as the start of the message body. The misbehaviour could have
3366 caused a number of strange effects, including loss of data in subsequent
3367 header lines, and spool format errors.
3369 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3370 client host had authenticated. This control can now be exercised by an ACL
3371 for more flexibility.
3373 69. By default, callouts do not happen when testing with -bh. There is now a
3374 variant, -bhc, which does actually run the callout code, including
3375 consulting and updating the callout cache.
3377 70. Added support for saslauthd authentication, courtesy of Alexander
3380 71. If statvfs() failed on the spool or log directories while checking their
3381 size for availability, Exim confusingly gave the error "space shortage".
3382 Furthermore, in debugging mode it crashed with a floating point exception.
3383 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3384 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3385 serious problem, Exim now writes to the main and panic logs when this
3386 happens, with details of the failure. It then refuses to accept the
3387 incoming message, giving the message "spool directory problem" or "log
3388 directory problem" with a 421 code for SMTP messages.
3390 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3391 0, 1, and 2 exist, because some OS complain for execs without them (see
3392 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3393 descriptors. However, the code omitted to check that the open succeeded,
3394 causing mysterious errors if for some reason the permissions on /dev/null
3395 got screwed. Now Exim writes a message to the main and panic logs, and
3396 bombs out if it can't open /dev/null.
3398 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3399 interact so that it is all more flexible. It is supposed to remain
3400 backwards compatible. Also added extra_local_interfaces.
3402 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3403 to bomb out with an assertion failure - to the client this appears as a
3404 connection drop. This problem occurs in the part of the code that was taken
3405 from the Samba project. Fortunately, the assertion is in a very simple
3406 function, so I have fixed this by reproducing the function inline in the
3407 one place where it is called, and arranging for authentication to fail
3408 instead of killing the process with assert().
3410 75. The SPA client code was not working when the server requested OEM rather
3411 than Unicode encoding.
3413 76. Added code to make require_files with a specific uid setting more usable in
3414 the case where statting the file as root fails - usually a non-root-mounted
3415 NFS file system. When this happens and the failure is EACCES, Exim now
3416 forks a subprocess and does the per-uid checking as the relevant uid.
3418 77. Added process_log_path.
3420 78. If log_file_path was not explicitly set, a setting of check_log_space or
3421 check_log_inodes was ignored.
3423 79. If a space check for the spool or log partitions fails, the incident is now
3424 logged. Of course, in the latter case the data may get lost...
3426 80. Added the %p formatting code to string_format() so that it can be used to
3427 print addresses in debug_print(). Adjusted all the address printing in the
3428 debugging in store.c to use %p rather than %d.
3430 81. There was a concern that a line of code in smtp_in.c could overflow a
3431 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3432 initially expressed, the concern was not well-founded, because trailing
3433 spaces are removed early. However, if the trailing spaces were followed by
3434 a NULL, they did not get removed, so the overflow was possible. Two fixes
3437 (a) I re-wrote the offending code in a cleaner fashion.
3438 (b) If an incoming SMTP command contains a NULL character, it is rejected
3441 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3442 runs initgroups(), so that if the Exim user is in any additional groups,
3443 they will be used during message reception.
3449 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / exim.git / blob