1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.245 2005/10/04 08:54:33 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
10 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
13 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
14 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
16 PH/03 Typo: missing ".o" in src/pcre/Makefile.
18 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
19 header line, restrict the check to what is listed in RFCs 2369 and 2929.
20 Also, for "Auto-Submitted", treat anything other than "no" as
21 non-personal, in accordance with RFC 3834. (Previously it treated
22 anything starting "auto-" as non-personal.)
24 TF/01 The control=submission/name=... option had a problem with syntax
25 errors if the name included a slash character. The /name= option
26 now slurps the rest of the string, so it can include any characters
27 but it must come last in the list of options (after /sender_retain
30 PH/05 Some modifications to the interface to the fake nameserver for the new
37 TK/01 Added the "success_on_redirect" address verification option. See
38 NewStuff for rationale and an example.
40 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
42 PH/02 Patch to exigrep to allow it to work on syslog lines.
44 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
45 fread() to skip over the body file's header line, because in Cygwin the
46 header line is locked and is inaccessible.
48 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
49 co-exist for some time) to make it clear that it is the Exim ID that is
50 referenced, not the Message-ID: header line.
52 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
53 string_format() function, because snprintf() does not exist on all
56 PH/06 The use of forbid_filter_existstest now also locks out the use of the
57 ${stat: expansion item.
59 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
60 protocol synchronization error", to keep the pedants happy.
62 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
63 well as for IRIX systems, when gcc is being used. See the host.c source
66 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
68 PH/10 Named domain lists were not working if used in a queue_smtp_domains
71 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
72 transport and to the smtp transport in LMTP mode.
74 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
76 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
77 run a filter in a subprocess. This could lead to confusion in subsequent
78 lookups in the parent process. There should also be a search_tidyup() at
79 the end of the subprocess.
81 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
82 only if the host matched helo_try_verify_hosts, which caused the
83 verification to occur when the EHLO/HELO command was issued. The ACL just
84 tested the remembered result. Now, if a previous verification attempt has
85 not happened, "verify = helo" does it there and then.
87 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
89 TK/03 Fix log output including CR from clamd.
91 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
92 exist provoked a memory error which could cause a segfault.
94 PH/15 Installed PCRE 6.2
96 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
98 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
99 of the problem. Specifically, suggested +O2 rather than +O1 for the
102 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
104 PH/20 If a delivery was routed to a non-standard port by means of an SRV
105 record, the port was not correctly logged when the outgoing_port log
106 selector was set (it logged the transort's default port).
108 PH/21 Added support for host-specific ports to manualroute, queryprogram,
109 fallback_hosts, and "hosts" in the smtp transport.
111 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
112 host errors such as "Connection refused".
114 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
115 authentication with radiusclient 0.4.9:
117 - Error returned from rc_read_config was caught wrongly
118 - Username/password not passed on to radius server due to wrong length.
120 The presumption is that some radiusclient API changes for 4.51/PH/17
121 were not taken care of correctly. The code is still untested by me (my
122 Linux distribution still has 0.3.2 of radiusclient), but it was
123 contributed by a Radius user.
125 PH/24 When doing a callout, the value of $domain wasn't set correctly when
126 expanding the "port" option of the smtp transport.
128 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
129 while reading a MIME header. Thanks to Tom Hughes for a patch.
131 PH/24 Include config.h inside local_scan.h so that configuration settings are
134 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
135 that in an ACL for RCPT (for example), you can examine exactly what was
138 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
139 commands, but it was not correctly comparing the address with the actual
140 client host address. Thus, it would show the EHLO address in Received:
141 header lines when this was not necessary.
143 PH/27 Added the % operator to ${eval:}.
145 PH/28 Exim tries to create and chdir to its spool directory when it starts;
146 it should be ignoring failures (because with -C, for example, it has lost
147 privilege). It wasn't ignoring creation failures other than "already
150 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
153 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
154 because an input line was too long, either on its own, or by virtue of
155 too many continuations, the temporary file was not being removed, and the
156 return code was incorrect.
158 PH/31 Missing "BOOL" in function definition in filtertest.c.
160 PH/32 Applied Sieve patches from the maintainer.
162 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
164 PH/33 Added "verify = not_blind".
166 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
167 Local/Makefile (with some defaults set). These are used in built scripts
168 such as exicyclog, but they have never been used in the exim_install
169 script (though there are many overriding facilities there). I have
170 arranged that the exim_install script now takes note of these two
173 PH/35 Installed configuration files for Dragonfly.
175 PH/36 When a locally submitted message by a trusted user did not contain a
176 From: header, and the sender address was obtained from -f or from an SMTP
177 MAIL command, and the trusted user did not use -F to supply a sender
178 name, $originator_name was incorrectly used when constructing a From:
179 header. Furthermore, $originator_name was used for submission mode
180 messages from external hosts without From: headers in a similar way,
181 which is clearly wrong.
183 PH/37 Added control=suppress_local_fixups.
185 PH/38 When log_selector = +received_sender was set, and the addition of the
186 sender made the log line's construction buffer exactly full, or one byte
187 less than full, an overflow happened when the terminating "\n" was
190 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
191 when the result of a list match is failure because a DNS lookup failed.
193 PH/40 RM_COMMAND is now used in the building process.
195 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
196 the "build-* directories that it finds.
198 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
199 address in a domain literal was a prefix of an interface address.
201 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
202 when verifying a sender address, unless rewrite_headers is false.
204 PH/44 Wrote a long comment about why errors_to addresses are verified as
205 recipients, not senders.
207 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
208 the ratelimit ACL was added.
210 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
212 PH/46 Added extra information about PostgreSQL errors to the error string.
214 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
215 avoiding the need to install special zones in a real server. This is
216 backwards compatible; if it can't find the fake resolver, it drops back.
217 Thus, both old and new test suites can be run.
219 TF/02 Added util/ratelimit.pl
221 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
222 clock is set back in time.
224 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
225 Candler <B.Candler@pobox.com>.
227 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
228 OK for addresses that are the result of redirections.
230 PH/48 A number of further additions for the benefit of the new test suite,
231 including a fake gethostbyname() that interfaces to the fake DNS resolver
234 TF/06 The fix for widen_domains has also been applied to qualify_single and
235 search_parents which are the other dnslookup options that can cause
238 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
241 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
243 TF/07 Exim produced the error message "an SRV record indicated no SMTP
244 service" if it encountered an MX record with an empty target hostname.
245 The message is now "an MX or SRV record indicated no SMTP service".
247 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
248 if the DNS of the sending site is misconfigured. This is quite a
249 common situation. This change restores the behaviour of treating a
250 helo verification defer as a failure.
252 PH/51 If self=fail was set on a router, the bounce message did not include the
253 actual error message.
259 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
261 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
262 command itself ended in error, the underlying message about the transport
263 filter timeout was being overwritten with the pipe command error. Now the
264 underlying error message should be appended to the second error message.
266 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
268 PH/02 Reduce the amount of output that "make" produces by default. Full output
269 can still be requested.
271 PH/03 The warning log line about a condition test deferring for a "warn" verb
272 was being output only once per connection, rather than after each
273 occurrence (because it was using the same function as for successful
274 "warn" verbs). This seems wrong, so I have changed it.
276 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
277 it should not have, which might have caused a crash in the right
278 circumstances, but probably never did.
280 PH/04 Installed a modified version of Tony Finch's patch to make submission
281 mode fix the return path as well as the Sender: header line, and to
282 add a /name= option so that you can make the user's friendly name appear
285 TF/03 Added the control = fakedefer ACL modifier.
287 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
288 Mark Lowes for thorough testing.
290 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
292 TK/03 Merged latest SRS patch from Miles Wilton.
294 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
295 with the definition in sysexits.h (which is #included earlier).
296 Fortunately, Exim does not actually use EX_OK. The code used to try to
297 preserve the sysexits.h value, by assumimg that macro definitions were
298 scanned for macro replacements. I have been disabused of this notion,
299 so now the code just undefines EX_OK before #including unistd.h.
301 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
302 in the smtp transport. When a block could not be written in a single
303 write() function, the timeout was being re-applied to each part-write.
304 This seems wrong - if the receiver was accepting one byte at a time it
305 would take for ever. The timeout is now adjusted when this happens. It
306 doesn't have to be particularly precise.
308 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
309 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
311 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
312 without a domain if the check to <postmaster@domain> fails.
314 SC/01 Eximstats: added -xls and the ability to specify output files
315 (patch written by Frank Heydlauf).
317 SC/02 Eximstats: use FileHandles for outputing results.
319 SC/03 Eximstats: allow any combination of xls, txt, and html output.
321 SC/04 Eximstats: fixed display of large numbers with -nvr option
323 SC/05 Eximstats: fixed merging of reports with empty tables.
325 SC/06 Eximstats: added the -include_original_destination flag
327 SC/07 Eximstats: removed tabs and trailing whitespace.
329 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
331 TK/06 MBOX spool code: Add real "From " MBOX separator line
332 so the .eml file is really in mbox format (even though
333 most programs do not really care). Patch from Alex Miller.
335 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
336 The latter is generated from $received_to and is only set if the
337 message has one envelope recipient. SA can use these headers,
338 obviously out-of-the-box. Patch from Alex Miller.
340 PH/08 The ${def test on a variable was returning false if the variable's
341 value was "0", contrary to what the specification has always said!
342 The result should be true unless the variable is empty.
344 PH/09 The syntax error of a character other than { following "${if
345 def:variable_name" (after optional whitespace) was not being diagnosed.
346 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
347 accidental colon was present, for example, could give incorrect results.
349 PH/10 Tidied the code in a number of places where the st_size field of a stat()
350 result is used (not including appendfile, where other changes are about
353 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
354 This involved changing a lot of size variables from int to off_t. It
355 should work with maildirs and everything.
357 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
358 spamd dying while we are connected to it.
360 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
363 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
364 function for BDB 4.3.
366 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
368 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
369 the sources for PCRE have been re-arranged and more files are now
372 PH/15 The code I had for printing potentially long long variables in PH/11
373 above was not the best (it lost precision). The length of off_t variables
374 is now inspected at build time, and an appropriate printing format (%ld
375 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
376 to be "long long int" or "long int". This is needed for the internal
377 formatting function string_vformat().
379 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
380 the configuration file to be ":syslog", then the script "guesses" where
381 the logs files are, rather than using the compiled in default. In our
382 case the guess is not the same as the compiled default, so the script
383 suddenly stopped working when I started to use syslog. The patch checks
384 to see if log_file_path is "". If so, it attempts to read it from exim
385 with no configuration file to get the compiled in version, before it
386 falls back to the previous guessing code."
388 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
389 implementing BATV in an Exim configuration. See NewStuff for the gory
392 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
393 Makefile that are specific to HP-UX.
395 PH/18 If the "use_postmaster" option was set for a recipient callout together
396 with the "random" option, the postmaster address was used as the MAIL
397 FROM address for the random test, but not for the subsequent recipient
398 test. It is now used for both.
400 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
401 patch removes a few documentation additions to RFC 3028, because the
402 latest draft now contains them. It adds the new en;ascii-case comparator
403 and a new error check for 8bit text in MIME parts. Comparator and
404 require names are now matched exactly. I enabled the subaddress
405 extension, but it is not well tested yet (read: it works for me)."
407 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
408 rework some of the code of TK/09 above to avoid the hardwired use of
409 "%lld" and "long long". Replaced the call to snprintf() with a call to
412 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
413 records point to non-existent hosts", "retry timeout exceeded", and
414 "retry time not reached for any host after a long failure period".
416 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
417 experimental DomainKeys support:
419 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
420 (2) On an error such as an illegally used "control", the wrong name for
421 the control was given.
423 These problems did NOT occur unless DomainKeys support was compiled.
425 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
427 PH/24 Added ${if match_ip condition.
429 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
430 databases so that it will be absolutely obvious if a crash occurs in the
431 DB library. This is a regular occurrence (often caused by mis-matched
434 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
435 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
436 picked up on a user's system that detects such things. There doesn't seem
437 to be a gcc warning option for this - only an attribute that has to be
438 put on the function's prototype. It seems that in Fedora Core 4 they have
439 set this on a number of new functions. No doubt there will be more in due
442 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
443 specify a transport. However, if an address that was verified by such a
444 router was the subject of a callout, Exim crashed because it tried to
445 read the rcpt_include_affixes from the non-existent transport. Now it
446 just assumes that the setting of that option is false. This bug was
447 introduced by 4.51/PH/31.
449 PH/28 Changed -d+all to exclude +memory, because that information is very
450 rarely of interest, but it makes the output a lot bigger. People tend to
451 do -d+all out of habit.
453 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
454 code in os-type was giving problems when libc.so lives in lib64, like on
457 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
458 aren't the modern standard, and it seems that some systems' include files
459 don't always have them. Exim was already checking for some of the newer
460 ones like T_AAAA, and defining it itself. I've added checks for all the
461 record types that Exim uses.
463 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
464 not automatically generating a new one, as it is supposed to. This
465 prevented TLS from working. If the file did exist, but contained invalid
466 data, a new version was generated, as expected. It was only the case of a
467 non-existent file that was broken.
469 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
470 with a change in libdomainkeys > 0.64.
472 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
473 from DNS. If the selector record carries the flag, it now has
474 precedence over the domain-wide flag.
476 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
478 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
479 the use of a port name that isn't defined in /etc/services) occurred, the
480 message was deferred as in a normal delivery, and thus remained on the
481 spool, instead of being failed because of the mua_wrapper setting. This
482 is now fixed, and I tidied up some of the mua_wrapper messages at the
485 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
486 the messages in a hash of arrays rather than using individual hashes.
487 This is a bit cleaner and results in dramatic memory savings, albeit
488 at a slight CPU cost.
490 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
491 as requested by Marc Sherman.
493 SC/10 Eximstats: added histograms for user specified patterns as requested
496 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
498 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
499 fopen() in the content-scanning modules that did not already have it.
505 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
506 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
508 TK/02 Fix ACL "control" statement not being available in MIME ACL.
510 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
512 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
513 to test Sieve filters that use "vacation".
515 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
516 that changes the way the GnuTLS parameters are stored in the cache file.
517 The new format can be generated externally. For backward compatibility,
518 if the data in the cache doesn't make sense, Exim assumes it has read an
519 old-format file, and it generates new data and writes a new file. This
520 means that you can't go back to an older release without removing the
523 PH/03 A redirect router that has both "unseen" and "one_time" set does not
524 work if there are any delivery delays because "one_time" forces the
525 parent to be marked "delivered", so its unseen clone is never tried
526 again. For this reason, Exim now forbids the simultaneous setting of
529 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
530 redirected to themselves ("homonym" addresses). Read the long ChangeLog
531 entry if you want to know the details. The fix, however, neglected to
532 consider the case when local delivery batching is involved. The test for
533 "previously delivered" was not happening when checking to see if an
534 address could be batched with a previous (undelivered) one; under
535 certain circumstances this could lead to multiple deliveries to the same
538 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
539 in its include files, and this causes problems building Exim.
541 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
542 header_syntax) but Exim was just ignoring anything given after a slash.
543 In particular, this caused confusion with an attempt to use "verify =
544 reverse_host_lookup/defer_ok". An error is now given when options are
545 supplied for verify items that do not have them. (Maybe reverse_host_
546 lookup should have a defer_ok option, but that's a different point.)
548 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
549 defined by RFC 821) to 2048, because there were problems with some AUTH
550 commands, and RFC 1869 says the size should be increased for extended
551 SMTP commands that take arguments.
553 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
556 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
557 "unknown" error; now it says that the functionality isn't in the binary.
559 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
560 an address' error message when a string expansion fails (syntax or
561 whatever). Otherwise the password may appear in the log. Following change
562 PH/42 below, there is no longer a chance of it appearing in a bounce
565 PH/11 Installed exipick version 20050225.0 from John Jetmore.
567 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
568 of its addresses was ever tried. (Bugzilla bug #2.)
570 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
571 the result incorrectly in the debug output. (It correctly added a newline
572 to what was transported.)
574 TF/01 Added $received_time.
576 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
577 commented out examples of how to interface to a virus scanner and to
578 SpamAssassin. Also added commented examples of av_scanner and
579 spamd_address settings.
581 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
582 and controls are allowed in which ACLs. There were a couple of minor
583 errors. Some of the entries in the conditions table (which is a table of
584 where they are NOT allowed) were getting very unwieldy; rewrote them as a
585 negation of where the condition IS allowed.
587 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
589 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
590 header file does not have a version number, so I've had to invent a new
591 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
592 API. The code is untested by me (my Linux distribution still has 0.3.2 of
593 radiusclient), but it was contributed by a Radius user.
595 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
596 files or directories, for OpenSSL.
598 PH/19 When an Exim process that is running as root has to create an Exim log
599 file, it does so in a subprocess that runs as exim:exim so as to get the
600 ownership right at creation (otherwise, other Exim processes might see
601 the file with the wrong ownership). There was no test for failure of this
602 fork() call, which would lead to the process getting stuck as it waited
603 for a non-existent subprocess. Forks do occasionally fail when resources
604 run out. I reviewed all the other calls to fork(); they all seem to check
607 PH/20 When checking for unexpected SMTP input at connect time (before writing
608 the banner), Exim was not dealing correctly with a non-positive return
609 from the read() function. If the client had disconnected by this time,
610 the result was a log entry for a synchronization error with an empty
611 string after "input=" when read() returned zero. If read() returned -1
612 (an event I could not check), uninitialized data bytes were printed.
613 There were reports of junk text (parts of files, etc) appearing after
616 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
618 PH/22 Added support for macro redefinition, and (re)definition in between
619 driver and ACL definitions.
621 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
622 forgetting to use the resulting value; it was using the unexpanded value.
624 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
625 hadn't been configured. The fix is from Juergen Kreileder, who
626 understands it better than I do:
628 "Here's what I see happening with three configured cyrus_sasl
629 authenticators configured (plain, login, cram-md5):
631 On startup auth_cyrus_sasl_init() gets called for each of these.
632 This means three calls to sasl_listmech() without a specified mech_list.
633 => SASL tests which mechs of all available mechs actually work
634 => three warnings about OTP not working
635 => the returned list contains: plain, login, cram-md5, digest-md5, ...
637 With the patch, sasl_listmech() also gets called three times. But now
638 SASL's mech_list option is set to the server_mech specified in the the
639 authenticator. Or in other words, the answer from sasl_listmech()
640 gets limited to just the mech you're testing for (which is different
642 => the return list contains just 'plain' or 'login', 'cram-md5' or
643 nothing depending on the value of ob->server_mech.
645 I've just tested the patch: Authentication still works fine,
646 unavailable mechs specified in the exim configuration are still
647 caught, and the auth.log warnings about OTP are gone."
649 PH/25 When debugging is enabled, the contents of the command line are added
650 to the debugging output, even when log_selector=+arguments is not
653 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
654 answer is "GNU", and only if the return is "GNU/something" is the answer
657 PH/27 $acl_verify_message is now set immediately after the failure of a
658 verification in an ACL, and so is available in subsequent modifiers. In
659 particular, the message can be preserved by coding like this:
661 warn !verify = sender
662 set acl_m0 = $acl_verify_message
664 Previously, $acl_verify_message was set only while expanding "message"
665 and "log_message" when a very denied access.
667 PH/28 Modified OS/os.c-Linux with
669 -#ifndef OS_LOAD_AVERAGE
670 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
672 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
673 nomenclature these days.)
675 PH/29 Installed patch from the Sieve maintainer that adds the options
676 sieve_useraddress and sieve_subaddress to the redirect router.
678 PH/30 In these circumstances:
679 . Two addresses routed to the same list of hosts;
680 . First host does not offer TLS;
681 . First host accepts first address;
682 . First host gives temporary error to second address;
683 . Second host offers TLS and a TLS session is established;
684 . Second host accepts second address.
685 Exim incorrectly logged both deliveries with the TLS parameters (cipher
686 and peerdn, if requested) that were in fact used only for the second
689 PH/31 When doing a callout as part of verifying an address, Exim was not paying
690 attention to any local part prefix or suffix that was matched by the
691 router that accepted the address. It now behaves in the same way as it
692 does for delivery: the affixes are removed from the local part unless
693 rcpt_include_affixes is set on the transport.
695 PH/32 Add the sender address, as F=<...>, to the log line when logging a
696 timeout during the DATA phase of an incoming message.
698 PH/33 Sieve envelope tests were broken for match types other than :is. I have
699 applied a patch sanctioned by the Sieve maintainer.
701 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
702 the uid or gid is negative. A case of a negative gid caused this to be
703 noticed. The fix allows for either to be negative.
705 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
706 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
709 PH/36 The change PH/12 above was broken. Fixed it.
711 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
712 the grounds that routing the same address twice would always produce the
713 same answer. This might have been true once, but it is certainly no
714 longer true now. Routing a child address may depend on the previous
715 routing that produced that child. Some complicated redirection strategies
716 went wrong when messages had multiple recipients, and made Exim's
717 behaviour dependent on the order in which the addresses were given.
719 I have moved the duplicate checking until after the routing is complete.
720 Exim scans the addresses that are assigned to local and remote
721 transports, and removes any duplicates. This means that more work will be
722 done, as duplicates will always all be routed, but duplicates are
723 presumably rare, so I don't expect this is of any significance.
725 For deliveries to pipes, files, and autoreplies, the duplicate checking
726 still happens during the routing process, since they are not going to be
729 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
730 It corrects a timeout issue with spamd. This is Ian's comment: "The
731 background is that sometimes spamd either never reads data from a
732 connection it has accepted, or it never writes response data. The exiscan
733 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
734 blindly assumes that writes won't block so it may never time out."
736 PH/39 Allow G after quota size as well as K and M.
738 PH/40 The value set for $authenticated_id in an authenticator may not contain
739 binary zeroes or newlines because the value is written to log lines and
740 to spool files. There was no check on this. Now the value is run through
741 the string_printing() function so that such characters are converted to
742 printable escape sequences.
744 PH/41 $message_linecount is a new variable that contains the total number of
745 lines in the message. Compare $body_linecount, which is the count for the
748 PH/42 Exim no longer gives details of delivery errors for specific addresses in
749 bounce and delay warning messages, except in certain special cases, which
752 (a) An SMTP error message from a remote host;
753 (b) A message specified in a :fail: redirection;
754 (c) A message specified in a "fail" command in a system filter;
755 (d) A message specified in a FAIL return from the queryprogram router;
756 (e) A message specified by the cannot_route_message router option.
758 In these cases only, Exim does include the error details in bounce and
759 warning messages. There are also a few cases where bland messages such
760 as "unrouteable address" or "local delivery error" are given.
762 PH/43 $value is now also set for the "else" part of a ${run expansion.
764 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
765 being worked on, but at least Exim now implements the latest version to
768 PH/45 In a pipe transport, although a timeout while waiting for the pipe
769 process to complete was treated as a delivery failure, a timeout while
770 writing the message to the pipe was logged, but erroneously treated as a
771 successful delivery. Such timeouts include transport filter timeouts. For
772 consistency with the overall process timeout, these timeouts are now
773 treated as errors, giving rise to delivery failures by default. However,
774 there is now a new Boolean option for the pipe transport called
775 timeout_defer, which, if set TRUE, converts the failures into defers for
776 both kinds of timeout. A transport filter timeout is now identified in
779 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
780 systems where "make" and "gmake" are different, calling "gmake" at top
781 level broke things. I've arranged for the value of $(MAKE) to be passed
782 from the Makefile to this script so that it can call the same version of
786 A note about Exim versions 4.44 and 4.50
787 ----------------------------------------
789 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
790 changes of various kinds. As a consequence, a big documentation update was
791 needed. This delayed the release for rather longer than seemed good, especially
792 in the light of a couple of (minor) security issues. Therefore, the changes
793 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
794 release. So 4.44 and 4.50 are in effect two different branches that both start
797 I have left the 4.50 change log unchanged; it contains all the changes since
798 4.43. The change log for 4.44 is below; many of its items are identical to
799 those for 4.50. This seems to be the most sensible way to preserve the
800 historical information.
806 1. Minor wording change to the doc/README.SIEVE file.
808 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
809 computation of the current number of files was incorrect.
811 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
812 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
815 4. Give more explanation in the error message when the command for a transport
816 filter fails to execute.
818 5. There are several places where Exim runs a non-Exim command in a
819 subprocess. The SIGUSR1 signal should be disabled for these processes. This
820 was being done only for the command run by the queryprogram router. It is
821 now done for all such subprocesses. The other cases are: ${run, transport
822 filters, and the commands run by the lmtp and pipe transports.
824 6. Added CONFIGURE_GROUP build-time option.
826 7. Some older OS have a limit of 256 on the maximum number of file
827 descriptors. Exim was using setrlimit() to set 1000 as a large value
828 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
829 systems. I've change it so that if it can't get 1000, it tries for 256.
831 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
832 was an oversight, and furthermore, ever since the addition of extra
833 controls (e.g. 4.43/32), the checks on when to allow different forms of
834 "control" were broken. There should now be diagnostics for all cases when a
835 control that does not make sense is encountered.
837 9. Added the /retain_sender option to "control=submission".
839 10. $recipients is now available in the predata ACL (oversight).
841 11. Tidy the search cache before the fork to do a delivery from a message
842 received from the command line. Otherwise the child will trigger a lookup
843 failure and thereby defer the delivery if it tries to use (for example) a
844 cached ldap connection that the parent has called unbind on.
846 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
847 of $address_data from the recipient verification was clobbered by the
850 13. The value of address_data from a sender verification is now available in
851 $sender_address_data in subsequent conditions in the ACL statement.
853 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
855 15. Added a new option "connect=<time>" to callout options, to set a different
858 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
859 was its contents. (It was OK if the option was not defined at all.)
861 17. A "Completed" log line is now written for messages that are removed from
862 the spool by the -Mrm option.
864 18. New variables $sender_verify_failure and $recipient_verify_failure contain
865 information about exactly what failed.
867 19. Added -dd to debug only the daemon process.
869 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
870 handles timeouts, both on the server side and network timeouts. Renamed the
871 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
873 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
874 It is now set to "smtps".
876 22. $host_address is now set to the target address during the checking of
879 23. When checking ignore_target_hosts for an ipliteral router, no host name was
880 being passed; this would have caused $sender_host_name to have been used if
881 matching the list had actually called for a host name (not very likely,
882 since this list is usually IP addresses). A host name is now passed as
885 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
886 code that specifies a non-restarting handler (typically sigaction() in
887 modern systems) in an attempt to fix a rare and obscure crash bug.
889 25. Narrowed the window for a race in the daemon that could cause it to ignore
890 SIGCHLD signals. This is not a major problem, because they are used only to
891 wake it up if nothing else does.
893 26. A malformed maildirsize file could cause Exim to calculate negative values
894 for the mailbox size or file count. Odd effects could occur as a result.
895 The maildirsize information is now recalculated if the size or filecount
898 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
899 support for a long time. Removed HAVE_SYS_VFS_H.
901 28. Installed the latest version of exipick from John Jetmore.
903 29. In an address list, if the pattern was not a regular expression, an empty
904 subject address (from a bounce message) matched only if the pattern was an
905 empty string. Non-empty patterns were not even tested. This was the wrong
906 because it is perfectly reasonable to use an empty address as part of a
907 database query. An empty address is now tested by patterns that are
908 lookups. However, all the other forms of pattern expect the subject to
909 contain a local part and a domain, and therefore, for them, an empty
910 address still always fails if the pattern is not itself empty.
912 30. Exim went into a mad DNS loop when attempting to do a callout where the
913 host was specified on an smtp transport, and looking it up yielded more
916 31. Re-factored the code for checking spool and log partition space into a
917 function that finds that data and another that does the check. The former
918 is then used to implement four new variables: $spool_space, $log_space,
919 $spool_inodes, and $log_inodes.
921 32. The RFC2047 encoding function was originally intended for short strings
922 such as real names; it was not keeping to the 75-character limit for
923 encoded words that the RFC imposes. It now respects the limit, and
924 generates multiple encoded words if necessary. To be on the safe side, I
925 have increased the buffer size for the ${rfc2047: expansion operator from
928 33. It is now permitted to omit both strings after an "if" condition; if the
929 condition is true, the result is "true". As before, when the second string
930 is omitted, a false condition yields an empty string. This makes it less
931 cumbersome to write custom ACL and router conditions.
933 34. Failure to deliver a bounce message always caused it to be frozen, even if
934 there was an errors_to setting on the router. The errors_to setting is now
937 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
938 canonical form (fully expanded) before being placed in
939 $sender_host_address.
941 36. The table in the code that translates DNS record types into text (T_A to
942 "A" for instance) was missing entries for NS and CNAME. It is just possible
943 that this could have caused confusion if both these types were looked up
944 for the same domain, because the text type is used as part of Exim's
945 per-process caching. But the chance of anyone hitting this buglet seems
948 37. The dnsdb lookup has been extended in a number of ways.
950 (1) There is a new type, "zns", which walks up the domain tree until it
951 finds some nameserver records. It should be used with care.
953 (2) There is a new type, "mxh", which is like "mx" except that it returns
954 just the host names, not the priorities.
956 (3) It is now possible to give a list of domains (or IP addresses) to be
957 looked up. The behaviour when one of the lookups defers can be
958 controlled by a keyword.
960 (4) It is now possible to specify the separator character for use when
961 multiple records are returned.
963 38. The dnslists ACL condition has been extended: it is now possible to supply
964 a list of IP addresses and/or domains to be looked up in a particular DNS
967 39. Added log_selector=+queue_time_overall.
969 40. When running the queue in the test harness, wait just a tad after forking a
970 delivery process, to get repeatability of debugging output.
972 41. Include certificate and key file names in error message when GnuTLS fails
973 to set them up, because the GnuTLS error message doesn't include the name
974 of the failing file when there is a problem reading it.
976 42. Allow both -bf and -bF in the same test run.
978 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
980 44. The "Exiscan patch" is now merged into the mainline Exim source.
982 45. Sometimes the final signoff response after QUIT could fail to get
983 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
984 < 0 before doing a fflush(). This bug looks as though it goes back to the
985 introduction of TLS in release 3.20, but "sometimes" must have been rare
986 because the tests only now provoked it.
988 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
989 (this can affect the format of dates).
991 47. exim_tidydb, when checking for the continued existence of a message for
992 which it has found a message-specific retry record, was not finding
993 messages that were in split spool directories. Consequently, it was
994 deleting retry records that should have stayed in existence.
996 48. Steve fixed some bugs in eximstats.
998 49. The SPA authentication driver was not abandoning authentication and moving
999 on to the next authenticator when an expansion was forced to fail,
1000 contradicting the general specification for all authenticators. Instead it
1001 was generating a temporary error. It now behaves as specified.
1003 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
1004 (the order specifies the preference for clients). The order is now AES256,
1005 AES128, 3DES, ARCFOUR128.
1007 51. Small patch to Sieve code - explicitly set From: when generating an
1010 52. Exim crashed if a remote delivery caused a very long error message to be
1011 recorded - for instance if somebody sent an entire SpamAssassin report back
1012 as a large number of 550 error lines. This bug was coincidentally fixed by
1013 increasing the size of one of Exim's internal buffers (big_buffer) that
1014 happened as part of the Exiscan merge. However, to be on the safe side, I
1015 have made the code more robust (and fixed the comments that describe what
1018 53. Now that there can be additional text after "Completed" in log lines (if
1019 the queue_time_overall log selector is set), a one-byte patch to exigrep
1020 was needed to allow it to recognize "Completed" as not the last thing in
1023 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1024 patch that reportedly fixes this has been added. I am not expert enough to
1025 create a test for it. This is what the patch creator wrote:
1027 "I found a little strange behaviour of ldap code when working with
1028 Windows 2003 AD Domain, where users was placed in more than one
1029 Organization Units. When I tried to give exim partial DN, the exit code
1030 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1031 But simultaneously result of request was absolutely normal ldap result,
1032 so I produce this patch..."
1034 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1035 REFERENCE, so I have modified the code to exclude the patch when that macro
1038 55. Some experimental protocols are using DNS PTR records for new purposes. The
1039 keys for these records are domain names, not reversed IP addresses. The
1040 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1041 leaves it alone. Component reversal etc. now happens only for IP addresses.
1044 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1046 57. Double the size of the debug message buffer (to 2048) so that more of very
1047 long debug lines gets shown.
1049 58. The exicyclog utility now does better if the number of log files to keep
1050 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1052 59. Two changes related to the smtp_active_hostname option:
1054 (1) $smtp_active_hostname is now available as a variable.
1055 (2) The default for smtp_banner uses $smtp_active_hostname instead
1056 of $primary_hostname.
1058 60. The host_aton() function is supposed to be passed a string that is known
1059 to be a valid IP address. However, in the case of IPv6 addresses, it was
1060 not checking this. This is a hostage to fortune. Exim now panics and dies
1061 if the condition is not met. A case was found where this could be provoked
1062 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1063 components; fortuitously, this particular loophole had already been fixed
1064 by change 4.50/55 above.
1066 If there are any other similar loopholes, the new check in host_aton()
1067 itself should stop them being exploited. The report I received stated that
1068 data on the command line could provoke the exploit when Exim was running as
1069 exim, but did not say which command line option was involved. All I could
1070 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1071 running as the user.
1074 61. There was a buffer overflow vulnerability in the SPA authentication code
1075 (which came originally from the Samba project). I have added a test to the
1076 spa_base64_to_bits() function which I hope fixes it.
1079 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1080 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1082 63. The daemon start-up calls getloadavg() while still root for those OS that
1083 need the first call to be done as root, but it missed one case: when
1084 deliver_queue_load_max is set with deliver_drop_privilege. This is
1085 necessary for the benefit of the queue runner, because there is no re-exec
1086 when deliver_drop_privilege is set.
1088 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1089 This has been fixed.
1091 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1092 was in use, was not putting the data itself into the right store pool;
1093 consequently, it could be overwritten for a subsequent message in the same
1094 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1097 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1099 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1100 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1101 the function were treating the return as a boolean value, which happened to
1102 work because 0=false and not-0=true, but is not correct code.
1104 68. The host_aton() function was not handling scoped IPv6 addresses (those
1105 with, for example, "%eth0" on the end) correctly.
1107 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1108 negated items (that is, ~something) in unsigned ints. Some compilers
1109 apparently mutter when there is no cast.
1111 70. If an address verification called from an ACL failed, and did not produce a
1112 user-specific message (i.e. there was only a "system" message), nothing was
1113 put in $acl_verify_message. In this situation, it now puts the system
1116 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1117 session; change 4.31/43 added the unwanted input to the log line - except
1118 that it did not do this in the start of session case. It now does.
1120 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1121 This is wrong and can cause the other end to generate a synchronization
1122 error if it is another Exim or anything else that does the synchronization
1123 check. A QUIT command is no longer sent after a timeout.
1125 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1126 during host lookups.
1128 74. The defer_ok option of callout verification was not working if it was used
1129 when verifying addresses in header lines, that is, for this case:
1131 verify = header_sender/callout=defer_ok
1133 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1134 those file descriptors could be used for SMTP connections. If anything
1135 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1136 could be sent to the SMTP client, causing chaos. The daemon now opens
1137 stdin, stdout, and stderr to /dev/null when it puts itself into the
1140 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1141 log by default. The user can override this with suitable Perl magic.
1143 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1144 the log message when discard triggers, was not working for the DATA ACL or
1145 for the non-SMTP ACL.
1147 78. Error message wording change in sieve.c.
1149 79. If smtp_accept_max_per_host was set, the number of connections could be
1150 restricted to fewer than expected, because the daemon was trying to set up
1151 a new connection before checking whether the processes handling previous
1152 connections had finished. The check for completed processes is now done
1153 earlier. On busy systems, this bug wouldn't be noticed because something
1154 else would have woken the daemon, and it would have reaped the completed
1157 80. If a message was submitted locally by a user whose login name contained one
1158 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1159 It caused a spool format error. I have fixed the spool reading code. A
1160 related problem was that the "from" clause in the Received: line became
1161 illegal because of the space(s). It is now covered by ${quote_local_part.
1163 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1166 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1168 83. Patch from Sieve maintainer to fix unterminated string problem in
1169 "vacation" handling.
1171 84. Some minor changes to the Linux configuration files to help with other
1172 OS variants using glibc.
1174 85. One more patch for Sieve to update vacation handling to latest spec.
1177 ----------------------------------------------------
1178 See the note above about the 4.44 and 4.50 releases.
1179 ----------------------------------------------------
1185 1. Change 4.43/35 introduced a bug that caused file counts to be
1186 incorrectly computed when quota_filecount was set in an appendfile
1189 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1190 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1193 3. Give more explanation in the error message when the command for a transport
1194 filter fails to execute.
1196 4. There are several places where Exim runs a non-Exim command in a
1197 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1198 was being done only for the command run by the queryprogram router. It is
1199 now done for all such subprocesses. The other cases are: ${run, transport
1200 filters, and the commands run by the lmtp and pipe transports.
1202 5. Some older OS have a limit of 256 on the maximum number of file
1203 descriptors. Exim was using setrlimit() to set 1000 as a large value
1204 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1205 systems. I've change it so that if it can't get 1000, it tries for 256.
1207 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1208 was an oversight, and furthermore, ever since the addition of extra
1209 controls (e.g. 4.43/32), the checks on when to allow different forms of
1210 "control" were broken. There should now be diagnostics for all cases when a
1211 control that does not make sense is encountered.
1213 7. $recipients is now available in the predata ACL (oversight).
1215 8. Tidy the search cache before the fork to do a delivery from a message
1216 received from the command line. Otherwise the child will trigger a lookup
1217 failure and thereby defer the delivery if it tries to use (for example) a
1218 cached ldap connection that the parent has called unbind on.
1220 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1221 of $address_data from the recipient verification was clobbered by the
1222 sender verification.
1224 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1225 was its contents. (It was OK if the option was not defined at all.)
1227 11. A "Completed" log line is now written for messages that are removed from
1228 the spool by the -Mrm option.
1230 12. $host_address is now set to the target address during the checking of
1231 ignore_target_hosts.
1233 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1234 being passed; this would have caused $sender_host_name to have been used if
1235 matching the list had actually called for a host name (not very likely,
1236 since this list is usually IP addresses). A host name is now passed as
1239 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1240 code that specifies a non-restarting handler (typically sigaction() in
1241 modern systems) in an attempt to fix a rare and obscure crash bug.
1243 15. Narrowed the window for a race in the daemon that could cause it to ignore
1244 SIGCHLD signals. This is not a major problem, because they are used only to
1245 wake it up if nothing else does.
1247 16. A malformed maildirsize file could cause Exim to calculate negative values
1248 for the mailbox size or file count. Odd effects could occur as a result.
1249 The maildirsize information is now recalculated if the size or filecount
1252 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1253 support for a long time. Removed HAVE_SYS_VFS_H.
1255 18. Updated exipick to current release from John Jetmore.
1257 19. Allow an empty sender to be matched against a lookup in an address list.
1258 Previously the only cases considered were a regular expression, or an
1261 20. Exim went into a mad DNS lookup loop when doing a callout where the
1262 host was specified on the transport, if the DNS lookup yielded more than
1265 21. The RFC2047 encoding function was originally intended for short strings
1266 such as real names; it was not keeping to the 75-character limit for
1267 encoded words that the RFC imposes. It now respects the limit, and
1268 generates multiple encoded words if necessary. To be on the safe side, I
1269 have increased the buffer size for the ${rfc2047: expansion operator from
1272 22. Failure to deliver a bounce message always caused it to be frozen, even if
1273 there was an errors_to setting on the router. The errors_to setting is now
1276 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1277 canonical form (fully expanded) before being placed in
1278 $sender_host_address.
1280 24. Updated eximstats to version 1.33
1282 25. Include certificate and key file names in error message when GnuTLS fails
1283 to set them up, because the GnuTLS error message doesn't include the name
1284 of the failing file when there is a problem reading it.
1286 26. Expand error message when OpenSSL has problems setting up cert/key files.
1289 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1290 (this can affect the format of dates).
1292 28. exim_tidydb, when checking for the continued existence of a message for
1293 which it has found a message-specific retry record, was not finding
1294 messages that were in split spool directories. Consequently, it was
1295 deleting retry records that should have stayed in existence.
1297 29. eximstats updated to version 1.35
1298 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1299 1.35 - bugfix such that pie charts by volume are generated correctly
1301 30. The SPA authentication driver was not abandoning authentication and moving
1302 on to the next authenticator when an expansion was forced to fail,
1303 contradicting the general specification for all authenticators. Instead it
1304 was generating a temporary error. It now behaves as specified.
1306 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1307 (the order specifies the preference for clients). The order is now AES256,
1308 AES128, 3DES, ARCFOUR128.
1310 31. Small patch to Sieve code - explicitly set From: when generating an
1313 32. Exim crashed if a remote delivery caused a very long error message to be
1314 recorded - for instance if somebody sent an entire SpamAssassin report back
1315 as a large number of 550 error lines. This bug was coincidentally fixed by
1316 increasing the size of one of Exim's internal buffers (big_buffer) that
1317 happened as part of the Exiscan merge. However, to be on the safe side, I
1318 have made the code more robust (and fixed the comments that describe what
1321 33. Some experimental protocols are using DNS PTR records for new purposes. The
1322 keys for these records are domain names, not reversed IP addresses. The
1323 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1324 leaves it alone. Component reversal etc. now happens only for IP addresses.
1327 34. The host_aton() function is supposed to be passed a string that is known
1328 to be a valid IP address. However, in the case of IPv6 addresses, it was
1329 not checking this. This is a hostage to fortune. Exim now panics and dies
1330 if the condition is not met. A case was found where this could be provoked
1331 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1332 components; fortuitously, this particular loophole had already been fixed
1333 by change 4.50/55 or 4.44/33 above.
1335 If there are any other similar loopholes, the new check in host_aton()
1336 itself should stop them being exploited. The report I received stated that
1337 data on the command line could provoke the exploit when Exim was running as
1338 exim, but did not say which command line option was involved. All I could
1339 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1340 running as the user.
1343 35. There was a buffer overflow vulnerability in the SPA authentication code
1344 (which came originally from the Samba project). I have added a test to the
1345 spa_base64_to_bits() function which I hope fixes it.
1348 36. The daemon start-up calls getloadavg() while still root for those OS that
1349 need the first call to be done as root, but it missed one case: when
1350 deliver_queue_load_max is set with deliver_drop_privilege. This is
1351 necessary for the benefit of the queue runner, because there is no re-exec
1352 when deliver_drop_privilege is set.
1354 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1355 was in use, was not putting the data itself into the right store pool;
1356 consequently, it could be overwritten for a subsequent message in the same
1357 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1360 38. Sometimes the final signoff response after QUIT could fail to get
1361 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1362 < 0 before doing a fflush(). This bug looks as though it goes back to the
1363 introduction of TLS in release 3.20, but "sometimes" must have been rare
1364 because the tests only now provoked it.
1370 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1371 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1372 Now it returns an integer. A number of calls were still expecting a T/F
1373 return. Fortuitously, in all cases, the tests worked in OK situations,
1374 which is the norm. However, things would have gone wrong on any write
1375 failures on the smtp file descriptor. This function is used when sending
1376 messages over SMTP and also when doing verify callouts.
1378 2. When Exim is called to do synchronous delivery of a locally submitted
1379 message (the -odf or -odi options), it no longer closes stderr before doing
1382 3. Implemented the mua_wrapper option.
1384 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1386 5. Implemented the functions header_remove(), header_testname(),
1387 header_add_at_position(), and receive_remove_recipient(), and exported them
1390 6. If an ACL "warn" statement specified the addition of headers, Exim already
1391 inserted X-ACL-Warn: at the start if there was no header name. However, it
1392 was not making this test for the second and subsequent header lines if
1393 there were newlines in the string. This meant that an invalid header could
1394 be inserted if Exim was badly configured.
1396 7. Allow an ACL "warn" statement to add header lines at the start or after all
1397 the Received: headers, as well as at the end.
1399 8. Added the rcpt_4xx retry error code.
1401 9. Added postmaster_mailfrom=xxx to callout verification option.
1403 10. Added mailfrom=xxxx to the callout verification option, for verify=
1406 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1407 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1409 12. Inserted some casts to stop certain compilers warning when using pointer
1410 differences as field lengths or precisions in printf-type calls (mostly
1411 affecting debugging statements).
1413 13. Added optional readline() support for -be (dynamically loaded).
1415 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1416 same clock tick as a message's arrival, so that its received time was the
1417 same as the "first fail" time on the retry record, and that message
1418 remained on the queue past the ultimate address timeout, every queue runner
1419 would try a delivery (because it was past the ultimate address timeout) but
1420 after another failure, the ultimate address timeout, which should have then
1421 bounced the address, did not kick in. This was a "< instead of <=" error;
1422 in most cases the first failure would have been in the next clock tick
1423 after the received time, and all would be well.
1425 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1426 being recognized when the domain list was tested by the match_domain
1427 condition in an expansion string.
1429 16. Added the ${str2b64: operator.
1431 17. Exim was always calling setrlimit() to set a large limit for the number of
1432 processes, without checking whether the existing limit was already
1433 adequate. (It did check for the limit on file descriptors.) Furthermore,
1434 errors from getrlimit() and setrlimit() were being ignored. Now they are
1435 logged to the main and panic logs, but Exim does carry on, to try to do its
1436 job under whatever limits there are.
1438 18. Imported PCRE 5.0.
1440 19. Trivial typo in log message " temporarily refused connection" (the leading
1443 20. If the log selector return_path_on_delivery was set and an address was
1444 redirected to /dev/null, the delivery process crashed because it assumed
1445 that a return path would always be set for a "successful" delivery. In this
1446 case, the whole delivery is bypassed as an optimization, and therefore no
1449 21. Internal re-arrangement: the function for sending a challenge and reading
1450 a response while authentication was assuming a zero-terminated challenge
1451 string. It's now changed to take a pointer and a length, to allow for
1452 binary data in such strings.
1454 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1456 23. Exim was not respecting finduser_retries when seeking the login of the
1457 uid under which it was called; it was always trying 10 times. (The default
1458 setting of finduser_retries is zero.) Also, it was sleeping after the final
1459 failure, which is pointless.
1461 24. Implemented tls_on_connect_ports.
1463 25. Implemented acl_smtp_predata.
1465 26. If the domain in control=submission is set empty, Exim assumes that the
1466 authenticated id is a complete email address when it generates From: or
1467 Sender: header lines.
1469 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1470 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1471 chown and chgrp in /bin and hostname in /usr/bin.
1473 28. Exim was keeping the "process log" file open after each use, just as it
1474 does for the main log. This opens the possibility of it remaining open for
1475 long periods when the USR1 signal hits a daemon. Occasional processlog
1476 errors were reported, that could have been caused by this. Anyway, it seems
1477 much more sensible not to leave this file open at all, so that is what now
1480 29. The long-running daemon process does not normally write to the log once it
1481 has entered its main loop, and it closes the log before doing so. This is
1482 so that log files can straightforwardly be renamed and moved. However,
1483 there are a couple of unusual error situations where the daemon does write
1484 log entries, and I had neglected to close the log afterwards.
1486 30. The text of an SMTP error response that was received during a remote
1487 delivery was being truncated at 512 bytes. This is too short for some of
1488 the long messages that one sometimes sees. I've increased the limit to
1491 31. It is now possible to make retry rules that apply only when a message has a
1492 specific sender, in particular, an empty sender.
1494 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1495 it possible to be selective about when SMTP synchronization is enforced.
1497 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1499 32. Implemented hosts_connection_nolog.
1501 33. Added an ACL for QUIT.
1503 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1506 35. Added mailbox_size and mailbox_filecount to appendfile.
1508 36. Added control = no_multiline_responses to ACLs.
1510 37. There was a bug in the logic of the code that waits for the clock to tick
1511 in the case where the clock went backwards by a substantial amount such
1512 that the microsecond fraction of "now" was more than the microsecond
1513 fraction of "then" (but the whole seconds number was less).
1515 38. Added support for the libradius Radius client library this is found on
1516 FreeBSD (previously only the radiusclient library was supported).
1522 1. When certain lookups returned multiple values in the form name=value, the
1523 quoting of the values was not always being done properly. Specifically:
1524 (a) If the value started with a double quote, but contained no whitespace,
1526 (b) If the value contained whitespace other than a space character (i.e.
1527 tabs or newlines or carriage returns) it was not quoted.
1528 This fix has been applied to the mysql and pgsql lookups by writing a
1529 separate quoting function and calling it from the lookup code. The fix
1530 should probably also be applied to nisplus, ibase and oracle lookups, but
1531 since I cannot test any of those, I have not disturbed their existing code.
1533 2. A hit in the callout cache for a specific address caused a log line with no
1534 reason for rejecting RCPT. Now it says "Previous (cached) callout
1535 verification failure".
1537 3. There was an off-by-one bug in the queryprogram router. An over-long
1538 return line was truncated at 256 instead of 255 characters, thereby
1539 overflowing its buffer with the terminating zero. As well as fixing this, I
1540 have increased the buffer size to 1024 (and made a note to document this).
1542 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1543 when Exim is waiting for an SMTP response from a remote server, Exim
1544 restarts its select() call on the socket, thereby resetting its timeout.
1545 This is not a problem when such interrupts are rare. Somebody set up a cron
1546 job to run exiwhat every 2 minutes, which is less than the normal select()
1547 timeout (5 or 10 minutes). This meant that the select() timeout never
1548 kicked in because it was always reset. I have fixed this by comparing the
1549 time when an interrupt arrives with the time at the start of the first call
1550 to select(). If more time than the timeout has elapsed, the interrupt is
1551 treated as a timeout.
1553 5. Some internal re-factoring in preparation for the addition of Sieve
1554 extensions (by MH). In particular, the "personal" test is moved to a
1555 separate function, and given an option for scanning Cc: and Bcc: (which is
1556 not set for Exim filters).
1558 6. When Exim created an email address using the login of the caller as the
1559 local part (e.g. when creating a From: or Sender: header line), it was not
1560 quoting the local part when it contained special characters such as @.
1562 7. Installed new OpenBSD configuration files.
1564 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1565 try to make them clearer.
1567 9. Callout options, other than the timeout value, were being ignored when
1568 verifying sender addresses in header lines. For example, when using
1570 verify = header_sender/callout=no_cache
1572 the cache was (incorrectly) being used.
1574 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1575 only the Cygwin environment.
1577 11. When return_path_on_delivery was set as a log selector, if different remote
1578 addresses in the same message used different return paths and parallel
1579 remote delivery occurred, the wrong values would sometimes be logged.
1580 (Whenever a remote delivery process finished, the return path value from
1581 the most recently started remote delivery process was logged.)
1583 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1584 lines when AUTH and/or TLS are in use. This is the "received protocol"
1585 field. Exim used to use "asmtp" for authenticated SMTP, without any
1586 indication (in the protocol name) for TLS use. Now it follows the RFC and
1587 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1588 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1589 names appear in log lines as well as in Received: header lines.
1591 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1592 extensions, and comparison tests, and to fix some bugs.
1594 14. Changes to the "personal" filter test:
1596 (1) The test was buggy in that it was just doing the equivalent of
1597 "contains" tests on header lines. For example, if a user's address was
1598 anne@some.where, the "personal" test would incorrectly be true for
1600 To: susanne@some.where
1602 This test is now done by extracting each address from the header in turn,
1603 and checking the entire address. Other tests that are part of "personal"
1604 are now done using regular expressions (for example, to check local parts
1605 of addresses in From: header lines).
1607 (2) The list of non-personal local parts in From: addresses has been
1608 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1609 taken from the Sieve specification recommendations.
1611 (3) If the message contains any header line starting with "List-" it is
1612 treated as non-personal.
1614 (4) The test for "circular" in the Subject: header line has been removed
1615 because it now seems ill-conceived.
1617 15. Minor typos in src/EDITME comments corrected.
1619 16. Installed latest exipick from John Jetmore.
1621 17. If headers_add on a router specified a text string that was too long for
1622 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1623 of string_sprintf() is now avoided.
1625 18. $message_body_size was not set (it was always zero) when running the DATA
1626 ACL and the local_scan() function.
1628 19. For the "mail" command in an Exim filter, no default was being set for
1629 the once_repeat time, causing a random time value to be used if "once" was
1630 specified. (If the value happened to be <= 0, no repeat happened.) The
1631 default is now 0s, meaning "never repeat". The "vacation" command was OK
1632 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1633 (I found it when inspecting the code).
1635 20. There is now an overall timeout for performing a callout verification. It
1636 defaults to 4 times the callout timeout, which applies to individual SMTP
1637 commands during the callout. The overall timeout applies when there is more
1638 than one host that can be tried. The timeout is checked before trying the
1639 next host. This prevents very long delays if there are a large number of
1640 hosts and all are timing out (e.g. when the network connections are timing
1641 out). The value of the overall timeout can be changed by specifying an
1642 additional sub-option for "callout", called "maxwait". For example:
1644 verify = sender/callout=5s,maxwait=20s
1646 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1647 to the end before writing, but this should make it even safer).
1649 22. Exim was forgetting that it had advertised PIPELINING for the second and
1650 subsequent messages on an SMTP connection. It was also not resetting its
1651 memory on STARTTLS and an internal HELO.
1653 23. When Exim logs an SMTP synchronization error within a session, it now
1654 records whether PIPELINING has been advertised or not.
1656 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1657 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1658 rather than long int.
1660 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1662 26. Added the never_mail option to autoreply.
1668 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1669 crash if the getsockname() call failed; this can happen if a connection is
1670 closed very soon after it is established. The problem was simply in the
1671 order in which certain operations were done, causing Exim to try to write
1672 to the SMTP stream before it had set up the file descriptor. The bug has
1673 been fixed by making things happen in the correct order.
1679 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1680 before the connection was closed, thus losing the rejection response.
1682 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1683 some early Solaris releases, but causes trouble in current releases where
1684 socklen_t is defined.
1686 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1689 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1692 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1693 that releasing the top of it at the end releases what was used for sub-
1694 expansions (unless the block got too big). However, discard this block if
1695 the first thing is a variable or header, so that we can use its block when
1696 it is dynamic (useful for very large $message_headers, for example).
1698 6. Lookups now cache *every* query, not just the most recent. A new, separate
1699 store pool is used for this. It can be recovered when all lookup caches are
1700 flushed. Lookups now release memory at the end of their result strings.
1701 This has involved some general refactoring of the lookup sources.
1703 7. Some code has been added to the store_xxx() functions to reduce the amount
1704 of flapping under certain conditions.
1706 8. log_incoming_interface used to affect only the <= reception log lines. Now
1707 it causes the local interface and port to be added to several more SMTP log
1708 lines, for example "SMTP connection from", and rejection lines.
1710 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1712 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1714 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1715 could be overwritten at the end of the current message (or the start of a
1716 new message if it was set in a HELO ACL). The value is now preserved for
1717 the duration of the SMTP connection.
1719 12. If a transport had a headers_rewrite setting, and a matching header line
1720 contained an unqualified address, that address was qualified, even if it
1721 did not match any rewriting rules. The underlying bug was that the values
1722 of the flags that permit the existence of unqualified sender and recipient
1723 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1724 non-local messages, and by -bnq for local messages) were not being
1725 preserved with the message after it was received.
1727 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1728 "next input=" as part of the text comprising the host identity instead of
1729 the correct text. The code was using the same buffer for two different
1730 strings. However, depending on which order the printing function evaluated
1731 its arguments, the bug did not always show up. Under Linux, for example, my
1732 test suite worked just fine.
1734 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1735 doesn't work with some older versions of Perl. It has been changed to "my",
1736 which in any case is probably the better facility to use.
1738 15. A really picky compiler found some instances of statements for creating
1739 error messages that either had too many or two few arguments for the format
1742 16. The size of the buffer for calls to the DNS resolver has been increased
1743 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1744 for addresses that have a lot of PTR records. This alleviates a problem; it
1745 does not fully solve it.
1747 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1748 buffer now truncates the list and logs the incident, which is the same
1749 action as happens when Exim is looking up a host name and its aliases.
1750 Previously in this situation something unpredictable would happen;
1751 sometimes it was "internal error: store_reset failed".
1753 18. If a server dropped the connection unexpectedly when an Exim client was
1754 using GnuTLS and trying to read a response, the client delivery process
1755 crashed while trying to generate an error log message.
1757 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1758 string, for example:
1760 warn message = H1: something\nH2: something
1762 the text was added as a single header line from Exim's point of view
1763 though it ended up OK in the delivered message. However, searching for the
1764 second and subsequent header lines using $h_h2: did not work. This has been
1765 fixed. Similarly, if a system filter added multiple headers in this way,
1766 the routers could not see them.
1768 20. Expanded the error message when iplsearch is called with an invalid key to
1769 suggest using net-iplsearch in a host list.
1771 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1772 ACLs are no longer actually imposed (and a message to that effect is
1775 22. If a "gecos" field in a passwd entry contained escaped characters, in
1776 particular, if it contained a \" sequence, Exim got it wrong when building
1777 a From: or a Sender: header from that name. A second bug also caused
1778 incorrect handling when an unquoted " was present following a character
1779 that needed quoting.
1781 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1782 was not being matched caselessly.
1784 24. Arranged for all hyphens in the exim.8 source to be escaped with
1787 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1788 properly. Recipient callout cache records were still being keyed to include
1789 the sender, even when use_sender was set false. This led to far more
1790 callouts that were necessary. The sender is no longer included in the key
1791 when use_sender is false.
1793 26. Added "control = submission" modifier to ACLs.
1795 27. Added the ${base62d: operator to decode base 62 numbers.
1797 28. dnsdb lookups can now access SRV records.
1799 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1800 the configuration file.
1802 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1803 (-v) mode. This makes the output for a verbose queue run more intelligible.
1805 31. Added a use_postmaster feature to recipient callouts.
1807 32. Added the $body_zerocount variable, containing the number of binary zero
1808 bytes in the message body.
1810 33. The time of last modification of the "new" subdirectory is now used as the
1811 "mailbox time last read" when there is a quota error for a maildir
1814 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1816 35. Added +ignore_unknown as a special item in host lists.
1818 36. Code for decoding IPv6 addresses in host lists is now included, even if
1819 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1820 address was recognized as an IP address, but was then not correctly decoded
1821 into binary, causing unexpected and incorrect effects when compared with
1828 1. Very minor rewording of debugging text in manualroute to say "list of
1829 hosts" instead of "hostlist".
1831 2. If verify=header_syntax was set, and a header line with an unqualified
1832 address (no domain) and a large number of spaces between the end of the
1833 name and the colon was received, the reception process suffered a buffer
1834 overflow, and (when I tested it) crashed. This was caused by some obsolete
1835 code that should have been removed. The fix is to remove it!
1837 3. When running in the test harness, delay a bit after writing a bounce
1838 message to get a bit more predictability in the log output.
1840 4. Added a call to search_tidyup() just before forking a reception process. In
1841 theory, someone could use a lookup in the expansion of smtp_accept_max_
1842 per_host which, without the tidyup, could leave open a database connection.
1844 5. Added the variables $recipient_data and $sender_data which get set from a
1845 lookup success in an ACL "recipients" or "senders" condition, or a router
1846 "senders" option, similar to $domain_data and $local_part_data.
1848 6. Moved the writing of debug_print from before to after the "senders" test
1851 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1852 problems for message scanning, either using a data ACL, or using
1853 local_scan() because the Received: header was not generated till after they
1854 were called (in order to set the time as the time of reception completion).
1855 I have revised the way this works. The header is now generated after the
1856 body is received, but before the ACL or local_scan() are called. After they
1857 are run, the timestamp in the header is updated.
1863 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1864 before starting a queue runner without re-exec. This happened only when
1865 deliver_drop_privilege was set or when the Exim user was set to root. The
1866 effect of the bug was that timeouts during subsequent deliveries caused
1867 crashes instead of being properly handled. The handler is now left at its
1868 default (and expected) setting.
1870 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1871 message, again when deliver_drop_privilege is set or Exim is run as root.
1872 The bug described in (1) was not present in this case, but the tidying up
1873 of the other signals was missing. I have made the two cases consistent.
1875 3. The ignore_target_hosts setting on a manualroute router was being ignored
1876 for hosts that were looked up using the /MX notation.
1878 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1881 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1882 operated on the sender address. After changing the $sender_address to <>
1883 for the sender address verify, Exim was re-instated it as the original
1884 (before rewriting) address, but remembering that it had rewritten it, so it
1885 wasn't rewriting it again. This bug also had the effect of breaking the
1886 sender address verification caching when the sender address was rewritten.
1888 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1889 This has been changed so that if the ip literal address matches
1890 ignore_target_hosts, the router declines.
1892 7. Added expansion conditions match_domain, match_address, and match_local_
1893 part (NOT match_host).
1895 8. The placeholder for the Received: header didn't have a length field set.
1897 9. Added code to Exim itself and to exim_lock to test for a specific race
1898 condition that could lead to file corruption when using MBX delivery. The
1899 issue is with the lockfile that is created in /tmp. If this file is removed
1900 after a process has opened it but before that process has acquired a lock,
1901 there is the potential for a second process to recreate the file and also
1902 acquire a lock. This could lead to two Exim processes writing to the file
1903 at the same time. The added code performs the same test as UW imapd; it
1904 checks after acquiring the lock that its file descriptor still refers to
1905 the same named file.
1907 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1908 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1911 11. Added the smtp_active_hostname option. If used, this will typically be made
1912 to depend on the incoming interface address. Because $interface_address is
1913 not set up until the daemon has forked a reception process, error responses
1914 that can happen earlier (such as "too many connections") no longer contain
1917 12. If an expansion in a condition on a "warn" statement fails because a lookup
1918 defers, the "warn" statement is abandoned, and the next ACL statement is
1919 processed. Previously this caused the whole ACL to be aborted.
1921 13. Added the iplsearch lookup type.
1923 14. Added ident_timeout as a log selector.
1925 15. Added tls_certificate_verified as a log selector.
1927 16. Added a global option tls_require_ciphers (compare the smtp transport
1928 option of the same name). This controls incoming TLS connections.
1930 17. I finally figured out how to make tls_require_ciphers do a similar thing
1931 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1932 before starting the TLS session.
1934 18. Tabs are now shown as \t in -bP output.
1936 19. If the log selector return_path_on_delivery was set, Exim crashed when
1937 bouncing a message because it had too many Received: header lines.
1939 20. If two routers both had headers_remove settings, and the first one included
1940 a superfluous trailing colon, the final name in the first list and the
1941 first name in the second list were incorrectly joined into one item (with a
1942 colon in the middle).
1948 1. Added -C and -D options to the exinext utility, mainly to make it easier
1949 to include in the automated testing, but these could be helpful when
1950 multiple configurations are in use.
1952 2. The exinext utility was not formatting the output nicely when there was
1953 an alternate port involved in the retry record key, nor when there was a
1954 message id as well (for retries that were specific to a specific message
1955 and a specific host). It was also confused by IPv6 addresses, because of
1956 the additional colons they contain. I have fixed the IPv4 problem, and
1957 patched it up to do a reasonable job for IPv6.
1959 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1960 delivery, the log line now contains "pipelined" if PIPELINING was used.
1962 4. An SMTP transport process used to panic and die if the bind() call to set
1963 an explicit outgoing interface failed. This has been changed; it is now
1964 treated in the same way as a connect() failure.
1966 5. A reference to $sender_host_name in the part of a conditional expansion
1967 that was being skipped was still causing a DNS lookup. This no longer
1970 6. The def: expansion condition was not recognizing references to header lines
1971 that used bh_ and bheader_.
1973 7. Added the _cache feature to named lists.
1975 8. The code for checking quota_filecount in the appendfile transport was
1976 allowing one more file than it should have been.
1978 9. For compatibility with Sendmail, the command line option
1986 and sets the incoming protocol and host name (for trusted callers). The
1987 host name and its colon can be omitted when only the protocol is to be set.
1988 Note the Exim already has two private options, -pd and -ps, that refer to
1989 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1990 "s", but I don't think that's a major issue.
1992 10. A number of refactoring changes to the code, none of which should affect
1995 (a) The number of logging options was getting close to filling up the
1996 32-bit word that was used as a bit map. I have split them into two classes:
1997 those that are passed in the argument to log_write(), and those that are
1998 only ever tested independently outside of that function. These are now in
1999 separate 32-bit words, so there is plenty of room for expansion again.
2000 There is no change in the user interface or the logging behaviour.
2002 (b) When building, for example, log lines, the code previously used a
2003 macro that called string_cat() twice, in order to add two strings. This is
2004 not really sufficiently general. Furthermore, there was one instance where
2005 it was actually wrong because one of the argument was used twice, and in
2006 one call a function was used. (As it happened, calling the function twice
2007 did not affect the overall behaviour.) The macro has been replaced by a
2008 function that can join an arbitrary number of extra strings onto a growing
2011 (c) The code for expansion conditions now uses a table and a binary chop
2012 instead of a serial search (which was left over from when there were very
2013 few conditions). Also, it now recognizes conditions like "pam" even when
2014 the relevant support is not compiled in: a suitably worded error message is
2015 given if an attempt is made to use such a condition.
2017 11. Added ${time_interval:xxxxx}.
2019 12. A bug was causing one of the ddress fields not to be passed back correctly
2020 from remote delivery subprocesses. The field in question was not being
2021 subsequently used, so this caused to problems in practice.
2023 13. Added new log selectors queue_time and deliver_time.
2025 14. Might have fixed a bug in maildirsizefile handling that threw up
2026 "unexpected character" debug warnings, and recalculated the data
2027 unnecessarily. In any case, I expanded the warning message to give more
2030 15. Added the message "Restricted characters in address" to the statements in
2031 the default ACL that block characters like @ and % in local parts.
2033 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2034 Three changes have been made:
2036 (a) There was a serious bug; a negative response to MAIL caused the whole
2037 recipient domain to be cached as invalid, thereby blocking all messages
2038 to all local parts at the same domain, from all senders. This bug has
2039 been fixed. The domain is no longer cached after a negative response to
2040 MAIL if the sender used is not empty.
2042 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2045 (c) A new callout option, "use_sender" has been added for people who want
2046 the modified behaviour.
2052 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2055 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2056 indeed breaks things for older releases.
2058 3. Added additional logging to the case where there is a problem reading data
2059 from a filter that is running in a subprocess using a pipe, in order to
2060 try to track down a specific problem.
2062 4. Testing facility fudge: when running in the test harness and attempting
2063 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2064 getting "No route to host". Convert this to a timeout.
2066 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2069 6. Some OS don't have socklen_t but use size_t instead. This affects the
2070 fifth argument of getsockopt() amongst other things. This is now
2071 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2072 can be set for individual OS. I have set it for SunOS5, OSF1, and
2073 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2074 some earlier ones do not.
2076 7. Change 4.30/15 was not doing the test caselessly.
2078 8. The standard form for an IPv6 address literal was being rejected by address
2079 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2080 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2081 this, as well as the form without the "IPv6" on the front (but only when
2082 address literals are enabled, of course).
2084 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2086 10. Exim crashed if a message with an empty sender address specified by -f
2087 encountered a router with an errors_to setting. This could be provoked only
2088 by a command such as
2092 where an empty string was supplied; "<>" did not hit this bug.
2094 11. Installed PCRE release 4.5.
2096 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2097 remained set. It is now erased.
2099 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2100 times from message ids (which are base 36 rather than the normal 62).
2102 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2103 were being counted as actual protocol errors, and logged if the log
2104 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2105 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2106 and DATA following a set of rejected RCPTs do not count as protocol errors.
2107 In other words, Exim assumes they were pipelined, though this may not
2108 actually be the case. Of course, in all cases the client gets an
2109 appropriate error code.
2111 15. If a lookup fails in an ACL condition, a message about the failure may
2112 be available; it is used if testing the ACL cannot continue, because most
2113 such messages specify what the cause of the deferral is. However, some
2114 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2115 that caused an old message to be retained and used if a later statement
2116 caused a defer, replacing the real cause of the deferral.
2118 16. If an IP address had so many PTR records that the DNS lookup buffer
2119 was not large enough to hold them, Exim could crash while trying to process
2120 the truncated data. It now detects and logs this case.
2122 17. Further to 4.21/58, another change has been made: if (and only if) the
2123 first line of a message (the first header line) ends with CRLF, a bare LF
2124 in a subsequent header line has a space inserted after it, so as not to
2125 terminate the header.
2127 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2128 unnecessarily, used atoi() instead of strtol(), and didn't check the
2129 termination when getting file sizes from file names by regex.
2131 19. Completely re-implemented the support for maildirsize files, in the light
2132 of a number of problems with the previous contributed implementation
2133 (4.30/29). In particular:
2135 . If the quota is zero, the maildirsize file is maintained, but no quota is
2138 . If the maildir directory does not exist, it is created before any attempt
2139 to write a maildirsize file.
2141 . The quota value in the file is just a cache; if the quota is changed in
2142 the transport, the new value overrides.
2144 . A regular expression is available for excluding directories from the
2147 20. The autoreply transport checks the characters in options that define the
2148 message's headers; it allows continued headers, but it was checking with
2149 isspace() after an embedded newline instead of explicitly looking for a
2152 21. If all the "regular" hosts to which an address was routed had passed their
2153 expiry times, and had not reached their retry times, the address was
2154 bounced, even if fallback hosts were defined. Now Exim should go on to try
2157 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2158 equivalent code in the SMTP transport. Some hosts send humungous responses
2159 to HELO/EHLO, more than 1024 it seems.
2161 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2162 gives compiler warnings in some environments. I've now done it "properly",
2165 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2166 (because of problems with the built-in one) was declared to return uschar *
2167 instead of char *, causing compiler failure.
2169 25. Fixed a file descriptor leak when processing alias/forward files.
2171 26. Fixed a minor format string issue in dbfn.c.
2173 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2175 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2176 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2179 29. When a sender address is verified, it is cached, to save repeating the test
2180 when there is more than one recipient in a message. However, when the
2181 verification involves a callout, it is possible for different callout
2182 options to be set for different recipients. It is too complicated to keep
2183 track of this in the cache, so now Exim always runs a verification when a
2184 callout is required, relying on the callout cache for the optimization.
2185 The overhead is duplication of the address routing, but this should not be
2188 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2189 to be verified with callout=postmaster, and the main callout worked but the
2190 postmaster check failed, the verification correctly failed. However, if a
2191 subsequent RCPT command asked for sender verification *without* the
2192 postmaster check, incorrect caching caused this verification also to fail,
2195 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2196 it was not caching the DNS options (qualify_single, search_parents) that
2197 were used when the lookup failed. A subsequent lookup with different
2198 options therefore always gave the same answer, though there were cases
2199 where it should not have. (Example: a "domains = !$mx_any" option on a
2200 dnslookup router: the "domains" option is always processed without any
2201 widening, but the router might have qualify_single set.) Now Exim uses the
2202 cached value only when the same options are set.
2204 32. Added John Jetmore's "exipick" utility to the distribution.
2206 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2207 than a timeout (e.g. a certificate is required, and is not provided), an
2208 Exim server now closes the connection immediately. Previously it waited for
2209 the client to close - but if the client is SSL, it seems that they each
2210 wait for each other, leading to a delay before one of them times out.
2212 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2213 maintained 0.8.x compatibility because I don't think many are using it, and
2214 it is clearly obsolete.
2216 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2219 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2220 client certificate was expired. A simple patch fixes this, though I don't
2221 understand the full logic of why the verify callback is called multiple
2224 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2225 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2226 which causes problems with some clients (such as the Certicom SSL Plus
2227 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2228 disables the coutermeasure allowing Eudora to connect."
2230 38. Exim was not checking that a write() to a log file succeeded. This could
2231 lead to Bad Things if a log got too big, in particular if it hit a file
2232 size limit. Exim now panics and dies if it cannot write to a log file, just
2233 as it does if it cannot open a log file.
2235 39. Modified OS/Makefile-Linux so that it now contains
2237 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2239 The two -D definitions ensure that Exim is compiled with large file
2240 support, which makes it possible to handle log files that are bigger than
2243 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2244 instance) a domain was checked against a named list that involved a lookup,
2245 causing $domain_data to be set, then another domain was checked against the
2246 same list, then the first domain was re-checked, the value of $domain_data
2247 after the final check could be wrong. In particular, if the second check
2248 failed, it could be set empty. This bug probably also applied to
2251 41. The strip_trailing_dot option was not being applied to the address given
2252 with the -f command-line option.
2254 42. The code for reading a message's header from the spool was incrementing
2255 $received_count, but never initializing it. This meant that the value was
2256 incorrect (doubled) while delivering a message in the same process in which
2257 it was received. In the most common configuration of Exim, this never
2258 happens - a fresh exec is done - but it can happen when
2259 deliver_drop_privilege is set.
2261 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2262 it now includes up to 150 characters of the unexpected data in the log
2265 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2266 and building data strings. The size of both of these buffers was 10 000
2267 bytes - far larger than anybody would *ever* want, thought I. Needless to
2268 say, somebody hit the limit. I have increased the maximum line length to
2269 20 000 and the maximum data length of concatenated lines to 100 000. I have
2270 also fixed two bugs, because there was no checking on these buffers. Tsk,
2271 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2272 buffer is too small.
2274 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2275 lsearch lookups. Now it does.
2277 46. When parsing a route_list item in a manualroute router, a fixed-length
2278 buffer was used for the list of hosts. I made this 1024 bytes long,
2279 thinking that nobody would ever have a list of hosts that long. Wrong.
2280 Somebody had a whole pile of complicated expansion conditions, and the
2281 string was silently truncated, leading to an expansion error. It turns out
2282 that it is easier to change to an unlimited length (owing to other changes
2283 that have happened since this code was originally written) than to build
2284 structure for giving a limitation error. The length of the item that
2285 expands into the list of hosts is now unlimited.
2287 47. The lsearch lookup could not handle data where the length of text line was
2288 more than 4095 characters. Such lines were truncated, leading to shortened
2289 data being returned. It should now handle lines of any length.
2291 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2292 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2295 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2296 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2297 no longer happens in comment lines. A list of replacements is now placed
2298 at the head of all of the source files, except those whose only change is
2299 to replace PERL_COMMAND in the very first #! line.
2301 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2302 messages on the queue, with a bottom-up merge sort, using code contributed
2303 by Michael Haardt. This should make operations like -bp somewhat faster on
2304 large queues. It won't affect queue runners, except when queue_run_in_order
2307 51. Installed eximstats 1.31 in the distribution.
2309 52. Added support for SRV lookups to the dnslookup router.
2311 53. If an ACL referred to $message_body or $message_body_end, the value was not
2312 reset for any messages that followed in the same SMTP session.
2314 54. The store-handling optimization for building very long strings was not
2315 differentiating between the different store pools. I don't think this
2316 actually made any difference in practice, but I've tidied it.
2318 55. While running the routers to verify a sender address, $sender_address
2319 was still set to the sender address. This is wrong, because when routing to
2320 send a bounce to the sender, it would be empty. Therefore, I have changed
2321 it so that, while verifying a sender address, $sender_address is set to <>.
2322 (There is no change to what happens when verifying a recipient address.)
2324 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2325 target A or AAAA records (if not already returned) without resetting the
2326 qualify_single or search_parents options of the DNS resolver. These are
2327 inappropriate in this case because the targets of MX and SRV records must
2328 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2329 target that, when qualified, matched something in the local domain. These
2330 two options are now turned off when doing these lookups.
2332 57. It seems that at least some releases of Reiserfs (which does not have the
2333 concept of a fixed number of inodes) returns zero and not -1 for the
2334 number of available inodes. This interacted badly with check_spool_inodes,
2335 which assumed that -1 was the "no such thing" setting. What I have done is
2336 to check that the total number of inodes is greater than zero before doing
2337 the test of how many are available.
2339 58. When a "warn" ACL statement has a log_message modifier, the message is
2340 remembered, and not repeated. This is to avoid a lot of repetition when a
2341 message has many recipients that cause the same warning to be written.
2342 Howewer, Exim was preserving the list of already written lines for an
2343 entire SMTP session, which doesn't seem right. The memory is now reset if a
2344 new message is started.
2346 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2347 debugging output unless log_rewrite was also set.
2349 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2350 of (int)(handle) when we know that handle contains (void *)(-1).
2352 61. The Exim daemon panic-logs an error return when it closes the incoming
2353 connection. However "connection reset by peer" seems to be common, and
2354 isn't really an error worthy of noting specially, so that particular error
2357 62. When Exim is trying to find all the local interfaces, it used to panic and
2358 die if the ioctl to get the interface flags failed. However, it seems that
2359 on at least one OS (Solaris 9) it is possible to have an interface that is
2360 included in the list of interfaces, but for which you get a failure error
2361 for this call. This happens when the interface is not "plumbed" into a
2362 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2363 failure of the "get flags" call assumes that the interface is down.
2365 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2366 makes life easier for people who are doing arithmetic on fields extracted
2367 from dates, where you often get leading zeros that should not be
2368 interpreted as octal.
2370 64. Added qualify_domain to the redirect router, to override the global
2373 65. If a pathologically long header line contained very many addresses (the
2374 report of this problem mentioned 10 000) and each of them was rewritten,
2375 Exim could use up a very large amount of memory. (It kept on making new
2376 copies of the header line as it rewrote, and never released the old ones.)
2377 At the expense of a bit more processing, the header rewriting function has
2378 been changed so that it no longer eats memory in this way.
2380 66. The generation of the Received: header has been moved from the time that a
2381 message starts to be received, to the time that it finishes. The timestamp
2382 in the Received: header should now be very close to that of the <= log
2383 line. There are two side-effects of this change:
2385 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2386 logged header lines no longer include the local Received: line, because
2387 it has not yet been created. The same applies to a copy of the message
2388 that is returned to a non-SMTP sender when a message is rejected.
2390 (b) When a filter file is tested using -bf, no additional Received: header
2391 is added to the test message. After some thought, I decided that this
2394 This change does not affect the value of $received_for. It is still set
2395 after address rewriting, but before local_scan() is called.
2397 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2399 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2400 gave an unhelpful panic error message, and a defer error. I have managed to
2401 change this behaviour so that it now rejects any supplied certificate,
2402 which seems right, as the list of acceptable certificates is empty.
2404 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2405 gave an unhelpful defer error. I have not managed to make this reject any
2406 supplied certificates, but the error message it gives is "no certificate
2407 supplied", which is not helpful.
2409 70. exigrep's output now also includes lines that are not associated with any
2410 message, but which match the given pattern. Implemented by a patch from
2411 Martin Sluka, which also tidied up the Perl a bit.
2413 71. Recipient callout verification, like sender verification, was using <> in
2414 the MAIL FROM command. This isn't really the right thing, since the actual
2415 sender may affect whether the remote host accepts the recipient or not. I
2416 have changed it to use the actual sender in the callout; this means that
2417 the cache record is now keyed on a recipient/sender pair, not just the
2418 recipient address. There doesn't seem to be a real danger of callout loops,
2419 since a callout by the remote host to check the sender would use <>.
2420 [SEE ABOVE: changed after hitting problems.]
2422 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2423 temporary errors. However, in the case of such a code being given after
2424 the end of a data transmission (i.e. after ".") Exim was failing to write
2425 a retry record for the message. (Yes, there was some broken host that was
2426 actually sending 8xx at this point.)
2428 73. An unknown lookup type in a host list could cause Exim to panic-die when
2429 the list was checked. (An example that provoked this was putting <; in the
2430 middle of a list instead of at the start.) If this happened during a DATA
2431 ACL check, a -D file could be left lying around. This kind of configuration
2432 error no longer causes Exim to die; instead it causes a defer errror. The
2433 incident is still logged to the main and panic logs.
2435 74. Buglet left over from Exim 3 conversion. The message "too many messages
2436 in one connection" was written to the rejectlog but not the mainlog, except
2437 when address rewriting (yes!) was being logged.
2439 75. Added write_rejectlog option.
2441 76. When a system filter was run not as root (that is, when system_filter_user
2442 was set), the values of the $n variables were not being returned to the
2443 main process; thus, they were not subsequently available in the $sn
2446 77. Added +return_path_on_delivery log selector.
2448 78. A connection timeout was being treated differently from recipients deferred
2449 when testing hosts_max_try with a message that was older than the host's
2450 retry timeout. (The host should not be counted, thus allowing all hosts to
2451 be tried at least once before bouncing.) This may have been the cause of an
2452 occasionally reported bug whereby a message would remain on the queue
2453 longer than the retry timeout, but would be bounced if a delivery was
2454 forced. I say "may" because I never totally pinned down the problem;
2455 setting up timeout/retry tests is difficult. See also the next item.
2457 79. The ultimate address timeout was not being applied to errors that involved
2458 a combination of host plus message (for example, a timeout on a MAIL
2459 command). When an address resolved to a number of possible hosts, and they
2460 were not all tried for each delivery (e.g. because of hosts_max_try), a
2461 message could remain on the queue longer than the retry timeout.
2463 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2466 81. Fixed an obscure SMTP outgoing bug which required at least the following
2467 conditions: (a) there was another message waiting for the same server;
2468 (b) the server returned 5xx to all RCPT commands in the first message so
2469 that the message was not completed; (c) the server dropped the connection
2470 or gave a negative response to the RSET that Exim sends to abort the
2471 transaction. The observed case was a dropped connection after DATA that had
2472 been sent in pipelining mode. That is, the server had advertised PIPELINING
2473 but was not implementing it correctly. The effect of the bug was incorrect
2474 behaviour, such as trying another host, and this could lead to a crash.
2480 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2481 and daemon.c were passed as pointers to ints; they should have been
2482 pointers to socklen_t variables (which are typically unsigned ints).
2484 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2487 3. Fixed a really odd bug that affected only the testing scheme; patching a
2488 certain fixed string in the binary changed the value of another string that
2489 happened to be identical to the end of the original first string.
2491 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2492 name", it returns that address as the IP address. On some operating
2493 systems (e.g. Solaris), it also passes back the IP address string as the
2494 "host name". However, on others (e.g. Linux), it passes back an empty
2495 string. Exim wasn't checking for this, and was changing the host name to an
2496 empty string, assuming it had been canonicized.
2498 5. Although rare, it is permitted to have more than one PTR record for a given
2499 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2500 all the names associated with an address, because they do in Solaris.
2501 However, it seems that they do not in Linux for data that comes from the
2502 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2503 I found this out when I moved to a new Linux workstation and tried to run
2504 the Exim test suite.
2506 To get round this problem I have changed the code so that it now does its
2507 own call to the DNS to look up PTR records when searching for a host name.
2508 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2509 addresses that are only in /etc/hosts are still found.
2511 This behaviour is, however, controlled by an option called host_lookup_
2512 order, which defaults to "bydns:byaddr". If people want to use the other
2513 order, or indeed, just use one or the other means of lookup, they can
2514 specify it in this variable.
2516 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2517 some operating systems, this comes back from gethostbyaddr() as an empty
2518 string, and this is what Exim used to test for. However, it seems that in
2519 other systems, "." is yielded. Exim now tests for this case too.
2521 7. The values of check_spool_space and check_log_space are now held internally
2522 as a number of kilobytes instead of an absolute number of bytes. If a
2523 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2524 kilobyte. This means that much larger values can be stored.
2526 8. Exim monitor: an attempt to get the action menu when not actually pointing
2527 at a message produces an empty menu entitled "No message selected". This
2528 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2529 no entries in it ("Shell widget menu has zero width and/or height"). So I
2530 have added a single, blank menu entry in this case.
2532 9. Added ${quote_local_part.
2534 10. MIME decoding is now applied to the contents of Subject: header lines when
2537 11. Now that a reference to $sender_host_address automatically causes a reverse
2538 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2539 host lookup before query-style lookups in lists that might use this
2540 variable. This has therefore been abolished, and the "net-" prefix is no
2541 longer necessary for query-style lookups.
2543 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2544 have been a typo for LFLAGS, so it has been changed.
2546 13. The install script calls Exim with "-C /dev/null" in order to find the
2547 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2548 to be output. Howeve, since Exim outputs its version number before the
2549 error, it didn't break the script. It just looked ugly. I fixed this by
2550 always allowing "-C /dev/null" if the caller is root.
2552 14. Ignore overlarge ACL variable number when reading spool file - insurance
2553 against a later release with more variables having written the file.
2555 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2556 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2557 this, as well as the form without the "IPv6" on the front.
2559 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2560 OS/Makefile-Darwin file.
2562 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2563 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2565 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2566 variable. However, some OS use ints for inodes. I've added cast to long int
2567 to get rid of the compiler warning.
2569 19. I had forgotten to lock out "/../" in configuration file names when
2570 ALT_CONFIG_PREFIX was set.
2572 20. Routers used for verification do not need to specify transports. However,
2573 if such a router generated a host list, and callout was configured, Exim
2574 crashed, because it could not find a port number from the (non-existent)
2575 transport. It now assumes port 25 in this circumstance.
2577 21. Added the -t option to exigrep.
2579 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2580 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2581 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2582 them is set, LOOKUP_LSEARCH is forced.
2584 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2586 24. Added sender and host information to the "rejected by local_scan()" log
2587 line; previously there was no indication of these.
2589 25. Added .include_if_exists.
2591 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2592 Linux. It turns out that not all file systems support this. Apparently some
2593 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2594 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2595 not supported on the file descriptor, is now ignored when Exim is trying to
2596 sync a directory. This applies only to Linux.
2598 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2600 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2601 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2604 29. Added support for maildirsize files from supplied patch (modified a bit).
2606 30. The use of :fail: followed by an empty string could lead Exim to respond to
2607 sender verification failures with (e.g.):
2609 550 Verification failed for <xxx>
2610 550 Sender verify failed
2612 where the first response line was missing the '-' that indicates it is not
2613 the final line of the response.
2615 31. The loop for finding the name of the user that called Exim had a hardwired
2616 limit of 10; it now uses the value of finduser_retries, which is used for
2617 all other user lookups.
2619 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2622 33. Exim was neglecting to zero errno before one call of strtol() when
2623 expanding a string and expecting an integer value. On some systems this
2624 resulted in spurious "integer overflow" errors. Also, it was casting the
2625 result into an int without checking.
2627 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2628 did not work. The code looks as if it has *never* worked, though it appears
2629 to have been documented since at least releast 1.62. I have made it work.
2631 35. The "timeout_DNS" error in retry rules, also documented since at least
2632 1.62, also never worked. As it isn't clear exactly what this means, and
2633 clearly it isn't a major issue, I have abolished the feature by treating it
2634 as "timeout", and writing a warning to the main and panic logs.
2636 36. The display of retry rules for -brt wasn't always showing the error code
2639 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2640 timeout_connect_A, timeout_connect_MX.
2642 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2643 to the empty sender.
2645 39. The daemon was not analysing the content of -oX till after it had closed
2646 stderr and disconnected from the controlling terminal. This meant that any
2647 syntax errors were only noted on the panic log, and the return code from
2648 the command was 0. By re-arranging the code a little, I've made the
2649 decoding happen first, so such errors now appear on stderr, and the return
2650 code is 1. However, the actual setting up of the sockets still happens in
2651 the disconnected process, so errors there are still only recorded on the
2654 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2655 connections (as happens on some IP stacks) was logged at start up time as
2656 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2657 it from "IPv6 and IPv4", which means that two separate sockets are being
2660 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2661 says whether AF_INET or AF_INET6 was passed as an argument.
2663 42. Exiwhat output was messed up when time zones were included in log
2666 43. Exiwhat now gives more information about the daemon's listening ports,
2667 and whether -tls-on-connect was used.
2669 44. The "port" option of the smtp transport is now expanded.
2671 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2672 silently ignored. Now an error message is written to the main and panic
2675 46. There's a new ACL modifier called "logwrite" which writes to a log file
2676 as soon as it is encountered.
2678 47. Added $local_user_uid and $local_user_gid at routing time.
2680 48. Exim crashed when trying to verify a sender address that was being
2683 49. Exim was recognizing only a space character after ".include". It now also
2684 recognizes a tab character.
2686 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2687 extracting the relevant information from the specification. The man page no
2688 longer contains scrambled data for the -d option, and I've added a section
2689 at the front about calling Exim under different names.
2691 51. Added "extra_headers" argument to the "mail" command in filter files.
2693 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2696 53. Installed eximstats 1.29.
2698 54. Added transport_filter_timeout as a generic transport option.
2700 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2701 Cc: header lines. This was required by RFC 822, but it not required by RFC
2704 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2705 incoming messages that did not have them. Now it does so only if the
2706 message originates locally, that is, if there is no associated remote host
2707 address. When Resent- header lines are present, this applies to the Resent-
2708 lines rather than the non-Resent- lines.
2710 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2711 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2713 58. Messages for configuration errors now include the name of the main
2714 configuration file - useful now that there may be more than one file in a
2715 list (.included file names were always shown).
2717 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2718 for those rare installations that do not start the daemon as root or run it
2719 setuid root. I've cut out the call to initgroups() if the daemon is not
2722 60. The Exim user and group can now be bound into the binary as text strings
2723 that are looked up at the start of Exim's processing.
2725 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2727 62. Added $mailstore_basename variable.
2729 63. Installed patch to sieve.c from Michael Haardt.
2731 64. When Exim failed to open the panic log after failing to open the main log,
2732 the original message it was trying to log was written to stderr and debug
2733 output, but if they were not available (the usual case in production), it
2734 was lost. Now it is written to syslog before the two lines that record the
2735 failures to open the logs.
2737 65. Users' Exim filters run in subprocesses under the user's uid. It is
2738 possible for a "deliver" command or an alias in a "personal" command to
2739 provoke an address rewrite. If logging of address rewriting is configured,
2740 this fails because the process is not running as root or exim. There may be
2741 a better way of dealing with this, but for the moment (because 4.30 needs
2742 to be released), I have disabled address rewrite logging when running a
2743 filter in a non-root, non-exim process.
2749 1. The buildconfig auxiliary program wasn't quoting the value set for
2750 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2751 not defined. This bug was masked in 4.22 by the effect that was fixed in
2754 2. Some messages that were rejected after a message id was allocated were
2755 shown as "incomplete" by exigrep. It no longer does this for messages that
2756 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2758 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2759 have allow_domain_literals set, the ID did not get logged in the <= line.
2760 Domain literals are now always recognized in Message-ID: header lines.
2762 4. The first argument for a ${extract expansion item is the key name or field
2763 number. Leading and trailing spaces in this item were not being ignored,
2764 causing some misleading effects.
2766 5. When deliver_drop_privilege was set, single queue runner processes started
2767 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2768 same command in the process it spins off) were not dropping privilege.
2770 6. When the daemon running as "exim" started a queue runner, it always
2771 re-executed Exim in the spun-off process. This is a waste of effort when
2772 deliver_drop_privilege is set. The new process now just calls the
2773 queue-runner function directly.
2779 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2782 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2783 ignored. Though the use of -oP was forcing the writing of a pid file, it
2784 was always written to the default place.
2786 3. If the message "no IP address found for host xxxx" is generated during
2787 incoming verification, it is now followed by identification of the incoming
2788 connection (so you can more easily find what provoked it).
2790 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2792 5. Added some features to "harden" Exim a bit more against certain attacks:
2794 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2795 be put in Local/Makefile. This is like the never_users runtime option,
2796 but it cannot be overridden. The default setting is "root".
2798 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2799 prefix string with which any file named in a -C command line option
2802 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2803 is retained for -C and -D only if the caller of Exim is root. Without
2804 it, the exim user may also use -C and -D and retain privilege.
2806 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2807 command line option is disabled.
2809 6. Macro names set by the -D option must start with an upper case letter, just
2810 like macro names defined in the configuration file.
2812 7. Added "dereference=" facility to LDAP.
2814 8. Two instances of the typo "uknown" in the source files are fixed.
2816 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2817 the Configure-Makefile script screwed up while processing it.
2819 10. Incorporated PCRE 4.4.
2821 11. The SMTP synchronization check was not operating right at the start of an
2822 SMTP session. For example, it could not catch a HELO sent before the client
2823 waited for the greeting. There is now a check for outstanding input at the
2824 point when the greeting is written. Because of the duplex, asynchronous
2825 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2826 way, but not yet received, when the check is performed.
2828 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2829 on TCP/IP sockets, because this apparently causes some broken clients to
2832 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2833 unchanged) from the Cygwin maintainer.
2835 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2836 was supported instead of testing for mbx. Effectively a typo.
2838 15. The spa authenticator server code was not checking that the input it
2839 received was valid base64.
2841 16. The debug output line for the "set" modifier in ACLs was not showing the
2842 name of the variable that was being set.
2844 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2846 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2847 lookup on its own. Something else was needed to trigger the lookup. For
2848 example, a match in host_lookup or the need for a host name in a host list.
2849 Now, if $sender_host_name is referenced and the host name has not yet been
2850 looked up, a lookup is performed. If the lookup fails, the variable remains
2851 empty, and $host_lookup_failed is set to "1".
2853 19. Added "eqi" as a case-independent comparison operator.
2855 20. The saslauthd authentication condition could segfault if neither service
2856 nor realm was specified.
2858 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2859 error message that was logged was misleading, and incoming SMTP
2860 connections were dropped. The message is now more accurate, and temporary
2861 errors are given to SMTP connections.
2863 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2864 (except RSET) with a 421 error, until QUIT is received. However, it was
2865 failing to send a response to QUIT.
2867 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2868 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2869 ACL is now run after the helo_try_verify_hosts code.
2871 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2872 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2873 case-independent, so other case variants are also recognized). Apparently
2874 some systems use these upper case variants.
2876 25. If more than two messages were waiting for the same host, and a transport
2877 filter was specified for the transport, Exim sent two messages over the
2878 same TCP/IP connection, and then failed with "socket operation on non-
2879 socket" when it tried to send the third.
2881 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2883 27. The extern definition of crypt16() in expand.c was not being excluded when
2884 the OS had its own crypt16() function.
2886 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2887 as a preferred synonym for return_size_limit, both as an option and as an
2890 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2892 30. Changed the default configuration ACL to relax the local part checking rule
2893 for addresses that are not in any local domains. For these addresses,
2894 slashes and pipe symbols are allowed within local parts, but the sequence
2895 /../ is explicitly forbidden.
2897 31. SPA server authentication was not clearing the challenge buffer before
2900 32. log_message in a "warn" ACL statement was writing to the reject log as
2901 well as to the main log, which contradicts the documentation and doesn't
2902 seem right (because no rejection is happening). So I have stopped it.
2904 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2905 However, I am unable to do any testing of this.
2907 34. Fixed an infelicity in the appendfile transport. When checking directories
2908 for a mailbox, to see if any needed to be created, it was accidentally
2909 using path names with one or more superfluous leading slashes; tracing
2910 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2912 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2913 contents are added to the log line that is written for every discarded
2914 recipient. (Previously a log_message setting was ignored.)
2916 36. The ${quote: operator now quotes the string if it is empty.
2918 37. The install script runs exim in order to find its version number. If for
2919 some reason other than non-existence or emptiness, which it checks, it
2920 could not run './exim', it was installing it with an empty version number,
2921 i.e. as "exim-". This error state is now caught, and the installation is
2924 38. An argument was missing from the function that creates an error message
2925 when Exim fails to connect to the socket for saslauthd authentication.
2926 This could cause Exim to crash, or give a corrupted message.
2928 39. Added isip, isip4, and isip6 to ${if conditions.
2930 40. The ACL variables $acl_xx are now saved with the message, and can be
2931 accessed later in routers, transports, and filters.
2933 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2934 strings in the file are not string-expanded.
2936 42. If a MAIL command specified a SIZE value that was too large to fit into an
2937 int variable, the check against message_size_limit failed. Such values are
2938 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2939 day this will have to be increased, but I don't think I want to be around
2940 when emails are that large.
2947 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2948 iconv() is not standard in FreeBSD.
2950 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2951 IPv6 enabled. The observed symptom was a segmentation fault on return from
2952 the function os_common_find_running_interfaces() in src/os.c.
2954 3. In the check_special_case() function in daemon.c I had used "errno" as an
2955 argument name, which causes warnings on some systems. This was basically a
2956 typo, since it was named "eno" in the comments!
2958 4. The code that waits for the clock to tick (at a resolution of some fraction
2959 of a second) so as to ensure message-id uniqueness was always waiting for
2960 at least one whole tick, when it could have waited for less. [This is
2961 almost certainly not relevant at current processor speeds, where it is
2962 unlikely to ever wait at all. But we try to future-proof.]
2964 5. The function that sleeps for a time interval that includes fractions of a
2965 second contained a race. It did not block SIGALRM between setting the
2966 timer, and suspending (a couple of lines later). If the interval was short
2967 and the sigsuspend() was delayed until after it had expired, the suspension
2968 never ended. On busy systems this could lead to processes getting stuck for
2971 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2972 process, before it forks any delivery processes. The open lookup caching
2973 mechanism meant that the open file or database connection was passed into
2974 the delivery process. The problem was that delivery processes always tidy
2975 up cached lookup data. This could cause a problem for the next delivery
2976 process started by the queue runner, because the external queue runner
2977 process does not know about the closure. So the next delivery process
2978 still has data in the lookup cache. In the case of a file lookup, there was
2979 no problem because closing a file descriptor in a subprocess doesn't affect
2980 the parent. However, if the lookup was caching a connection to a database,
2981 the connection was closed, and the second delivery process was likely to
2982 see errors such as "PGSQL: query failed: server closed the connection
2983 unexpectedly". The problem has been fixed by closing all cached lookups
2984 in a queue runner before running a delivery process.
2986 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2987 seem to have the "const" qualifier which it has on other OS. I've
2990 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2991 *of the same type* (client or server) with the same public name that an
2992 error should be diagnosed.
2994 9. When Exim looked up a host name for an IP address, but failed to find the
2995 original IP address when looking up the host name (a safety check), it
2996 output the message "<ip address> does not match any IP for NULL", which was
2997 confusing, to say the least. The bug was that the host name should have
2998 appeared instead of "NULL".
3000 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
3001 user that is built into the binary, and the -C or -D options is used, root
3002 privilege is dropped before the configuration file is read. In addition,
3003 logging is switched to stderr instead of the normal log files. If the
3004 configuration then re-defines the Exim user, the unprivileged environment
3005 is probably not what is expected, so Exim logs a panic warning message (but
3008 However, if deliver_drop_privilege is set, the unprivileged state may well
3009 be exactly what is intended, so the warning has been cut out in that case,
3010 and Exim is allowed to try to write to its normal log files.
3016 1. smtp_return_error_details was not giving details for temporary sender
3017 or receiver verification errors.
3019 2. Diagnose a configuration error if two authenticators have the same public
3022 3. Exim used not to create the message log file for a message until the first
3023 delivery attempt. This could be confusing when incoming messages were held
3024 for policy or load reasons. The message log file is now created at the time
3025 the message is received, and an initial "Received" line is written to it.
3027 4. The automatically generated man page for command line options had a minor
3028 bug that caused no ill effects; however, a more serious problem was that
3029 the procedure for building the man page automatically didn't always
3030 operate. Consequently, release 4.20 contains an out-of-date version. This
3031 shouldn't happen again.
3033 5. When building Exim with embedded Perl support, the script that builds the
3034 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3035 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3037 6. The freeze_tell option was not being used for messages that were frozen on
3038 arrival, either by an ACL or by local_scan().
3040 7. Added the smtp_incomplete_transaction log selector.
3042 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3043 was accepting AUTH without a new EHLO.
3045 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3046 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3049 10. Logging of TCP/IP connections (when configured) now happens in the main
3050 daemon process instead of the child process, so that the TCP/IP connection
3051 count is more accurate (but it can never be perfect).
3053 11. The use of "drop" in a nested ACL was not being handled correctly in the
3054 outer ACL. Now, if condition failure induced by the nested "drop" causes
3055 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3056 or "require"), the connection is dropped.
3058 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3059 that yield "discard" can now be used with an "accept" or a "discard" verb,
3060 but an error is generated for any others (because I can't see a useful way
3061 to define what should happen).
3063 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3064 are now processed in the same way as lines in the Exim configuration file.
3065 In particular, continuation lines are supported.
3067 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3069 15. Added -ti meaning -t -i.
3071 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3072 domains, and warn by logging if others are found.
3074 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3075 returned from GIFCONF when Exim is trying to find the list of interfaces on
3076 a host. The code in os.c has been modified to copy each ifreq to an aligned
3077 structure in all cases.
3079 Also, in some cases, the returned ifreq's were being copied to a 'struct
3080 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3081 means the last couple of bytes of an IPv6 address could be chopped if the
3082 ifreq contained only a normal sockaddr (14 bytes storage).
3084 18. Named domain lists were not supported in the hosts_treat_as_local option.
3085 An entry such as +xxxx was not recognized, and was treated as a literal
3088 19. Ensure that header lines added by a DATA ACL are included in the reject log
3089 if the ACL subsequently rejects the message.
3091 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3092 MD5 (which is deprecated).
3094 21. When testing a filter file using -bf, Exim was writing a message when it
3095 took the sender from a "From " line in the message, but it was not doing so
3096 when it took $return_path from a Return-Path: header line. It now does.
3098 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3099 with a valid header line field name (a series of printing characters
3100 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3102 23. Changed "disc" in the source to "disk" to conform to the documentation and
3103 the book and for uniformity.
3105 24. Ignore Sendmail's -Ooption=value command line item.
3107 25. When execve() failed while trying to run a command in a pipe transport,
3108 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3109 could be confused with a return value of 69 from the command itself. This
3110 has been changed to 127, the value the shell returns if it is asked to run
3111 a non-existent command. The wording for the related log line suggests a
3112 non-existent command as the problem.
3114 26. If received_header_text expands to an empty string, do not add a Received:
3115 header line to the message. (Well, it adds a token one on the spool, but
3116 marks it "old" so that it doesn't get used or transmitted.)
3118 27. Installed eximstats 1.28 (addition of -nt option).
3120 28. There was no check for failure on the call to getsockname() in the daemon
3121 code. This can fail if there is a shortage of resources on the system, with
3122 ENOMEM, for example. A temporary error is now given on failure.
3124 29. Contrary to the C standard, it seems that in some environments, the
3125 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3126 program. Exim now does this explicitly; it affects the formatting of
3127 timestamps using strftime().
3129 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3130 complaints. I've now initialized all the variables, to avoid this.
3132 32. Header lines added by a system filter were not being "seen" during
3133 transport-time rewrites.
3135 33. The info_callback() function passed to OpenSSL is set up with type void
3136 (*)(SSL *, int, int), as described somewhere. However, when calling the
3137 function (actually a macro) that sets it up, the type void(*)() is
3138 expected. I've put in a cast to prevent warnings from picky compilers.
3140 34. If a DNS black list lookup found a CNAME record, but there were no A
3141 records associated with the domain it pointed at, Exim crashed.
3143 35. If a DNS black list lookup returned more than one A record, Exim ignored
3144 all but the first. It now scans all returned addresses if a particular IP
3145 value is being sought. In this situation, the contents of the
3146 $dnslist_value variable are a list of all the addresses, separated by a
3149 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3150 to accept a host name and all its aliases if the forward lookup for any of
3151 them yielded the IP address of the incoming connection. Now it accepts only
3152 those names whose forward lookup yields the correct IP address. Any other
3153 names are discarded. This closes a loophole whereby a rogue DNS
3154 administrator could create reverse DNS records to break through a
3155 wildcarded host restriction in an ACL.
3157 37. If a user filter or a system filter that ran in a subprocess used any of
3158 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3159 the wrong values were passed to the pipe command ($thisaddress had the
3160 value of $0, $0 had the value of $1, etc). This bug was introduced by
3161 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3163 38. Improved the line breaking for long SMTP error messages from ACLs.
3164 Previously, if there was no break point between 40 and 75 characters, Exim
3165 left the rest of the message alone. Two changes have been made: (a) I've
3166 reduced the minimum length to 35 characters; (b) if it can't find a break
3167 point between 35 and 75 characters, it looks ahead and uses the first one
3168 that it finds. This may give the occasional overlong line, but at least the
3169 remaining text gets split now.
3171 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3172 file descriptors might be low, and that setting 1000 would always raise it.
3173 It turns out that in some environments, the limit is already over 1000 and
3174 that lowering it causes trouble. So now Exim takes care not to decrease it.
3176 40. When delivering a message, the value of $return_path is set to $sender_
3177 address at the start of routing (routers may change the value). By an
3178 oversight, this default was not being set up when an address was tested by
3179 -bt or -bv, which affected the outcome if any router or filter referred to
3182 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3183 log only when "message" or "log_message" are set. However, if one of the
3184 conditions was an address verification, or a call to a nested ACL, the
3185 messages generated by the underlying test were being passed through. This
3186 no longer happens. The underlying message is available in $acl_verify_
3187 message for both "message" and "log_message" expansions, so it can be
3188 passed through if needed.
3190 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3191 new expansion $bh_ to give the encoded byte string without charset
3192 translation. Translation happens only if iconv() is available; HAVE_ICONV
3193 indicates this at build time. HEADERS_CHARSET gives the charset to
3194 translate to; headers_charset can change it in the configuration, and
3195 "headers charset" can change it in an individual filter file.
3197 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3198 that creates RFC 2047 encoded "words" labels them as that charset instead
3199 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3200 expansion operator; (ii) when Exim creates a From: line for a local
3201 message; (iii) when a header line is rewritten to include a "phrase" part.
3203 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3204 buggy, causing it to skip the first lines of messages whose message ID
3205 ended in 'D'. This would not have bitten before Exim release 4.14, because
3206 message IDs were unlikely to end in 'D' before then. The effect was to have
3207 incorrect size information for certain domains.
3209 45. #include "config.h" was missing at the start of the crypt16.c module. This
3210 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3213 46. If there was a timeout during a "random" callout check, Exim treated it as
3214 a failure of the random address, and carried on sending RSET and the real
3215 address. If the delay was just some slowness somewhere, the response to the
3216 original RCPT would be taken as a response to RSET and so on, causing
3217 mayhem of various kinds.
3219 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3220 when I implemented it. It didn't allow for the fact that some option values
3221 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3222 the right test for positive values.
3224 48. Domain names in DNS records are case-independent. Exim always looks them up
3225 in lower case. Some resolvers return domain names in exactly the case they
3226 appear in the zone file, that is, they may contain uppercase letters. Not
3227 all resolvers do this - some return always lower case. Exim was treating a
3228 change of case by a resolver as a change of domain, similar to a widening
3229 of a domain abbreviation. This triggered its re-routing code and so it was
3230 trying to route what was effectively the same domain again. This normally
3231 caused routing to fail (because the router wouldn't handle the domain
3232 twice). Now Exim checks for this case specially, and just changes the
3233 casing of the domain that it ultimately uses when it transmits the message
3236 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3239 50. If a filter generated a file delivery with a non-absolute name (possible if
3240 no home directory exists for the router), the forbid_file option was not
3243 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3244 the existing equality matching.
3246 52. Exim was using ints instead of ino_t variables in some places where it was
3247 dealing with inode numbers.
3249 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3250 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3251 called TMPDIR, and if it finds it is different, it changes its value.
3253 54. The smtp_printf() function is now made available to local_scan() so
3254 additional output lines can be written before returning. There is also an
3255 smtp_fflush() function to enable the detection of a dropped connection.
3256 The variables smtp_input and smtp_batched_input are exported to
3259 55. Changed the default runtime configuration: the message "Unknown user"
3260 has been removed from the ACL, and instead placed on the localuser router,
3261 using the cannot_route_message feature. This means that any verification
3262 failures that generate their own messages won't get overridden. Similarly,
3263 the "Unrouteable address" message that was in the ACL for unverifiable
3264 relay addresses has also been removed.
3266 56. Added hosts_avoid_esmtp to the smtp transport.
3268 57. The exicyclog script was not checking for the esoteric option
3269 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3270 will work only if exicyclog is run under the appropriate euid.
3272 58. Following a discussion on the list, the rules by which Exim recognises line
3273 endings on incoming messages have been changed. The -dropcr and drop_cr
3274 options are now no-ops, retained only for backwards compatibility. The
3275 following line terminators are recognized: LF CRLF CR. However, special
3276 processing applies to CR:
3278 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3279 nor a local message in the state where . is a terminator.
3281 (ii) If a bare CR is encountered in a header line, an extra space is added
3282 after the line terminator so as not to end the header. The reasoning
3283 behind this is that bare CRs in header lines are most likely either
3284 to be mistakes, or people trying to play silly games.
3286 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3287 was being incorrectly given as 18 bytes larger than it should have been.
3288 This is a VOB (very old bug).
3290 60. This may never have affected anything current, but just in case it has:
3291 When the local host is found other than at the start of a list of hosts,
3292 the local host, those with the same MX, and any that follow, are discarded.
3293 When the list in question was part of a longer list of hosts, the following
3294 hosts (not currently being processed) were also being discarded. This no
3295 longer happens. I'm not sure if this situation could ever has previously
3298 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3301 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3302 header. This is something that is recommended in a new Internet Draft, and
3303 is something that is documented as being done by Sendmail. There are two
3304 possible values. For messages generated by the autoreply transport, Exim
3307 Auto-Submitted: auto-replied
3309 whereas for all other generated messages (e.g. bounces) it adds
3311 Auto-Submitted: auto-generated
3313 63. The "personal" condition in filters now includes a test for the
3314 Auto-Submitted: header. If it contains the string "auto-" the message it
3315 not considered personal.
3317 64. Added rcpt_include_affixes as a generic transport option.
3319 65. Added queue_only_override (default true).
3321 66. Added the syslog_duplication option.
3323 67. If what should have been the first header line of a message consisted of
3324 a space followed by a colon, Exim was mis-interpreting it as a header line.
3325 It isn't of course - it is syntactically invalid and should therefore be
3326 treated as the start of the message body. The misbehaviour could have
3327 caused a number of strange effects, including loss of data in subsequent
3328 header lines, and spool format errors.
3330 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3331 client host had authenticated. This control can now be exercised by an ACL
3332 for more flexibility.
3334 69. By default, callouts do not happen when testing with -bh. There is now a
3335 variant, -bhc, which does actually run the callout code, including
3336 consulting and updating the callout cache.
3338 70. Added support for saslauthd authentication, courtesy of Alexander
3341 71. If statvfs() failed on the spool or log directories while checking their
3342 size for availability, Exim confusingly gave the error "space shortage".
3343 Furthermore, in debugging mode it crashed with a floating point exception.
3344 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3345 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3346 serious problem, Exim now writes to the main and panic logs when this
3347 happens, with details of the failure. It then refuses to accept the
3348 incoming message, giving the message "spool directory problem" or "log
3349 directory problem" with a 421 code for SMTP messages.
3351 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3352 0, 1, and 2 exist, because some OS complain for execs without them (see
3353 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3354 descriptors. However, the code omitted to check that the open succeeded,
3355 causing mysterious errors if for some reason the permissions on /dev/null
3356 got screwed. Now Exim writes a message to the main and panic logs, and
3357 bombs out if it can't open /dev/null.
3359 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3360 interact so that it is all more flexible. It is supposed to remain
3361 backwards compatible. Also added extra_local_interfaces.
3363 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3364 to bomb out with an assertion failure - to the client this appears as a
3365 connection drop. This problem occurs in the part of the code that was taken
3366 from the Samba project. Fortunately, the assertion is in a very simple
3367 function, so I have fixed this by reproducing the function inline in the
3368 one place where it is called, and arranging for authentication to fail
3369 instead of killing the process with assert().
3371 75. The SPA client code was not working when the server requested OEM rather
3372 than Unicode encoding.
3374 76. Added code to make require_files with a specific uid setting more usable in
3375 the case where statting the file as root fails - usually a non-root-mounted
3376 NFS file system. When this happens and the failure is EACCES, Exim now
3377 forks a subprocess and does the per-uid checking as the relevant uid.
3379 77. Added process_log_path.
3381 78. If log_file_path was not explicitly set, a setting of check_log_space or
3382 check_log_inodes was ignored.
3384 79. If a space check for the spool or log partitions fails, the incident is now
3385 logged. Of course, in the latter case the data may get lost...
3387 80. Added the %p formatting code to string_format() so that it can be used to
3388 print addresses in debug_print(). Adjusted all the address printing in the
3389 debugging in store.c to use %p rather than %d.
3391 81. There was a concern that a line of code in smtp_in.c could overflow a
3392 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3393 initially expressed, the concern was not well-founded, because trailing
3394 spaces are removed early. However, if the trailing spaces were followed by
3395 a NULL, they did not get removed, so the overflow was possible. Two fixes
3398 (a) I re-wrote the offending code in a cleaner fashion.
3399 (b) If an incoming SMTP command contains a NULL character, it is rejected
3402 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3403 runs initgroups(), so that if the Exim user is in any additional groups,
3404 they will be used during message reception.
3410 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / exim.git / blob