1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.244 2005/10/03 13:25:32 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
34 TK/01 Added the "success_on_redirect" address verification option. See
35 NewStuff for rationale and an example.
37 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
39 PH/02 Patch to exigrep to allow it to work on syslog lines.
41 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
42 fread() to skip over the body file's header line, because in Cygwin the
43 header line is locked and is inaccessible.
45 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
46 co-exist for some time) to make it clear that it is the Exim ID that is
47 referenced, not the Message-ID: header line.
49 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
50 string_format() function, because snprintf() does not exist on all
53 PH/06 The use of forbid_filter_existstest now also locks out the use of the
54 ${stat: expansion item.
56 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
57 protocol synchronization error", to keep the pedants happy.
59 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
60 well as for IRIX systems, when gcc is being used. See the host.c source
63 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
65 PH/10 Named domain lists were not working if used in a queue_smtp_domains
68 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
69 transport and to the smtp transport in LMTP mode.
71 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
73 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
74 run a filter in a subprocess. This could lead to confusion in subsequent
75 lookups in the parent process. There should also be a search_tidyup() at
76 the end of the subprocess.
78 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
79 only if the host matched helo_try_verify_hosts, which caused the
80 verification to occur when the EHLO/HELO command was issued. The ACL just
81 tested the remembered result. Now, if a previous verification attempt has
82 not happened, "verify = helo" does it there and then.
84 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
86 TK/03 Fix log output including CR from clamd.
88 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
89 exist provoked a memory error which could cause a segfault.
91 PH/15 Installed PCRE 6.2
93 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
95 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
96 of the problem. Specifically, suggested +O2 rather than +O1 for the
99 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
101 PH/20 If a delivery was routed to a non-standard port by means of an SRV
102 record, the port was not correctly logged when the outgoing_port log
103 selector was set (it logged the transort's default port).
105 PH/21 Added support for host-specific ports to manualroute, queryprogram,
106 fallback_hosts, and "hosts" in the smtp transport.
108 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
109 host errors such as "Connection refused".
111 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
112 authentication with radiusclient 0.4.9:
114 - Error returned from rc_read_config was caught wrongly
115 - Username/password not passed on to radius server due to wrong length.
117 The presumption is that some radiusclient API changes for 4.51/PH/17
118 were not taken care of correctly. The code is still untested by me (my
119 Linux distribution still has 0.3.2 of radiusclient), but it was
120 contributed by a Radius user.
122 PH/24 When doing a callout, the value of $domain wasn't set correctly when
123 expanding the "port" option of the smtp transport.
125 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
126 while reading a MIME header. Thanks to Tom Hughes for a patch.
128 PH/24 Include config.h inside local_scan.h so that configuration settings are
131 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
132 that in an ACL for RCPT (for example), you can examine exactly what was
135 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
136 commands, but it was not correctly comparing the address with the actual
137 client host address. Thus, it would show the EHLO address in Received:
138 header lines when this was not necessary.
140 PH/27 Added the % operator to ${eval:}.
142 PH/28 Exim tries to create and chdir to its spool directory when it starts;
143 it should be ignoring failures (because with -C, for example, it has lost
144 privilege). It wasn't ignoring creation failures other than "already
147 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
150 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
151 because an input line was too long, either on its own, or by virtue of
152 too many continuations, the temporary file was not being removed, and the
153 return code was incorrect.
155 PH/31 Missing "BOOL" in function definition in filtertest.c.
157 PH/32 Applied Sieve patches from the maintainer.
159 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
161 PH/33 Added "verify = not_blind".
163 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
164 Local/Makefile (with some defaults set). These are used in built scripts
165 such as exicyclog, but they have never been used in the exim_install
166 script (though there are many overriding facilities there). I have
167 arranged that the exim_install script now takes note of these two
170 PH/35 Installed configuration files for Dragonfly.
172 PH/36 When a locally submitted message by a trusted user did not contain a
173 From: header, and the sender address was obtained from -f or from an SMTP
174 MAIL command, and the trusted user did not use -F to supply a sender
175 name, $originator_name was incorrectly used when constructing a From:
176 header. Furthermore, $originator_name was used for submission mode
177 messages from external hosts without From: headers in a similar way,
178 which is clearly wrong.
180 PH/37 Added control=suppress_local_fixups.
182 PH/38 When log_selector = +received_sender was set, and the addition of the
183 sender made the log line's construction buffer exactly full, or one byte
184 less than full, an overflow happened when the terminating "\n" was
187 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
188 when the result of a list match is failure because a DNS lookup failed.
190 PH/40 RM_COMMAND is now used in the building process.
192 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
193 the "build-* directories that it finds.
195 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
196 address in a domain literal was a prefix of an interface address.
198 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
199 when verifying a sender address, unless rewrite_headers is false.
201 PH/44 Wrote a long comment about why errors_to addresses are verified as
202 recipients, not senders.
204 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
205 the ratelimit ACL was added.
207 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
209 PH/46 Added extra information about PostgreSQL errors to the error string.
211 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
212 avoiding the need to install special zones in a real server. This is
213 backwards compatible; if it can't find the fake resolver, it drops back.
214 Thus, both old and new test suites can be run.
216 TF/02 Added util/ratelimit.pl
218 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
219 clock is set back in time.
221 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
222 Candler <B.Candler@pobox.com>.
224 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
225 OK for addresses that are the result of redirections.
227 PH/48 A number of further additions for the benefit of the new test suite,
228 including a fake gethostbyname() that interfaces to the fake DNS resolver
231 TF/06 The fix for widen_domains has also been applied to qualify_single and
232 search_parents which are the other dnslookup options that can cause
235 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
238 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
240 TF/07 Exim produced the error message "an SRV record indicated no SMTP
241 service" if it encountered an MX record with an empty target hostname.
242 The message is now "an MX or SRV record indicated no SMTP service".
244 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
245 if the DNS of the sending site is misconfigured. This is quite a
246 common situation. This change restores the behaviour of treating a
247 helo verification defer as a failure.
249 PH/51 If self=fail was set on a router, the bounce message did not include the
250 actual error message.
256 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
258 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
259 command itself ended in error, the underlying message about the transport
260 filter timeout was being overwritten with the pipe command error. Now the
261 underlying error message should be appended to the second error message.
263 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
265 PH/02 Reduce the amount of output that "make" produces by default. Full output
266 can still be requested.
268 PH/03 The warning log line about a condition test deferring for a "warn" verb
269 was being output only once per connection, rather than after each
270 occurrence (because it was using the same function as for successful
271 "warn" verbs). This seems wrong, so I have changed it.
273 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
274 it should not have, which might have caused a crash in the right
275 circumstances, but probably never did.
277 PH/04 Installed a modified version of Tony Finch's patch to make submission
278 mode fix the return path as well as the Sender: header line, and to
279 add a /name= option so that you can make the user's friendly name appear
282 TF/03 Added the control = fakedefer ACL modifier.
284 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
285 Mark Lowes for thorough testing.
287 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
289 TK/03 Merged latest SRS patch from Miles Wilton.
291 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
292 with the definition in sysexits.h (which is #included earlier).
293 Fortunately, Exim does not actually use EX_OK. The code used to try to
294 preserve the sysexits.h value, by assumimg that macro definitions were
295 scanned for macro replacements. I have been disabused of this notion,
296 so now the code just undefines EX_OK before #including unistd.h.
298 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
299 in the smtp transport. When a block could not be written in a single
300 write() function, the timeout was being re-applied to each part-write.
301 This seems wrong - if the receiver was accepting one byte at a time it
302 would take for ever. The timeout is now adjusted when this happens. It
303 doesn't have to be particularly precise.
305 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
306 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
308 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
309 without a domain if the check to <postmaster@domain> fails.
311 SC/01 Eximstats: added -xls and the ability to specify output files
312 (patch written by Frank Heydlauf).
314 SC/02 Eximstats: use FileHandles for outputing results.
316 SC/03 Eximstats: allow any combination of xls, txt, and html output.
318 SC/04 Eximstats: fixed display of large numbers with -nvr option
320 SC/05 Eximstats: fixed merging of reports with empty tables.
322 SC/06 Eximstats: added the -include_original_destination flag
324 SC/07 Eximstats: removed tabs and trailing whitespace.
326 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
328 TK/06 MBOX spool code: Add real "From " MBOX separator line
329 so the .eml file is really in mbox format (even though
330 most programs do not really care). Patch from Alex Miller.
332 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
333 The latter is generated from $received_to and is only set if the
334 message has one envelope recipient. SA can use these headers,
335 obviously out-of-the-box. Patch from Alex Miller.
337 PH/08 The ${def test on a variable was returning false if the variable's
338 value was "0", contrary to what the specification has always said!
339 The result should be true unless the variable is empty.
341 PH/09 The syntax error of a character other than { following "${if
342 def:variable_name" (after optional whitespace) was not being diagnosed.
343 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
344 accidental colon was present, for example, could give incorrect results.
346 PH/10 Tidied the code in a number of places where the st_size field of a stat()
347 result is used (not including appendfile, where other changes are about
350 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
351 This involved changing a lot of size variables from int to off_t. It
352 should work with maildirs and everything.
354 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
355 spamd dying while we are connected to it.
357 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
360 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
361 function for BDB 4.3.
363 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
365 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
366 the sources for PCRE have been re-arranged and more files are now
369 PH/15 The code I had for printing potentially long long variables in PH/11
370 above was not the best (it lost precision). The length of off_t variables
371 is now inspected at build time, and an appropriate printing format (%ld
372 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
373 to be "long long int" or "long int". This is needed for the internal
374 formatting function string_vformat().
376 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
377 the configuration file to be ":syslog", then the script "guesses" where
378 the logs files are, rather than using the compiled in default. In our
379 case the guess is not the same as the compiled default, so the script
380 suddenly stopped working when I started to use syslog. The patch checks
381 to see if log_file_path is "". If so, it attempts to read it from exim
382 with no configuration file to get the compiled in version, before it
383 falls back to the previous guessing code."
385 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
386 implementing BATV in an Exim configuration. See NewStuff for the gory
389 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
390 Makefile that are specific to HP-UX.
392 PH/18 If the "use_postmaster" option was set for a recipient callout together
393 with the "random" option, the postmaster address was used as the MAIL
394 FROM address for the random test, but not for the subsequent recipient
395 test. It is now used for both.
397 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
398 patch removes a few documentation additions to RFC 3028, because the
399 latest draft now contains them. It adds the new en;ascii-case comparator
400 and a new error check for 8bit text in MIME parts. Comparator and
401 require names are now matched exactly. I enabled the subaddress
402 extension, but it is not well tested yet (read: it works for me)."
404 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
405 rework some of the code of TK/09 above to avoid the hardwired use of
406 "%lld" and "long long". Replaced the call to snprintf() with a call to
409 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
410 records point to non-existent hosts", "retry timeout exceeded", and
411 "retry time not reached for any host after a long failure period".
413 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
414 experimental DomainKeys support:
416 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
417 (2) On an error such as an illegally used "control", the wrong name for
418 the control was given.
420 These problems did NOT occur unless DomainKeys support was compiled.
422 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
424 PH/24 Added ${if match_ip condition.
426 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
427 databases so that it will be absolutely obvious if a crash occurs in the
428 DB library. This is a regular occurrence (often caused by mis-matched
431 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
432 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
433 picked up on a user's system that detects such things. There doesn't seem
434 to be a gcc warning option for this - only an attribute that has to be
435 put on the function's prototype. It seems that in Fedora Core 4 they have
436 set this on a number of new functions. No doubt there will be more in due
439 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
440 specify a transport. However, if an address that was verified by such a
441 router was the subject of a callout, Exim crashed because it tried to
442 read the rcpt_include_affixes from the non-existent transport. Now it
443 just assumes that the setting of that option is false. This bug was
444 introduced by 4.51/PH/31.
446 PH/28 Changed -d+all to exclude +memory, because that information is very
447 rarely of interest, but it makes the output a lot bigger. People tend to
448 do -d+all out of habit.
450 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
451 code in os-type was giving problems when libc.so lives in lib64, like on
454 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
455 aren't the modern standard, and it seems that some systems' include files
456 don't always have them. Exim was already checking for some of the newer
457 ones like T_AAAA, and defining it itself. I've added checks for all the
458 record types that Exim uses.
460 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
461 not automatically generating a new one, as it is supposed to. This
462 prevented TLS from working. If the file did exist, but contained invalid
463 data, a new version was generated, as expected. It was only the case of a
464 non-existent file that was broken.
466 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
467 with a change in libdomainkeys > 0.64.
469 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
470 from DNS. If the selector record carries the flag, it now has
471 precedence over the domain-wide flag.
473 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
475 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
476 the use of a port name that isn't defined in /etc/services) occurred, the
477 message was deferred as in a normal delivery, and thus remained on the
478 spool, instead of being failed because of the mua_wrapper setting. This
479 is now fixed, and I tidied up some of the mua_wrapper messages at the
482 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
483 the messages in a hash of arrays rather than using individual hashes.
484 This is a bit cleaner and results in dramatic memory savings, albeit
485 at a slight CPU cost.
487 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
488 as requested by Marc Sherman.
490 SC/10 Eximstats: added histograms for user specified patterns as requested
493 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
495 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
496 fopen() in the content-scanning modules that did not already have it.
502 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
503 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
505 TK/02 Fix ACL "control" statement not being available in MIME ACL.
507 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
509 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
510 to test Sieve filters that use "vacation".
512 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
513 that changes the way the GnuTLS parameters are stored in the cache file.
514 The new format can be generated externally. For backward compatibility,
515 if the data in the cache doesn't make sense, Exim assumes it has read an
516 old-format file, and it generates new data and writes a new file. This
517 means that you can't go back to an older release without removing the
520 PH/03 A redirect router that has both "unseen" and "one_time" set does not
521 work if there are any delivery delays because "one_time" forces the
522 parent to be marked "delivered", so its unseen clone is never tried
523 again. For this reason, Exim now forbids the simultaneous setting of
526 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
527 redirected to themselves ("homonym" addresses). Read the long ChangeLog
528 entry if you want to know the details. The fix, however, neglected to
529 consider the case when local delivery batching is involved. The test for
530 "previously delivered" was not happening when checking to see if an
531 address could be batched with a previous (undelivered) one; under
532 certain circumstances this could lead to multiple deliveries to the same
535 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
536 in its include files, and this causes problems building Exim.
538 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
539 header_syntax) but Exim was just ignoring anything given after a slash.
540 In particular, this caused confusion with an attempt to use "verify =
541 reverse_host_lookup/defer_ok". An error is now given when options are
542 supplied for verify items that do not have them. (Maybe reverse_host_
543 lookup should have a defer_ok option, but that's a different point.)
545 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
546 defined by RFC 821) to 2048, because there were problems with some AUTH
547 commands, and RFC 1869 says the size should be increased for extended
548 SMTP commands that take arguments.
550 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
553 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
554 "unknown" error; now it says that the functionality isn't in the binary.
556 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
557 an address' error message when a string expansion fails (syntax or
558 whatever). Otherwise the password may appear in the log. Following change
559 PH/42 below, there is no longer a chance of it appearing in a bounce
562 PH/11 Installed exipick version 20050225.0 from John Jetmore.
564 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
565 of its addresses was ever tried. (Bugzilla bug #2.)
567 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
568 the result incorrectly in the debug output. (It correctly added a newline
569 to what was transported.)
571 TF/01 Added $received_time.
573 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
574 commented out examples of how to interface to a virus scanner and to
575 SpamAssassin. Also added commented examples of av_scanner and
576 spamd_address settings.
578 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
579 and controls are allowed in which ACLs. There were a couple of minor
580 errors. Some of the entries in the conditions table (which is a table of
581 where they are NOT allowed) were getting very unwieldy; rewrote them as a
582 negation of where the condition IS allowed.
584 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
586 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
587 header file does not have a version number, so I've had to invent a new
588 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
589 API. The code is untested by me (my Linux distribution still has 0.3.2 of
590 radiusclient), but it was contributed by a Radius user.
592 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
593 files or directories, for OpenSSL.
595 PH/19 When an Exim process that is running as root has to create an Exim log
596 file, it does so in a subprocess that runs as exim:exim so as to get the
597 ownership right at creation (otherwise, other Exim processes might see
598 the file with the wrong ownership). There was no test for failure of this
599 fork() call, which would lead to the process getting stuck as it waited
600 for a non-existent subprocess. Forks do occasionally fail when resources
601 run out. I reviewed all the other calls to fork(); they all seem to check
604 PH/20 When checking for unexpected SMTP input at connect time (before writing
605 the banner), Exim was not dealing correctly with a non-positive return
606 from the read() function. If the client had disconnected by this time,
607 the result was a log entry for a synchronization error with an empty
608 string after "input=" when read() returned zero. If read() returned -1
609 (an event I could not check), uninitialized data bytes were printed.
610 There were reports of junk text (parts of files, etc) appearing after
613 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
615 PH/22 Added support for macro redefinition, and (re)definition in between
616 driver and ACL definitions.
618 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
619 forgetting to use the resulting value; it was using the unexpanded value.
621 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
622 hadn't been configured. The fix is from Juergen Kreileder, who
623 understands it better than I do:
625 "Here's what I see happening with three configured cyrus_sasl
626 authenticators configured (plain, login, cram-md5):
628 On startup auth_cyrus_sasl_init() gets called for each of these.
629 This means three calls to sasl_listmech() without a specified mech_list.
630 => SASL tests which mechs of all available mechs actually work
631 => three warnings about OTP not working
632 => the returned list contains: plain, login, cram-md5, digest-md5, ...
634 With the patch, sasl_listmech() also gets called three times. But now
635 SASL's mech_list option is set to the server_mech specified in the the
636 authenticator. Or in other words, the answer from sasl_listmech()
637 gets limited to just the mech you're testing for (which is different
639 => the return list contains just 'plain' or 'login', 'cram-md5' or
640 nothing depending on the value of ob->server_mech.
642 I've just tested the patch: Authentication still works fine,
643 unavailable mechs specified in the exim configuration are still
644 caught, and the auth.log warnings about OTP are gone."
646 PH/25 When debugging is enabled, the contents of the command line are added
647 to the debugging output, even when log_selector=+arguments is not
650 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
651 answer is "GNU", and only if the return is "GNU/something" is the answer
654 PH/27 $acl_verify_message is now set immediately after the failure of a
655 verification in an ACL, and so is available in subsequent modifiers. In
656 particular, the message can be preserved by coding like this:
658 warn !verify = sender
659 set acl_m0 = $acl_verify_message
661 Previously, $acl_verify_message was set only while expanding "message"
662 and "log_message" when a very denied access.
664 PH/28 Modified OS/os.c-Linux with
666 -#ifndef OS_LOAD_AVERAGE
667 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
669 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
670 nomenclature these days.)
672 PH/29 Installed patch from the Sieve maintainer that adds the options
673 sieve_useraddress and sieve_subaddress to the redirect router.
675 PH/30 In these circumstances:
676 . Two addresses routed to the same list of hosts;
677 . First host does not offer TLS;
678 . First host accepts first address;
679 . First host gives temporary error to second address;
680 . Second host offers TLS and a TLS session is established;
681 . Second host accepts second address.
682 Exim incorrectly logged both deliveries with the TLS parameters (cipher
683 and peerdn, if requested) that were in fact used only for the second
686 PH/31 When doing a callout as part of verifying an address, Exim was not paying
687 attention to any local part prefix or suffix that was matched by the
688 router that accepted the address. It now behaves in the same way as it
689 does for delivery: the affixes are removed from the local part unless
690 rcpt_include_affixes is set on the transport.
692 PH/32 Add the sender address, as F=<...>, to the log line when logging a
693 timeout during the DATA phase of an incoming message.
695 PH/33 Sieve envelope tests were broken for match types other than :is. I have
696 applied a patch sanctioned by the Sieve maintainer.
698 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
699 the uid or gid is negative. A case of a negative gid caused this to be
700 noticed. The fix allows for either to be negative.
702 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
703 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
706 PH/36 The change PH/12 above was broken. Fixed it.
708 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
709 the grounds that routing the same address twice would always produce the
710 same answer. This might have been true once, but it is certainly no
711 longer true now. Routing a child address may depend on the previous
712 routing that produced that child. Some complicated redirection strategies
713 went wrong when messages had multiple recipients, and made Exim's
714 behaviour dependent on the order in which the addresses were given.
716 I have moved the duplicate checking until after the routing is complete.
717 Exim scans the addresses that are assigned to local and remote
718 transports, and removes any duplicates. This means that more work will be
719 done, as duplicates will always all be routed, but duplicates are
720 presumably rare, so I don't expect this is of any significance.
722 For deliveries to pipes, files, and autoreplies, the duplicate checking
723 still happens during the routing process, since they are not going to be
726 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
727 It corrects a timeout issue with spamd. This is Ian's comment: "The
728 background is that sometimes spamd either never reads data from a
729 connection it has accepted, or it never writes response data. The exiscan
730 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
731 blindly assumes that writes won't block so it may never time out."
733 PH/39 Allow G after quota size as well as K and M.
735 PH/40 The value set for $authenticated_id in an authenticator may not contain
736 binary zeroes or newlines because the value is written to log lines and
737 to spool files. There was no check on this. Now the value is run through
738 the string_printing() function so that such characters are converted to
739 printable escape sequences.
741 PH/41 $message_linecount is a new variable that contains the total number of
742 lines in the message. Compare $body_linecount, which is the count for the
745 PH/42 Exim no longer gives details of delivery errors for specific addresses in
746 bounce and delay warning messages, except in certain special cases, which
749 (a) An SMTP error message from a remote host;
750 (b) A message specified in a :fail: redirection;
751 (c) A message specified in a "fail" command in a system filter;
752 (d) A message specified in a FAIL return from the queryprogram router;
753 (e) A message specified by the cannot_route_message router option.
755 In these cases only, Exim does include the error details in bounce and
756 warning messages. There are also a few cases where bland messages such
757 as "unrouteable address" or "local delivery error" are given.
759 PH/43 $value is now also set for the "else" part of a ${run expansion.
761 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
762 being worked on, but at least Exim now implements the latest version to
765 PH/45 In a pipe transport, although a timeout while waiting for the pipe
766 process to complete was treated as a delivery failure, a timeout while
767 writing the message to the pipe was logged, but erroneously treated as a
768 successful delivery. Such timeouts include transport filter timeouts. For
769 consistency with the overall process timeout, these timeouts are now
770 treated as errors, giving rise to delivery failures by default. However,
771 there is now a new Boolean option for the pipe transport called
772 timeout_defer, which, if set TRUE, converts the failures into defers for
773 both kinds of timeout. A transport filter timeout is now identified in
776 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
777 systems where "make" and "gmake" are different, calling "gmake" at top
778 level broke things. I've arranged for the value of $(MAKE) to be passed
779 from the Makefile to this script so that it can call the same version of
783 A note about Exim versions 4.44 and 4.50
784 ----------------------------------------
786 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
787 changes of various kinds. As a consequence, a big documentation update was
788 needed. This delayed the release for rather longer than seemed good, especially
789 in the light of a couple of (minor) security issues. Therefore, the changes
790 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
791 release. So 4.44 and 4.50 are in effect two different branches that both start
794 I have left the 4.50 change log unchanged; it contains all the changes since
795 4.43. The change log for 4.44 is below; many of its items are identical to
796 those for 4.50. This seems to be the most sensible way to preserve the
797 historical information.
803 1. Minor wording change to the doc/README.SIEVE file.
805 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
806 computation of the current number of files was incorrect.
808 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
809 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
812 4. Give more explanation in the error message when the command for a transport
813 filter fails to execute.
815 5. There are several places where Exim runs a non-Exim command in a
816 subprocess. The SIGUSR1 signal should be disabled for these processes. This
817 was being done only for the command run by the queryprogram router. It is
818 now done for all such subprocesses. The other cases are: ${run, transport
819 filters, and the commands run by the lmtp and pipe transports.
821 6. Added CONFIGURE_GROUP build-time option.
823 7. Some older OS have a limit of 256 on the maximum number of file
824 descriptors. Exim was using setrlimit() to set 1000 as a large value
825 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
826 systems. I've change it so that if it can't get 1000, it tries for 256.
828 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
829 was an oversight, and furthermore, ever since the addition of extra
830 controls (e.g. 4.43/32), the checks on when to allow different forms of
831 "control" were broken. There should now be diagnostics for all cases when a
832 control that does not make sense is encountered.
834 9. Added the /retain_sender option to "control=submission".
836 10. $recipients is now available in the predata ACL (oversight).
838 11. Tidy the search cache before the fork to do a delivery from a message
839 received from the command line. Otherwise the child will trigger a lookup
840 failure and thereby defer the delivery if it tries to use (for example) a
841 cached ldap connection that the parent has called unbind on.
843 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
844 of $address_data from the recipient verification was clobbered by the
847 13. The value of address_data from a sender verification is now available in
848 $sender_address_data in subsequent conditions in the ACL statement.
850 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
852 15. Added a new option "connect=<time>" to callout options, to set a different
855 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
856 was its contents. (It was OK if the option was not defined at all.)
858 17. A "Completed" log line is now written for messages that are removed from
859 the spool by the -Mrm option.
861 18. New variables $sender_verify_failure and $recipient_verify_failure contain
862 information about exactly what failed.
864 19. Added -dd to debug only the daemon process.
866 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
867 handles timeouts, both on the server side and network timeouts. Renamed the
868 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
870 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
871 It is now set to "smtps".
873 22. $host_address is now set to the target address during the checking of
876 23. When checking ignore_target_hosts for an ipliteral router, no host name was
877 being passed; this would have caused $sender_host_name to have been used if
878 matching the list had actually called for a host name (not very likely,
879 since this list is usually IP addresses). A host name is now passed as
882 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
883 code that specifies a non-restarting handler (typically sigaction() in
884 modern systems) in an attempt to fix a rare and obscure crash bug.
886 25. Narrowed the window for a race in the daemon that could cause it to ignore
887 SIGCHLD signals. This is not a major problem, because they are used only to
888 wake it up if nothing else does.
890 26. A malformed maildirsize file could cause Exim to calculate negative values
891 for the mailbox size or file count. Odd effects could occur as a result.
892 The maildirsize information is now recalculated if the size or filecount
895 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
896 support for a long time. Removed HAVE_SYS_VFS_H.
898 28. Installed the latest version of exipick from John Jetmore.
900 29. In an address list, if the pattern was not a regular expression, an empty
901 subject address (from a bounce message) matched only if the pattern was an
902 empty string. Non-empty patterns were not even tested. This was the wrong
903 because it is perfectly reasonable to use an empty address as part of a
904 database query. An empty address is now tested by patterns that are
905 lookups. However, all the other forms of pattern expect the subject to
906 contain a local part and a domain, and therefore, for them, an empty
907 address still always fails if the pattern is not itself empty.
909 30. Exim went into a mad DNS loop when attempting to do a callout where the
910 host was specified on an smtp transport, and looking it up yielded more
913 31. Re-factored the code for checking spool and log partition space into a
914 function that finds that data and another that does the check. The former
915 is then used to implement four new variables: $spool_space, $log_space,
916 $spool_inodes, and $log_inodes.
918 32. The RFC2047 encoding function was originally intended for short strings
919 such as real names; it was not keeping to the 75-character limit for
920 encoded words that the RFC imposes. It now respects the limit, and
921 generates multiple encoded words if necessary. To be on the safe side, I
922 have increased the buffer size for the ${rfc2047: expansion operator from
925 33. It is now permitted to omit both strings after an "if" condition; if the
926 condition is true, the result is "true". As before, when the second string
927 is omitted, a false condition yields an empty string. This makes it less
928 cumbersome to write custom ACL and router conditions.
930 34. Failure to deliver a bounce message always caused it to be frozen, even if
931 there was an errors_to setting on the router. The errors_to setting is now
934 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
935 canonical form (fully expanded) before being placed in
936 $sender_host_address.
938 36. The table in the code that translates DNS record types into text (T_A to
939 "A" for instance) was missing entries for NS and CNAME. It is just possible
940 that this could have caused confusion if both these types were looked up
941 for the same domain, because the text type is used as part of Exim's
942 per-process caching. But the chance of anyone hitting this buglet seems
945 37. The dnsdb lookup has been extended in a number of ways.
947 (1) There is a new type, "zns", which walks up the domain tree until it
948 finds some nameserver records. It should be used with care.
950 (2) There is a new type, "mxh", which is like "mx" except that it returns
951 just the host names, not the priorities.
953 (3) It is now possible to give a list of domains (or IP addresses) to be
954 looked up. The behaviour when one of the lookups defers can be
955 controlled by a keyword.
957 (4) It is now possible to specify the separator character for use when
958 multiple records are returned.
960 38. The dnslists ACL condition has been extended: it is now possible to supply
961 a list of IP addresses and/or domains to be looked up in a particular DNS
964 39. Added log_selector=+queue_time_overall.
966 40. When running the queue in the test harness, wait just a tad after forking a
967 delivery process, to get repeatability of debugging output.
969 41. Include certificate and key file names in error message when GnuTLS fails
970 to set them up, because the GnuTLS error message doesn't include the name
971 of the failing file when there is a problem reading it.
973 42. Allow both -bf and -bF in the same test run.
975 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
977 44. The "Exiscan patch" is now merged into the mainline Exim source.
979 45. Sometimes the final signoff response after QUIT could fail to get
980 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
981 < 0 before doing a fflush(). This bug looks as though it goes back to the
982 introduction of TLS in release 3.20, but "sometimes" must have been rare
983 because the tests only now provoked it.
985 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
986 (this can affect the format of dates).
988 47. exim_tidydb, when checking for the continued existence of a message for
989 which it has found a message-specific retry record, was not finding
990 messages that were in split spool directories. Consequently, it was
991 deleting retry records that should have stayed in existence.
993 48. Steve fixed some bugs in eximstats.
995 49. The SPA authentication driver was not abandoning authentication and moving
996 on to the next authenticator when an expansion was forced to fail,
997 contradicting the general specification for all authenticators. Instead it
998 was generating a temporary error. It now behaves as specified.
1000 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
1001 (the order specifies the preference for clients). The order is now AES256,
1002 AES128, 3DES, ARCFOUR128.
1004 51. Small patch to Sieve code - explicitly set From: when generating an
1007 52. Exim crashed if a remote delivery caused a very long error message to be
1008 recorded - for instance if somebody sent an entire SpamAssassin report back
1009 as a large number of 550 error lines. This bug was coincidentally fixed by
1010 increasing the size of one of Exim's internal buffers (big_buffer) that
1011 happened as part of the Exiscan merge. However, to be on the safe side, I
1012 have made the code more robust (and fixed the comments that describe what
1015 53. Now that there can be additional text after "Completed" in log lines (if
1016 the queue_time_overall log selector is set), a one-byte patch to exigrep
1017 was needed to allow it to recognize "Completed" as not the last thing in
1020 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1021 patch that reportedly fixes this has been added. I am not expert enough to
1022 create a test for it. This is what the patch creator wrote:
1024 "I found a little strange behaviour of ldap code when working with
1025 Windows 2003 AD Domain, where users was placed in more than one
1026 Organization Units. When I tried to give exim partial DN, the exit code
1027 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1028 But simultaneously result of request was absolutely normal ldap result,
1029 so I produce this patch..."
1031 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1032 REFERENCE, so I have modified the code to exclude the patch when that macro
1035 55. Some experimental protocols are using DNS PTR records for new purposes. The
1036 keys for these records are domain names, not reversed IP addresses. The
1037 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1038 leaves it alone. Component reversal etc. now happens only for IP addresses.
1041 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1043 57. Double the size of the debug message buffer (to 2048) so that more of very
1044 long debug lines gets shown.
1046 58. The exicyclog utility now does better if the number of log files to keep
1047 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1049 59. Two changes related to the smtp_active_hostname option:
1051 (1) $smtp_active_hostname is now available as a variable.
1052 (2) The default for smtp_banner uses $smtp_active_hostname instead
1053 of $primary_hostname.
1055 60. The host_aton() function is supposed to be passed a string that is known
1056 to be a valid IP address. However, in the case of IPv6 addresses, it was
1057 not checking this. This is a hostage to fortune. Exim now panics and dies
1058 if the condition is not met. A case was found where this could be provoked
1059 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1060 components; fortuitously, this particular loophole had already been fixed
1061 by change 4.50/55 above.
1063 If there are any other similar loopholes, the new check in host_aton()
1064 itself should stop them being exploited. The report I received stated that
1065 data on the command line could provoke the exploit when Exim was running as
1066 exim, but did not say which command line option was involved. All I could
1067 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1068 running as the user.
1071 61. There was a buffer overflow vulnerability in the SPA authentication code
1072 (which came originally from the Samba project). I have added a test to the
1073 spa_base64_to_bits() function which I hope fixes it.
1076 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1077 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1079 63. The daemon start-up calls getloadavg() while still root for those OS that
1080 need the first call to be done as root, but it missed one case: when
1081 deliver_queue_load_max is set with deliver_drop_privilege. This is
1082 necessary for the benefit of the queue runner, because there is no re-exec
1083 when deliver_drop_privilege is set.
1085 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1086 This has been fixed.
1088 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1089 was in use, was not putting the data itself into the right store pool;
1090 consequently, it could be overwritten for a subsequent message in the same
1091 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1094 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1096 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1097 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1098 the function were treating the return as a boolean value, which happened to
1099 work because 0=false and not-0=true, but is not correct code.
1101 68. The host_aton() function was not handling scoped IPv6 addresses (those
1102 with, for example, "%eth0" on the end) correctly.
1104 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1105 negated items (that is, ~something) in unsigned ints. Some compilers
1106 apparently mutter when there is no cast.
1108 70. If an address verification called from an ACL failed, and did not produce a
1109 user-specific message (i.e. there was only a "system" message), nothing was
1110 put in $acl_verify_message. In this situation, it now puts the system
1113 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1114 session; change 4.31/43 added the unwanted input to the log line - except
1115 that it did not do this in the start of session case. It now does.
1117 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1118 This is wrong and can cause the other end to generate a synchronization
1119 error if it is another Exim or anything else that does the synchronization
1120 check. A QUIT command is no longer sent after a timeout.
1122 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1123 during host lookups.
1125 74. The defer_ok option of callout verification was not working if it was used
1126 when verifying addresses in header lines, that is, for this case:
1128 verify = header_sender/callout=defer_ok
1130 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1131 those file descriptors could be used for SMTP connections. If anything
1132 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1133 could be sent to the SMTP client, causing chaos. The daemon now opens
1134 stdin, stdout, and stderr to /dev/null when it puts itself into the
1137 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1138 log by default. The user can override this with suitable Perl magic.
1140 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1141 the log message when discard triggers, was not working for the DATA ACL or
1142 for the non-SMTP ACL.
1144 78. Error message wording change in sieve.c.
1146 79. If smtp_accept_max_per_host was set, the number of connections could be
1147 restricted to fewer than expected, because the daemon was trying to set up
1148 a new connection before checking whether the processes handling previous
1149 connections had finished. The check for completed processes is now done
1150 earlier. On busy systems, this bug wouldn't be noticed because something
1151 else would have woken the daemon, and it would have reaped the completed
1154 80. If a message was submitted locally by a user whose login name contained one
1155 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1156 It caused a spool format error. I have fixed the spool reading code. A
1157 related problem was that the "from" clause in the Received: line became
1158 illegal because of the space(s). It is now covered by ${quote_local_part.
1160 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1163 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1165 83. Patch from Sieve maintainer to fix unterminated string problem in
1166 "vacation" handling.
1168 84. Some minor changes to the Linux configuration files to help with other
1169 OS variants using glibc.
1171 85. One more patch for Sieve to update vacation handling to latest spec.
1174 ----------------------------------------------------
1175 See the note above about the 4.44 and 4.50 releases.
1176 ----------------------------------------------------
1182 1. Change 4.43/35 introduced a bug that caused file counts to be
1183 incorrectly computed when quota_filecount was set in an appendfile
1186 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1187 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1190 3. Give more explanation in the error message when the command for a transport
1191 filter fails to execute.
1193 4. There are several places where Exim runs a non-Exim command in a
1194 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1195 was being done only for the command run by the queryprogram router. It is
1196 now done for all such subprocesses. The other cases are: ${run, transport
1197 filters, and the commands run by the lmtp and pipe transports.
1199 5. Some older OS have a limit of 256 on the maximum number of file
1200 descriptors. Exim was using setrlimit() to set 1000 as a large value
1201 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1202 systems. I've change it so that if it can't get 1000, it tries for 256.
1204 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1205 was an oversight, and furthermore, ever since the addition of extra
1206 controls (e.g. 4.43/32), the checks on when to allow different forms of
1207 "control" were broken. There should now be diagnostics for all cases when a
1208 control that does not make sense is encountered.
1210 7. $recipients is now available in the predata ACL (oversight).
1212 8. Tidy the search cache before the fork to do a delivery from a message
1213 received from the command line. Otherwise the child will trigger a lookup
1214 failure and thereby defer the delivery if it tries to use (for example) a
1215 cached ldap connection that the parent has called unbind on.
1217 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1218 of $address_data from the recipient verification was clobbered by the
1219 sender verification.
1221 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1222 was its contents. (It was OK if the option was not defined at all.)
1224 11. A "Completed" log line is now written for messages that are removed from
1225 the spool by the -Mrm option.
1227 12. $host_address is now set to the target address during the checking of
1228 ignore_target_hosts.
1230 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1231 being passed; this would have caused $sender_host_name to have been used if
1232 matching the list had actually called for a host name (not very likely,
1233 since this list is usually IP addresses). A host name is now passed as
1236 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1237 code that specifies a non-restarting handler (typically sigaction() in
1238 modern systems) in an attempt to fix a rare and obscure crash bug.
1240 15. Narrowed the window for a race in the daemon that could cause it to ignore
1241 SIGCHLD signals. This is not a major problem, because they are used only to
1242 wake it up if nothing else does.
1244 16. A malformed maildirsize file could cause Exim to calculate negative values
1245 for the mailbox size or file count. Odd effects could occur as a result.
1246 The maildirsize information is now recalculated if the size or filecount
1249 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1250 support for a long time. Removed HAVE_SYS_VFS_H.
1252 18. Updated exipick to current release from John Jetmore.
1254 19. Allow an empty sender to be matched against a lookup in an address list.
1255 Previously the only cases considered were a regular expression, or an
1258 20. Exim went into a mad DNS lookup loop when doing a callout where the
1259 host was specified on the transport, if the DNS lookup yielded more than
1262 21. The RFC2047 encoding function was originally intended for short strings
1263 such as real names; it was not keeping to the 75-character limit for
1264 encoded words that the RFC imposes. It now respects the limit, and
1265 generates multiple encoded words if necessary. To be on the safe side, I
1266 have increased the buffer size for the ${rfc2047: expansion operator from
1269 22. Failure to deliver a bounce message always caused it to be frozen, even if
1270 there was an errors_to setting on the router. The errors_to setting is now
1273 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1274 canonical form (fully expanded) before being placed in
1275 $sender_host_address.
1277 24. Updated eximstats to version 1.33
1279 25. Include certificate and key file names in error message when GnuTLS fails
1280 to set them up, because the GnuTLS error message doesn't include the name
1281 of the failing file when there is a problem reading it.
1283 26. Expand error message when OpenSSL has problems setting up cert/key files.
1286 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1287 (this can affect the format of dates).
1289 28. exim_tidydb, when checking for the continued existence of a message for
1290 which it has found a message-specific retry record, was not finding
1291 messages that were in split spool directories. Consequently, it was
1292 deleting retry records that should have stayed in existence.
1294 29. eximstats updated to version 1.35
1295 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1296 1.35 - bugfix such that pie charts by volume are generated correctly
1298 30. The SPA authentication driver was not abandoning authentication and moving
1299 on to the next authenticator when an expansion was forced to fail,
1300 contradicting the general specification for all authenticators. Instead it
1301 was generating a temporary error. It now behaves as specified.
1303 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1304 (the order specifies the preference for clients). The order is now AES256,
1305 AES128, 3DES, ARCFOUR128.
1307 31. Small patch to Sieve code - explicitly set From: when generating an
1310 32. Exim crashed if a remote delivery caused a very long error message to be
1311 recorded - for instance if somebody sent an entire SpamAssassin report back
1312 as a large number of 550 error lines. This bug was coincidentally fixed by
1313 increasing the size of one of Exim's internal buffers (big_buffer) that
1314 happened as part of the Exiscan merge. However, to be on the safe side, I
1315 have made the code more robust (and fixed the comments that describe what
1318 33. Some experimental protocols are using DNS PTR records for new purposes. The
1319 keys for these records are domain names, not reversed IP addresses. The
1320 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1321 leaves it alone. Component reversal etc. now happens only for IP addresses.
1324 34. The host_aton() function is supposed to be passed a string that is known
1325 to be a valid IP address. However, in the case of IPv6 addresses, it was
1326 not checking this. This is a hostage to fortune. Exim now panics and dies
1327 if the condition is not met. A case was found where this could be provoked
1328 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1329 components; fortuitously, this particular loophole had already been fixed
1330 by change 4.50/55 or 4.44/33 above.
1332 If there are any other similar loopholes, the new check in host_aton()
1333 itself should stop them being exploited. The report I received stated that
1334 data on the command line could provoke the exploit when Exim was running as
1335 exim, but did not say which command line option was involved. All I could
1336 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1337 running as the user.
1340 35. There was a buffer overflow vulnerability in the SPA authentication code
1341 (which came originally from the Samba project). I have added a test to the
1342 spa_base64_to_bits() function which I hope fixes it.
1345 36. The daemon start-up calls getloadavg() while still root for those OS that
1346 need the first call to be done as root, but it missed one case: when
1347 deliver_queue_load_max is set with deliver_drop_privilege. This is
1348 necessary for the benefit of the queue runner, because there is no re-exec
1349 when deliver_drop_privilege is set.
1351 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1352 was in use, was not putting the data itself into the right store pool;
1353 consequently, it could be overwritten for a subsequent message in the same
1354 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1357 38. Sometimes the final signoff response after QUIT could fail to get
1358 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1359 < 0 before doing a fflush(). This bug looks as though it goes back to the
1360 introduction of TLS in release 3.20, but "sometimes" must have been rare
1361 because the tests only now provoked it.
1367 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1368 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1369 Now it returns an integer. A number of calls were still expecting a T/F
1370 return. Fortuitously, in all cases, the tests worked in OK situations,
1371 which is the norm. However, things would have gone wrong on any write
1372 failures on the smtp file descriptor. This function is used when sending
1373 messages over SMTP and also when doing verify callouts.
1375 2. When Exim is called to do synchronous delivery of a locally submitted
1376 message (the -odf or -odi options), it no longer closes stderr before doing
1379 3. Implemented the mua_wrapper option.
1381 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1383 5. Implemented the functions header_remove(), header_testname(),
1384 header_add_at_position(), and receive_remove_recipient(), and exported them
1387 6. If an ACL "warn" statement specified the addition of headers, Exim already
1388 inserted X-ACL-Warn: at the start if there was no header name. However, it
1389 was not making this test for the second and subsequent header lines if
1390 there were newlines in the string. This meant that an invalid header could
1391 be inserted if Exim was badly configured.
1393 7. Allow an ACL "warn" statement to add header lines at the start or after all
1394 the Received: headers, as well as at the end.
1396 8. Added the rcpt_4xx retry error code.
1398 9. Added postmaster_mailfrom=xxx to callout verification option.
1400 10. Added mailfrom=xxxx to the callout verification option, for verify=
1403 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1404 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1406 12. Inserted some casts to stop certain compilers warning when using pointer
1407 differences as field lengths or precisions in printf-type calls (mostly
1408 affecting debugging statements).
1410 13. Added optional readline() support for -be (dynamically loaded).
1412 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1413 same clock tick as a message's arrival, so that its received time was the
1414 same as the "first fail" time on the retry record, and that message
1415 remained on the queue past the ultimate address timeout, every queue runner
1416 would try a delivery (because it was past the ultimate address timeout) but
1417 after another failure, the ultimate address timeout, which should have then
1418 bounced the address, did not kick in. This was a "< instead of <=" error;
1419 in most cases the first failure would have been in the next clock tick
1420 after the received time, and all would be well.
1422 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1423 being recognized when the domain list was tested by the match_domain
1424 condition in an expansion string.
1426 16. Added the ${str2b64: operator.
1428 17. Exim was always calling setrlimit() to set a large limit for the number of
1429 processes, without checking whether the existing limit was already
1430 adequate. (It did check for the limit on file descriptors.) Furthermore,
1431 errors from getrlimit() and setrlimit() were being ignored. Now they are
1432 logged to the main and panic logs, but Exim does carry on, to try to do its
1433 job under whatever limits there are.
1435 18. Imported PCRE 5.0.
1437 19. Trivial typo in log message " temporarily refused connection" (the leading
1440 20. If the log selector return_path_on_delivery was set and an address was
1441 redirected to /dev/null, the delivery process crashed because it assumed
1442 that a return path would always be set for a "successful" delivery. In this
1443 case, the whole delivery is bypassed as an optimization, and therefore no
1446 21. Internal re-arrangement: the function for sending a challenge and reading
1447 a response while authentication was assuming a zero-terminated challenge
1448 string. It's now changed to take a pointer and a length, to allow for
1449 binary data in such strings.
1451 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1453 23. Exim was not respecting finduser_retries when seeking the login of the
1454 uid under which it was called; it was always trying 10 times. (The default
1455 setting of finduser_retries is zero.) Also, it was sleeping after the final
1456 failure, which is pointless.
1458 24. Implemented tls_on_connect_ports.
1460 25. Implemented acl_smtp_predata.
1462 26. If the domain in control=submission is set empty, Exim assumes that the
1463 authenticated id is a complete email address when it generates From: or
1464 Sender: header lines.
1466 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1467 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1468 chown and chgrp in /bin and hostname in /usr/bin.
1470 28. Exim was keeping the "process log" file open after each use, just as it
1471 does for the main log. This opens the possibility of it remaining open for
1472 long periods when the USR1 signal hits a daemon. Occasional processlog
1473 errors were reported, that could have been caused by this. Anyway, it seems
1474 much more sensible not to leave this file open at all, so that is what now
1477 29. The long-running daemon process does not normally write to the log once it
1478 has entered its main loop, and it closes the log before doing so. This is
1479 so that log files can straightforwardly be renamed and moved. However,
1480 there are a couple of unusual error situations where the daemon does write
1481 log entries, and I had neglected to close the log afterwards.
1483 30. The text of an SMTP error response that was received during a remote
1484 delivery was being truncated at 512 bytes. This is too short for some of
1485 the long messages that one sometimes sees. I've increased the limit to
1488 31. It is now possible to make retry rules that apply only when a message has a
1489 specific sender, in particular, an empty sender.
1491 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1492 it possible to be selective about when SMTP synchronization is enforced.
1494 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1496 32. Implemented hosts_connection_nolog.
1498 33. Added an ACL for QUIT.
1500 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1503 35. Added mailbox_size and mailbox_filecount to appendfile.
1505 36. Added control = no_multiline_responses to ACLs.
1507 37. There was a bug in the logic of the code that waits for the clock to tick
1508 in the case where the clock went backwards by a substantial amount such
1509 that the microsecond fraction of "now" was more than the microsecond
1510 fraction of "then" (but the whole seconds number was less).
1512 38. Added support for the libradius Radius client library this is found on
1513 FreeBSD (previously only the radiusclient library was supported).
1519 1. When certain lookups returned multiple values in the form name=value, the
1520 quoting of the values was not always being done properly. Specifically:
1521 (a) If the value started with a double quote, but contained no whitespace,
1523 (b) If the value contained whitespace other than a space character (i.e.
1524 tabs or newlines or carriage returns) it was not quoted.
1525 This fix has been applied to the mysql and pgsql lookups by writing a
1526 separate quoting function and calling it from the lookup code. The fix
1527 should probably also be applied to nisplus, ibase and oracle lookups, but
1528 since I cannot test any of those, I have not disturbed their existing code.
1530 2. A hit in the callout cache for a specific address caused a log line with no
1531 reason for rejecting RCPT. Now it says "Previous (cached) callout
1532 verification failure".
1534 3. There was an off-by-one bug in the queryprogram router. An over-long
1535 return line was truncated at 256 instead of 255 characters, thereby
1536 overflowing its buffer with the terminating zero. As well as fixing this, I
1537 have increased the buffer size to 1024 (and made a note to document this).
1539 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1540 when Exim is waiting for an SMTP response from a remote server, Exim
1541 restarts its select() call on the socket, thereby resetting its timeout.
1542 This is not a problem when such interrupts are rare. Somebody set up a cron
1543 job to run exiwhat every 2 minutes, which is less than the normal select()
1544 timeout (5 or 10 minutes). This meant that the select() timeout never
1545 kicked in because it was always reset. I have fixed this by comparing the
1546 time when an interrupt arrives with the time at the start of the first call
1547 to select(). If more time than the timeout has elapsed, the interrupt is
1548 treated as a timeout.
1550 5. Some internal re-factoring in preparation for the addition of Sieve
1551 extensions (by MH). In particular, the "personal" test is moved to a
1552 separate function, and given an option for scanning Cc: and Bcc: (which is
1553 not set for Exim filters).
1555 6. When Exim created an email address using the login of the caller as the
1556 local part (e.g. when creating a From: or Sender: header line), it was not
1557 quoting the local part when it contained special characters such as @.
1559 7. Installed new OpenBSD configuration files.
1561 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1562 try to make them clearer.
1564 9. Callout options, other than the timeout value, were being ignored when
1565 verifying sender addresses in header lines. For example, when using
1567 verify = header_sender/callout=no_cache
1569 the cache was (incorrectly) being used.
1571 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1572 only the Cygwin environment.
1574 11. When return_path_on_delivery was set as a log selector, if different remote
1575 addresses in the same message used different return paths and parallel
1576 remote delivery occurred, the wrong values would sometimes be logged.
1577 (Whenever a remote delivery process finished, the return path value from
1578 the most recently started remote delivery process was logged.)
1580 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1581 lines when AUTH and/or TLS are in use. This is the "received protocol"
1582 field. Exim used to use "asmtp" for authenticated SMTP, without any
1583 indication (in the protocol name) for TLS use. Now it follows the RFC and
1584 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1585 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1586 names appear in log lines as well as in Received: header lines.
1588 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1589 extensions, and comparison tests, and to fix some bugs.
1591 14. Changes to the "personal" filter test:
1593 (1) The test was buggy in that it was just doing the equivalent of
1594 "contains" tests on header lines. For example, if a user's address was
1595 anne@some.where, the "personal" test would incorrectly be true for
1597 To: susanne@some.where
1599 This test is now done by extracting each address from the header in turn,
1600 and checking the entire address. Other tests that are part of "personal"
1601 are now done using regular expressions (for example, to check local parts
1602 of addresses in From: header lines).
1604 (2) The list of non-personal local parts in From: addresses has been
1605 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1606 taken from the Sieve specification recommendations.
1608 (3) If the message contains any header line starting with "List-" it is
1609 treated as non-personal.
1611 (4) The test for "circular" in the Subject: header line has been removed
1612 because it now seems ill-conceived.
1614 15. Minor typos in src/EDITME comments corrected.
1616 16. Installed latest exipick from John Jetmore.
1618 17. If headers_add on a router specified a text string that was too long for
1619 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1620 of string_sprintf() is now avoided.
1622 18. $message_body_size was not set (it was always zero) when running the DATA
1623 ACL and the local_scan() function.
1625 19. For the "mail" command in an Exim filter, no default was being set for
1626 the once_repeat time, causing a random time value to be used if "once" was
1627 specified. (If the value happened to be <= 0, no repeat happened.) The
1628 default is now 0s, meaning "never repeat". The "vacation" command was OK
1629 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1630 (I found it when inspecting the code).
1632 20. There is now an overall timeout for performing a callout verification. It
1633 defaults to 4 times the callout timeout, which applies to individual SMTP
1634 commands during the callout. The overall timeout applies when there is more
1635 than one host that can be tried. The timeout is checked before trying the
1636 next host. This prevents very long delays if there are a large number of
1637 hosts and all are timing out (e.g. when the network connections are timing
1638 out). The value of the overall timeout can be changed by specifying an
1639 additional sub-option for "callout", called "maxwait". For example:
1641 verify = sender/callout=5s,maxwait=20s
1643 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1644 to the end before writing, but this should make it even safer).
1646 22. Exim was forgetting that it had advertised PIPELINING for the second and
1647 subsequent messages on an SMTP connection. It was also not resetting its
1648 memory on STARTTLS and an internal HELO.
1650 23. When Exim logs an SMTP synchronization error within a session, it now
1651 records whether PIPELINING has been advertised or not.
1653 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1654 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1655 rather than long int.
1657 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1659 26. Added the never_mail option to autoreply.
1665 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1666 crash if the getsockname() call failed; this can happen if a connection is
1667 closed very soon after it is established. The problem was simply in the
1668 order in which certain operations were done, causing Exim to try to write
1669 to the SMTP stream before it had set up the file descriptor. The bug has
1670 been fixed by making things happen in the correct order.
1676 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1677 before the connection was closed, thus losing the rejection response.
1679 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1680 some early Solaris releases, but causes trouble in current releases where
1681 socklen_t is defined.
1683 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1686 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1689 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1690 that releasing the top of it at the end releases what was used for sub-
1691 expansions (unless the block got too big). However, discard this block if
1692 the first thing is a variable or header, so that we can use its block when
1693 it is dynamic (useful for very large $message_headers, for example).
1695 6. Lookups now cache *every* query, not just the most recent. A new, separate
1696 store pool is used for this. It can be recovered when all lookup caches are
1697 flushed. Lookups now release memory at the end of their result strings.
1698 This has involved some general refactoring of the lookup sources.
1700 7. Some code has been added to the store_xxx() functions to reduce the amount
1701 of flapping under certain conditions.
1703 8. log_incoming_interface used to affect only the <= reception log lines. Now
1704 it causes the local interface and port to be added to several more SMTP log
1705 lines, for example "SMTP connection from", and rejection lines.
1707 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1709 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1711 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1712 could be overwritten at the end of the current message (or the start of a
1713 new message if it was set in a HELO ACL). The value is now preserved for
1714 the duration of the SMTP connection.
1716 12. If a transport had a headers_rewrite setting, and a matching header line
1717 contained an unqualified address, that address was qualified, even if it
1718 did not match any rewriting rules. The underlying bug was that the values
1719 of the flags that permit the existence of unqualified sender and recipient
1720 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1721 non-local messages, and by -bnq for local messages) were not being
1722 preserved with the message after it was received.
1724 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1725 "next input=" as part of the text comprising the host identity instead of
1726 the correct text. The code was using the same buffer for two different
1727 strings. However, depending on which order the printing function evaluated
1728 its arguments, the bug did not always show up. Under Linux, for example, my
1729 test suite worked just fine.
1731 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1732 doesn't work with some older versions of Perl. It has been changed to "my",
1733 which in any case is probably the better facility to use.
1735 15. A really picky compiler found some instances of statements for creating
1736 error messages that either had too many or two few arguments for the format
1739 16. The size of the buffer for calls to the DNS resolver has been increased
1740 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1741 for addresses that have a lot of PTR records. This alleviates a problem; it
1742 does not fully solve it.
1744 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1745 buffer now truncates the list and logs the incident, which is the same
1746 action as happens when Exim is looking up a host name and its aliases.
1747 Previously in this situation something unpredictable would happen;
1748 sometimes it was "internal error: store_reset failed".
1750 18. If a server dropped the connection unexpectedly when an Exim client was
1751 using GnuTLS and trying to read a response, the client delivery process
1752 crashed while trying to generate an error log message.
1754 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1755 string, for example:
1757 warn message = H1: something\nH2: something
1759 the text was added as a single header line from Exim's point of view
1760 though it ended up OK in the delivered message. However, searching for the
1761 second and subsequent header lines using $h_h2: did not work. This has been
1762 fixed. Similarly, if a system filter added multiple headers in this way,
1763 the routers could not see them.
1765 20. Expanded the error message when iplsearch is called with an invalid key to
1766 suggest using net-iplsearch in a host list.
1768 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1769 ACLs are no longer actually imposed (and a message to that effect is
1772 22. If a "gecos" field in a passwd entry contained escaped characters, in
1773 particular, if it contained a \" sequence, Exim got it wrong when building
1774 a From: or a Sender: header from that name. A second bug also caused
1775 incorrect handling when an unquoted " was present following a character
1776 that needed quoting.
1778 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1779 was not being matched caselessly.
1781 24. Arranged for all hyphens in the exim.8 source to be escaped with
1784 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1785 properly. Recipient callout cache records were still being keyed to include
1786 the sender, even when use_sender was set false. This led to far more
1787 callouts that were necessary. The sender is no longer included in the key
1788 when use_sender is false.
1790 26. Added "control = submission" modifier to ACLs.
1792 27. Added the ${base62d: operator to decode base 62 numbers.
1794 28. dnsdb lookups can now access SRV records.
1796 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1797 the configuration file.
1799 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1800 (-v) mode. This makes the output for a verbose queue run more intelligible.
1802 31. Added a use_postmaster feature to recipient callouts.
1804 32. Added the $body_zerocount variable, containing the number of binary zero
1805 bytes in the message body.
1807 33. The time of last modification of the "new" subdirectory is now used as the
1808 "mailbox time last read" when there is a quota error for a maildir
1811 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1813 35. Added +ignore_unknown as a special item in host lists.
1815 36. Code for decoding IPv6 addresses in host lists is now included, even if
1816 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1817 address was recognized as an IP address, but was then not correctly decoded
1818 into binary, causing unexpected and incorrect effects when compared with
1825 1. Very minor rewording of debugging text in manualroute to say "list of
1826 hosts" instead of "hostlist".
1828 2. If verify=header_syntax was set, and a header line with an unqualified
1829 address (no domain) and a large number of spaces between the end of the
1830 name and the colon was received, the reception process suffered a buffer
1831 overflow, and (when I tested it) crashed. This was caused by some obsolete
1832 code that should have been removed. The fix is to remove it!
1834 3. When running in the test harness, delay a bit after writing a bounce
1835 message to get a bit more predictability in the log output.
1837 4. Added a call to search_tidyup() just before forking a reception process. In
1838 theory, someone could use a lookup in the expansion of smtp_accept_max_
1839 per_host which, without the tidyup, could leave open a database connection.
1841 5. Added the variables $recipient_data and $sender_data which get set from a
1842 lookup success in an ACL "recipients" or "senders" condition, or a router
1843 "senders" option, similar to $domain_data and $local_part_data.
1845 6. Moved the writing of debug_print from before to after the "senders" test
1848 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1849 problems for message scanning, either using a data ACL, or using
1850 local_scan() because the Received: header was not generated till after they
1851 were called (in order to set the time as the time of reception completion).
1852 I have revised the way this works. The header is now generated after the
1853 body is received, but before the ACL or local_scan() are called. After they
1854 are run, the timestamp in the header is updated.
1860 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1861 before starting a queue runner without re-exec. This happened only when
1862 deliver_drop_privilege was set or when the Exim user was set to root. The
1863 effect of the bug was that timeouts during subsequent deliveries caused
1864 crashes instead of being properly handled. The handler is now left at its
1865 default (and expected) setting.
1867 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1868 message, again when deliver_drop_privilege is set or Exim is run as root.
1869 The bug described in (1) was not present in this case, but the tidying up
1870 of the other signals was missing. I have made the two cases consistent.
1872 3. The ignore_target_hosts setting on a manualroute router was being ignored
1873 for hosts that were looked up using the /MX notation.
1875 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1878 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1879 operated on the sender address. After changing the $sender_address to <>
1880 for the sender address verify, Exim was re-instated it as the original
1881 (before rewriting) address, but remembering that it had rewritten it, so it
1882 wasn't rewriting it again. This bug also had the effect of breaking the
1883 sender address verification caching when the sender address was rewritten.
1885 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1886 This has been changed so that if the ip literal address matches
1887 ignore_target_hosts, the router declines.
1889 7. Added expansion conditions match_domain, match_address, and match_local_
1890 part (NOT match_host).
1892 8. The placeholder for the Received: header didn't have a length field set.
1894 9. Added code to Exim itself and to exim_lock to test for a specific race
1895 condition that could lead to file corruption when using MBX delivery. The
1896 issue is with the lockfile that is created in /tmp. If this file is removed
1897 after a process has opened it but before that process has acquired a lock,
1898 there is the potential for a second process to recreate the file and also
1899 acquire a lock. This could lead to two Exim processes writing to the file
1900 at the same time. The added code performs the same test as UW imapd; it
1901 checks after acquiring the lock that its file descriptor still refers to
1902 the same named file.
1904 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1905 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1908 11. Added the smtp_active_hostname option. If used, this will typically be made
1909 to depend on the incoming interface address. Because $interface_address is
1910 not set up until the daemon has forked a reception process, error responses
1911 that can happen earlier (such as "too many connections") no longer contain
1914 12. If an expansion in a condition on a "warn" statement fails because a lookup
1915 defers, the "warn" statement is abandoned, and the next ACL statement is
1916 processed. Previously this caused the whole ACL to be aborted.
1918 13. Added the iplsearch lookup type.
1920 14. Added ident_timeout as a log selector.
1922 15. Added tls_certificate_verified as a log selector.
1924 16. Added a global option tls_require_ciphers (compare the smtp transport
1925 option of the same name). This controls incoming TLS connections.
1927 17. I finally figured out how to make tls_require_ciphers do a similar thing
1928 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1929 before starting the TLS session.
1931 18. Tabs are now shown as \t in -bP output.
1933 19. If the log selector return_path_on_delivery was set, Exim crashed when
1934 bouncing a message because it had too many Received: header lines.
1936 20. If two routers both had headers_remove settings, and the first one included
1937 a superfluous trailing colon, the final name in the first list and the
1938 first name in the second list were incorrectly joined into one item (with a
1939 colon in the middle).
1945 1. Added -C and -D options to the exinext utility, mainly to make it easier
1946 to include in the automated testing, but these could be helpful when
1947 multiple configurations are in use.
1949 2. The exinext utility was not formatting the output nicely when there was
1950 an alternate port involved in the retry record key, nor when there was a
1951 message id as well (for retries that were specific to a specific message
1952 and a specific host). It was also confused by IPv6 addresses, because of
1953 the additional colons they contain. I have fixed the IPv4 problem, and
1954 patched it up to do a reasonable job for IPv6.
1956 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1957 delivery, the log line now contains "pipelined" if PIPELINING was used.
1959 4. An SMTP transport process used to panic and die if the bind() call to set
1960 an explicit outgoing interface failed. This has been changed; it is now
1961 treated in the same way as a connect() failure.
1963 5. A reference to $sender_host_name in the part of a conditional expansion
1964 that was being skipped was still causing a DNS lookup. This no longer
1967 6. The def: expansion condition was not recognizing references to header lines
1968 that used bh_ and bheader_.
1970 7. Added the _cache feature to named lists.
1972 8. The code for checking quota_filecount in the appendfile transport was
1973 allowing one more file than it should have been.
1975 9. For compatibility with Sendmail, the command line option
1983 and sets the incoming protocol and host name (for trusted callers). The
1984 host name and its colon can be omitted when only the protocol is to be set.
1985 Note the Exim already has two private options, -pd and -ps, that refer to
1986 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1987 "s", but I don't think that's a major issue.
1989 10. A number of refactoring changes to the code, none of which should affect
1992 (a) The number of logging options was getting close to filling up the
1993 32-bit word that was used as a bit map. I have split them into two classes:
1994 those that are passed in the argument to log_write(), and those that are
1995 only ever tested independently outside of that function. These are now in
1996 separate 32-bit words, so there is plenty of room for expansion again.
1997 There is no change in the user interface or the logging behaviour.
1999 (b) When building, for example, log lines, the code previously used a
2000 macro that called string_cat() twice, in order to add two strings. This is
2001 not really sufficiently general. Furthermore, there was one instance where
2002 it was actually wrong because one of the argument was used twice, and in
2003 one call a function was used. (As it happened, calling the function twice
2004 did not affect the overall behaviour.) The macro has been replaced by a
2005 function that can join an arbitrary number of extra strings onto a growing
2008 (c) The code for expansion conditions now uses a table and a binary chop
2009 instead of a serial search (which was left over from when there were very
2010 few conditions). Also, it now recognizes conditions like "pam" even when
2011 the relevant support is not compiled in: a suitably worded error message is
2012 given if an attempt is made to use such a condition.
2014 11. Added ${time_interval:xxxxx}.
2016 12. A bug was causing one of the ddress fields not to be passed back correctly
2017 from remote delivery subprocesses. The field in question was not being
2018 subsequently used, so this caused to problems in practice.
2020 13. Added new log selectors queue_time and deliver_time.
2022 14. Might have fixed a bug in maildirsizefile handling that threw up
2023 "unexpected character" debug warnings, and recalculated the data
2024 unnecessarily. In any case, I expanded the warning message to give more
2027 15. Added the message "Restricted characters in address" to the statements in
2028 the default ACL that block characters like @ and % in local parts.
2030 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2031 Three changes have been made:
2033 (a) There was a serious bug; a negative response to MAIL caused the whole
2034 recipient domain to be cached as invalid, thereby blocking all messages
2035 to all local parts at the same domain, from all senders. This bug has
2036 been fixed. The domain is no longer cached after a negative response to
2037 MAIL if the sender used is not empty.
2039 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2042 (c) A new callout option, "use_sender" has been added for people who want
2043 the modified behaviour.
2049 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2052 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2053 indeed breaks things for older releases.
2055 3. Added additional logging to the case where there is a problem reading data
2056 from a filter that is running in a subprocess using a pipe, in order to
2057 try to track down a specific problem.
2059 4. Testing facility fudge: when running in the test harness and attempting
2060 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2061 getting "No route to host". Convert this to a timeout.
2063 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2066 6. Some OS don't have socklen_t but use size_t instead. This affects the
2067 fifth argument of getsockopt() amongst other things. This is now
2068 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2069 can be set for individual OS. I have set it for SunOS5, OSF1, and
2070 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2071 some earlier ones do not.
2073 7. Change 4.30/15 was not doing the test caselessly.
2075 8. The standard form for an IPv6 address literal was being rejected by address
2076 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2077 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2078 this, as well as the form without the "IPv6" on the front (but only when
2079 address literals are enabled, of course).
2081 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2083 10. Exim crashed if a message with an empty sender address specified by -f
2084 encountered a router with an errors_to setting. This could be provoked only
2085 by a command such as
2089 where an empty string was supplied; "<>" did not hit this bug.
2091 11. Installed PCRE release 4.5.
2093 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2094 remained set. It is now erased.
2096 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2097 times from message ids (which are base 36 rather than the normal 62).
2099 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2100 were being counted as actual protocol errors, and logged if the log
2101 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2102 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2103 and DATA following a set of rejected RCPTs do not count as protocol errors.
2104 In other words, Exim assumes they were pipelined, though this may not
2105 actually be the case. Of course, in all cases the client gets an
2106 appropriate error code.
2108 15. If a lookup fails in an ACL condition, a message about the failure may
2109 be available; it is used if testing the ACL cannot continue, because most
2110 such messages specify what the cause of the deferral is. However, some
2111 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2112 that caused an old message to be retained and used if a later statement
2113 caused a defer, replacing the real cause of the deferral.
2115 16. If an IP address had so many PTR records that the DNS lookup buffer
2116 was not large enough to hold them, Exim could crash while trying to process
2117 the truncated data. It now detects and logs this case.
2119 17. Further to 4.21/58, another change has been made: if (and only if) the
2120 first line of a message (the first header line) ends with CRLF, a bare LF
2121 in a subsequent header line has a space inserted after it, so as not to
2122 terminate the header.
2124 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2125 unnecessarily, used atoi() instead of strtol(), and didn't check the
2126 termination when getting file sizes from file names by regex.
2128 19. Completely re-implemented the support for maildirsize files, in the light
2129 of a number of problems with the previous contributed implementation
2130 (4.30/29). In particular:
2132 . If the quota is zero, the maildirsize file is maintained, but no quota is
2135 . If the maildir directory does not exist, it is created before any attempt
2136 to write a maildirsize file.
2138 . The quota value in the file is just a cache; if the quota is changed in
2139 the transport, the new value overrides.
2141 . A regular expression is available for excluding directories from the
2144 20. The autoreply transport checks the characters in options that define the
2145 message's headers; it allows continued headers, but it was checking with
2146 isspace() after an embedded newline instead of explicitly looking for a
2149 21. If all the "regular" hosts to which an address was routed had passed their
2150 expiry times, and had not reached their retry times, the address was
2151 bounced, even if fallback hosts were defined. Now Exim should go on to try
2154 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2155 equivalent code in the SMTP transport. Some hosts send humungous responses
2156 to HELO/EHLO, more than 1024 it seems.
2158 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2159 gives compiler warnings in some environments. I've now done it "properly",
2162 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2163 (because of problems with the built-in one) was declared to return uschar *
2164 instead of char *, causing compiler failure.
2166 25. Fixed a file descriptor leak when processing alias/forward files.
2168 26. Fixed a minor format string issue in dbfn.c.
2170 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2172 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2173 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2176 29. When a sender address is verified, it is cached, to save repeating the test
2177 when there is more than one recipient in a message. However, when the
2178 verification involves a callout, it is possible for different callout
2179 options to be set for different recipients. It is too complicated to keep
2180 track of this in the cache, so now Exim always runs a verification when a
2181 callout is required, relying on the callout cache for the optimization.
2182 The overhead is duplication of the address routing, but this should not be
2185 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2186 to be verified with callout=postmaster, and the main callout worked but the
2187 postmaster check failed, the verification correctly failed. However, if a
2188 subsequent RCPT command asked for sender verification *without* the
2189 postmaster check, incorrect caching caused this verification also to fail,
2192 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2193 it was not caching the DNS options (qualify_single, search_parents) that
2194 were used when the lookup failed. A subsequent lookup with different
2195 options therefore always gave the same answer, though there were cases
2196 where it should not have. (Example: a "domains = !$mx_any" option on a
2197 dnslookup router: the "domains" option is always processed without any
2198 widening, but the router might have qualify_single set.) Now Exim uses the
2199 cached value only when the same options are set.
2201 32. Added John Jetmore's "exipick" utility to the distribution.
2203 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2204 than a timeout (e.g. a certificate is required, and is not provided), an
2205 Exim server now closes the connection immediately. Previously it waited for
2206 the client to close - but if the client is SSL, it seems that they each
2207 wait for each other, leading to a delay before one of them times out.
2209 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2210 maintained 0.8.x compatibility because I don't think many are using it, and
2211 it is clearly obsolete.
2213 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2216 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2217 client certificate was expired. A simple patch fixes this, though I don't
2218 understand the full logic of why the verify callback is called multiple
2221 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2222 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2223 which causes problems with some clients (such as the Certicom SSL Plus
2224 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2225 disables the coutermeasure allowing Eudora to connect."
2227 38. Exim was not checking that a write() to a log file succeeded. This could
2228 lead to Bad Things if a log got too big, in particular if it hit a file
2229 size limit. Exim now panics and dies if it cannot write to a log file, just
2230 as it does if it cannot open a log file.
2232 39. Modified OS/Makefile-Linux so that it now contains
2234 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2236 The two -D definitions ensure that Exim is compiled with large file
2237 support, which makes it possible to handle log files that are bigger than
2240 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2241 instance) a domain was checked against a named list that involved a lookup,
2242 causing $domain_data to be set, then another domain was checked against the
2243 same list, then the first domain was re-checked, the value of $domain_data
2244 after the final check could be wrong. In particular, if the second check
2245 failed, it could be set empty. This bug probably also applied to
2248 41. The strip_trailing_dot option was not being applied to the address given
2249 with the -f command-line option.
2251 42. The code for reading a message's header from the spool was incrementing
2252 $received_count, but never initializing it. This meant that the value was
2253 incorrect (doubled) while delivering a message in the same process in which
2254 it was received. In the most common configuration of Exim, this never
2255 happens - a fresh exec is done - but it can happen when
2256 deliver_drop_privilege is set.
2258 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2259 it now includes up to 150 characters of the unexpected data in the log
2262 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2263 and building data strings. The size of both of these buffers was 10 000
2264 bytes - far larger than anybody would *ever* want, thought I. Needless to
2265 say, somebody hit the limit. I have increased the maximum line length to
2266 20 000 and the maximum data length of concatenated lines to 100 000. I have
2267 also fixed two bugs, because there was no checking on these buffers. Tsk,
2268 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2269 buffer is too small.
2271 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2272 lsearch lookups. Now it does.
2274 46. When parsing a route_list item in a manualroute router, a fixed-length
2275 buffer was used for the list of hosts. I made this 1024 bytes long,
2276 thinking that nobody would ever have a list of hosts that long. Wrong.
2277 Somebody had a whole pile of complicated expansion conditions, and the
2278 string was silently truncated, leading to an expansion error. It turns out
2279 that it is easier to change to an unlimited length (owing to other changes
2280 that have happened since this code was originally written) than to build
2281 structure for giving a limitation error. The length of the item that
2282 expands into the list of hosts is now unlimited.
2284 47. The lsearch lookup could not handle data where the length of text line was
2285 more than 4095 characters. Such lines were truncated, leading to shortened
2286 data being returned. It should now handle lines of any length.
2288 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2289 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2292 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2293 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2294 no longer happens in comment lines. A list of replacements is now placed
2295 at the head of all of the source files, except those whose only change is
2296 to replace PERL_COMMAND in the very first #! line.
2298 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2299 messages on the queue, with a bottom-up merge sort, using code contributed
2300 by Michael Haardt. This should make operations like -bp somewhat faster on
2301 large queues. It won't affect queue runners, except when queue_run_in_order
2304 51. Installed eximstats 1.31 in the distribution.
2306 52. Added support for SRV lookups to the dnslookup router.
2308 53. If an ACL referred to $message_body or $message_body_end, the value was not
2309 reset for any messages that followed in the same SMTP session.
2311 54. The store-handling optimization for building very long strings was not
2312 differentiating between the different store pools. I don't think this
2313 actually made any difference in practice, but I've tidied it.
2315 55. While running the routers to verify a sender address, $sender_address
2316 was still set to the sender address. This is wrong, because when routing to
2317 send a bounce to the sender, it would be empty. Therefore, I have changed
2318 it so that, while verifying a sender address, $sender_address is set to <>.
2319 (There is no change to what happens when verifying a recipient address.)
2321 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2322 target A or AAAA records (if not already returned) without resetting the
2323 qualify_single or search_parents options of the DNS resolver. These are
2324 inappropriate in this case because the targets of MX and SRV records must
2325 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2326 target that, when qualified, matched something in the local domain. These
2327 two options are now turned off when doing these lookups.
2329 57. It seems that at least some releases of Reiserfs (which does not have the
2330 concept of a fixed number of inodes) returns zero and not -1 for the
2331 number of available inodes. This interacted badly with check_spool_inodes,
2332 which assumed that -1 was the "no such thing" setting. What I have done is
2333 to check that the total number of inodes is greater than zero before doing
2334 the test of how many are available.
2336 58. When a "warn" ACL statement has a log_message modifier, the message is
2337 remembered, and not repeated. This is to avoid a lot of repetition when a
2338 message has many recipients that cause the same warning to be written.
2339 Howewer, Exim was preserving the list of already written lines for an
2340 entire SMTP session, which doesn't seem right. The memory is now reset if a
2341 new message is started.
2343 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2344 debugging output unless log_rewrite was also set.
2346 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2347 of (int)(handle) when we know that handle contains (void *)(-1).
2349 61. The Exim daemon panic-logs an error return when it closes the incoming
2350 connection. However "connection reset by peer" seems to be common, and
2351 isn't really an error worthy of noting specially, so that particular error
2354 62. When Exim is trying to find all the local interfaces, it used to panic and
2355 die if the ioctl to get the interface flags failed. However, it seems that
2356 on at least one OS (Solaris 9) it is possible to have an interface that is
2357 included in the list of interfaces, but for which you get a failure error
2358 for this call. This happens when the interface is not "plumbed" into a
2359 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2360 failure of the "get flags" call assumes that the interface is down.
2362 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2363 makes life easier for people who are doing arithmetic on fields extracted
2364 from dates, where you often get leading zeros that should not be
2365 interpreted as octal.
2367 64. Added qualify_domain to the redirect router, to override the global
2370 65. If a pathologically long header line contained very many addresses (the
2371 report of this problem mentioned 10 000) and each of them was rewritten,
2372 Exim could use up a very large amount of memory. (It kept on making new
2373 copies of the header line as it rewrote, and never released the old ones.)
2374 At the expense of a bit more processing, the header rewriting function has
2375 been changed so that it no longer eats memory in this way.
2377 66. The generation of the Received: header has been moved from the time that a
2378 message starts to be received, to the time that it finishes. The timestamp
2379 in the Received: header should now be very close to that of the <= log
2380 line. There are two side-effects of this change:
2382 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2383 logged header lines no longer include the local Received: line, because
2384 it has not yet been created. The same applies to a copy of the message
2385 that is returned to a non-SMTP sender when a message is rejected.
2387 (b) When a filter file is tested using -bf, no additional Received: header
2388 is added to the test message. After some thought, I decided that this
2391 This change does not affect the value of $received_for. It is still set
2392 after address rewriting, but before local_scan() is called.
2394 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2396 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2397 gave an unhelpful panic error message, and a defer error. I have managed to
2398 change this behaviour so that it now rejects any supplied certificate,
2399 which seems right, as the list of acceptable certificates is empty.
2401 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2402 gave an unhelpful defer error. I have not managed to make this reject any
2403 supplied certificates, but the error message it gives is "no certificate
2404 supplied", which is not helpful.
2406 70. exigrep's output now also includes lines that are not associated with any
2407 message, but which match the given pattern. Implemented by a patch from
2408 Martin Sluka, which also tidied up the Perl a bit.
2410 71. Recipient callout verification, like sender verification, was using <> in
2411 the MAIL FROM command. This isn't really the right thing, since the actual
2412 sender may affect whether the remote host accepts the recipient or not. I
2413 have changed it to use the actual sender in the callout; this means that
2414 the cache record is now keyed on a recipient/sender pair, not just the
2415 recipient address. There doesn't seem to be a real danger of callout loops,
2416 since a callout by the remote host to check the sender would use <>.
2417 [SEE ABOVE: changed after hitting problems.]
2419 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2420 temporary errors. However, in the case of such a code being given after
2421 the end of a data transmission (i.e. after ".") Exim was failing to write
2422 a retry record for the message. (Yes, there was some broken host that was
2423 actually sending 8xx at this point.)
2425 73. An unknown lookup type in a host list could cause Exim to panic-die when
2426 the list was checked. (An example that provoked this was putting <; in the
2427 middle of a list instead of at the start.) If this happened during a DATA
2428 ACL check, a -D file could be left lying around. This kind of configuration
2429 error no longer causes Exim to die; instead it causes a defer errror. The
2430 incident is still logged to the main and panic logs.
2432 74. Buglet left over from Exim 3 conversion. The message "too many messages
2433 in one connection" was written to the rejectlog but not the mainlog, except
2434 when address rewriting (yes!) was being logged.
2436 75. Added write_rejectlog option.
2438 76. When a system filter was run not as root (that is, when system_filter_user
2439 was set), the values of the $n variables were not being returned to the
2440 main process; thus, they were not subsequently available in the $sn
2443 77. Added +return_path_on_delivery log selector.
2445 78. A connection timeout was being treated differently from recipients deferred
2446 when testing hosts_max_try with a message that was older than the host's
2447 retry timeout. (The host should not be counted, thus allowing all hosts to
2448 be tried at least once before bouncing.) This may have been the cause of an
2449 occasionally reported bug whereby a message would remain on the queue
2450 longer than the retry timeout, but would be bounced if a delivery was
2451 forced. I say "may" because I never totally pinned down the problem;
2452 setting up timeout/retry tests is difficult. See also the next item.
2454 79. The ultimate address timeout was not being applied to errors that involved
2455 a combination of host plus message (for example, a timeout on a MAIL
2456 command). When an address resolved to a number of possible hosts, and they
2457 were not all tried for each delivery (e.g. because of hosts_max_try), a
2458 message could remain on the queue longer than the retry timeout.
2460 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2463 81. Fixed an obscure SMTP outgoing bug which required at least the following
2464 conditions: (a) there was another message waiting for the same server;
2465 (b) the server returned 5xx to all RCPT commands in the first message so
2466 that the message was not completed; (c) the server dropped the connection
2467 or gave a negative response to the RSET that Exim sends to abort the
2468 transaction. The observed case was a dropped connection after DATA that had
2469 been sent in pipelining mode. That is, the server had advertised PIPELINING
2470 but was not implementing it correctly. The effect of the bug was incorrect
2471 behaviour, such as trying another host, and this could lead to a crash.
2477 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2478 and daemon.c were passed as pointers to ints; they should have been
2479 pointers to socklen_t variables (which are typically unsigned ints).
2481 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2484 3. Fixed a really odd bug that affected only the testing scheme; patching a
2485 certain fixed string in the binary changed the value of another string that
2486 happened to be identical to the end of the original first string.
2488 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2489 name", it returns that address as the IP address. On some operating
2490 systems (e.g. Solaris), it also passes back the IP address string as the
2491 "host name". However, on others (e.g. Linux), it passes back an empty
2492 string. Exim wasn't checking for this, and was changing the host name to an
2493 empty string, assuming it had been canonicized.
2495 5. Although rare, it is permitted to have more than one PTR record for a given
2496 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2497 all the names associated with an address, because they do in Solaris.
2498 However, it seems that they do not in Linux for data that comes from the
2499 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2500 I found this out when I moved to a new Linux workstation and tried to run
2501 the Exim test suite.
2503 To get round this problem I have changed the code so that it now does its
2504 own call to the DNS to look up PTR records when searching for a host name.
2505 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2506 addresses that are only in /etc/hosts are still found.
2508 This behaviour is, however, controlled by an option called host_lookup_
2509 order, which defaults to "bydns:byaddr". If people want to use the other
2510 order, or indeed, just use one or the other means of lookup, they can
2511 specify it in this variable.
2513 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2514 some operating systems, this comes back from gethostbyaddr() as an empty
2515 string, and this is what Exim used to test for. However, it seems that in
2516 other systems, "." is yielded. Exim now tests for this case too.
2518 7. The values of check_spool_space and check_log_space are now held internally
2519 as a number of kilobytes instead of an absolute number of bytes. If a
2520 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2521 kilobyte. This means that much larger values can be stored.
2523 8. Exim monitor: an attempt to get the action menu when not actually pointing
2524 at a message produces an empty menu entitled "No message selected". This
2525 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2526 no entries in it ("Shell widget menu has zero width and/or height"). So I
2527 have added a single, blank menu entry in this case.
2529 9. Added ${quote_local_part.
2531 10. MIME decoding is now applied to the contents of Subject: header lines when
2534 11. Now that a reference to $sender_host_address automatically causes a reverse
2535 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2536 host lookup before query-style lookups in lists that might use this
2537 variable. This has therefore been abolished, and the "net-" prefix is no
2538 longer necessary for query-style lookups.
2540 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2541 have been a typo for LFLAGS, so it has been changed.
2543 13. The install script calls Exim with "-C /dev/null" in order to find the
2544 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2545 to be output. Howeve, since Exim outputs its version number before the
2546 error, it didn't break the script. It just looked ugly. I fixed this by
2547 always allowing "-C /dev/null" if the caller is root.
2549 14. Ignore overlarge ACL variable number when reading spool file - insurance
2550 against a later release with more variables having written the file.
2552 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2553 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2554 this, as well as the form without the "IPv6" on the front.
2556 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2557 OS/Makefile-Darwin file.
2559 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2560 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2562 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2563 variable. However, some OS use ints for inodes. I've added cast to long int
2564 to get rid of the compiler warning.
2566 19. I had forgotten to lock out "/../" in configuration file names when
2567 ALT_CONFIG_PREFIX was set.
2569 20. Routers used for verification do not need to specify transports. However,
2570 if such a router generated a host list, and callout was configured, Exim
2571 crashed, because it could not find a port number from the (non-existent)
2572 transport. It now assumes port 25 in this circumstance.
2574 21. Added the -t option to exigrep.
2576 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2577 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2578 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2579 them is set, LOOKUP_LSEARCH is forced.
2581 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2583 24. Added sender and host information to the "rejected by local_scan()" log
2584 line; previously there was no indication of these.
2586 25. Added .include_if_exists.
2588 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2589 Linux. It turns out that not all file systems support this. Apparently some
2590 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2591 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2592 not supported on the file descriptor, is now ignored when Exim is trying to
2593 sync a directory. This applies only to Linux.
2595 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2597 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2598 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2601 29. Added support for maildirsize files from supplied patch (modified a bit).
2603 30. The use of :fail: followed by an empty string could lead Exim to respond to
2604 sender verification failures with (e.g.):
2606 550 Verification failed for <xxx>
2607 550 Sender verify failed
2609 where the first response line was missing the '-' that indicates it is not
2610 the final line of the response.
2612 31. The loop for finding the name of the user that called Exim had a hardwired
2613 limit of 10; it now uses the value of finduser_retries, which is used for
2614 all other user lookups.
2616 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2619 33. Exim was neglecting to zero errno before one call of strtol() when
2620 expanding a string and expecting an integer value. On some systems this
2621 resulted in spurious "integer overflow" errors. Also, it was casting the
2622 result into an int without checking.
2624 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2625 did not work. The code looks as if it has *never* worked, though it appears
2626 to have been documented since at least releast 1.62. I have made it work.
2628 35. The "timeout_DNS" error in retry rules, also documented since at least
2629 1.62, also never worked. As it isn't clear exactly what this means, and
2630 clearly it isn't a major issue, I have abolished the feature by treating it
2631 as "timeout", and writing a warning to the main and panic logs.
2633 36. The display of retry rules for -brt wasn't always showing the error code
2636 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2637 timeout_connect_A, timeout_connect_MX.
2639 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2640 to the empty sender.
2642 39. The daemon was not analysing the content of -oX till after it had closed
2643 stderr and disconnected from the controlling terminal. This meant that any
2644 syntax errors were only noted on the panic log, and the return code from
2645 the command was 0. By re-arranging the code a little, I've made the
2646 decoding happen first, so such errors now appear on stderr, and the return
2647 code is 1. However, the actual setting up of the sockets still happens in
2648 the disconnected process, so errors there are still only recorded on the
2651 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2652 connections (as happens on some IP stacks) was logged at start up time as
2653 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2654 it from "IPv6 and IPv4", which means that two separate sockets are being
2657 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2658 says whether AF_INET or AF_INET6 was passed as an argument.
2660 42. Exiwhat output was messed up when time zones were included in log
2663 43. Exiwhat now gives more information about the daemon's listening ports,
2664 and whether -tls-on-connect was used.
2666 44. The "port" option of the smtp transport is now expanded.
2668 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2669 silently ignored. Now an error message is written to the main and panic
2672 46. There's a new ACL modifier called "logwrite" which writes to a log file
2673 as soon as it is encountered.
2675 47. Added $local_user_uid and $local_user_gid at routing time.
2677 48. Exim crashed when trying to verify a sender address that was being
2680 49. Exim was recognizing only a space character after ".include". It now also
2681 recognizes a tab character.
2683 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2684 extracting the relevant information from the specification. The man page no
2685 longer contains scrambled data for the -d option, and I've added a section
2686 at the front about calling Exim under different names.
2688 51. Added "extra_headers" argument to the "mail" command in filter files.
2690 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2693 53. Installed eximstats 1.29.
2695 54. Added transport_filter_timeout as a generic transport option.
2697 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2698 Cc: header lines. This was required by RFC 822, but it not required by RFC
2701 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2702 incoming messages that did not have them. Now it does so only if the
2703 message originates locally, that is, if there is no associated remote host
2704 address. When Resent- header lines are present, this applies to the Resent-
2705 lines rather than the non-Resent- lines.
2707 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2708 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2710 58. Messages for configuration errors now include the name of the main
2711 configuration file - useful now that there may be more than one file in a
2712 list (.included file names were always shown).
2714 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2715 for those rare installations that do not start the daemon as root or run it
2716 setuid root. I've cut out the call to initgroups() if the daemon is not
2719 60. The Exim user and group can now be bound into the binary as text strings
2720 that are looked up at the start of Exim's processing.
2722 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2724 62. Added $mailstore_basename variable.
2726 63. Installed patch to sieve.c from Michael Haardt.
2728 64. When Exim failed to open the panic log after failing to open the main log,
2729 the original message it was trying to log was written to stderr and debug
2730 output, but if they were not available (the usual case in production), it
2731 was lost. Now it is written to syslog before the two lines that record the
2732 failures to open the logs.
2734 65. Users' Exim filters run in subprocesses under the user's uid. It is
2735 possible for a "deliver" command or an alias in a "personal" command to
2736 provoke an address rewrite. If logging of address rewriting is configured,
2737 this fails because the process is not running as root or exim. There may be
2738 a better way of dealing with this, but for the moment (because 4.30 needs
2739 to be released), I have disabled address rewrite logging when running a
2740 filter in a non-root, non-exim process.
2746 1. The buildconfig auxiliary program wasn't quoting the value set for
2747 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2748 not defined. This bug was masked in 4.22 by the effect that was fixed in
2751 2. Some messages that were rejected after a message id was allocated were
2752 shown as "incomplete" by exigrep. It no longer does this for messages that
2753 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2755 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2756 have allow_domain_literals set, the ID did not get logged in the <= line.
2757 Domain literals are now always recognized in Message-ID: header lines.
2759 4. The first argument for a ${extract expansion item is the key name or field
2760 number. Leading and trailing spaces in this item were not being ignored,
2761 causing some misleading effects.
2763 5. When deliver_drop_privilege was set, single queue runner processes started
2764 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2765 same command in the process it spins off) were not dropping privilege.
2767 6. When the daemon running as "exim" started a queue runner, it always
2768 re-executed Exim in the spun-off process. This is a waste of effort when
2769 deliver_drop_privilege is set. The new process now just calls the
2770 queue-runner function directly.
2776 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2779 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2780 ignored. Though the use of -oP was forcing the writing of a pid file, it
2781 was always written to the default place.
2783 3. If the message "no IP address found for host xxxx" is generated during
2784 incoming verification, it is now followed by identification of the incoming
2785 connection (so you can more easily find what provoked it).
2787 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2789 5. Added some features to "harden" Exim a bit more against certain attacks:
2791 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2792 be put in Local/Makefile. This is like the never_users runtime option,
2793 but it cannot be overridden. The default setting is "root".
2795 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2796 prefix string with which any file named in a -C command line option
2799 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2800 is retained for -C and -D only if the caller of Exim is root. Without
2801 it, the exim user may also use -C and -D and retain privilege.
2803 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2804 command line option is disabled.
2806 6. Macro names set by the -D option must start with an upper case letter, just
2807 like macro names defined in the configuration file.
2809 7. Added "dereference=" facility to LDAP.
2811 8. Two instances of the typo "uknown" in the source files are fixed.
2813 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2814 the Configure-Makefile script screwed up while processing it.
2816 10. Incorporated PCRE 4.4.
2818 11. The SMTP synchronization check was not operating right at the start of an
2819 SMTP session. For example, it could not catch a HELO sent before the client
2820 waited for the greeting. There is now a check for outstanding input at the
2821 point when the greeting is written. Because of the duplex, asynchronous
2822 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2823 way, but not yet received, when the check is performed.
2825 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2826 on TCP/IP sockets, because this apparently causes some broken clients to
2829 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2830 unchanged) from the Cygwin maintainer.
2832 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2833 was supported instead of testing for mbx. Effectively a typo.
2835 15. The spa authenticator server code was not checking that the input it
2836 received was valid base64.
2838 16. The debug output line for the "set" modifier in ACLs was not showing the
2839 name of the variable that was being set.
2841 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2843 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2844 lookup on its own. Something else was needed to trigger the lookup. For
2845 example, a match in host_lookup or the need for a host name in a host list.
2846 Now, if $sender_host_name is referenced and the host name has not yet been
2847 looked up, a lookup is performed. If the lookup fails, the variable remains
2848 empty, and $host_lookup_failed is set to "1".
2850 19. Added "eqi" as a case-independent comparison operator.
2852 20. The saslauthd authentication condition could segfault if neither service
2853 nor realm was specified.
2855 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2856 error message that was logged was misleading, and incoming SMTP
2857 connections were dropped. The message is now more accurate, and temporary
2858 errors are given to SMTP connections.
2860 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2861 (except RSET) with a 421 error, until QUIT is received. However, it was
2862 failing to send a response to QUIT.
2864 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2865 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2866 ACL is now run after the helo_try_verify_hosts code.
2868 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2869 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2870 case-independent, so other case variants are also recognized). Apparently
2871 some systems use these upper case variants.
2873 25. If more than two messages were waiting for the same host, and a transport
2874 filter was specified for the transport, Exim sent two messages over the
2875 same TCP/IP connection, and then failed with "socket operation on non-
2876 socket" when it tried to send the third.
2878 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2880 27. The extern definition of crypt16() in expand.c was not being excluded when
2881 the OS had its own crypt16() function.
2883 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2884 as a preferred synonym for return_size_limit, both as an option and as an
2887 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2889 30. Changed the default configuration ACL to relax the local part checking rule
2890 for addresses that are not in any local domains. For these addresses,
2891 slashes and pipe symbols are allowed within local parts, but the sequence
2892 /../ is explicitly forbidden.
2894 31. SPA server authentication was not clearing the challenge buffer before
2897 32. log_message in a "warn" ACL statement was writing to the reject log as
2898 well as to the main log, which contradicts the documentation and doesn't
2899 seem right (because no rejection is happening). So I have stopped it.
2901 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2902 However, I am unable to do any testing of this.
2904 34. Fixed an infelicity in the appendfile transport. When checking directories
2905 for a mailbox, to see if any needed to be created, it was accidentally
2906 using path names with one or more superfluous leading slashes; tracing
2907 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2909 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2910 contents are added to the log line that is written for every discarded
2911 recipient. (Previously a log_message setting was ignored.)
2913 36. The ${quote: operator now quotes the string if it is empty.
2915 37. The install script runs exim in order to find its version number. If for
2916 some reason other than non-existence or emptiness, which it checks, it
2917 could not run './exim', it was installing it with an empty version number,
2918 i.e. as "exim-". This error state is now caught, and the installation is
2921 38. An argument was missing from the function that creates an error message
2922 when Exim fails to connect to the socket for saslauthd authentication.
2923 This could cause Exim to crash, or give a corrupted message.
2925 39. Added isip, isip4, and isip6 to ${if conditions.
2927 40. The ACL variables $acl_xx are now saved with the message, and can be
2928 accessed later in routers, transports, and filters.
2930 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2931 strings in the file are not string-expanded.
2933 42. If a MAIL command specified a SIZE value that was too large to fit into an
2934 int variable, the check against message_size_limit failed. Such values are
2935 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2936 day this will have to be increased, but I don't think I want to be around
2937 when emails are that large.
2944 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2945 iconv() is not standard in FreeBSD.
2947 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2948 IPv6 enabled. The observed symptom was a segmentation fault on return from
2949 the function os_common_find_running_interfaces() in src/os.c.
2951 3. In the check_special_case() function in daemon.c I had used "errno" as an
2952 argument name, which causes warnings on some systems. This was basically a
2953 typo, since it was named "eno" in the comments!
2955 4. The code that waits for the clock to tick (at a resolution of some fraction
2956 of a second) so as to ensure message-id uniqueness was always waiting for
2957 at least one whole tick, when it could have waited for less. [This is
2958 almost certainly not relevant at current processor speeds, where it is
2959 unlikely to ever wait at all. But we try to future-proof.]
2961 5. The function that sleeps for a time interval that includes fractions of a
2962 second contained a race. It did not block SIGALRM between setting the
2963 timer, and suspending (a couple of lines later). If the interval was short
2964 and the sigsuspend() was delayed until after it had expired, the suspension
2965 never ended. On busy systems this could lead to processes getting stuck for
2968 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2969 process, before it forks any delivery processes. The open lookup caching
2970 mechanism meant that the open file or database connection was passed into
2971 the delivery process. The problem was that delivery processes always tidy
2972 up cached lookup data. This could cause a problem for the next delivery
2973 process started by the queue runner, because the external queue runner
2974 process does not know about the closure. So the next delivery process
2975 still has data in the lookup cache. In the case of a file lookup, there was
2976 no problem because closing a file descriptor in a subprocess doesn't affect
2977 the parent. However, if the lookup was caching a connection to a database,
2978 the connection was closed, and the second delivery process was likely to
2979 see errors such as "PGSQL: query failed: server closed the connection
2980 unexpectedly". The problem has been fixed by closing all cached lookups
2981 in a queue runner before running a delivery process.
2983 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2984 seem to have the "const" qualifier which it has on other OS. I've
2987 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2988 *of the same type* (client or server) with the same public name that an
2989 error should be diagnosed.
2991 9. When Exim looked up a host name for an IP address, but failed to find the
2992 original IP address when looking up the host name (a safety check), it
2993 output the message "<ip address> does not match any IP for NULL", which was
2994 confusing, to say the least. The bug was that the host name should have
2995 appeared instead of "NULL".
2997 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2998 user that is built into the binary, and the -C or -D options is used, root
2999 privilege is dropped before the configuration file is read. In addition,
3000 logging is switched to stderr instead of the normal log files. If the
3001 configuration then re-defines the Exim user, the unprivileged environment
3002 is probably not what is expected, so Exim logs a panic warning message (but
3005 However, if deliver_drop_privilege is set, the unprivileged state may well
3006 be exactly what is intended, so the warning has been cut out in that case,
3007 and Exim is allowed to try to write to its normal log files.
3013 1. smtp_return_error_details was not giving details for temporary sender
3014 or receiver verification errors.
3016 2. Diagnose a configuration error if two authenticators have the same public
3019 3. Exim used not to create the message log file for a message until the first
3020 delivery attempt. This could be confusing when incoming messages were held
3021 for policy or load reasons. The message log file is now created at the time
3022 the message is received, and an initial "Received" line is written to it.
3024 4. The automatically generated man page for command line options had a minor
3025 bug that caused no ill effects; however, a more serious problem was that
3026 the procedure for building the man page automatically didn't always
3027 operate. Consequently, release 4.20 contains an out-of-date version. This
3028 shouldn't happen again.
3030 5. When building Exim with embedded Perl support, the script that builds the
3031 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3032 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3034 6. The freeze_tell option was not being used for messages that were frozen on
3035 arrival, either by an ACL or by local_scan().
3037 7. Added the smtp_incomplete_transaction log selector.
3039 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3040 was accepting AUTH without a new EHLO.
3042 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3043 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3046 10. Logging of TCP/IP connections (when configured) now happens in the main
3047 daemon process instead of the child process, so that the TCP/IP connection
3048 count is more accurate (but it can never be perfect).
3050 11. The use of "drop" in a nested ACL was not being handled correctly in the
3051 outer ACL. Now, if condition failure induced by the nested "drop" causes
3052 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3053 or "require"), the connection is dropped.
3055 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3056 that yield "discard" can now be used with an "accept" or a "discard" verb,
3057 but an error is generated for any others (because I can't see a useful way
3058 to define what should happen).
3060 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3061 are now processed in the same way as lines in the Exim configuration file.
3062 In particular, continuation lines are supported.
3064 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3066 15. Added -ti meaning -t -i.
3068 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3069 domains, and warn by logging if others are found.
3071 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3072 returned from GIFCONF when Exim is trying to find the list of interfaces on
3073 a host. The code in os.c has been modified to copy each ifreq to an aligned
3074 structure in all cases.
3076 Also, in some cases, the returned ifreq's were being copied to a 'struct
3077 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3078 means the last couple of bytes of an IPv6 address could be chopped if the
3079 ifreq contained only a normal sockaddr (14 bytes storage).
3081 18. Named domain lists were not supported in the hosts_treat_as_local option.
3082 An entry such as +xxxx was not recognized, and was treated as a literal
3085 19. Ensure that header lines added by a DATA ACL are included in the reject log
3086 if the ACL subsequently rejects the message.
3088 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3089 MD5 (which is deprecated).
3091 21. When testing a filter file using -bf, Exim was writing a message when it
3092 took the sender from a "From " line in the message, but it was not doing so
3093 when it took $return_path from a Return-Path: header line. It now does.
3095 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3096 with a valid header line field name (a series of printing characters
3097 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3099 23. Changed "disc" in the source to "disk" to conform to the documentation and
3100 the book and for uniformity.
3102 24. Ignore Sendmail's -Ooption=value command line item.
3104 25. When execve() failed while trying to run a command in a pipe transport,
3105 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3106 could be confused with a return value of 69 from the command itself. This
3107 has been changed to 127, the value the shell returns if it is asked to run
3108 a non-existent command. The wording for the related log line suggests a
3109 non-existent command as the problem.
3111 26. If received_header_text expands to an empty string, do not add a Received:
3112 header line to the message. (Well, it adds a token one on the spool, but
3113 marks it "old" so that it doesn't get used or transmitted.)
3115 27. Installed eximstats 1.28 (addition of -nt option).
3117 28. There was no check for failure on the call to getsockname() in the daemon
3118 code. This can fail if there is a shortage of resources on the system, with
3119 ENOMEM, for example. A temporary error is now given on failure.
3121 29. Contrary to the C standard, it seems that in some environments, the
3122 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3123 program. Exim now does this explicitly; it affects the formatting of
3124 timestamps using strftime().
3126 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3127 complaints. I've now initialized all the variables, to avoid this.
3129 32. Header lines added by a system filter were not being "seen" during
3130 transport-time rewrites.
3132 33. The info_callback() function passed to OpenSSL is set up with type void
3133 (*)(SSL *, int, int), as described somewhere. However, when calling the
3134 function (actually a macro) that sets it up, the type void(*)() is
3135 expected. I've put in a cast to prevent warnings from picky compilers.
3137 34. If a DNS black list lookup found a CNAME record, but there were no A
3138 records associated with the domain it pointed at, Exim crashed.
3140 35. If a DNS black list lookup returned more than one A record, Exim ignored
3141 all but the first. It now scans all returned addresses if a particular IP
3142 value is being sought. In this situation, the contents of the
3143 $dnslist_value variable are a list of all the addresses, separated by a
3146 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3147 to accept a host name and all its aliases if the forward lookup for any of
3148 them yielded the IP address of the incoming connection. Now it accepts only
3149 those names whose forward lookup yields the correct IP address. Any other
3150 names are discarded. This closes a loophole whereby a rogue DNS
3151 administrator could create reverse DNS records to break through a
3152 wildcarded host restriction in an ACL.
3154 37. If a user filter or a system filter that ran in a subprocess used any of
3155 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3156 the wrong values were passed to the pipe command ($thisaddress had the
3157 value of $0, $0 had the value of $1, etc). This bug was introduced by
3158 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3160 38. Improved the line breaking for long SMTP error messages from ACLs.
3161 Previously, if there was no break point between 40 and 75 characters, Exim
3162 left the rest of the message alone. Two changes have been made: (a) I've
3163 reduced the minimum length to 35 characters; (b) if it can't find a break
3164 point between 35 and 75 characters, it looks ahead and uses the first one
3165 that it finds. This may give the occasional overlong line, but at least the
3166 remaining text gets split now.
3168 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3169 file descriptors might be low, and that setting 1000 would always raise it.
3170 It turns out that in some environments, the limit is already over 1000 and
3171 that lowering it causes trouble. So now Exim takes care not to decrease it.
3173 40. When delivering a message, the value of $return_path is set to $sender_
3174 address at the start of routing (routers may change the value). By an
3175 oversight, this default was not being set up when an address was tested by
3176 -bt or -bv, which affected the outcome if any router or filter referred to
3179 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3180 log only when "message" or "log_message" are set. However, if one of the
3181 conditions was an address verification, or a call to a nested ACL, the
3182 messages generated by the underlying test were being passed through. This
3183 no longer happens. The underlying message is available in $acl_verify_
3184 message for both "message" and "log_message" expansions, so it can be
3185 passed through if needed.
3187 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3188 new expansion $bh_ to give the encoded byte string without charset
3189 translation. Translation happens only if iconv() is available; HAVE_ICONV
3190 indicates this at build time. HEADERS_CHARSET gives the charset to
3191 translate to; headers_charset can change it in the configuration, and
3192 "headers charset" can change it in an individual filter file.
3194 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3195 that creates RFC 2047 encoded "words" labels them as that charset instead
3196 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3197 expansion operator; (ii) when Exim creates a From: line for a local
3198 message; (iii) when a header line is rewritten to include a "phrase" part.
3200 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3201 buggy, causing it to skip the first lines of messages whose message ID
3202 ended in 'D'. This would not have bitten before Exim release 4.14, because
3203 message IDs were unlikely to end in 'D' before then. The effect was to have
3204 incorrect size information for certain domains.
3206 45. #include "config.h" was missing at the start of the crypt16.c module. This
3207 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3210 46. If there was a timeout during a "random" callout check, Exim treated it as
3211 a failure of the random address, and carried on sending RSET and the real
3212 address. If the delay was just some slowness somewhere, the response to the
3213 original RCPT would be taken as a response to RSET and so on, causing
3214 mayhem of various kinds.
3216 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3217 when I implemented it. It didn't allow for the fact that some option values
3218 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3219 the right test for positive values.
3221 48. Domain names in DNS records are case-independent. Exim always looks them up
3222 in lower case. Some resolvers return domain names in exactly the case they
3223 appear in the zone file, that is, they may contain uppercase letters. Not
3224 all resolvers do this - some return always lower case. Exim was treating a
3225 change of case by a resolver as a change of domain, similar to a widening
3226 of a domain abbreviation. This triggered its re-routing code and so it was
3227 trying to route what was effectively the same domain again. This normally
3228 caused routing to fail (because the router wouldn't handle the domain
3229 twice). Now Exim checks for this case specially, and just changes the
3230 casing of the domain that it ultimately uses when it transmits the message
3233 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3236 50. If a filter generated a file delivery with a non-absolute name (possible if
3237 no home directory exists for the router), the forbid_file option was not
3240 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3241 the existing equality matching.
3243 52. Exim was using ints instead of ino_t variables in some places where it was
3244 dealing with inode numbers.
3246 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3247 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3248 called TMPDIR, and if it finds it is different, it changes its value.
3250 54. The smtp_printf() function is now made available to local_scan() so
3251 additional output lines can be written before returning. There is also an
3252 smtp_fflush() function to enable the detection of a dropped connection.
3253 The variables smtp_input and smtp_batched_input are exported to
3256 55. Changed the default runtime configuration: the message "Unknown user"
3257 has been removed from the ACL, and instead placed on the localuser router,
3258 using the cannot_route_message feature. This means that any verification
3259 failures that generate their own messages won't get overridden. Similarly,
3260 the "Unrouteable address" message that was in the ACL for unverifiable
3261 relay addresses has also been removed.
3263 56. Added hosts_avoid_esmtp to the smtp transport.
3265 57. The exicyclog script was not checking for the esoteric option
3266 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3267 will work only if exicyclog is run under the appropriate euid.
3269 58. Following a discussion on the list, the rules by which Exim recognises line
3270 endings on incoming messages have been changed. The -dropcr and drop_cr
3271 options are now no-ops, retained only for backwards compatibility. The
3272 following line terminators are recognized: LF CRLF CR. However, special
3273 processing applies to CR:
3275 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3276 nor a local message in the state where . is a terminator.
3278 (ii) If a bare CR is encountered in a header line, an extra space is added
3279 after the line terminator so as not to end the header. The reasoning
3280 behind this is that bare CRs in header lines are most likely either
3281 to be mistakes, or people trying to play silly games.
3283 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3284 was being incorrectly given as 18 bytes larger than it should have been.
3285 This is a VOB (very old bug).
3287 60. This may never have affected anything current, but just in case it has:
3288 When the local host is found other than at the start of a list of hosts,
3289 the local host, those with the same MX, and any that follow, are discarded.
3290 When the list in question was part of a longer list of hosts, the following
3291 hosts (not currently being processed) were also being discarded. This no
3292 longer happens. I'm not sure if this situation could ever has previously
3295 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3298 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3299 header. This is something that is recommended in a new Internet Draft, and
3300 is something that is documented as being done by Sendmail. There are two
3301 possible values. For messages generated by the autoreply transport, Exim
3304 Auto-Submitted: auto-replied
3306 whereas for all other generated messages (e.g. bounces) it adds
3308 Auto-Submitted: auto-generated
3310 63. The "personal" condition in filters now includes a test for the
3311 Auto-Submitted: header. If it contains the string "auto-" the message it
3312 not considered personal.
3314 64. Added rcpt_include_affixes as a generic transport option.
3316 65. Added queue_only_override (default true).
3318 66. Added the syslog_duplication option.
3320 67. If what should have been the first header line of a message consisted of
3321 a space followed by a colon, Exim was mis-interpreting it as a header line.
3322 It isn't of course - it is syntactically invalid and should therefore be
3323 treated as the start of the message body. The misbehaviour could have
3324 caused a number of strange effects, including loss of data in subsequent
3325 header lines, and spool format errors.
3327 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3328 client host had authenticated. This control can now be exercised by an ACL
3329 for more flexibility.
3331 69. By default, callouts do not happen when testing with -bh. There is now a
3332 variant, -bhc, which does actually run the callout code, including
3333 consulting and updating the callout cache.
3335 70. Added support for saslauthd authentication, courtesy of Alexander
3338 71. If statvfs() failed on the spool or log directories while checking their
3339 size for availability, Exim confusingly gave the error "space shortage".
3340 Furthermore, in debugging mode it crashed with a floating point exception.
3341 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3342 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3343 serious problem, Exim now writes to the main and panic logs when this
3344 happens, with details of the failure. It then refuses to accept the
3345 incoming message, giving the message "spool directory problem" or "log
3346 directory problem" with a 421 code for SMTP messages.
3348 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3349 0, 1, and 2 exist, because some OS complain for execs without them (see
3350 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3351 descriptors. However, the code omitted to check that the open succeeded,
3352 causing mysterious errors if for some reason the permissions on /dev/null
3353 got screwed. Now Exim writes a message to the main and panic logs, and
3354 bombs out if it can't open /dev/null.
3356 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3357 interact so that it is all more flexible. It is supposed to remain
3358 backwards compatible. Also added extra_local_interfaces.
3360 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3361 to bomb out with an assertion failure - to the client this appears as a
3362 connection drop. This problem occurs in the part of the code that was taken
3363 from the Samba project. Fortunately, the assertion is in a very simple
3364 function, so I have fixed this by reproducing the function inline in the
3365 one place where it is called, and arranging for authentication to fail
3366 instead of killing the process with assert().
3368 75. The SPA client code was not working when the server requested OEM rather
3369 than Unicode encoding.
3371 76. Added code to make require_files with a specific uid setting more usable in
3372 the case where statting the file as root fails - usually a non-root-mounted
3373 NFS file system. When this happens and the failure is EACCES, Exim now
3374 forks a subprocess and does the per-uid checking as the relevant uid.
3376 77. Added process_log_path.
3378 78. If log_file_path was not explicitly set, a setting of check_log_space or
3379 check_log_inodes was ignored.
3381 79. If a space check for the spool or log partitions fails, the incident is now
3382 logged. Of course, in the latter case the data may get lost...
3384 80. Added the %p formatting code to string_format() so that it can be used to
3385 print addresses in debug_print(). Adjusted all the address printing in the
3386 debugging in store.c to use %p rather than %d.
3388 81. There was a concern that a line of code in smtp_in.c could overflow a
3389 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3390 initially expressed, the concern was not well-founded, because trailing
3391 spaces are removed early. However, if the trailing spaces were followed by
3392 a NULL, they did not get removed, so the overflow was possible. Two fixes
3395 (a) I re-wrote the offending code in a cleaner fashion.
3396 (b) If an incoming SMTP command contains a NULL character, it is rejected
3399 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3400 runs initgroups(), so that if the Exim user is in any additional groups,
3401 they will be used during message reception.
3407 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob