1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.242 2005/10/03 09:56:42 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.
22 TK/01 Added the "success_on_redirect" address verification option. See
23 NewStuff for rationale and an example.
25 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
27 PH/02 Patch to exigrep to allow it to work on syslog lines.
29 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
30 fread() to skip over the body file's header line, because in Cygwin the
31 header line is locked and is inaccessible.
33 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
34 co-exist for some time) to make it clear that it is the Exim ID that is
35 referenced, not the Message-ID: header line.
37 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
38 string_format() function, because snprintf() does not exist on all
41 PH/06 The use of forbid_filter_existstest now also locks out the use of the
42 ${stat: expansion item.
44 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
45 protocol synchronization error", to keep the pedants happy.
47 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
48 well as for IRIX systems, when gcc is being used. See the host.c source
51 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
53 PH/10 Named domain lists were not working if used in a queue_smtp_domains
56 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
57 transport and to the smtp transport in LMTP mode.
59 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
61 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
62 run a filter in a subprocess. This could lead to confusion in subsequent
63 lookups in the parent process. There should also be a search_tidyup() at
64 the end of the subprocess.
66 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
67 only if the host matched helo_try_verify_hosts, which caused the
68 verification to occur when the EHLO/HELO command was issued. The ACL just
69 tested the remembered result. Now, if a previous verification attempt has
70 not happened, "verify = helo" does it there and then.
72 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
74 TK/03 Fix log output including CR from clamd.
76 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
77 exist provoked a memory error which could cause a segfault.
79 PH/15 Installed PCRE 6.2
81 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
83 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
84 of the problem. Specifically, suggested +O2 rather than +O1 for the
87 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
89 PH/20 If a delivery was routed to a non-standard port by means of an SRV
90 record, the port was not correctly logged when the outgoing_port log
91 selector was set (it logged the transort's default port).
93 PH/21 Added support for host-specific ports to manualroute, queryprogram,
94 fallback_hosts, and "hosts" in the smtp transport.
96 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
97 host errors such as "Connection refused".
99 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
100 authentication with radiusclient 0.4.9:
102 - Error returned from rc_read_config was caught wrongly
103 - Username/password not passed on to radius server due to wrong length.
105 The presumption is that some radiusclient API changes for 4.51/PH/17
106 were not taken care of correctly. The code is still untested by me (my
107 Linux distribution still has 0.3.2 of radiusclient), but it was
108 contributed by a Radius user.
110 PH/24 When doing a callout, the value of $domain wasn't set correctly when
111 expanding the "port" option of the smtp transport.
113 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
114 while reading a MIME header. Thanks to Tom Hughes for a patch.
116 PH/24 Include config.h inside local_scan.h so that configuration settings are
119 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
120 that in an ACL for RCPT (for example), you can examine exactly what was
123 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
124 commands, but it was not correctly comparing the address with the actual
125 client host address. Thus, it would show the EHLO address in Received:
126 header lines when this was not necessary.
128 PH/27 Added the % operator to ${eval:}.
130 PH/28 Exim tries to create and chdir to its spool directory when it starts;
131 it should be ignoring failures (because with -C, for example, it has lost
132 privilege). It wasn't ignoring creation failures other than "already
135 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
138 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
139 because an input line was too long, either on its own, or by virtue of
140 too many continuations, the temporary file was not being removed, and the
141 return code was incorrect.
143 PH/31 Missing "BOOL" in function definition in filtertest.c.
145 PH/32 Applied Sieve patches from the maintainer.
147 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
149 PH/33 Added "verify = not_blind".
151 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
152 Local/Makefile (with some defaults set). These are used in built scripts
153 such as exicyclog, but they have never been used in the exim_install
154 script (though there are many overriding facilities there). I have
155 arranged that the exim_install script now takes note of these two
158 PH/35 Installed configuration files for Dragonfly.
160 PH/36 When a locally submitted message by a trusted user did not contain a
161 From: header, and the sender address was obtained from -f or from an SMTP
162 MAIL command, and the trusted user did not use -F to supply a sender
163 name, $originator_name was incorrectly used when constructing a From:
164 header. Furthermore, $originator_name was used for submission mode
165 messages from external hosts without From: headers in a similar way,
166 which is clearly wrong.
168 PH/37 Added control=suppress_local_fixups.
170 PH/38 When log_selector = +received_sender was set, and the addition of the
171 sender made the log line's construction buffer exactly full, or one byte
172 less than full, an overflow happened when the terminating "\n" was
175 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
176 when the result of a list match is failure because a DNS lookup failed.
178 PH/40 RM_COMMAND is now used in the building process.
180 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
181 the "build-* directories that it finds.
183 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
184 address in a domain literal was a prefix of an interface address.
186 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
187 when verifying a sender address, unless rewrite_headers is false.
189 PH/44 Wrote a long comment about why errors_to addresses are verified as
190 recipients, not senders.
192 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
193 the ratelimit ACL was added.
195 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
197 PH/46 Added extra information about PostgreSQL errors to the error string.
199 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
200 avoiding the need to install special zones in a real server. This is
201 backwards compatible; if it can't find the fake resolver, it drops back.
202 Thus, both old and new test suites can be run.
204 TF/02 Added util/ratelimit.pl
206 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
207 clock is set back in time.
209 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
210 Candler <B.Candler@pobox.com>.
212 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
213 OK for addresses that are the result of redirections.
215 PH/48 A number of further additions for the benefit of the new test suite,
216 including a fake gethostbyname() that interfaces to the fake DNS resolver
219 TF/06 The fix for widen_domains has also been applied to qualify_single and
220 search_parents which are the other dnslookup options that can cause
223 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
226 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
228 TF/07 Exim produced the error message "an SRV record indicated no SMTP
229 service" if it encountered an MX record with an empty target hostname.
230 The message is now "an MX or SRV record indicated no SMTP service".
232 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
233 if the DNS of the sending site is misconfigured. This is quite a
234 common situation. This change restores the behaviour of treating a
235 helo verification defer as a failure.
237 PH/51 If self=fail was set on a router, the bounce message did not include the
238 actual error message.
244 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
246 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
247 command itself ended in error, the underlying message about the transport
248 filter timeout was being overwritten with the pipe command error. Now the
249 underlying error message should be appended to the second error message.
251 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
253 PH/02 Reduce the amount of output that "make" produces by default. Full output
254 can still be requested.
256 PH/03 The warning log line about a condition test deferring for a "warn" verb
257 was being output only once per connection, rather than after each
258 occurrence (because it was using the same function as for successful
259 "warn" verbs). This seems wrong, so I have changed it.
261 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
262 it should not have, which might have caused a crash in the right
263 circumstances, but probably never did.
265 PH/04 Installed a modified version of Tony Finch's patch to make submission
266 mode fix the return path as well as the Sender: header line, and to
267 add a /name= option so that you can make the user's friendly name appear
270 TF/03 Added the control = fakedefer ACL modifier.
272 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
273 Mark Lowes for thorough testing.
275 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
277 TK/03 Merged latest SRS patch from Miles Wilton.
279 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
280 with the definition in sysexits.h (which is #included earlier).
281 Fortunately, Exim does not actually use EX_OK. The code used to try to
282 preserve the sysexits.h value, by assumimg that macro definitions were
283 scanned for macro replacements. I have been disabused of this notion,
284 so now the code just undefines EX_OK before #including unistd.h.
286 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
287 in the smtp transport. When a block could not be written in a single
288 write() function, the timeout was being re-applied to each part-write.
289 This seems wrong - if the receiver was accepting one byte at a time it
290 would take for ever. The timeout is now adjusted when this happens. It
291 doesn't have to be particularly precise.
293 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
294 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
296 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
297 without a domain if the check to <postmaster@domain> fails.
299 SC/01 Eximstats: added -xls and the ability to specify output files
300 (patch written by Frank Heydlauf).
302 SC/02 Eximstats: use FileHandles for outputing results.
304 SC/03 Eximstats: allow any combination of xls, txt, and html output.
306 SC/04 Eximstats: fixed display of large numbers with -nvr option
308 SC/05 Eximstats: fixed merging of reports with empty tables.
310 SC/06 Eximstats: added the -include_original_destination flag
312 SC/07 Eximstats: removed tabs and trailing whitespace.
314 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
316 TK/06 MBOX spool code: Add real "From " MBOX separator line
317 so the .eml file is really in mbox format (even though
318 most programs do not really care). Patch from Alex Miller.
320 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
321 The latter is generated from $received_to and is only set if the
322 message has one envelope recipient. SA can use these headers,
323 obviously out-of-the-box. Patch from Alex Miller.
325 PH/08 The ${def test on a variable was returning false if the variable's
326 value was "0", contrary to what the specification has always said!
327 The result should be true unless the variable is empty.
329 PH/09 The syntax error of a character other than { following "${if
330 def:variable_name" (after optional whitespace) was not being diagnosed.
331 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
332 accidental colon was present, for example, could give incorrect results.
334 PH/10 Tidied the code in a number of places where the st_size field of a stat()
335 result is used (not including appendfile, where other changes are about
338 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
339 This involved changing a lot of size variables from int to off_t. It
340 should work with maildirs and everything.
342 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
343 spamd dying while we are connected to it.
345 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
348 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
349 function for BDB 4.3.
351 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
353 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
354 the sources for PCRE have been re-arranged and more files are now
357 PH/15 The code I had for printing potentially long long variables in PH/11
358 above was not the best (it lost precision). The length of off_t variables
359 is now inspected at build time, and an appropriate printing format (%ld
360 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
361 to be "long long int" or "long int". This is needed for the internal
362 formatting function string_vformat().
364 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
365 the configuration file to be ":syslog", then the script "guesses" where
366 the logs files are, rather than using the compiled in default. In our
367 case the guess is not the same as the compiled default, so the script
368 suddenly stopped working when I started to use syslog. The patch checks
369 to see if log_file_path is "". If so, it attempts to read it from exim
370 with no configuration file to get the compiled in version, before it
371 falls back to the previous guessing code."
373 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
374 implementing BATV in an Exim configuration. See NewStuff for the gory
377 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
378 Makefile that are specific to HP-UX.
380 PH/18 If the "use_postmaster" option was set for a recipient callout together
381 with the "random" option, the postmaster address was used as the MAIL
382 FROM address for the random test, but not for the subsequent recipient
383 test. It is now used for both.
385 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
386 patch removes a few documentation additions to RFC 3028, because the
387 latest draft now contains them. It adds the new en;ascii-case comparator
388 and a new error check for 8bit text in MIME parts. Comparator and
389 require names are now matched exactly. I enabled the subaddress
390 extension, but it is not well tested yet (read: it works for me)."
392 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
393 rework some of the code of TK/09 above to avoid the hardwired use of
394 "%lld" and "long long". Replaced the call to snprintf() with a call to
397 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
398 records point to non-existent hosts", "retry timeout exceeded", and
399 "retry time not reached for any host after a long failure period".
401 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
402 experimental DomainKeys support:
404 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
405 (2) On an error such as an illegally used "control", the wrong name for
406 the control was given.
408 These problems did NOT occur unless DomainKeys support was compiled.
410 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
412 PH/24 Added ${if match_ip condition.
414 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
415 databases so that it will be absolutely obvious if a crash occurs in the
416 DB library. This is a regular occurrence (often caused by mis-matched
419 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
420 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
421 picked up on a user's system that detects such things. There doesn't seem
422 to be a gcc warning option for this - only an attribute that has to be
423 put on the function's prototype. It seems that in Fedora Core 4 they have
424 set this on a number of new functions. No doubt there will be more in due
427 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
428 specify a transport. However, if an address that was verified by such a
429 router was the subject of a callout, Exim crashed because it tried to
430 read the rcpt_include_affixes from the non-existent transport. Now it
431 just assumes that the setting of that option is false. This bug was
432 introduced by 4.51/PH/31.
434 PH/28 Changed -d+all to exclude +memory, because that information is very
435 rarely of interest, but it makes the output a lot bigger. People tend to
436 do -d+all out of habit.
438 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
439 code in os-type was giving problems when libc.so lives in lib64, like on
442 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
443 aren't the modern standard, and it seems that some systems' include files
444 don't always have them. Exim was already checking for some of the newer
445 ones like T_AAAA, and defining it itself. I've added checks for all the
446 record types that Exim uses.
448 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
449 not automatically generating a new one, as it is supposed to. This
450 prevented TLS from working. If the file did exist, but contained invalid
451 data, a new version was generated, as expected. It was only the case of a
452 non-existent file that was broken.
454 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
455 with a change in libdomainkeys > 0.64.
457 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
458 from DNS. If the selector record carries the flag, it now has
459 precedence over the domain-wide flag.
461 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
463 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
464 the use of a port name that isn't defined in /etc/services) occurred, the
465 message was deferred as in a normal delivery, and thus remained on the
466 spool, instead of being failed because of the mua_wrapper setting. This
467 is now fixed, and I tidied up some of the mua_wrapper messages at the
470 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
471 the messages in a hash of arrays rather than using individual hashes.
472 This is a bit cleaner and results in dramatic memory savings, albeit
473 at a slight CPU cost.
475 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
476 as requested by Marc Sherman.
478 SC/10 Eximstats: added histograms for user specified patterns as requested
481 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
483 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
484 fopen() in the content-scanning modules that did not already have it.
490 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
491 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
493 TK/02 Fix ACL "control" statement not being available in MIME ACL.
495 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
497 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
498 to test Sieve filters that use "vacation".
500 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
501 that changes the way the GnuTLS parameters are stored in the cache file.
502 The new format can be generated externally. For backward compatibility,
503 if the data in the cache doesn't make sense, Exim assumes it has read an
504 old-format file, and it generates new data and writes a new file. This
505 means that you can't go back to an older release without removing the
508 PH/03 A redirect router that has both "unseen" and "one_time" set does not
509 work if there are any delivery delays because "one_time" forces the
510 parent to be marked "delivered", so its unseen clone is never tried
511 again. For this reason, Exim now forbids the simultaneous setting of
514 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
515 redirected to themselves ("homonym" addresses). Read the long ChangeLog
516 entry if you want to know the details. The fix, however, neglected to
517 consider the case when local delivery batching is involved. The test for
518 "previously delivered" was not happening when checking to see if an
519 address could be batched with a previous (undelivered) one; under
520 certain circumstances this could lead to multiple deliveries to the same
523 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
524 in its include files, and this causes problems building Exim.
526 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
527 header_syntax) but Exim was just ignoring anything given after a slash.
528 In particular, this caused confusion with an attempt to use "verify =
529 reverse_host_lookup/defer_ok". An error is now given when options are
530 supplied for verify items that do not have them. (Maybe reverse_host_
531 lookup should have a defer_ok option, but that's a different point.)
533 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
534 defined by RFC 821) to 2048, because there were problems with some AUTH
535 commands, and RFC 1869 says the size should be increased for extended
536 SMTP commands that take arguments.
538 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
541 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
542 "unknown" error; now it says that the functionality isn't in the binary.
544 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
545 an address' error message when a string expansion fails (syntax or
546 whatever). Otherwise the password may appear in the log. Following change
547 PH/42 below, there is no longer a chance of it appearing in a bounce
550 PH/11 Installed exipick version 20050225.0 from John Jetmore.
552 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
553 of its addresses was ever tried. (Bugzilla bug #2.)
555 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
556 the result incorrectly in the debug output. (It correctly added a newline
557 to what was transported.)
559 TF/01 Added $received_time.
561 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
562 commented out examples of how to interface to a virus scanner and to
563 SpamAssassin. Also added commented examples of av_scanner and
564 spamd_address settings.
566 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
567 and controls are allowed in which ACLs. There were a couple of minor
568 errors. Some of the entries in the conditions table (which is a table of
569 where they are NOT allowed) were getting very unwieldy; rewrote them as a
570 negation of where the condition IS allowed.
572 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
574 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
575 header file does not have a version number, so I've had to invent a new
576 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
577 API. The code is untested by me (my Linux distribution still has 0.3.2 of
578 radiusclient), but it was contributed by a Radius user.
580 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
581 files or directories, for OpenSSL.
583 PH/19 When an Exim process that is running as root has to create an Exim log
584 file, it does so in a subprocess that runs as exim:exim so as to get the
585 ownership right at creation (otherwise, other Exim processes might see
586 the file with the wrong ownership). There was no test for failure of this
587 fork() call, which would lead to the process getting stuck as it waited
588 for a non-existent subprocess. Forks do occasionally fail when resources
589 run out. I reviewed all the other calls to fork(); they all seem to check
592 PH/20 When checking for unexpected SMTP input at connect time (before writing
593 the banner), Exim was not dealing correctly with a non-positive return
594 from the read() function. If the client had disconnected by this time,
595 the result was a log entry for a synchronization error with an empty
596 string after "input=" when read() returned zero. If read() returned -1
597 (an event I could not check), uninitialized data bytes were printed.
598 There were reports of junk text (parts of files, etc) appearing after
601 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
603 PH/22 Added support for macro redefinition, and (re)definition in between
604 driver and ACL definitions.
606 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
607 forgetting to use the resulting value; it was using the unexpanded value.
609 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
610 hadn't been configured. The fix is from Juergen Kreileder, who
611 understands it better than I do:
613 "Here's what I see happening with three configured cyrus_sasl
614 authenticators configured (plain, login, cram-md5):
616 On startup auth_cyrus_sasl_init() gets called for each of these.
617 This means three calls to sasl_listmech() without a specified mech_list.
618 => SASL tests which mechs of all available mechs actually work
619 => three warnings about OTP not working
620 => the returned list contains: plain, login, cram-md5, digest-md5, ...
622 With the patch, sasl_listmech() also gets called three times. But now
623 SASL's mech_list option is set to the server_mech specified in the the
624 authenticator. Or in other words, the answer from sasl_listmech()
625 gets limited to just the mech you're testing for (which is different
627 => the return list contains just 'plain' or 'login', 'cram-md5' or
628 nothing depending on the value of ob->server_mech.
630 I've just tested the patch: Authentication still works fine,
631 unavailable mechs specified in the exim configuration are still
632 caught, and the auth.log warnings about OTP are gone."
634 PH/25 When debugging is enabled, the contents of the command line are added
635 to the debugging output, even when log_selector=+arguments is not
638 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
639 answer is "GNU", and only if the return is "GNU/something" is the answer
642 PH/27 $acl_verify_message is now set immediately after the failure of a
643 verification in an ACL, and so is available in subsequent modifiers. In
644 particular, the message can be preserved by coding like this:
646 warn !verify = sender
647 set acl_m0 = $acl_verify_message
649 Previously, $acl_verify_message was set only while expanding "message"
650 and "log_message" when a very denied access.
652 PH/28 Modified OS/os.c-Linux with
654 -#ifndef OS_LOAD_AVERAGE
655 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
657 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
658 nomenclature these days.)
660 PH/29 Installed patch from the Sieve maintainer that adds the options
661 sieve_useraddress and sieve_subaddress to the redirect router.
663 PH/30 In these circumstances:
664 . Two addresses routed to the same list of hosts;
665 . First host does not offer TLS;
666 . First host accepts first address;
667 . First host gives temporary error to second address;
668 . Second host offers TLS and a TLS session is established;
669 . Second host accepts second address.
670 Exim incorrectly logged both deliveries with the TLS parameters (cipher
671 and peerdn, if requested) that were in fact used only for the second
674 PH/31 When doing a callout as part of verifying an address, Exim was not paying
675 attention to any local part prefix or suffix that was matched by the
676 router that accepted the address. It now behaves in the same way as it
677 does for delivery: the affixes are removed from the local part unless
678 rcpt_include_affixes is set on the transport.
680 PH/32 Add the sender address, as F=<...>, to the log line when logging a
681 timeout during the DATA phase of an incoming message.
683 PH/33 Sieve envelope tests were broken for match types other than :is. I have
684 applied a patch sanctioned by the Sieve maintainer.
686 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
687 the uid or gid is negative. A case of a negative gid caused this to be
688 noticed. The fix allows for either to be negative.
690 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
691 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
694 PH/36 The change PH/12 above was broken. Fixed it.
696 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
697 the grounds that routing the same address twice would always produce the
698 same answer. This might have been true once, but it is certainly no
699 longer true now. Routing a child address may depend on the previous
700 routing that produced that child. Some complicated redirection strategies
701 went wrong when messages had multiple recipients, and made Exim's
702 behaviour dependent on the order in which the addresses were given.
704 I have moved the duplicate checking until after the routing is complete.
705 Exim scans the addresses that are assigned to local and remote
706 transports, and removes any duplicates. This means that more work will be
707 done, as duplicates will always all be routed, but duplicates are
708 presumably rare, so I don't expect this is of any significance.
710 For deliveries to pipes, files, and autoreplies, the duplicate checking
711 still happens during the routing process, since they are not going to be
714 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
715 It corrects a timeout issue with spamd. This is Ian's comment: "The
716 background is that sometimes spamd either never reads data from a
717 connection it has accepted, or it never writes response data. The exiscan
718 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
719 blindly assumes that writes won't block so it may never time out."
721 PH/39 Allow G after quota size as well as K and M.
723 PH/40 The value set for $authenticated_id in an authenticator may not contain
724 binary zeroes or newlines because the value is written to log lines and
725 to spool files. There was no check on this. Now the value is run through
726 the string_printing() function so that such characters are converted to
727 printable escape sequences.
729 PH/41 $message_linecount is a new variable that contains the total number of
730 lines in the message. Compare $body_linecount, which is the count for the
733 PH/42 Exim no longer gives details of delivery errors for specific addresses in
734 bounce and delay warning messages, except in certain special cases, which
737 (a) An SMTP error message from a remote host;
738 (b) A message specified in a :fail: redirection;
739 (c) A message specified in a "fail" command in a system filter;
740 (d) A message specified in a FAIL return from the queryprogram router;
741 (e) A message specified by the cannot_route_message router option.
743 In these cases only, Exim does include the error details in bounce and
744 warning messages. There are also a few cases where bland messages such
745 as "unrouteable address" or "local delivery error" are given.
747 PH/43 $value is now also set for the "else" part of a ${run expansion.
749 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
750 being worked on, but at least Exim now implements the latest version to
753 PH/45 In a pipe transport, although a timeout while waiting for the pipe
754 process to complete was treated as a delivery failure, a timeout while
755 writing the message to the pipe was logged, but erroneously treated as a
756 successful delivery. Such timeouts include transport filter timeouts. For
757 consistency with the overall process timeout, these timeouts are now
758 treated as errors, giving rise to delivery failures by default. However,
759 there is now a new Boolean option for the pipe transport called
760 timeout_defer, which, if set TRUE, converts the failures into defers for
761 both kinds of timeout. A transport filter timeout is now identified in
764 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
765 systems where "make" and "gmake" are different, calling "gmake" at top
766 level broke things. I've arranged for the value of $(MAKE) to be passed
767 from the Makefile to this script so that it can call the same version of
771 A note about Exim versions 4.44 and 4.50
772 ----------------------------------------
774 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
775 changes of various kinds. As a consequence, a big documentation update was
776 needed. This delayed the release for rather longer than seemed good, especially
777 in the light of a couple of (minor) security issues. Therefore, the changes
778 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
779 release. So 4.44 and 4.50 are in effect two different branches that both start
782 I have left the 4.50 change log unchanged; it contains all the changes since
783 4.43. The change log for 4.44 is below; many of its items are identical to
784 those for 4.50. This seems to be the most sensible way to preserve the
785 historical information.
791 1. Minor wording change to the doc/README.SIEVE file.
793 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
794 computation of the current number of files was incorrect.
796 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
797 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
800 4. Give more explanation in the error message when the command for a transport
801 filter fails to execute.
803 5. There are several places where Exim runs a non-Exim command in a
804 subprocess. The SIGUSR1 signal should be disabled for these processes. This
805 was being done only for the command run by the queryprogram router. It is
806 now done for all such subprocesses. The other cases are: ${run, transport
807 filters, and the commands run by the lmtp and pipe transports.
809 6. Added CONFIGURE_GROUP build-time option.
811 7. Some older OS have a limit of 256 on the maximum number of file
812 descriptors. Exim was using setrlimit() to set 1000 as a large value
813 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
814 systems. I've change it so that if it can't get 1000, it tries for 256.
816 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
817 was an oversight, and furthermore, ever since the addition of extra
818 controls (e.g. 4.43/32), the checks on when to allow different forms of
819 "control" were broken. There should now be diagnostics for all cases when a
820 control that does not make sense is encountered.
822 9. Added the /retain_sender option to "control=submission".
824 10. $recipients is now available in the predata ACL (oversight).
826 11. Tidy the search cache before the fork to do a delivery from a message
827 received from the command line. Otherwise the child will trigger a lookup
828 failure and thereby defer the delivery if it tries to use (for example) a
829 cached ldap connection that the parent has called unbind on.
831 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
832 of $address_data from the recipient verification was clobbered by the
835 13. The value of address_data from a sender verification is now available in
836 $sender_address_data in subsequent conditions in the ACL statement.
838 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
840 15. Added a new option "connect=<time>" to callout options, to set a different
843 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
844 was its contents. (It was OK if the option was not defined at all.)
846 17. A "Completed" log line is now written for messages that are removed from
847 the spool by the -Mrm option.
849 18. New variables $sender_verify_failure and $recipient_verify_failure contain
850 information about exactly what failed.
852 19. Added -dd to debug only the daemon process.
854 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
855 handles timeouts, both on the server side and network timeouts. Renamed the
856 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
858 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
859 It is now set to "smtps".
861 22. $host_address is now set to the target address during the checking of
864 23. When checking ignore_target_hosts for an ipliteral router, no host name was
865 being passed; this would have caused $sender_host_name to have been used if
866 matching the list had actually called for a host name (not very likely,
867 since this list is usually IP addresses). A host name is now passed as
870 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
871 code that specifies a non-restarting handler (typically sigaction() in
872 modern systems) in an attempt to fix a rare and obscure crash bug.
874 25. Narrowed the window for a race in the daemon that could cause it to ignore
875 SIGCHLD signals. This is not a major problem, because they are used only to
876 wake it up if nothing else does.
878 26. A malformed maildirsize file could cause Exim to calculate negative values
879 for the mailbox size or file count. Odd effects could occur as a result.
880 The maildirsize information is now recalculated if the size or filecount
883 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
884 support for a long time. Removed HAVE_SYS_VFS_H.
886 28. Installed the latest version of exipick from John Jetmore.
888 29. In an address list, if the pattern was not a regular expression, an empty
889 subject address (from a bounce message) matched only if the pattern was an
890 empty string. Non-empty patterns were not even tested. This was the wrong
891 because it is perfectly reasonable to use an empty address as part of a
892 database query. An empty address is now tested by patterns that are
893 lookups. However, all the other forms of pattern expect the subject to
894 contain a local part and a domain, and therefore, for them, an empty
895 address still always fails if the pattern is not itself empty.
897 30. Exim went into a mad DNS loop when attempting to do a callout where the
898 host was specified on an smtp transport, and looking it up yielded more
901 31. Re-factored the code for checking spool and log partition space into a
902 function that finds that data and another that does the check. The former
903 is then used to implement four new variables: $spool_space, $log_space,
904 $spool_inodes, and $log_inodes.
906 32. The RFC2047 encoding function was originally intended for short strings
907 such as real names; it was not keeping to the 75-character limit for
908 encoded words that the RFC imposes. It now respects the limit, and
909 generates multiple encoded words if necessary. To be on the safe side, I
910 have increased the buffer size for the ${rfc2047: expansion operator from
913 33. It is now permitted to omit both strings after an "if" condition; if the
914 condition is true, the result is "true". As before, when the second string
915 is omitted, a false condition yields an empty string. This makes it less
916 cumbersome to write custom ACL and router conditions.
918 34. Failure to deliver a bounce message always caused it to be frozen, even if
919 there was an errors_to setting on the router. The errors_to setting is now
922 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
923 canonical form (fully expanded) before being placed in
924 $sender_host_address.
926 36. The table in the code that translates DNS record types into text (T_A to
927 "A" for instance) was missing entries for NS and CNAME. It is just possible
928 that this could have caused confusion if both these types were looked up
929 for the same domain, because the text type is used as part of Exim's
930 per-process caching. But the chance of anyone hitting this buglet seems
933 37. The dnsdb lookup has been extended in a number of ways.
935 (1) There is a new type, "zns", which walks up the domain tree until it
936 finds some nameserver records. It should be used with care.
938 (2) There is a new type, "mxh", which is like "mx" except that it returns
939 just the host names, not the priorities.
941 (3) It is now possible to give a list of domains (or IP addresses) to be
942 looked up. The behaviour when one of the lookups defers can be
943 controlled by a keyword.
945 (4) It is now possible to specify the separator character for use when
946 multiple records are returned.
948 38. The dnslists ACL condition has been extended: it is now possible to supply
949 a list of IP addresses and/or domains to be looked up in a particular DNS
952 39. Added log_selector=+queue_time_overall.
954 40. When running the queue in the test harness, wait just a tad after forking a
955 delivery process, to get repeatability of debugging output.
957 41. Include certificate and key file names in error message when GnuTLS fails
958 to set them up, because the GnuTLS error message doesn't include the name
959 of the failing file when there is a problem reading it.
961 42. Allow both -bf and -bF in the same test run.
963 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
965 44. The "Exiscan patch" is now merged into the mainline Exim source.
967 45. Sometimes the final signoff response after QUIT could fail to get
968 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
969 < 0 before doing a fflush(). This bug looks as though it goes back to the
970 introduction of TLS in release 3.20, but "sometimes" must have been rare
971 because the tests only now provoked it.
973 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
974 (this can affect the format of dates).
976 47. exim_tidydb, when checking for the continued existence of a message for
977 which it has found a message-specific retry record, was not finding
978 messages that were in split spool directories. Consequently, it was
979 deleting retry records that should have stayed in existence.
981 48. Steve fixed some bugs in eximstats.
983 49. The SPA authentication driver was not abandoning authentication and moving
984 on to the next authenticator when an expansion was forced to fail,
985 contradicting the general specification for all authenticators. Instead it
986 was generating a temporary error. It now behaves as specified.
988 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
989 (the order specifies the preference for clients). The order is now AES256,
990 AES128, 3DES, ARCFOUR128.
992 51. Small patch to Sieve code - explicitly set From: when generating an
995 52. Exim crashed if a remote delivery caused a very long error message to be
996 recorded - for instance if somebody sent an entire SpamAssassin report back
997 as a large number of 550 error lines. This bug was coincidentally fixed by
998 increasing the size of one of Exim's internal buffers (big_buffer) that
999 happened as part of the Exiscan merge. However, to be on the safe side, I
1000 have made the code more robust (and fixed the comments that describe what
1003 53. Now that there can be additional text after "Completed" in log lines (if
1004 the queue_time_overall log selector is set), a one-byte patch to exigrep
1005 was needed to allow it to recognize "Completed" as not the last thing in
1008 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1009 patch that reportedly fixes this has been added. I am not expert enough to
1010 create a test for it. This is what the patch creator wrote:
1012 "I found a little strange behaviour of ldap code when working with
1013 Windows 2003 AD Domain, where users was placed in more than one
1014 Organization Units. When I tried to give exim partial DN, the exit code
1015 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1016 But simultaneously result of request was absolutely normal ldap result,
1017 so I produce this patch..."
1019 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1020 REFERENCE, so I have modified the code to exclude the patch when that macro
1023 55. Some experimental protocols are using DNS PTR records for new purposes. The
1024 keys for these records are domain names, not reversed IP addresses. The
1025 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1026 leaves it alone. Component reversal etc. now happens only for IP addresses.
1029 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1031 57. Double the size of the debug message buffer (to 2048) so that more of very
1032 long debug lines gets shown.
1034 58. The exicyclog utility now does better if the number of log files to keep
1035 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1037 59. Two changes related to the smtp_active_hostname option:
1039 (1) $smtp_active_hostname is now available as a variable.
1040 (2) The default for smtp_banner uses $smtp_active_hostname instead
1041 of $primary_hostname.
1043 60. The host_aton() function is supposed to be passed a string that is known
1044 to be a valid IP address. However, in the case of IPv6 addresses, it was
1045 not checking this. This is a hostage to fortune. Exim now panics and dies
1046 if the condition is not met. A case was found where this could be provoked
1047 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1048 components; fortuitously, this particular loophole had already been fixed
1049 by change 4.50/55 above.
1051 If there are any other similar loopholes, the new check in host_aton()
1052 itself should stop them being exploited. The report I received stated that
1053 data on the command line could provoke the exploit when Exim was running as
1054 exim, but did not say which command line option was involved. All I could
1055 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1056 running as the user.
1059 61. There was a buffer overflow vulnerability in the SPA authentication code
1060 (which came originally from the Samba project). I have added a test to the
1061 spa_base64_to_bits() function which I hope fixes it.
1064 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1065 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1067 63. The daemon start-up calls getloadavg() while still root for those OS that
1068 need the first call to be done as root, but it missed one case: when
1069 deliver_queue_load_max is set with deliver_drop_privilege. This is
1070 necessary for the benefit of the queue runner, because there is no re-exec
1071 when deliver_drop_privilege is set.
1073 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1074 This has been fixed.
1076 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1077 was in use, was not putting the data itself into the right store pool;
1078 consequently, it could be overwritten for a subsequent message in the same
1079 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1082 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1084 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1085 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1086 the function were treating the return as a boolean value, which happened to
1087 work because 0=false and not-0=true, but is not correct code.
1089 68. The host_aton() function was not handling scoped IPv6 addresses (those
1090 with, for example, "%eth0" on the end) correctly.
1092 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1093 negated items (that is, ~something) in unsigned ints. Some compilers
1094 apparently mutter when there is no cast.
1096 70. If an address verification called from an ACL failed, and did not produce a
1097 user-specific message (i.e. there was only a "system" message), nothing was
1098 put in $acl_verify_message. In this situation, it now puts the system
1101 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1102 session; change 4.31/43 added the unwanted input to the log line - except
1103 that it did not do this in the start of session case. It now does.
1105 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1106 This is wrong and can cause the other end to generate a synchronization
1107 error if it is another Exim or anything else that does the synchronization
1108 check. A QUIT command is no longer sent after a timeout.
1110 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1111 during host lookups.
1113 74. The defer_ok option of callout verification was not working if it was used
1114 when verifying addresses in header lines, that is, for this case:
1116 verify = header_sender/callout=defer_ok
1118 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1119 those file descriptors could be used for SMTP connections. If anything
1120 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1121 could be sent to the SMTP client, causing chaos. The daemon now opens
1122 stdin, stdout, and stderr to /dev/null when it puts itself into the
1125 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1126 log by default. The user can override this with suitable Perl magic.
1128 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1129 the log message when discard triggers, was not working for the DATA ACL or
1130 for the non-SMTP ACL.
1132 78. Error message wording change in sieve.c.
1134 79. If smtp_accept_max_per_host was set, the number of connections could be
1135 restricted to fewer than expected, because the daemon was trying to set up
1136 a new connection before checking whether the processes handling previous
1137 connections had finished. The check for completed processes is now done
1138 earlier. On busy systems, this bug wouldn't be noticed because something
1139 else would have woken the daemon, and it would have reaped the completed
1142 80. If a message was submitted locally by a user whose login name contained one
1143 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1144 It caused a spool format error. I have fixed the spool reading code. A
1145 related problem was that the "from" clause in the Received: line became
1146 illegal because of the space(s). It is now covered by ${quote_local_part.
1148 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1151 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1153 83. Patch from Sieve maintainer to fix unterminated string problem in
1154 "vacation" handling.
1156 84. Some minor changes to the Linux configuration files to help with other
1157 OS variants using glibc.
1159 85. One more patch for Sieve to update vacation handling to latest spec.
1162 ----------------------------------------------------
1163 See the note above about the 4.44 and 4.50 releases.
1164 ----------------------------------------------------
1170 1. Change 4.43/35 introduced a bug that caused file counts to be
1171 incorrectly computed when quota_filecount was set in an appendfile
1174 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1175 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1178 3. Give more explanation in the error message when the command for a transport
1179 filter fails to execute.
1181 4. There are several places where Exim runs a non-Exim command in a
1182 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1183 was being done only for the command run by the queryprogram router. It is
1184 now done for all such subprocesses. The other cases are: ${run, transport
1185 filters, and the commands run by the lmtp and pipe transports.
1187 5. Some older OS have a limit of 256 on the maximum number of file
1188 descriptors. Exim was using setrlimit() to set 1000 as a large value
1189 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1190 systems. I've change it so that if it can't get 1000, it tries for 256.
1192 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1193 was an oversight, and furthermore, ever since the addition of extra
1194 controls (e.g. 4.43/32), the checks on when to allow different forms of
1195 "control" were broken. There should now be diagnostics for all cases when a
1196 control that does not make sense is encountered.
1198 7. $recipients is now available in the predata ACL (oversight).
1200 8. Tidy the search cache before the fork to do a delivery from a message
1201 received from the command line. Otherwise the child will trigger a lookup
1202 failure and thereby defer the delivery if it tries to use (for example) a
1203 cached ldap connection that the parent has called unbind on.
1205 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1206 of $address_data from the recipient verification was clobbered by the
1207 sender verification.
1209 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1210 was its contents. (It was OK if the option was not defined at all.)
1212 11. A "Completed" log line is now written for messages that are removed from
1213 the spool by the -Mrm option.
1215 12. $host_address is now set to the target address during the checking of
1216 ignore_target_hosts.
1218 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1219 being passed; this would have caused $sender_host_name to have been used if
1220 matching the list had actually called for a host name (not very likely,
1221 since this list is usually IP addresses). A host name is now passed as
1224 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1225 code that specifies a non-restarting handler (typically sigaction() in
1226 modern systems) in an attempt to fix a rare and obscure crash bug.
1228 15. Narrowed the window for a race in the daemon that could cause it to ignore
1229 SIGCHLD signals. This is not a major problem, because they are used only to
1230 wake it up if nothing else does.
1232 16. A malformed maildirsize file could cause Exim to calculate negative values
1233 for the mailbox size or file count. Odd effects could occur as a result.
1234 The maildirsize information is now recalculated if the size or filecount
1237 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1238 support for a long time. Removed HAVE_SYS_VFS_H.
1240 18. Updated exipick to current release from John Jetmore.
1242 19. Allow an empty sender to be matched against a lookup in an address list.
1243 Previously the only cases considered were a regular expression, or an
1246 20. Exim went into a mad DNS lookup loop when doing a callout where the
1247 host was specified on the transport, if the DNS lookup yielded more than
1250 21. The RFC2047 encoding function was originally intended for short strings
1251 such as real names; it was not keeping to the 75-character limit for
1252 encoded words that the RFC imposes. It now respects the limit, and
1253 generates multiple encoded words if necessary. To be on the safe side, I
1254 have increased the buffer size for the ${rfc2047: expansion operator from
1257 22. Failure to deliver a bounce message always caused it to be frozen, even if
1258 there was an errors_to setting on the router. The errors_to setting is now
1261 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1262 canonical form (fully expanded) before being placed in
1263 $sender_host_address.
1265 24. Updated eximstats to version 1.33
1267 25. Include certificate and key file names in error message when GnuTLS fails
1268 to set them up, because the GnuTLS error message doesn't include the name
1269 of the failing file when there is a problem reading it.
1271 26. Expand error message when OpenSSL has problems setting up cert/key files.
1274 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1275 (this can affect the format of dates).
1277 28. exim_tidydb, when checking for the continued existence of a message for
1278 which it has found a message-specific retry record, was not finding
1279 messages that were in split spool directories. Consequently, it was
1280 deleting retry records that should have stayed in existence.
1282 29. eximstats updated to version 1.35
1283 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1284 1.35 - bugfix such that pie charts by volume are generated correctly
1286 30. The SPA authentication driver was not abandoning authentication and moving
1287 on to the next authenticator when an expansion was forced to fail,
1288 contradicting the general specification for all authenticators. Instead it
1289 was generating a temporary error. It now behaves as specified.
1291 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1292 (the order specifies the preference for clients). The order is now AES256,
1293 AES128, 3DES, ARCFOUR128.
1295 31. Small patch to Sieve code - explicitly set From: when generating an
1298 32. Exim crashed if a remote delivery caused a very long error message to be
1299 recorded - for instance if somebody sent an entire SpamAssassin report back
1300 as a large number of 550 error lines. This bug was coincidentally fixed by
1301 increasing the size of one of Exim's internal buffers (big_buffer) that
1302 happened as part of the Exiscan merge. However, to be on the safe side, I
1303 have made the code more robust (and fixed the comments that describe what
1306 33. Some experimental protocols are using DNS PTR records for new purposes. The
1307 keys for these records are domain names, not reversed IP addresses. The
1308 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1309 leaves it alone. Component reversal etc. now happens only for IP addresses.
1312 34. The host_aton() function is supposed to be passed a string that is known
1313 to be a valid IP address. However, in the case of IPv6 addresses, it was
1314 not checking this. This is a hostage to fortune. Exim now panics and dies
1315 if the condition is not met. A case was found where this could be provoked
1316 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1317 components; fortuitously, this particular loophole had already been fixed
1318 by change 4.50/55 or 4.44/33 above.
1320 If there are any other similar loopholes, the new check in host_aton()
1321 itself should stop them being exploited. The report I received stated that
1322 data on the command line could provoke the exploit when Exim was running as
1323 exim, but did not say which command line option was involved. All I could
1324 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1325 running as the user.
1328 35. There was a buffer overflow vulnerability in the SPA authentication code
1329 (which came originally from the Samba project). I have added a test to the
1330 spa_base64_to_bits() function which I hope fixes it.
1333 36. The daemon start-up calls getloadavg() while still root for those OS that
1334 need the first call to be done as root, but it missed one case: when
1335 deliver_queue_load_max is set with deliver_drop_privilege. This is
1336 necessary for the benefit of the queue runner, because there is no re-exec
1337 when deliver_drop_privilege is set.
1339 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1340 was in use, was not putting the data itself into the right store pool;
1341 consequently, it could be overwritten for a subsequent message in the same
1342 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1345 38. Sometimes the final signoff response after QUIT could fail to get
1346 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1347 < 0 before doing a fflush(). This bug looks as though it goes back to the
1348 introduction of TLS in release 3.20, but "sometimes" must have been rare
1349 because the tests only now provoked it.
1355 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1356 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1357 Now it returns an integer. A number of calls were still expecting a T/F
1358 return. Fortuitously, in all cases, the tests worked in OK situations,
1359 which is the norm. However, things would have gone wrong on any write
1360 failures on the smtp file descriptor. This function is used when sending
1361 messages over SMTP and also when doing verify callouts.
1363 2. When Exim is called to do synchronous delivery of a locally submitted
1364 message (the -odf or -odi options), it no longer closes stderr before doing
1367 3. Implemented the mua_wrapper option.
1369 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1371 5. Implemented the functions header_remove(), header_testname(),
1372 header_add_at_position(), and receive_remove_recipient(), and exported them
1375 6. If an ACL "warn" statement specified the addition of headers, Exim already
1376 inserted X-ACL-Warn: at the start if there was no header name. However, it
1377 was not making this test for the second and subsequent header lines if
1378 there were newlines in the string. This meant that an invalid header could
1379 be inserted if Exim was badly configured.
1381 7. Allow an ACL "warn" statement to add header lines at the start or after all
1382 the Received: headers, as well as at the end.
1384 8. Added the rcpt_4xx retry error code.
1386 9. Added postmaster_mailfrom=xxx to callout verification option.
1388 10. Added mailfrom=xxxx to the callout verification option, for verify=
1391 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1392 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1394 12. Inserted some casts to stop certain compilers warning when using pointer
1395 differences as field lengths or precisions in printf-type calls (mostly
1396 affecting debugging statements).
1398 13. Added optional readline() support for -be (dynamically loaded).
1400 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1401 same clock tick as a message's arrival, so that its received time was the
1402 same as the "first fail" time on the retry record, and that message
1403 remained on the queue past the ultimate address timeout, every queue runner
1404 would try a delivery (because it was past the ultimate address timeout) but
1405 after another failure, the ultimate address timeout, which should have then
1406 bounced the address, did not kick in. This was a "< instead of <=" error;
1407 in most cases the first failure would have been in the next clock tick
1408 after the received time, and all would be well.
1410 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1411 being recognized when the domain list was tested by the match_domain
1412 condition in an expansion string.
1414 16. Added the ${str2b64: operator.
1416 17. Exim was always calling setrlimit() to set a large limit for the number of
1417 processes, without checking whether the existing limit was already
1418 adequate. (It did check for the limit on file descriptors.) Furthermore,
1419 errors from getrlimit() and setrlimit() were being ignored. Now they are
1420 logged to the main and panic logs, but Exim does carry on, to try to do its
1421 job under whatever limits there are.
1423 18. Imported PCRE 5.0.
1425 19. Trivial typo in log message " temporarily refused connection" (the leading
1428 20. If the log selector return_path_on_delivery was set and an address was
1429 redirected to /dev/null, the delivery process crashed because it assumed
1430 that a return path would always be set for a "successful" delivery. In this
1431 case, the whole delivery is bypassed as an optimization, and therefore no
1434 21. Internal re-arrangement: the function for sending a challenge and reading
1435 a response while authentication was assuming a zero-terminated challenge
1436 string. It's now changed to take a pointer and a length, to allow for
1437 binary data in such strings.
1439 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1441 23. Exim was not respecting finduser_retries when seeking the login of the
1442 uid under which it was called; it was always trying 10 times. (The default
1443 setting of finduser_retries is zero.) Also, it was sleeping after the final
1444 failure, which is pointless.
1446 24. Implemented tls_on_connect_ports.
1448 25. Implemented acl_smtp_predata.
1450 26. If the domain in control=submission is set empty, Exim assumes that the
1451 authenticated id is a complete email address when it generates From: or
1452 Sender: header lines.
1454 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1455 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1456 chown and chgrp in /bin and hostname in /usr/bin.
1458 28. Exim was keeping the "process log" file open after each use, just as it
1459 does for the main log. This opens the possibility of it remaining open for
1460 long periods when the USR1 signal hits a daemon. Occasional processlog
1461 errors were reported, that could have been caused by this. Anyway, it seems
1462 much more sensible not to leave this file open at all, so that is what now
1465 29. The long-running daemon process does not normally write to the log once it
1466 has entered its main loop, and it closes the log before doing so. This is
1467 so that log files can straightforwardly be renamed and moved. However,
1468 there are a couple of unusual error situations where the daemon does write
1469 log entries, and I had neglected to close the log afterwards.
1471 30. The text of an SMTP error response that was received during a remote
1472 delivery was being truncated at 512 bytes. This is too short for some of
1473 the long messages that one sometimes sees. I've increased the limit to
1476 31. It is now possible to make retry rules that apply only when a message has a
1477 specific sender, in particular, an empty sender.
1479 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1480 it possible to be selective about when SMTP synchronization is enforced.
1482 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1484 32. Implemented hosts_connection_nolog.
1486 33. Added an ACL for QUIT.
1488 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1491 35. Added mailbox_size and mailbox_filecount to appendfile.
1493 36. Added control = no_multiline_responses to ACLs.
1495 37. There was a bug in the logic of the code that waits for the clock to tick
1496 in the case where the clock went backwards by a substantial amount such
1497 that the microsecond fraction of "now" was more than the microsecond
1498 fraction of "then" (but the whole seconds number was less).
1500 38. Added support for the libradius Radius client library this is found on
1501 FreeBSD (previously only the radiusclient library was supported).
1507 1. When certain lookups returned multiple values in the form name=value, the
1508 quoting of the values was not always being done properly. Specifically:
1509 (a) If the value started with a double quote, but contained no whitespace,
1511 (b) If the value contained whitespace other than a space character (i.e.
1512 tabs or newlines or carriage returns) it was not quoted.
1513 This fix has been applied to the mysql and pgsql lookups by writing a
1514 separate quoting function and calling it from the lookup code. The fix
1515 should probably also be applied to nisplus, ibase and oracle lookups, but
1516 since I cannot test any of those, I have not disturbed their existing code.
1518 2. A hit in the callout cache for a specific address caused a log line with no
1519 reason for rejecting RCPT. Now it says "Previous (cached) callout
1520 verification failure".
1522 3. There was an off-by-one bug in the queryprogram router. An over-long
1523 return line was truncated at 256 instead of 255 characters, thereby
1524 overflowing its buffer with the terminating zero. As well as fixing this, I
1525 have increased the buffer size to 1024 (and made a note to document this).
1527 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1528 when Exim is waiting for an SMTP response from a remote server, Exim
1529 restarts its select() call on the socket, thereby resetting its timeout.
1530 This is not a problem when such interrupts are rare. Somebody set up a cron
1531 job to run exiwhat every 2 minutes, which is less than the normal select()
1532 timeout (5 or 10 minutes). This meant that the select() timeout never
1533 kicked in because it was always reset. I have fixed this by comparing the
1534 time when an interrupt arrives with the time at the start of the first call
1535 to select(). If more time than the timeout has elapsed, the interrupt is
1536 treated as a timeout.
1538 5. Some internal re-factoring in preparation for the addition of Sieve
1539 extensions (by MH). In particular, the "personal" test is moved to a
1540 separate function, and given an option for scanning Cc: and Bcc: (which is
1541 not set for Exim filters).
1543 6. When Exim created an email address using the login of the caller as the
1544 local part (e.g. when creating a From: or Sender: header line), it was not
1545 quoting the local part when it contained special characters such as @.
1547 7. Installed new OpenBSD configuration files.
1549 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1550 try to make them clearer.
1552 9. Callout options, other than the timeout value, were being ignored when
1553 verifying sender addresses in header lines. For example, when using
1555 verify = header_sender/callout=no_cache
1557 the cache was (incorrectly) being used.
1559 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1560 only the Cygwin environment.
1562 11. When return_path_on_delivery was set as a log selector, if different remote
1563 addresses in the same message used different return paths and parallel
1564 remote delivery occurred, the wrong values would sometimes be logged.
1565 (Whenever a remote delivery process finished, the return path value from
1566 the most recently started remote delivery process was logged.)
1568 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1569 lines when AUTH and/or TLS are in use. This is the "received protocol"
1570 field. Exim used to use "asmtp" for authenticated SMTP, without any
1571 indication (in the protocol name) for TLS use. Now it follows the RFC and
1572 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1573 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1574 names appear in log lines as well as in Received: header lines.
1576 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1577 extensions, and comparison tests, and to fix some bugs.
1579 14. Changes to the "personal" filter test:
1581 (1) The test was buggy in that it was just doing the equivalent of
1582 "contains" tests on header lines. For example, if a user's address was
1583 anne@some.where, the "personal" test would incorrectly be true for
1585 To: susanne@some.where
1587 This test is now done by extracting each address from the header in turn,
1588 and checking the entire address. Other tests that are part of "personal"
1589 are now done using regular expressions (for example, to check local parts
1590 of addresses in From: header lines).
1592 (2) The list of non-personal local parts in From: addresses has been
1593 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1594 taken from the Sieve specification recommendations.
1596 (3) If the message contains any header line starting with "List-" it is
1597 treated as non-personal.
1599 (4) The test for "circular" in the Subject: header line has been removed
1600 because it now seems ill-conceived.
1602 15. Minor typos in src/EDITME comments corrected.
1604 16. Installed latest exipick from John Jetmore.
1606 17. If headers_add on a router specified a text string that was too long for
1607 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1608 of string_sprintf() is now avoided.
1610 18. $message_body_size was not set (it was always zero) when running the DATA
1611 ACL and the local_scan() function.
1613 19. For the "mail" command in an Exim filter, no default was being set for
1614 the once_repeat time, causing a random time value to be used if "once" was
1615 specified. (If the value happened to be <= 0, no repeat happened.) The
1616 default is now 0s, meaning "never repeat". The "vacation" command was OK
1617 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1618 (I found it when inspecting the code).
1620 20. There is now an overall timeout for performing a callout verification. It
1621 defaults to 4 times the callout timeout, which applies to individual SMTP
1622 commands during the callout. The overall timeout applies when there is more
1623 than one host that can be tried. The timeout is checked before trying the
1624 next host. This prevents very long delays if there are a large number of
1625 hosts and all are timing out (e.g. when the network connections are timing
1626 out). The value of the overall timeout can be changed by specifying an
1627 additional sub-option for "callout", called "maxwait". For example:
1629 verify = sender/callout=5s,maxwait=20s
1631 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1632 to the end before writing, but this should make it even safer).
1634 22. Exim was forgetting that it had advertised PIPELINING for the second and
1635 subsequent messages on an SMTP connection. It was also not resetting its
1636 memory on STARTTLS and an internal HELO.
1638 23. When Exim logs an SMTP synchronization error within a session, it now
1639 records whether PIPELINING has been advertised or not.
1641 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1642 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1643 rather than long int.
1645 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1647 26. Added the never_mail option to autoreply.
1653 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1654 crash if the getsockname() call failed; this can happen if a connection is
1655 closed very soon after it is established. The problem was simply in the
1656 order in which certain operations were done, causing Exim to try to write
1657 to the SMTP stream before it had set up the file descriptor. The bug has
1658 been fixed by making things happen in the correct order.
1664 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1665 before the connection was closed, thus losing the rejection response.
1667 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1668 some early Solaris releases, but causes trouble in current releases where
1669 socklen_t is defined.
1671 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1674 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1677 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1678 that releasing the top of it at the end releases what was used for sub-
1679 expansions (unless the block got too big). However, discard this block if
1680 the first thing is a variable or header, so that we can use its block when
1681 it is dynamic (useful for very large $message_headers, for example).
1683 6. Lookups now cache *every* query, not just the most recent. A new, separate
1684 store pool is used for this. It can be recovered when all lookup caches are
1685 flushed. Lookups now release memory at the end of their result strings.
1686 This has involved some general refactoring of the lookup sources.
1688 7. Some code has been added to the store_xxx() functions to reduce the amount
1689 of flapping under certain conditions.
1691 8. log_incoming_interface used to affect only the <= reception log lines. Now
1692 it causes the local interface and port to be added to several more SMTP log
1693 lines, for example "SMTP connection from", and rejection lines.
1695 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1697 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1699 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1700 could be overwritten at the end of the current message (or the start of a
1701 new message if it was set in a HELO ACL). The value is now preserved for
1702 the duration of the SMTP connection.
1704 12. If a transport had a headers_rewrite setting, and a matching header line
1705 contained an unqualified address, that address was qualified, even if it
1706 did not match any rewriting rules. The underlying bug was that the values
1707 of the flags that permit the existence of unqualified sender and recipient
1708 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1709 non-local messages, and by -bnq for local messages) were not being
1710 preserved with the message after it was received.
1712 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1713 "next input=" as part of the text comprising the host identity instead of
1714 the correct text. The code was using the same buffer for two different
1715 strings. However, depending on which order the printing function evaluated
1716 its arguments, the bug did not always show up. Under Linux, for example, my
1717 test suite worked just fine.
1719 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1720 doesn't work with some older versions of Perl. It has been changed to "my",
1721 which in any case is probably the better facility to use.
1723 15. A really picky compiler found some instances of statements for creating
1724 error messages that either had too many or two few arguments for the format
1727 16. The size of the buffer for calls to the DNS resolver has been increased
1728 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1729 for addresses that have a lot of PTR records. This alleviates a problem; it
1730 does not fully solve it.
1732 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1733 buffer now truncates the list and logs the incident, which is the same
1734 action as happens when Exim is looking up a host name and its aliases.
1735 Previously in this situation something unpredictable would happen;
1736 sometimes it was "internal error: store_reset failed".
1738 18. If a server dropped the connection unexpectedly when an Exim client was
1739 using GnuTLS and trying to read a response, the client delivery process
1740 crashed while trying to generate an error log message.
1742 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1743 string, for example:
1745 warn message = H1: something\nH2: something
1747 the text was added as a single header line from Exim's point of view
1748 though it ended up OK in the delivered message. However, searching for the
1749 second and subsequent header lines using $h_h2: did not work. This has been
1750 fixed. Similarly, if a system filter added multiple headers in this way,
1751 the routers could not see them.
1753 20. Expanded the error message when iplsearch is called with an invalid key to
1754 suggest using net-iplsearch in a host list.
1756 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1757 ACLs are no longer actually imposed (and a message to that effect is
1760 22. If a "gecos" field in a passwd entry contained escaped characters, in
1761 particular, if it contained a \" sequence, Exim got it wrong when building
1762 a From: or a Sender: header from that name. A second bug also caused
1763 incorrect handling when an unquoted " was present following a character
1764 that needed quoting.
1766 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1767 was not being matched caselessly.
1769 24. Arranged for all hyphens in the exim.8 source to be escaped with
1772 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1773 properly. Recipient callout cache records were still being keyed to include
1774 the sender, even when use_sender was set false. This led to far more
1775 callouts that were necessary. The sender is no longer included in the key
1776 when use_sender is false.
1778 26. Added "control = submission" modifier to ACLs.
1780 27. Added the ${base62d: operator to decode base 62 numbers.
1782 28. dnsdb lookups can now access SRV records.
1784 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1785 the configuration file.
1787 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1788 (-v) mode. This makes the output for a verbose queue run more intelligible.
1790 31. Added a use_postmaster feature to recipient callouts.
1792 32. Added the $body_zerocount variable, containing the number of binary zero
1793 bytes in the message body.
1795 33. The time of last modification of the "new" subdirectory is now used as the
1796 "mailbox time last read" when there is a quota error for a maildir
1799 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1801 35. Added +ignore_unknown as a special item in host lists.
1803 36. Code for decoding IPv6 addresses in host lists is now included, even if
1804 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1805 address was recognized as an IP address, but was then not correctly decoded
1806 into binary, causing unexpected and incorrect effects when compared with
1813 1. Very minor rewording of debugging text in manualroute to say "list of
1814 hosts" instead of "hostlist".
1816 2. If verify=header_syntax was set, and a header line with an unqualified
1817 address (no domain) and a large number of spaces between the end of the
1818 name and the colon was received, the reception process suffered a buffer
1819 overflow, and (when I tested it) crashed. This was caused by some obsolete
1820 code that should have been removed. The fix is to remove it!
1822 3. When running in the test harness, delay a bit after writing a bounce
1823 message to get a bit more predictability in the log output.
1825 4. Added a call to search_tidyup() just before forking a reception process. In
1826 theory, someone could use a lookup in the expansion of smtp_accept_max_
1827 per_host which, without the tidyup, could leave open a database connection.
1829 5. Added the variables $recipient_data and $sender_data which get set from a
1830 lookup success in an ACL "recipients" or "senders" condition, or a router
1831 "senders" option, similar to $domain_data and $local_part_data.
1833 6. Moved the writing of debug_print from before to after the "senders" test
1836 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1837 problems for message scanning, either using a data ACL, or using
1838 local_scan() because the Received: header was not generated till after they
1839 were called (in order to set the time as the time of reception completion).
1840 I have revised the way this works. The header is now generated after the
1841 body is received, but before the ACL or local_scan() are called. After they
1842 are run, the timestamp in the header is updated.
1848 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1849 before starting a queue runner without re-exec. This happened only when
1850 deliver_drop_privilege was set or when the Exim user was set to root. The
1851 effect of the bug was that timeouts during subsequent deliveries caused
1852 crashes instead of being properly handled. The handler is now left at its
1853 default (and expected) setting.
1855 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1856 message, again when deliver_drop_privilege is set or Exim is run as root.
1857 The bug described in (1) was not present in this case, but the tidying up
1858 of the other signals was missing. I have made the two cases consistent.
1860 3. The ignore_target_hosts setting on a manualroute router was being ignored
1861 for hosts that were looked up using the /MX notation.
1863 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1866 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1867 operated on the sender address. After changing the $sender_address to <>
1868 for the sender address verify, Exim was re-instated it as the original
1869 (before rewriting) address, but remembering that it had rewritten it, so it
1870 wasn't rewriting it again. This bug also had the effect of breaking the
1871 sender address verification caching when the sender address was rewritten.
1873 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1874 This has been changed so that if the ip literal address matches
1875 ignore_target_hosts, the router declines.
1877 7. Added expansion conditions match_domain, match_address, and match_local_
1878 part (NOT match_host).
1880 8. The placeholder for the Received: header didn't have a length field set.
1882 9. Added code to Exim itself and to exim_lock to test for a specific race
1883 condition that could lead to file corruption when using MBX delivery. The
1884 issue is with the lockfile that is created in /tmp. If this file is removed
1885 after a process has opened it but before that process has acquired a lock,
1886 there is the potential for a second process to recreate the file and also
1887 acquire a lock. This could lead to two Exim processes writing to the file
1888 at the same time. The added code performs the same test as UW imapd; it
1889 checks after acquiring the lock that its file descriptor still refers to
1890 the same named file.
1892 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1893 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1896 11. Added the smtp_active_hostname option. If used, this will typically be made
1897 to depend on the incoming interface address. Because $interface_address is
1898 not set up until the daemon has forked a reception process, error responses
1899 that can happen earlier (such as "too many connections") no longer contain
1902 12. If an expansion in a condition on a "warn" statement fails because a lookup
1903 defers, the "warn" statement is abandoned, and the next ACL statement is
1904 processed. Previously this caused the whole ACL to be aborted.
1906 13. Added the iplsearch lookup type.
1908 14. Added ident_timeout as a log selector.
1910 15. Added tls_certificate_verified as a log selector.
1912 16. Added a global option tls_require_ciphers (compare the smtp transport
1913 option of the same name). This controls incoming TLS connections.
1915 17. I finally figured out how to make tls_require_ciphers do a similar thing
1916 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1917 before starting the TLS session.
1919 18. Tabs are now shown as \t in -bP output.
1921 19. If the log selector return_path_on_delivery was set, Exim crashed when
1922 bouncing a message because it had too many Received: header lines.
1924 20. If two routers both had headers_remove settings, and the first one included
1925 a superfluous trailing colon, the final name in the first list and the
1926 first name in the second list were incorrectly joined into one item (with a
1927 colon in the middle).
1933 1. Added -C and -D options to the exinext utility, mainly to make it easier
1934 to include in the automated testing, but these could be helpful when
1935 multiple configurations are in use.
1937 2. The exinext utility was not formatting the output nicely when there was
1938 an alternate port involved in the retry record key, nor when there was a
1939 message id as well (for retries that were specific to a specific message
1940 and a specific host). It was also confused by IPv6 addresses, because of
1941 the additional colons they contain. I have fixed the IPv4 problem, and
1942 patched it up to do a reasonable job for IPv6.
1944 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1945 delivery, the log line now contains "pipelined" if PIPELINING was used.
1947 4. An SMTP transport process used to panic and die if the bind() call to set
1948 an explicit outgoing interface failed. This has been changed; it is now
1949 treated in the same way as a connect() failure.
1951 5. A reference to $sender_host_name in the part of a conditional expansion
1952 that was being skipped was still causing a DNS lookup. This no longer
1955 6. The def: expansion condition was not recognizing references to header lines
1956 that used bh_ and bheader_.
1958 7. Added the _cache feature to named lists.
1960 8. The code for checking quota_filecount in the appendfile transport was
1961 allowing one more file than it should have been.
1963 9. For compatibility with Sendmail, the command line option
1971 and sets the incoming protocol and host name (for trusted callers). The
1972 host name and its colon can be omitted when only the protocol is to be set.
1973 Note the Exim already has two private options, -pd and -ps, that refer to
1974 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1975 "s", but I don't think that's a major issue.
1977 10. A number of refactoring changes to the code, none of which should affect
1980 (a) The number of logging options was getting close to filling up the
1981 32-bit word that was used as a bit map. I have split them into two classes:
1982 those that are passed in the argument to log_write(), and those that are
1983 only ever tested independently outside of that function. These are now in
1984 separate 32-bit words, so there is plenty of room for expansion again.
1985 There is no change in the user interface or the logging behaviour.
1987 (b) When building, for example, log lines, the code previously used a
1988 macro that called string_cat() twice, in order to add two strings. This is
1989 not really sufficiently general. Furthermore, there was one instance where
1990 it was actually wrong because one of the argument was used twice, and in
1991 one call a function was used. (As it happened, calling the function twice
1992 did not affect the overall behaviour.) The macro has been replaced by a
1993 function that can join an arbitrary number of extra strings onto a growing
1996 (c) The code for expansion conditions now uses a table and a binary chop
1997 instead of a serial search (which was left over from when there were very
1998 few conditions). Also, it now recognizes conditions like "pam" even when
1999 the relevant support is not compiled in: a suitably worded error message is
2000 given if an attempt is made to use such a condition.
2002 11. Added ${time_interval:xxxxx}.
2004 12. A bug was causing one of the ddress fields not to be passed back correctly
2005 from remote delivery subprocesses. The field in question was not being
2006 subsequently used, so this caused to problems in practice.
2008 13. Added new log selectors queue_time and deliver_time.
2010 14. Might have fixed a bug in maildirsizefile handling that threw up
2011 "unexpected character" debug warnings, and recalculated the data
2012 unnecessarily. In any case, I expanded the warning message to give more
2015 15. Added the message "Restricted characters in address" to the statements in
2016 the default ACL that block characters like @ and % in local parts.
2018 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2019 Three changes have been made:
2021 (a) There was a serious bug; a negative response to MAIL caused the whole
2022 recipient domain to be cached as invalid, thereby blocking all messages
2023 to all local parts at the same domain, from all senders. This bug has
2024 been fixed. The domain is no longer cached after a negative response to
2025 MAIL if the sender used is not empty.
2027 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2030 (c) A new callout option, "use_sender" has been added for people who want
2031 the modified behaviour.
2037 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2040 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2041 indeed breaks things for older releases.
2043 3. Added additional logging to the case where there is a problem reading data
2044 from a filter that is running in a subprocess using a pipe, in order to
2045 try to track down a specific problem.
2047 4. Testing facility fudge: when running in the test harness and attempting
2048 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2049 getting "No route to host". Convert this to a timeout.
2051 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2054 6. Some OS don't have socklen_t but use size_t instead. This affects the
2055 fifth argument of getsockopt() amongst other things. This is now
2056 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2057 can be set for individual OS. I have set it for SunOS5, OSF1, and
2058 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2059 some earlier ones do not.
2061 7. Change 4.30/15 was not doing the test caselessly.
2063 8. The standard form for an IPv6 address literal was being rejected by address
2064 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2065 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2066 this, as well as the form without the "IPv6" on the front (but only when
2067 address literals are enabled, of course).
2069 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2071 10. Exim crashed if a message with an empty sender address specified by -f
2072 encountered a router with an errors_to setting. This could be provoked only
2073 by a command such as
2077 where an empty string was supplied; "<>" did not hit this bug.
2079 11. Installed PCRE release 4.5.
2081 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2082 remained set. It is now erased.
2084 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2085 times from message ids (which are base 36 rather than the normal 62).
2087 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2088 were being counted as actual protocol errors, and logged if the log
2089 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2090 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2091 and DATA following a set of rejected RCPTs do not count as protocol errors.
2092 In other words, Exim assumes they were pipelined, though this may not
2093 actually be the case. Of course, in all cases the client gets an
2094 appropriate error code.
2096 15. If a lookup fails in an ACL condition, a message about the failure may
2097 be available; it is used if testing the ACL cannot continue, because most
2098 such messages specify what the cause of the deferral is. However, some
2099 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2100 that caused an old message to be retained and used if a later statement
2101 caused a defer, replacing the real cause of the deferral.
2103 16. If an IP address had so many PTR records that the DNS lookup buffer
2104 was not large enough to hold them, Exim could crash while trying to process
2105 the truncated data. It now detects and logs this case.
2107 17. Further to 4.21/58, another change has been made: if (and only if) the
2108 first line of a message (the first header line) ends with CRLF, a bare LF
2109 in a subsequent header line has a space inserted after it, so as not to
2110 terminate the header.
2112 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2113 unnecessarily, used atoi() instead of strtol(), and didn't check the
2114 termination when getting file sizes from file names by regex.
2116 19. Completely re-implemented the support for maildirsize files, in the light
2117 of a number of problems with the previous contributed implementation
2118 (4.30/29). In particular:
2120 . If the quota is zero, the maildirsize file is maintained, but no quota is
2123 . If the maildir directory does not exist, it is created before any attempt
2124 to write a maildirsize file.
2126 . The quota value in the file is just a cache; if the quota is changed in
2127 the transport, the new value overrides.
2129 . A regular expression is available for excluding directories from the
2132 20. The autoreply transport checks the characters in options that define the
2133 message's headers; it allows continued headers, but it was checking with
2134 isspace() after an embedded newline instead of explicitly looking for a
2137 21. If all the "regular" hosts to which an address was routed had passed their
2138 expiry times, and had not reached their retry times, the address was
2139 bounced, even if fallback hosts were defined. Now Exim should go on to try
2142 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2143 equivalent code in the SMTP transport. Some hosts send humungous responses
2144 to HELO/EHLO, more than 1024 it seems.
2146 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2147 gives compiler warnings in some environments. I've now done it "properly",
2150 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2151 (because of problems with the built-in one) was declared to return uschar *
2152 instead of char *, causing compiler failure.
2154 25. Fixed a file descriptor leak when processing alias/forward files.
2156 26. Fixed a minor format string issue in dbfn.c.
2158 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2160 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2161 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2164 29. When a sender address is verified, it is cached, to save repeating the test
2165 when there is more than one recipient in a message. However, when the
2166 verification involves a callout, it is possible for different callout
2167 options to be set for different recipients. It is too complicated to keep
2168 track of this in the cache, so now Exim always runs a verification when a
2169 callout is required, relying on the callout cache for the optimization.
2170 The overhead is duplication of the address routing, but this should not be
2173 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2174 to be verified with callout=postmaster, and the main callout worked but the
2175 postmaster check failed, the verification correctly failed. However, if a
2176 subsequent RCPT command asked for sender verification *without* the
2177 postmaster check, incorrect caching caused this verification also to fail,
2180 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2181 it was not caching the DNS options (qualify_single, search_parents) that
2182 were used when the lookup failed. A subsequent lookup with different
2183 options therefore always gave the same answer, though there were cases
2184 where it should not have. (Example: a "domains = !$mx_any" option on a
2185 dnslookup router: the "domains" option is always processed without any
2186 widening, but the router might have qualify_single set.) Now Exim uses the
2187 cached value only when the same options are set.
2189 32. Added John Jetmore's "exipick" utility to the distribution.
2191 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2192 than a timeout (e.g. a certificate is required, and is not provided), an
2193 Exim server now closes the connection immediately. Previously it waited for
2194 the client to close - but if the client is SSL, it seems that they each
2195 wait for each other, leading to a delay before one of them times out.
2197 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2198 maintained 0.8.x compatibility because I don't think many are using it, and
2199 it is clearly obsolete.
2201 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2204 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2205 client certificate was expired. A simple patch fixes this, though I don't
2206 understand the full logic of why the verify callback is called multiple
2209 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2210 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2211 which causes problems with some clients (such as the Certicom SSL Plus
2212 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2213 disables the coutermeasure allowing Eudora to connect."
2215 38. Exim was not checking that a write() to a log file succeeded. This could
2216 lead to Bad Things if a log got too big, in particular if it hit a file
2217 size limit. Exim now panics and dies if it cannot write to a log file, just
2218 as it does if it cannot open a log file.
2220 39. Modified OS/Makefile-Linux so that it now contains
2222 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2224 The two -D definitions ensure that Exim is compiled with large file
2225 support, which makes it possible to handle log files that are bigger than
2228 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2229 instance) a domain was checked against a named list that involved a lookup,
2230 causing $domain_data to be set, then another domain was checked against the
2231 same list, then the first domain was re-checked, the value of $domain_data
2232 after the final check could be wrong. In particular, if the second check
2233 failed, it could be set empty. This bug probably also applied to
2236 41. The strip_trailing_dot option was not being applied to the address given
2237 with the -f command-line option.
2239 42. The code for reading a message's header from the spool was incrementing
2240 $received_count, but never initializing it. This meant that the value was
2241 incorrect (doubled) while delivering a message in the same process in which
2242 it was received. In the most common configuration of Exim, this never
2243 happens - a fresh exec is done - but it can happen when
2244 deliver_drop_privilege is set.
2246 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2247 it now includes up to 150 characters of the unexpected data in the log
2250 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2251 and building data strings. The size of both of these buffers was 10 000
2252 bytes - far larger than anybody would *ever* want, thought I. Needless to
2253 say, somebody hit the limit. I have increased the maximum line length to
2254 20 000 and the maximum data length of concatenated lines to 100 000. I have
2255 also fixed two bugs, because there was no checking on these buffers. Tsk,
2256 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2257 buffer is too small.
2259 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2260 lsearch lookups. Now it does.
2262 46. When parsing a route_list item in a manualroute router, a fixed-length
2263 buffer was used for the list of hosts. I made this 1024 bytes long,
2264 thinking that nobody would ever have a list of hosts that long. Wrong.
2265 Somebody had a whole pile of complicated expansion conditions, and the
2266 string was silently truncated, leading to an expansion error. It turns out
2267 that it is easier to change to an unlimited length (owing to other changes
2268 that have happened since this code was originally written) than to build
2269 structure for giving a limitation error. The length of the item that
2270 expands into the list of hosts is now unlimited.
2272 47. The lsearch lookup could not handle data where the length of text line was
2273 more than 4095 characters. Such lines were truncated, leading to shortened
2274 data being returned. It should now handle lines of any length.
2276 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2277 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2280 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2281 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2282 no longer happens in comment lines. A list of replacements is now placed
2283 at the head of all of the source files, except those whose only change is
2284 to replace PERL_COMMAND in the very first #! line.
2286 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2287 messages on the queue, with a bottom-up merge sort, using code contributed
2288 by Michael Haardt. This should make operations like -bp somewhat faster on
2289 large queues. It won't affect queue runners, except when queue_run_in_order
2292 51. Installed eximstats 1.31 in the distribution.
2294 52. Added support for SRV lookups to the dnslookup router.
2296 53. If an ACL referred to $message_body or $message_body_end, the value was not
2297 reset for any messages that followed in the same SMTP session.
2299 54. The store-handling optimization for building very long strings was not
2300 differentiating between the different store pools. I don't think this
2301 actually made any difference in practice, but I've tidied it.
2303 55. While running the routers to verify a sender address, $sender_address
2304 was still set to the sender address. This is wrong, because when routing to
2305 send a bounce to the sender, it would be empty. Therefore, I have changed
2306 it so that, while verifying a sender address, $sender_address is set to <>.
2307 (There is no change to what happens when verifying a recipient address.)
2309 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2310 target A or AAAA records (if not already returned) without resetting the
2311 qualify_single or search_parents options of the DNS resolver. These are
2312 inappropriate in this case because the targets of MX and SRV records must
2313 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2314 target that, when qualified, matched something in the local domain. These
2315 two options are now turned off when doing these lookups.
2317 57. It seems that at least some releases of Reiserfs (which does not have the
2318 concept of a fixed number of inodes) returns zero and not -1 for the
2319 number of available inodes. This interacted badly with check_spool_inodes,
2320 which assumed that -1 was the "no such thing" setting. What I have done is
2321 to check that the total number of inodes is greater than zero before doing
2322 the test of how many are available.
2324 58. When a "warn" ACL statement has a log_message modifier, the message is
2325 remembered, and not repeated. This is to avoid a lot of repetition when a
2326 message has many recipients that cause the same warning to be written.
2327 Howewer, Exim was preserving the list of already written lines for an
2328 entire SMTP session, which doesn't seem right. The memory is now reset if a
2329 new message is started.
2331 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2332 debugging output unless log_rewrite was also set.
2334 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2335 of (int)(handle) when we know that handle contains (void *)(-1).
2337 61. The Exim daemon panic-logs an error return when it closes the incoming
2338 connection. However "connection reset by peer" seems to be common, and
2339 isn't really an error worthy of noting specially, so that particular error
2342 62. When Exim is trying to find all the local interfaces, it used to panic and
2343 die if the ioctl to get the interface flags failed. However, it seems that
2344 on at least one OS (Solaris 9) it is possible to have an interface that is
2345 included in the list of interfaces, but for which you get a failure error
2346 for this call. This happens when the interface is not "plumbed" into a
2347 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2348 failure of the "get flags" call assumes that the interface is down.
2350 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2351 makes life easier for people who are doing arithmetic on fields extracted
2352 from dates, where you often get leading zeros that should not be
2353 interpreted as octal.
2355 64. Added qualify_domain to the redirect router, to override the global
2358 65. If a pathologically long header line contained very many addresses (the
2359 report of this problem mentioned 10 000) and each of them was rewritten,
2360 Exim could use up a very large amount of memory. (It kept on making new
2361 copies of the header line as it rewrote, and never released the old ones.)
2362 At the expense of a bit more processing, the header rewriting function has
2363 been changed so that it no longer eats memory in this way.
2365 66. The generation of the Received: header has been moved from the time that a
2366 message starts to be received, to the time that it finishes. The timestamp
2367 in the Received: header should now be very close to that of the <= log
2368 line. There are two side-effects of this change:
2370 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2371 logged header lines no longer include the local Received: line, because
2372 it has not yet been created. The same applies to a copy of the message
2373 that is returned to a non-SMTP sender when a message is rejected.
2375 (b) When a filter file is tested using -bf, no additional Received: header
2376 is added to the test message. After some thought, I decided that this
2379 This change does not affect the value of $received_for. It is still set
2380 after address rewriting, but before local_scan() is called.
2382 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2384 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2385 gave an unhelpful panic error message, and a defer error. I have managed to
2386 change this behaviour so that it now rejects any supplied certificate,
2387 which seems right, as the list of acceptable certificates is empty.
2389 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2390 gave an unhelpful defer error. I have not managed to make this reject any
2391 supplied certificates, but the error message it gives is "no certificate
2392 supplied", which is not helpful.
2394 70. exigrep's output now also includes lines that are not associated with any
2395 message, but which match the given pattern. Implemented by a patch from
2396 Martin Sluka, which also tidied up the Perl a bit.
2398 71. Recipient callout verification, like sender verification, was using <> in
2399 the MAIL FROM command. This isn't really the right thing, since the actual
2400 sender may affect whether the remote host accepts the recipient or not. I
2401 have changed it to use the actual sender in the callout; this means that
2402 the cache record is now keyed on a recipient/sender pair, not just the
2403 recipient address. There doesn't seem to be a real danger of callout loops,
2404 since a callout by the remote host to check the sender would use <>.
2405 [SEE ABOVE: changed after hitting problems.]
2407 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2408 temporary errors. However, in the case of such a code being given after
2409 the end of a data transmission (i.e. after ".") Exim was failing to write
2410 a retry record for the message. (Yes, there was some broken host that was
2411 actually sending 8xx at this point.)
2413 73. An unknown lookup type in a host list could cause Exim to panic-die when
2414 the list was checked. (An example that provoked this was putting <; in the
2415 middle of a list instead of at the start.) If this happened during a DATA
2416 ACL check, a -D file could be left lying around. This kind of configuration
2417 error no longer causes Exim to die; instead it causes a defer errror. The
2418 incident is still logged to the main and panic logs.
2420 74. Buglet left over from Exim 3 conversion. The message "too many messages
2421 in one connection" was written to the rejectlog but not the mainlog, except
2422 when address rewriting (yes!) was being logged.
2424 75. Added write_rejectlog option.
2426 76. When a system filter was run not as root (that is, when system_filter_user
2427 was set), the values of the $n variables were not being returned to the
2428 main process; thus, they were not subsequently available in the $sn
2431 77. Added +return_path_on_delivery log selector.
2433 78. A connection timeout was being treated differently from recipients deferred
2434 when testing hosts_max_try with a message that was older than the host's
2435 retry timeout. (The host should not be counted, thus allowing all hosts to
2436 be tried at least once before bouncing.) This may have been the cause of an
2437 occasionally reported bug whereby a message would remain on the queue
2438 longer than the retry timeout, but would be bounced if a delivery was
2439 forced. I say "may" because I never totally pinned down the problem;
2440 setting up timeout/retry tests is difficult. See also the next item.
2442 79. The ultimate address timeout was not being applied to errors that involved
2443 a combination of host plus message (for example, a timeout on a MAIL
2444 command). When an address resolved to a number of possible hosts, and they
2445 were not all tried for each delivery (e.g. because of hosts_max_try), a
2446 message could remain on the queue longer than the retry timeout.
2448 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2451 81. Fixed an obscure SMTP outgoing bug which required at least the following
2452 conditions: (a) there was another message waiting for the same server;
2453 (b) the server returned 5xx to all RCPT commands in the first message so
2454 that the message was not completed; (c) the server dropped the connection
2455 or gave a negative response to the RSET that Exim sends to abort the
2456 transaction. The observed case was a dropped connection after DATA that had
2457 been sent in pipelining mode. That is, the server had advertised PIPELINING
2458 but was not implementing it correctly. The effect of the bug was incorrect
2459 behaviour, such as trying another host, and this could lead to a crash.
2465 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2466 and daemon.c were passed as pointers to ints; they should have been
2467 pointers to socklen_t variables (which are typically unsigned ints).
2469 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2472 3. Fixed a really odd bug that affected only the testing scheme; patching a
2473 certain fixed string in the binary changed the value of another string that
2474 happened to be identical to the end of the original first string.
2476 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2477 name", it returns that address as the IP address. On some operating
2478 systems (e.g. Solaris), it also passes back the IP address string as the
2479 "host name". However, on others (e.g. Linux), it passes back an empty
2480 string. Exim wasn't checking for this, and was changing the host name to an
2481 empty string, assuming it had been canonicized.
2483 5. Although rare, it is permitted to have more than one PTR record for a given
2484 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2485 all the names associated with an address, because they do in Solaris.
2486 However, it seems that they do not in Linux for data that comes from the
2487 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2488 I found this out when I moved to a new Linux workstation and tried to run
2489 the Exim test suite.
2491 To get round this problem I have changed the code so that it now does its
2492 own call to the DNS to look up PTR records when searching for a host name.
2493 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2494 addresses that are only in /etc/hosts are still found.
2496 This behaviour is, however, controlled by an option called host_lookup_
2497 order, which defaults to "bydns:byaddr". If people want to use the other
2498 order, or indeed, just use one or the other means of lookup, they can
2499 specify it in this variable.
2501 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2502 some operating systems, this comes back from gethostbyaddr() as an empty
2503 string, and this is what Exim used to test for. However, it seems that in
2504 other systems, "." is yielded. Exim now tests for this case too.
2506 7. The values of check_spool_space and check_log_space are now held internally
2507 as a number of kilobytes instead of an absolute number of bytes. If a
2508 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2509 kilobyte. This means that much larger values can be stored.
2511 8. Exim monitor: an attempt to get the action menu when not actually pointing
2512 at a message produces an empty menu entitled "No message selected". This
2513 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2514 no entries in it ("Shell widget menu has zero width and/or height"). So I
2515 have added a single, blank menu entry in this case.
2517 9. Added ${quote_local_part.
2519 10. MIME decoding is now applied to the contents of Subject: header lines when
2522 11. Now that a reference to $sender_host_address automatically causes a reverse
2523 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2524 host lookup before query-style lookups in lists that might use this
2525 variable. This has therefore been abolished, and the "net-" prefix is no
2526 longer necessary for query-style lookups.
2528 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2529 have been a typo for LFLAGS, so it has been changed.
2531 13. The install script calls Exim with "-C /dev/null" in order to find the
2532 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2533 to be output. Howeve, since Exim outputs its version number before the
2534 error, it didn't break the script. It just looked ugly. I fixed this by
2535 always allowing "-C /dev/null" if the caller is root.
2537 14. Ignore overlarge ACL variable number when reading spool file - insurance
2538 against a later release with more variables having written the file.
2540 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2541 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2542 this, as well as the form without the "IPv6" on the front.
2544 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2545 OS/Makefile-Darwin file.
2547 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2548 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2550 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2551 variable. However, some OS use ints for inodes. I've added cast to long int
2552 to get rid of the compiler warning.
2554 19. I had forgotten to lock out "/../" in configuration file names when
2555 ALT_CONFIG_PREFIX was set.
2557 20. Routers used for verification do not need to specify transports. However,
2558 if such a router generated a host list, and callout was configured, Exim
2559 crashed, because it could not find a port number from the (non-existent)
2560 transport. It now assumes port 25 in this circumstance.
2562 21. Added the -t option to exigrep.
2564 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2565 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2566 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2567 them is set, LOOKUP_LSEARCH is forced.
2569 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2571 24. Added sender and host information to the "rejected by local_scan()" log
2572 line; previously there was no indication of these.
2574 25. Added .include_if_exists.
2576 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2577 Linux. It turns out that not all file systems support this. Apparently some
2578 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2579 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2580 not supported on the file descriptor, is now ignored when Exim is trying to
2581 sync a directory. This applies only to Linux.
2583 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2585 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2586 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2589 29. Added support for maildirsize files from supplied patch (modified a bit).
2591 30. The use of :fail: followed by an empty string could lead Exim to respond to
2592 sender verification failures with (e.g.):
2594 550 Verification failed for <xxx>
2595 550 Sender verify failed
2597 where the first response line was missing the '-' that indicates it is not
2598 the final line of the response.
2600 31. The loop for finding the name of the user that called Exim had a hardwired
2601 limit of 10; it now uses the value of finduser_retries, which is used for
2602 all other user lookups.
2604 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2607 33. Exim was neglecting to zero errno before one call of strtol() when
2608 expanding a string and expecting an integer value. On some systems this
2609 resulted in spurious "integer overflow" errors. Also, it was casting the
2610 result into an int without checking.
2612 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2613 did not work. The code looks as if it has *never* worked, though it appears
2614 to have been documented since at least releast 1.62. I have made it work.
2616 35. The "timeout_DNS" error in retry rules, also documented since at least
2617 1.62, also never worked. As it isn't clear exactly what this means, and
2618 clearly it isn't a major issue, I have abolished the feature by treating it
2619 as "timeout", and writing a warning to the main and panic logs.
2621 36. The display of retry rules for -brt wasn't always showing the error code
2624 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2625 timeout_connect_A, timeout_connect_MX.
2627 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2628 to the empty sender.
2630 39. The daemon was not analysing the content of -oX till after it had closed
2631 stderr and disconnected from the controlling terminal. This meant that any
2632 syntax errors were only noted on the panic log, and the return code from
2633 the command was 0. By re-arranging the code a little, I've made the
2634 decoding happen first, so such errors now appear on stderr, and the return
2635 code is 1. However, the actual setting up of the sockets still happens in
2636 the disconnected process, so errors there are still only recorded on the
2639 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2640 connections (as happens on some IP stacks) was logged at start up time as
2641 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2642 it from "IPv6 and IPv4", which means that two separate sockets are being
2645 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2646 says whether AF_INET or AF_INET6 was passed as an argument.
2648 42. Exiwhat output was messed up when time zones were included in log
2651 43. Exiwhat now gives more information about the daemon's listening ports,
2652 and whether -tls-on-connect was used.
2654 44. The "port" option of the smtp transport is now expanded.
2656 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2657 silently ignored. Now an error message is written to the main and panic
2660 46. There's a new ACL modifier called "logwrite" which writes to a log file
2661 as soon as it is encountered.
2663 47. Added $local_user_uid and $local_user_gid at routing time.
2665 48. Exim crashed when trying to verify a sender address that was being
2668 49. Exim was recognizing only a space character after ".include". It now also
2669 recognizes a tab character.
2671 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2672 extracting the relevant information from the specification. The man page no
2673 longer contains scrambled data for the -d option, and I've added a section
2674 at the front about calling Exim under different names.
2676 51. Added "extra_headers" argument to the "mail" command in filter files.
2678 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2681 53. Installed eximstats 1.29.
2683 54. Added transport_filter_timeout as a generic transport option.
2685 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2686 Cc: header lines. This was required by RFC 822, but it not required by RFC
2689 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2690 incoming messages that did not have them. Now it does so only if the
2691 message originates locally, that is, if there is no associated remote host
2692 address. When Resent- header lines are present, this applies to the Resent-
2693 lines rather than the non-Resent- lines.
2695 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2696 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2698 58. Messages for configuration errors now include the name of the main
2699 configuration file - useful now that there may be more than one file in a
2700 list (.included file names were always shown).
2702 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2703 for those rare installations that do not start the daemon as root or run it
2704 setuid root. I've cut out the call to initgroups() if the daemon is not
2707 60. The Exim user and group can now be bound into the binary as text strings
2708 that are looked up at the start of Exim's processing.
2710 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2712 62. Added $mailstore_basename variable.
2714 63. Installed patch to sieve.c from Michael Haardt.
2716 64. When Exim failed to open the panic log after failing to open the main log,
2717 the original message it was trying to log was written to stderr and debug
2718 output, but if they were not available (the usual case in production), it
2719 was lost. Now it is written to syslog before the two lines that record the
2720 failures to open the logs.
2722 65. Users' Exim filters run in subprocesses under the user's uid. It is
2723 possible for a "deliver" command or an alias in a "personal" command to
2724 provoke an address rewrite. If logging of address rewriting is configured,
2725 this fails because the process is not running as root or exim. There may be
2726 a better way of dealing with this, but for the moment (because 4.30 needs
2727 to be released), I have disabled address rewrite logging when running a
2728 filter in a non-root, non-exim process.
2734 1. The buildconfig auxiliary program wasn't quoting the value set for
2735 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2736 not defined. This bug was masked in 4.22 by the effect that was fixed in
2739 2. Some messages that were rejected after a message id was allocated were
2740 shown as "incomplete" by exigrep. It no longer does this for messages that
2741 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2743 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2744 have allow_domain_literals set, the ID did not get logged in the <= line.
2745 Domain literals are now always recognized in Message-ID: header lines.
2747 4. The first argument for a ${extract expansion item is the key name or field
2748 number. Leading and trailing spaces in this item were not being ignored,
2749 causing some misleading effects.
2751 5. When deliver_drop_privilege was set, single queue runner processes started
2752 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2753 same command in the process it spins off) were not dropping privilege.
2755 6. When the daemon running as "exim" started a queue runner, it always
2756 re-executed Exim in the spun-off process. This is a waste of effort when
2757 deliver_drop_privilege is set. The new process now just calls the
2758 queue-runner function directly.
2764 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2767 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2768 ignored. Though the use of -oP was forcing the writing of a pid file, it
2769 was always written to the default place.
2771 3. If the message "no IP address found for host xxxx" is generated during
2772 incoming verification, it is now followed by identification of the incoming
2773 connection (so you can more easily find what provoked it).
2775 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2777 5. Added some features to "harden" Exim a bit more against certain attacks:
2779 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2780 be put in Local/Makefile. This is like the never_users runtime option,
2781 but it cannot be overridden. The default setting is "root".
2783 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2784 prefix string with which any file named in a -C command line option
2787 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2788 is retained for -C and -D only if the caller of Exim is root. Without
2789 it, the exim user may also use -C and -D and retain privilege.
2791 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2792 command line option is disabled.
2794 6. Macro names set by the -D option must start with an upper case letter, just
2795 like macro names defined in the configuration file.
2797 7. Added "dereference=" facility to LDAP.
2799 8. Two instances of the typo "uknown" in the source files are fixed.
2801 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2802 the Configure-Makefile script screwed up while processing it.
2804 10. Incorporated PCRE 4.4.
2806 11. The SMTP synchronization check was not operating right at the start of an
2807 SMTP session. For example, it could not catch a HELO sent before the client
2808 waited for the greeting. There is now a check for outstanding input at the
2809 point when the greeting is written. Because of the duplex, asynchronous
2810 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2811 way, but not yet received, when the check is performed.
2813 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2814 on TCP/IP sockets, because this apparently causes some broken clients to
2817 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2818 unchanged) from the Cygwin maintainer.
2820 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2821 was supported instead of testing for mbx. Effectively a typo.
2823 15. The spa authenticator server code was not checking that the input it
2824 received was valid base64.
2826 16. The debug output line for the "set" modifier in ACLs was not showing the
2827 name of the variable that was being set.
2829 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2831 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2832 lookup on its own. Something else was needed to trigger the lookup. For
2833 example, a match in host_lookup or the need for a host name in a host list.
2834 Now, if $sender_host_name is referenced and the host name has not yet been
2835 looked up, a lookup is performed. If the lookup fails, the variable remains
2836 empty, and $host_lookup_failed is set to "1".
2838 19. Added "eqi" as a case-independent comparison operator.
2840 20. The saslauthd authentication condition could segfault if neither service
2841 nor realm was specified.
2843 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2844 error message that was logged was misleading, and incoming SMTP
2845 connections were dropped. The message is now more accurate, and temporary
2846 errors are given to SMTP connections.
2848 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2849 (except RSET) with a 421 error, until QUIT is received. However, it was
2850 failing to send a response to QUIT.
2852 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2853 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2854 ACL is now run after the helo_try_verify_hosts code.
2856 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2857 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2858 case-independent, so other case variants are also recognized). Apparently
2859 some systems use these upper case variants.
2861 25. If more than two messages were waiting for the same host, and a transport
2862 filter was specified for the transport, Exim sent two messages over the
2863 same TCP/IP connection, and then failed with "socket operation on non-
2864 socket" when it tried to send the third.
2866 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2868 27. The extern definition of crypt16() in expand.c was not being excluded when
2869 the OS had its own crypt16() function.
2871 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2872 as a preferred synonym for return_size_limit, both as an option and as an
2875 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2877 30. Changed the default configuration ACL to relax the local part checking rule
2878 for addresses that are not in any local domains. For these addresses,
2879 slashes and pipe symbols are allowed within local parts, but the sequence
2880 /../ is explicitly forbidden.
2882 31. SPA server authentication was not clearing the challenge buffer before
2885 32. log_message in a "warn" ACL statement was writing to the reject log as
2886 well as to the main log, which contradicts the documentation and doesn't
2887 seem right (because no rejection is happening). So I have stopped it.
2889 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2890 However, I am unable to do any testing of this.
2892 34. Fixed an infelicity in the appendfile transport. When checking directories
2893 for a mailbox, to see if any needed to be created, it was accidentally
2894 using path names with one or more superfluous leading slashes; tracing
2895 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2897 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2898 contents are added to the log line that is written for every discarded
2899 recipient. (Previously a log_message setting was ignored.)
2901 36. The ${quote: operator now quotes the string if it is empty.
2903 37. The install script runs exim in order to find its version number. If for
2904 some reason other than non-existence or emptiness, which it checks, it
2905 could not run './exim', it was installing it with an empty version number,
2906 i.e. as "exim-". This error state is now caught, and the installation is
2909 38. An argument was missing from the function that creates an error message
2910 when Exim fails to connect to the socket for saslauthd authentication.
2911 This could cause Exim to crash, or give a corrupted message.
2913 39. Added isip, isip4, and isip6 to ${if conditions.
2915 40. The ACL variables $acl_xx are now saved with the message, and can be
2916 accessed later in routers, transports, and filters.
2918 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2919 strings in the file are not string-expanded.
2921 42. If a MAIL command specified a SIZE value that was too large to fit into an
2922 int variable, the check against message_size_limit failed. Such values are
2923 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2924 day this will have to be increased, but I don't think I want to be around
2925 when emails are that large.
2932 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2933 iconv() is not standard in FreeBSD.
2935 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2936 IPv6 enabled. The observed symptom was a segmentation fault on return from
2937 the function os_common_find_running_interfaces() in src/os.c.
2939 3. In the check_special_case() function in daemon.c I had used "errno" as an
2940 argument name, which causes warnings on some systems. This was basically a
2941 typo, since it was named "eno" in the comments!
2943 4. The code that waits for the clock to tick (at a resolution of some fraction
2944 of a second) so as to ensure message-id uniqueness was always waiting for
2945 at least one whole tick, when it could have waited for less. [This is
2946 almost certainly not relevant at current processor speeds, where it is
2947 unlikely to ever wait at all. But we try to future-proof.]
2949 5. The function that sleeps for a time interval that includes fractions of a
2950 second contained a race. It did not block SIGALRM between setting the
2951 timer, and suspending (a couple of lines later). If the interval was short
2952 and the sigsuspend() was delayed until after it had expired, the suspension
2953 never ended. On busy systems this could lead to processes getting stuck for
2956 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2957 process, before it forks any delivery processes. The open lookup caching
2958 mechanism meant that the open file or database connection was passed into
2959 the delivery process. The problem was that delivery processes always tidy
2960 up cached lookup data. This could cause a problem for the next delivery
2961 process started by the queue runner, because the external queue runner
2962 process does not know about the closure. So the next delivery process
2963 still has data in the lookup cache. In the case of a file lookup, there was
2964 no problem because closing a file descriptor in a subprocess doesn't affect
2965 the parent. However, if the lookup was caching a connection to a database,
2966 the connection was closed, and the second delivery process was likely to
2967 see errors such as "PGSQL: query failed: server closed the connection
2968 unexpectedly". The problem has been fixed by closing all cached lookups
2969 in a queue runner before running a delivery process.
2971 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2972 seem to have the "const" qualifier which it has on other OS. I've
2975 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2976 *of the same type* (client or server) with the same public name that an
2977 error should be diagnosed.
2979 9. When Exim looked up a host name for an IP address, but failed to find the
2980 original IP address when looking up the host name (a safety check), it
2981 output the message "<ip address> does not match any IP for NULL", which was
2982 confusing, to say the least. The bug was that the host name should have
2983 appeared instead of "NULL".
2985 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2986 user that is built into the binary, and the -C or -D options is used, root
2987 privilege is dropped before the configuration file is read. In addition,
2988 logging is switched to stderr instead of the normal log files. If the
2989 configuration then re-defines the Exim user, the unprivileged environment
2990 is probably not what is expected, so Exim logs a panic warning message (but
2993 However, if deliver_drop_privilege is set, the unprivileged state may well
2994 be exactly what is intended, so the warning has been cut out in that case,
2995 and Exim is allowed to try to write to its normal log files.
3001 1. smtp_return_error_details was not giving details for temporary sender
3002 or receiver verification errors.
3004 2. Diagnose a configuration error if two authenticators have the same public
3007 3. Exim used not to create the message log file for a message until the first
3008 delivery attempt. This could be confusing when incoming messages were held
3009 for policy or load reasons. The message log file is now created at the time
3010 the message is received, and an initial "Received" line is written to it.
3012 4. The automatically generated man page for command line options had a minor
3013 bug that caused no ill effects; however, a more serious problem was that
3014 the procedure for building the man page automatically didn't always
3015 operate. Consequently, release 4.20 contains an out-of-date version. This
3016 shouldn't happen again.
3018 5. When building Exim with embedded Perl support, the script that builds the
3019 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3020 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3022 6. The freeze_tell option was not being used for messages that were frozen on
3023 arrival, either by an ACL or by local_scan().
3025 7. Added the smtp_incomplete_transaction log selector.
3027 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3028 was accepting AUTH without a new EHLO.
3030 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3031 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3034 10. Logging of TCP/IP connections (when configured) now happens in the main
3035 daemon process instead of the child process, so that the TCP/IP connection
3036 count is more accurate (but it can never be perfect).
3038 11. The use of "drop" in a nested ACL was not being handled correctly in the
3039 outer ACL. Now, if condition failure induced by the nested "drop" causes
3040 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3041 or "require"), the connection is dropped.
3043 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3044 that yield "discard" can now be used with an "accept" or a "discard" verb,
3045 but an error is generated for any others (because I can't see a useful way
3046 to define what should happen).
3048 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3049 are now processed in the same way as lines in the Exim configuration file.
3050 In particular, continuation lines are supported.
3052 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3054 15. Added -ti meaning -t -i.
3056 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3057 domains, and warn by logging if others are found.
3059 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3060 returned from GIFCONF when Exim is trying to find the list of interfaces on
3061 a host. The code in os.c has been modified to copy each ifreq to an aligned
3062 structure in all cases.
3064 Also, in some cases, the returned ifreq's were being copied to a 'struct
3065 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3066 means the last couple of bytes of an IPv6 address could be chopped if the
3067 ifreq contained only a normal sockaddr (14 bytes storage).
3069 18. Named domain lists were not supported in the hosts_treat_as_local option.
3070 An entry such as +xxxx was not recognized, and was treated as a literal
3073 19. Ensure that header lines added by a DATA ACL are included in the reject log
3074 if the ACL subsequently rejects the message.
3076 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3077 MD5 (which is deprecated).
3079 21. When testing a filter file using -bf, Exim was writing a message when it
3080 took the sender from a "From " line in the message, but it was not doing so
3081 when it took $return_path from a Return-Path: header line. It now does.
3083 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3084 with a valid header line field name (a series of printing characters
3085 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3087 23. Changed "disc" in the source to "disk" to conform to the documentation and
3088 the book and for uniformity.
3090 24. Ignore Sendmail's -Ooption=value command line item.
3092 25. When execve() failed while trying to run a command in a pipe transport,
3093 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3094 could be confused with a return value of 69 from the command itself. This
3095 has been changed to 127, the value the shell returns if it is asked to run
3096 a non-existent command. The wording for the related log line suggests a
3097 non-existent command as the problem.
3099 26. If received_header_text expands to an empty string, do not add a Received:
3100 header line to the message. (Well, it adds a token one on the spool, but
3101 marks it "old" so that it doesn't get used or transmitted.)
3103 27. Installed eximstats 1.28 (addition of -nt option).
3105 28. There was no check for failure on the call to getsockname() in the daemon
3106 code. This can fail if there is a shortage of resources on the system, with
3107 ENOMEM, for example. A temporary error is now given on failure.
3109 29. Contrary to the C standard, it seems that in some environments, the
3110 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3111 program. Exim now does this explicitly; it affects the formatting of
3112 timestamps using strftime().
3114 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3115 complaints. I've now initialized all the variables, to avoid this.
3117 32. Header lines added by a system filter were not being "seen" during
3118 transport-time rewrites.
3120 33. The info_callback() function passed to OpenSSL is set up with type void
3121 (*)(SSL *, int, int), as described somewhere. However, when calling the
3122 function (actually a macro) that sets it up, the type void(*)() is
3123 expected. I've put in a cast to prevent warnings from picky compilers.
3125 34. If a DNS black list lookup found a CNAME record, but there were no A
3126 records associated with the domain it pointed at, Exim crashed.
3128 35. If a DNS black list lookup returned more than one A record, Exim ignored
3129 all but the first. It now scans all returned addresses if a particular IP
3130 value is being sought. In this situation, the contents of the
3131 $dnslist_value variable are a list of all the addresses, separated by a
3134 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3135 to accept a host name and all its aliases if the forward lookup for any of
3136 them yielded the IP address of the incoming connection. Now it accepts only
3137 those names whose forward lookup yields the correct IP address. Any other
3138 names are discarded. This closes a loophole whereby a rogue DNS
3139 administrator could create reverse DNS records to break through a
3140 wildcarded host restriction in an ACL.
3142 37. If a user filter or a system filter that ran in a subprocess used any of
3143 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3144 the wrong values were passed to the pipe command ($thisaddress had the
3145 value of $0, $0 had the value of $1, etc). This bug was introduced by
3146 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3148 38. Improved the line breaking for long SMTP error messages from ACLs.
3149 Previously, if there was no break point between 40 and 75 characters, Exim
3150 left the rest of the message alone. Two changes have been made: (a) I've
3151 reduced the minimum length to 35 characters; (b) if it can't find a break
3152 point between 35 and 75 characters, it looks ahead and uses the first one
3153 that it finds. This may give the occasional overlong line, but at least the
3154 remaining text gets split now.
3156 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3157 file descriptors might be low, and that setting 1000 would always raise it.
3158 It turns out that in some environments, the limit is already over 1000 and
3159 that lowering it causes trouble. So now Exim takes care not to decrease it.
3161 40. When delivering a message, the value of $return_path is set to $sender_
3162 address at the start of routing (routers may change the value). By an
3163 oversight, this default was not being set up when an address was tested by
3164 -bt or -bv, which affected the outcome if any router or filter referred to
3167 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3168 log only when "message" or "log_message" are set. However, if one of the
3169 conditions was an address verification, or a call to a nested ACL, the
3170 messages generated by the underlying test were being passed through. This
3171 no longer happens. The underlying message is available in $acl_verify_
3172 message for both "message" and "log_message" expansions, so it can be
3173 passed through if needed.
3175 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3176 new expansion $bh_ to give the encoded byte string without charset
3177 translation. Translation happens only if iconv() is available; HAVE_ICONV
3178 indicates this at build time. HEADERS_CHARSET gives the charset to
3179 translate to; headers_charset can change it in the configuration, and
3180 "headers charset" can change it in an individual filter file.
3182 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3183 that creates RFC 2047 encoded "words" labels them as that charset instead
3184 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3185 expansion operator; (ii) when Exim creates a From: line for a local
3186 message; (iii) when a header line is rewritten to include a "phrase" part.
3188 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3189 buggy, causing it to skip the first lines of messages whose message ID
3190 ended in 'D'. This would not have bitten before Exim release 4.14, because
3191 message IDs were unlikely to end in 'D' before then. The effect was to have
3192 incorrect size information for certain domains.
3194 45. #include "config.h" was missing at the start of the crypt16.c module. This
3195 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3198 46. If there was a timeout during a "random" callout check, Exim treated it as
3199 a failure of the random address, and carried on sending RSET and the real
3200 address. If the delay was just some slowness somewhere, the response to the
3201 original RCPT would be taken as a response to RSET and so on, causing
3202 mayhem of various kinds.
3204 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3205 when I implemented it. It didn't allow for the fact that some option values
3206 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3207 the right test for positive values.
3209 48. Domain names in DNS records are case-independent. Exim always looks them up
3210 in lower case. Some resolvers return domain names in exactly the case they
3211 appear in the zone file, that is, they may contain uppercase letters. Not
3212 all resolvers do this - some return always lower case. Exim was treating a
3213 change of case by a resolver as a change of domain, similar to a widening
3214 of a domain abbreviation. This triggered its re-routing code and so it was
3215 trying to route what was effectively the same domain again. This normally
3216 caused routing to fail (because the router wouldn't handle the domain
3217 twice). Now Exim checks for this case specially, and just changes the
3218 casing of the domain that it ultimately uses when it transmits the message
3221 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3224 50. If a filter generated a file delivery with a non-absolute name (possible if
3225 no home directory exists for the router), the forbid_file option was not
3228 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3229 the existing equality matching.
3231 52. Exim was using ints instead of ino_t variables in some places where it was
3232 dealing with inode numbers.
3234 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3235 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3236 called TMPDIR, and if it finds it is different, it changes its value.
3238 54. The smtp_printf() function is now made available to local_scan() so
3239 additional output lines can be written before returning. There is also an
3240 smtp_fflush() function to enable the detection of a dropped connection.
3241 The variables smtp_input and smtp_batched_input are exported to
3244 55. Changed the default runtime configuration: the message "Unknown user"
3245 has been removed from the ACL, and instead placed on the localuser router,
3246 using the cannot_route_message feature. This means that any verification
3247 failures that generate their own messages won't get overridden. Similarly,
3248 the "Unrouteable address" message that was in the ACL for unverifiable
3249 relay addresses has also been removed.
3251 56. Added hosts_avoid_esmtp to the smtp transport.
3253 57. The exicyclog script was not checking for the esoteric option
3254 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3255 will work only if exicyclog is run under the appropriate euid.
3257 58. Following a discussion on the list, the rules by which Exim recognises line
3258 endings on incoming messages have been changed. The -dropcr and drop_cr
3259 options are now no-ops, retained only for backwards compatibility. The
3260 following line terminators are recognized: LF CRLF CR. However, special
3261 processing applies to CR:
3263 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3264 nor a local message in the state where . is a terminator.
3266 (ii) If a bare CR is encountered in a header line, an extra space is added
3267 after the line terminator so as not to end the header. The reasoning
3268 behind this is that bare CRs in header lines are most likely either
3269 to be mistakes, or people trying to play silly games.
3271 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3272 was being incorrectly given as 18 bytes larger than it should have been.
3273 This is a VOB (very old bug).
3275 60. This may never have affected anything current, but just in case it has:
3276 When the local host is found other than at the start of a list of hosts,
3277 the local host, those with the same MX, and any that follow, are discarded.
3278 When the list in question was part of a longer list of hosts, the following
3279 hosts (not currently being processed) were also being discarded. This no
3280 longer happens. I'm not sure if this situation could ever has previously
3283 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3286 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3287 header. This is something that is recommended in a new Internet Draft, and
3288 is something that is documented as being done by Sendmail. There are two
3289 possible values. For messages generated by the autoreply transport, Exim
3292 Auto-Submitted: auto-replied
3294 whereas for all other generated messages (e.g. bounces) it adds
3296 Auto-Submitted: auto-generated
3298 63. The "personal" condition in filters now includes a test for the
3299 Auto-Submitted: header. If it contains the string "auto-" the message it
3300 not considered personal.
3302 64. Added rcpt_include_affixes as a generic transport option.
3304 65. Added queue_only_override (default true).
3306 66. Added the syslog_duplication option.
3308 67. If what should have been the first header line of a message consisted of
3309 a space followed by a colon, Exim was mis-interpreting it as a header line.
3310 It isn't of course - it is syntactically invalid and should therefore be
3311 treated as the start of the message body. The misbehaviour could have
3312 caused a number of strange effects, including loss of data in subsequent
3313 header lines, and spool format errors.
3315 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3316 client host had authenticated. This control can now be exercised by an ACL
3317 for more flexibility.
3319 69. By default, callouts do not happen when testing with -bh. There is now a
3320 variant, -bhc, which does actually run the callout code, including
3321 consulting and updating the callout cache.
3323 70. Added support for saslauthd authentication, courtesy of Alexander
3326 71. If statvfs() failed on the spool or log directories while checking their
3327 size for availability, Exim confusingly gave the error "space shortage".
3328 Furthermore, in debugging mode it crashed with a floating point exception.
3329 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3330 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3331 serious problem, Exim now writes to the main and panic logs when this
3332 happens, with details of the failure. It then refuses to accept the
3333 incoming message, giving the message "spool directory problem" or "log
3334 directory problem" with a 421 code for SMTP messages.
3336 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3337 0, 1, and 2 exist, because some OS complain for execs without them (see
3338 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3339 descriptors. However, the code omitted to check that the open succeeded,
3340 causing mysterious errors if for some reason the permissions on /dev/null
3341 got screwed. Now Exim writes a message to the main and panic logs, and
3342 bombs out if it can't open /dev/null.
3344 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3345 interact so that it is all more flexible. It is supposed to remain
3346 backwards compatible. Also added extra_local_interfaces.
3348 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3349 to bomb out with an assertion failure - to the client this appears as a
3350 connection drop. This problem occurs in the part of the code that was taken
3351 from the Samba project. Fortunately, the assertion is in a very simple
3352 function, so I have fixed this by reproducing the function inline in the
3353 one place where it is called, and arranging for authentication to fail
3354 instead of killing the process with assert().
3356 75. The SPA client code was not working when the server requested OEM rather
3357 than Unicode encoding.
3359 76. Added code to make require_files with a specific uid setting more usable in
3360 the case where statting the file as root fails - usually a non-root-mounted
3361 NFS file system. When this happens and the failure is EACCES, Exim now
3362 forks a subprocess and does the per-uid checking as the relevant uid.
3364 77. Added process_log_path.
3366 78. If log_file_path was not explicitly set, a setting of check_log_space or
3367 check_log_inodes was ignored.
3369 79. If a space check for the spool or log partitions fails, the incident is now
3370 logged. Of course, in the latter case the data may get lost...
3372 80. Added the %p formatting code to string_format() so that it can be used to
3373 print addresses in debug_print(). Adjusted all the address printing in the
3374 debugging in store.c to use %p rather than %d.
3376 81. There was a concern that a line of code in smtp_in.c could overflow a
3377 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3378 initially expressed, the concern was not well-founded, because trailing
3379 spaces are removed early. However, if the trailing spaces were followed by
3380 a NULL, they did not get removed, so the overflow was possible. Two fixes
3383 (a) I re-wrote the offending code in a cleaner fashion.
3384 (b) If an incoming SMTP command contains a NULL character, it is rejected
3387 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3388 runs initgroups(), so that if the Exim user is in any additional groups,
3389 they will be used during message reception.
3395 The change log for 4.20 and earlier releases has been archived.