1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.235 2005/09/19 09:41:37 fanf2 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 TK/01 Added the "success_on_redirect" address verification option. See
10 NewStuff for rationale and an example.
12 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
14 PH/02 Patch to exigrep to allow it to work on syslog lines.
16 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
17 fread() to skip over the body file's header line, because in Cygwin the
18 header line is locked and is inaccessible.
20 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
21 co-exist for some time) to make it clear that it is the Exim ID that is
22 referenced, not the Message-ID: header line.
24 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
25 string_format() function, because snprintf() does not exist on all
28 PH/06 The use of forbid_filter_existstest now also locks out the use of the
29 ${stat: expansion item.
31 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
32 protocol synchronization error", to keep the pedants happy.
34 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
35 well as for IRIX systems, when gcc is being used. See the host.c source
38 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
40 PH/10 Named domain lists were not working if used in a queue_smtp_domains
43 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
44 transport and to the smtp transport in LMTP mode.
46 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
48 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
49 run a filter in a subprocess. This could lead to confusion in subsequent
50 lookups in the parent process. There should also be a search_tidyup() at
51 the end of the subprocess.
53 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
54 only if the host matched helo_try_verify_hosts, which caused the
55 verification to occur when the EHLO/HELO command was issued. The ACL just
56 tested the remembered result. Now, if a previous verification attempt has
57 not happened, "verify = helo" does it there and then.
59 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
61 TK/03 Fix log output including CR from clamd.
63 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
64 exist provoked a memory error which could cause a segfault.
66 PH/15 Installed PCRE 6.2
68 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
70 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
71 of the problem. Specifically, suggested +O2 rather than +O1 for the
74 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
76 PH/20 If a delivery was routed to a non-standard port by means of an SRV
77 record, the port was not correctly logged when the outgoing_port log
78 selector was set (it logged the transort's default port).
80 PH/21 Added support for host-specific ports to manualroute, queryprogram,
81 fallback_hosts, and "hosts" in the smtp transport.
83 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
84 host errors such as "Connection refused".
86 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
87 authentication with radiusclient 0.4.9:
89 - Error returned from rc_read_config was caught wrongly
90 - Username/password not passed on to radius server due to wrong length.
92 The presumption is that some radiusclient API changes for 4.51/PH/17
93 were not taken care of correctly. The code is still untested by me (my
94 Linux distribution still has 0.3.2 of radiusclient), but it was
95 contributed by a Radius user.
97 PH/24 When doing a callout, the value of $domain wasn't set correctly when
98 expanding the "port" option of the smtp transport.
100 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
101 while reading a MIME header. Thanks to Tom Hughes for a patch.
103 PH/24 Include config.h inside local_scan.h so that configuration settings are
106 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
107 that in an ACL for RCPT (for example), you can examine exactly what was
110 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
111 commands, but it was not correctly comparing the address with the actual
112 client host address. Thus, it would show the EHLO address in Received:
113 header lines when this was not necessary.
115 PH/27 Added the % operator to ${eval:}.
117 PH/28 Exim tries to create and chdir to its spool directory when it starts;
118 it should be ignoring failures (because with -C, for example, it has lost
119 privilege). It wasn't ignoring creation failures other than "already
122 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
125 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
126 because an input line was too long, either on its own, or by virtue of
127 too many continuations, the temporary file was not being removed, and the
128 return code was incorrect.
130 PH/31 Missing "BOOL" in function definition in filtertest.c.
132 PH/32 Applied Sieve patches from the maintainer.
134 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
136 PH/33 Added "verify = not_blind".
138 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
139 Local/Makefile (with some defaults set). These are used in built scripts
140 such as exicyclog, but they have never been used in the exim_install
141 script (though there are many overriding facilities there). I have
142 arranged that the exim_install script now takes note of these two
145 PH/35 Installed configuration files for Dragonfly.
147 PH/36 When a locally submitted message by a trusted user did not contain a
148 From: header, and the sender address was obtained from -f or from an SMTP
149 MAIL command, and the trusted user did not use -F to supply a sender
150 name, $originator_name was incorrectly used when constructing a From:
151 header. Furthermore, $originator_name was used for submission mode
152 messages from external hosts without From: headers in a similar way,
153 which is clearly wrong.
155 PH/37 Added control=suppress_local_fixups.
157 PH/38 When log_selector = +received_sender was set, and the addition of the
158 sender made the log line's construction buffer exactly full, or one byte
159 less than full, an overflow happened when the terminating "\n" was
162 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
163 when the result of a list match is failure because a DNS lookup failed.
165 PH/40 RM_COMMAND is now used in the building process.
167 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
168 the "build-* directories that it finds.
170 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
171 address in a domain literal was a prefix of an interface address.
173 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
174 when verifying a sender address, unless rewrite_headers is false.
176 PH/44 Wrote a long comment about why errors_to addresses are verified as
177 recipients, not senders.
179 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
180 the ratelimit ACL was added.
182 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
184 PH/46 Added extra information about PostgreSQL errors to the error string.
186 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
187 avoiding the need to install special zones in a real server. This is
188 backwards compatible; if it can't find the fake resolver, it drops back.
189 Thus, both old and new test suites can be run.
191 TF/02 Added util/ratelimit.pl
193 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
194 clock is set back in time.
196 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
197 Candler <B.Candler@pobox.com>.
199 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
200 OK for addresses that are the result of redirections.
202 PH/48 A number of further additions for the benefit of the new test suite,
203 including a fake gethostbyname() that interfaces to the fake DNS resolver
206 TF/06 The fix for widen_domains has also been applied to qualify_single and
207 search_parents which are the other dnslookup options that can cause
214 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
216 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
217 command itself ended in error, the underlying message about the transport
218 filter timeout was being overwritten with the pipe command error. Now the
219 underlying error message should be appended to the second error message.
221 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
223 PH/02 Reduce the amount of output that "make" produces by default. Full output
224 can still be requested.
226 PH/03 The warning log line about a condition test deferring for a "warn" verb
227 was being output only once per connection, rather than after each
228 occurrence (because it was using the same function as for successful
229 "warn" verbs). This seems wrong, so I have changed it.
231 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
232 it should not have, which might have caused a crash in the right
233 circumstances, but probably never did.
235 PH/04 Installed a modified version of Tony Finch's patch to make submission
236 mode fix the return path as well as the Sender: header line, and to
237 add a /name= option so that you can make the user's friendly name appear
240 TF/03 Added the control = fakedefer ACL modifier.
242 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
243 Mark Lowes for thorough testing.
245 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
247 TK/03 Merged latest SRS patch from Miles Wilton.
249 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
250 with the definition in sysexits.h (which is #included earlier).
251 Fortunately, Exim does not actually use EX_OK. The code used to try to
252 preserve the sysexits.h value, by assumimg that macro definitions were
253 scanned for macro replacements. I have been disabused of this notion,
254 so now the code just undefines EX_OK before #including unistd.h.
256 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
257 in the smtp transport. When a block could not be written in a single
258 write() function, the timeout was being re-applied to each part-write.
259 This seems wrong - if the receiver was accepting one byte at a time it
260 would take for ever. The timeout is now adjusted when this happens. It
261 doesn't have to be particularly precise.
263 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
264 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
266 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
267 without a domain if the check to <postmaster@domain> fails.
269 SC/01 Eximstats: added -xls and the ability to specify output files
270 (patch written by Frank Heydlauf).
272 SC/02 Eximstats: use FileHandles for outputing results.
274 SC/03 Eximstats: allow any combination of xls, txt, and html output.
276 SC/04 Eximstats: fixed display of large numbers with -nvr option
278 SC/05 Eximstats: fixed merging of reports with empty tables.
280 SC/06 Eximstats: added the -include_original_destination flag
282 SC/07 Eximstats: removed tabs and trailing whitespace.
284 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
286 TK/06 MBOX spool code: Add real "From " MBOX separator line
287 so the .eml file is really in mbox format (even though
288 most programs do not really care). Patch from Alex Miller.
290 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
291 The latter is generated from $received_to and is only set if the
292 message has one envelope recipient. SA can use these headers,
293 obviously out-of-the-box. Patch from Alex Miller.
295 PH/08 The ${def test on a variable was returning false if the variable's
296 value was "0", contrary to what the specification has always said!
297 The result should be true unless the variable is empty.
299 PH/09 The syntax error of a character other than { following "${if
300 def:variable_name" (after optional whitespace) was not being diagnosed.
301 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
302 accidental colon was present, for example, could give incorrect results.
304 PH/10 Tidied the code in a number of places where the st_size field of a stat()
305 result is used (not including appendfile, where other changes are about
308 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
309 This involved changing a lot of size variables from int to off_t. It
310 should work with maildirs and everything.
312 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
313 spamd dying while we are connected to it.
315 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
318 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
319 function for BDB 4.3.
321 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
323 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
324 the sources for PCRE have been re-arranged and more files are now
327 PH/15 The code I had for printing potentially long long variables in PH/11
328 above was not the best (it lost precision). The length of off_t variables
329 is now inspected at build time, and an appropriate printing format (%ld
330 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
331 to be "long long int" or "long int". This is needed for the internal
332 formatting function string_vformat().
334 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
335 the configuration file to be ":syslog", then the script "guesses" where
336 the logs files are, rather than using the compiled in default. In our
337 case the guess is not the same as the compiled default, so the script
338 suddenly stopped working when I started to use syslog. The patch checks
339 to see if log_file_path is "". If so, it attempts to read it from exim
340 with no configuration file to get the compiled in version, before it
341 falls back to the previous guessing code."
343 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
344 implementing BATV in an Exim configuration. See NewStuff for the gory
347 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
348 Makefile that are specific to HP-UX.
350 PH/18 If the "use_postmaster" option was set for a recipient callout together
351 with the "random" option, the postmaster address was used as the MAIL
352 FROM address for the random test, but not for the subsequent recipient
353 test. It is now used for both.
355 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
356 patch removes a few documentation additions to RFC 3028, because the
357 latest draft now contains them. It adds the new en;ascii-case comparator
358 and a new error check for 8bit text in MIME parts. Comparator and
359 require names are now matched exactly. I enabled the subaddress
360 extension, but it is not well tested yet (read: it works for me)."
362 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
363 rework some of the code of TK/09 above to avoid the hardwired use of
364 "%lld" and "long long". Replaced the call to snprintf() with a call to
367 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
368 records point to non-existent hosts", "retry timeout exceeded", and
369 "retry time not reached for any host after a long failure period".
371 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
372 experimental DomainKeys support:
374 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
375 (2) On an error such as an illegally used "control", the wrong name for
376 the control was given.
378 These problems did NOT occur unless DomainKeys support was compiled.
380 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
382 PH/24 Added ${if match_ip condition.
384 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
385 databases so that it will be absolutely obvious if a crash occurs in the
386 DB library. This is a regular occurrence (often caused by mis-matched
389 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
390 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
391 picked up on a user's system that detects such things. There doesn't seem
392 to be a gcc warning option for this - only an attribute that has to be
393 put on the function's prototype. It seems that in Fedora Core 4 they have
394 set this on a number of new functions. No doubt there will be more in due
397 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
398 specify a transport. However, if an address that was verified by such a
399 router was the subject of a callout, Exim crashed because it tried to
400 read the rcpt_include_affixes from the non-existent transport. Now it
401 just assumes that the setting of that option is false. This bug was
402 introduced by 4.51/PH/31.
404 PH/28 Changed -d+all to exclude +memory, because that information is very
405 rarely of interest, but it makes the output a lot bigger. People tend to
406 do -d+all out of habit.
408 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
409 code in os-type was giving problems when libc.so lives in lib64, like on
412 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
413 aren't the modern standard, and it seems that some systems' include files
414 don't always have them. Exim was already checking for some of the newer
415 ones like T_AAAA, and defining it itself. I've added checks for all the
416 record types that Exim uses.
418 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
419 not automatically generating a new one, as it is supposed to. This
420 prevented TLS from working. If the file did exist, but contained invalid
421 data, a new version was generated, as expected. It was only the case of a
422 non-existent file that was broken.
424 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
425 with a change in libdomainkeys > 0.64.
427 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
428 from DNS. If the selector record carries the flag, it now has
429 precedence over the domain-wide flag.
431 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
433 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
434 the use of a port name that isn't defined in /etc/services) occurred, the
435 message was deferred as in a normal delivery, and thus remained on the
436 spool, instead of being failed because of the mua_wrapper setting. This
437 is now fixed, and I tidied up some of the mua_wrapper messages at the
440 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
441 the messages in a hash of arrays rather than using individual hashes.
442 This is a bit cleaner and results in dramatic memory savings, albeit
443 at a slight CPU cost.
445 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
446 as requested by Marc Sherman.
448 SC/10 Eximstats: added histograms for user specified patterns as requested
451 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
453 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
454 fopen() in the content-scanning modules that did not already have it.
460 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
461 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
463 TK/02 Fix ACL "control" statement not being available in MIME ACL.
465 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
467 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
468 to test Sieve filters that use "vacation".
470 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
471 that changes the way the GnuTLS parameters are stored in the cache file.
472 The new format can be generated externally. For backward compatibility,
473 if the data in the cache doesn't make sense, Exim assumes it has read an
474 old-format file, and it generates new data and writes a new file. This
475 means that you can't go back to an older release without removing the
478 PH/03 A redirect router that has both "unseen" and "one_time" set does not
479 work if there are any delivery delays because "one_time" forces the
480 parent to be marked "delivered", so its unseen clone is never tried
481 again. For this reason, Exim now forbids the simultaneous setting of
484 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
485 redirected to themselves ("homonym" addresses). Read the long ChangeLog
486 entry if you want to know the details. The fix, however, neglected to
487 consider the case when local delivery batching is involved. The test for
488 "previously delivered" was not happening when checking to see if an
489 address could be batched with a previous (undelivered) one; under
490 certain circumstances this could lead to multiple deliveries to the same
493 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
494 in its include files, and this causes problems building Exim.
496 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
497 header_syntax) but Exim was just ignoring anything given after a slash.
498 In particular, this caused confusion with an attempt to use "verify =
499 reverse_host_lookup/defer_ok". An error is now given when options are
500 supplied for verify items that do not have them. (Maybe reverse_host_
501 lookup should have a defer_ok option, but that's a different point.)
503 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
504 defined by RFC 821) to 2048, because there were problems with some AUTH
505 commands, and RFC 1869 says the size should be increased for extended
506 SMTP commands that take arguments.
508 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
511 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
512 "unknown" error; now it says that the functionality isn't in the binary.
514 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
515 an address' error message when a string expansion fails (syntax or
516 whatever). Otherwise the password may appear in the log. Following change
517 PH/42 below, there is no longer a chance of it appearing in a bounce
520 PH/11 Installed exipick version 20050225.0 from John Jetmore.
522 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
523 of its addresses was ever tried. (Bugzilla bug #2.)
525 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
526 the result incorrectly in the debug output. (It correctly added a newline
527 to what was transported.)
529 TF/01 Added $received_time.
531 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
532 commented out examples of how to interface to a virus scanner and to
533 SpamAssassin. Also added commented examples of av_scanner and
534 spamd_address settings.
536 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
537 and controls are allowed in which ACLs. There were a couple of minor
538 errors. Some of the entries in the conditions table (which is a table of
539 where they are NOT allowed) were getting very unwieldy; rewrote them as a
540 negation of where the condition IS allowed.
542 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
544 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
545 header file does not have a version number, so I've had to invent a new
546 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
547 API. The code is untested by me (my Linux distribution still has 0.3.2 of
548 radiusclient), but it was contributed by a Radius user.
550 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
551 files or directories, for OpenSSL.
553 PH/19 When an Exim process that is running as root has to create an Exim log
554 file, it does so in a subprocess that runs as exim:exim so as to get the
555 ownership right at creation (otherwise, other Exim processes might see
556 the file with the wrong ownership). There was no test for failure of this
557 fork() call, which would lead to the process getting stuck as it waited
558 for a non-existent subprocess. Forks do occasionally fail when resources
559 run out. I reviewed all the other calls to fork(); they all seem to check
562 PH/20 When checking for unexpected SMTP input at connect time (before writing
563 the banner), Exim was not dealing correctly with a non-positive return
564 from the read() function. If the client had disconnected by this time,
565 the result was a log entry for a synchronization error with an empty
566 string after "input=" when read() returned zero. If read() returned -1
567 (an event I could not check), uninitialized data bytes were printed.
568 There were reports of junk text (parts of files, etc) appearing after
571 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
573 PH/22 Added support for macro redefinition, and (re)definition in between
574 driver and ACL definitions.
576 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
577 forgetting to use the resulting value; it was using the unexpanded value.
579 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
580 hadn't been configured. The fix is from Juergen Kreileder, who
581 understands it better than I do:
583 "Here's what I see happening with three configured cyrus_sasl
584 authenticators configured (plain, login, cram-md5):
586 On startup auth_cyrus_sasl_init() gets called for each of these.
587 This means three calls to sasl_listmech() without a specified mech_list.
588 => SASL tests which mechs of all available mechs actually work
589 => three warnings about OTP not working
590 => the returned list contains: plain, login, cram-md5, digest-md5, ...
592 With the patch, sasl_listmech() also gets called three times. But now
593 SASL's mech_list option is set to the server_mech specified in the the
594 authenticator. Or in other words, the answer from sasl_listmech()
595 gets limited to just the mech you're testing for (which is different
597 => the return list contains just 'plain' or 'login', 'cram-md5' or
598 nothing depending on the value of ob->server_mech.
600 I've just tested the patch: Authentication still works fine,
601 unavailable mechs specified in the exim configuration are still
602 caught, and the auth.log warnings about OTP are gone."
604 PH/25 When debugging is enabled, the contents of the command line are added
605 to the debugging output, even when log_selector=+arguments is not
608 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
609 answer is "GNU", and only if the return is "GNU/something" is the answer
612 PH/27 $acl_verify_message is now set immediately after the failure of a
613 verification in an ACL, and so is available in subsequent modifiers. In
614 particular, the message can be preserved by coding like this:
616 warn !verify = sender
617 set acl_m0 = $acl_verify_message
619 Previously, $acl_verify_message was set only while expanding "message"
620 and "log_message" when a very denied access.
622 PH/28 Modified OS/os.c-Linux with
624 -#ifndef OS_LOAD_AVERAGE
625 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
627 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
628 nomenclature these days.)
630 PH/29 Installed patch from the Sieve maintainer that adds the options
631 sieve_useraddress and sieve_subaddress to the redirect router.
633 PH/30 In these circumstances:
634 . Two addresses routed to the same list of hosts;
635 . First host does not offer TLS;
636 . First host accepts first address;
637 . First host gives temporary error to second address;
638 . Second host offers TLS and a TLS session is established;
639 . Second host accepts second address.
640 Exim incorrectly logged both deliveries with the TLS parameters (cipher
641 and peerdn, if requested) that were in fact used only for the second
644 PH/31 When doing a callout as part of verifying an address, Exim was not paying
645 attention to any local part prefix or suffix that was matched by the
646 router that accepted the address. It now behaves in the same way as it
647 does for delivery: the affixes are removed from the local part unless
648 rcpt_include_affixes is set on the transport.
650 PH/32 Add the sender address, as F=<...>, to the log line when logging a
651 timeout during the DATA phase of an incoming message.
653 PH/33 Sieve envelope tests were broken for match types other than :is. I have
654 applied a patch sanctioned by the Sieve maintainer.
656 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
657 the uid or gid is negative. A case of a negative gid caused this to be
658 noticed. The fix allows for either to be negative.
660 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
661 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
664 PH/36 The change PH/12 above was broken. Fixed it.
666 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
667 the grounds that routing the same address twice would always produce the
668 same answer. This might have been true once, but it is certainly no
669 longer true now. Routing a child address may depend on the previous
670 routing that produced that child. Some complicated redirection strategies
671 went wrong when messages had multiple recipients, and made Exim's
672 behaviour dependent on the order in which the addresses were given.
674 I have moved the duplicate checking until after the routing is complete.
675 Exim scans the addresses that are assigned to local and remote
676 transports, and removes any duplicates. This means that more work will be
677 done, as duplicates will always all be routed, but duplicates are
678 presumably rare, so I don't expect this is of any significance.
680 For deliveries to pipes, files, and autoreplies, the duplicate checking
681 still happens during the routing process, since they are not going to be
684 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
685 It corrects a timeout issue with spamd. This is Ian's comment: "The
686 background is that sometimes spamd either never reads data from a
687 connection it has accepted, or it never writes response data. The exiscan
688 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
689 blindly assumes that writes won't block so it may never time out."
691 PH/39 Allow G after quota size as well as K and M.
693 PH/40 The value set for $authenticated_id in an authenticator may not contain
694 binary zeroes or newlines because the value is written to log lines and
695 to spool files. There was no check on this. Now the value is run through
696 the string_printing() function so that such characters are converted to
697 printable escape sequences.
699 PH/41 $message_linecount is a new variable that contains the total number of
700 lines in the message. Compare $body_linecount, which is the count for the
703 PH/42 Exim no longer gives details of delivery errors for specific addresses in
704 bounce and delay warning messages, except in certain special cases, which
707 (a) An SMTP error message from a remote host;
708 (b) A message specified in a :fail: redirection;
709 (c) A message specified in a "fail" command in a system filter;
710 (d) A message specified in a FAIL return from the queryprogram router;
711 (e) A message specified by the cannot_route_message router option.
713 In these cases only, Exim does include the error details in bounce and
714 warning messages. There are also a few cases where bland messages such
715 as "unrouteable address" or "local delivery error" are given.
717 PH/43 $value is now also set for the "else" part of a ${run expansion.
719 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
720 being worked on, but at least Exim now implements the latest version to
723 PH/45 In a pipe transport, although a timeout while waiting for the pipe
724 process to complete was treated as a delivery failure, a timeout while
725 writing the message to the pipe was logged, but erroneously treated as a
726 successful delivery. Such timeouts include transport filter timeouts. For
727 consistency with the overall process timeout, these timeouts are now
728 treated as errors, giving rise to delivery failures by default. However,
729 there is now a new Boolean option for the pipe transport called
730 timeout_defer, which, if set TRUE, converts the failures into defers for
731 both kinds of timeout. A transport filter timeout is now identified in
734 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
735 systems where "make" and "gmake" are different, calling "gmake" at top
736 level broke things. I've arranged for the value of $(MAKE) to be passed
737 from the Makefile to this script so that it can call the same version of
741 A note about Exim versions 4.44 and 4.50
742 ----------------------------------------
744 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
745 changes of various kinds. As a consequence, a big documentation update was
746 needed. This delayed the release for rather longer than seemed good, especially
747 in the light of a couple of (minor) security issues. Therefore, the changes
748 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
749 release. So 4.44 and 4.50 are in effect two different branches that both start
752 I have left the 4.50 change log unchanged; it contains all the changes since
753 4.43. The change log for 4.44 is below; many of its items are identical to
754 those for 4.50. This seems to be the most sensible way to preserve the
755 historical information.
761 1. Minor wording change to the doc/README.SIEVE file.
763 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
764 computation of the current number of files was incorrect.
766 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
767 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
770 4. Give more explanation in the error message when the command for a transport
771 filter fails to execute.
773 5. There are several places where Exim runs a non-Exim command in a
774 subprocess. The SIGUSR1 signal should be disabled for these processes. This
775 was being done only for the command run by the queryprogram router. It is
776 now done for all such subprocesses. The other cases are: ${run, transport
777 filters, and the commands run by the lmtp and pipe transports.
779 6. Added CONFIGURE_GROUP build-time option.
781 7. Some older OS have a limit of 256 on the maximum number of file
782 descriptors. Exim was using setrlimit() to set 1000 as a large value
783 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
784 systems. I've change it so that if it can't get 1000, it tries for 256.
786 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
787 was an oversight, and furthermore, ever since the addition of extra
788 controls (e.g. 4.43/32), the checks on when to allow different forms of
789 "control" were broken. There should now be diagnostics for all cases when a
790 control that does not make sense is encountered.
792 9. Added the /retain_sender option to "control=submission".
794 10. $recipients is now available in the predata ACL (oversight).
796 11. Tidy the search cache before the fork to do a delivery from a message
797 received from the command line. Otherwise the child will trigger a lookup
798 failure and thereby defer the delivery if it tries to use (for example) a
799 cached ldap connection that the parent has called unbind on.
801 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
802 of $address_data from the recipient verification was clobbered by the
805 13. The value of address_data from a sender verification is now available in
806 $sender_address_data in subsequent conditions in the ACL statement.
808 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
810 15. Added a new option "connect=<time>" to callout options, to set a different
813 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
814 was its contents. (It was OK if the option was not defined at all.)
816 17. A "Completed" log line is now written for messages that are removed from
817 the spool by the -Mrm option.
819 18. New variables $sender_verify_failure and $recipient_verify_failure contain
820 information about exactly what failed.
822 19. Added -dd to debug only the daemon process.
824 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
825 handles timeouts, both on the server side and network timeouts. Renamed the
826 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
828 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
829 It is now set to "smtps".
831 22. $host_address is now set to the target address during the checking of
834 23. When checking ignore_target_hosts for an ipliteral router, no host name was
835 being passed; this would have caused $sender_host_name to have been used if
836 matching the list had actually called for a host name (not very likely,
837 since this list is usually IP addresses). A host name is now passed as
840 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
841 code that specifies a non-restarting handler (typically sigaction() in
842 modern systems) in an attempt to fix a rare and obscure crash bug.
844 25. Narrowed the window for a race in the daemon that could cause it to ignore
845 SIGCHLD signals. This is not a major problem, because they are used only to
846 wake it up if nothing else does.
848 26. A malformed maildirsize file could cause Exim to calculate negative values
849 for the mailbox size or file count. Odd effects could occur as a result.
850 The maildirsize information is now recalculated if the size or filecount
853 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
854 support for a long time. Removed HAVE_SYS_VFS_H.
856 28. Installed the latest version of exipick from John Jetmore.
858 29. In an address list, if the pattern was not a regular expression, an empty
859 subject address (from a bounce message) matched only if the pattern was an
860 empty string. Non-empty patterns were not even tested. This was the wrong
861 because it is perfectly reasonable to use an empty address as part of a
862 database query. An empty address is now tested by patterns that are
863 lookups. However, all the other forms of pattern expect the subject to
864 contain a local part and a domain, and therefore, for them, an empty
865 address still always fails if the pattern is not itself empty.
867 30. Exim went into a mad DNS loop when attempting to do a callout where the
868 host was specified on an smtp transport, and looking it up yielded more
871 31. Re-factored the code for checking spool and log partition space into a
872 function that finds that data and another that does the check. The former
873 is then used to implement four new variables: $spool_space, $log_space,
874 $spool_inodes, and $log_inodes.
876 32. The RFC2047 encoding function was originally intended for short strings
877 such as real names; it was not keeping to the 75-character limit for
878 encoded words that the RFC imposes. It now respects the limit, and
879 generates multiple encoded words if necessary. To be on the safe side, I
880 have increased the buffer size for the ${rfc2047: expansion operator from
883 33. It is now permitted to omit both strings after an "if" condition; if the
884 condition is true, the result is "true". As before, when the second string
885 is omitted, a false condition yields an empty string. This makes it less
886 cumbersome to write custom ACL and router conditions.
888 34. Failure to deliver a bounce message always caused it to be frozen, even if
889 there was an errors_to setting on the router. The errors_to setting is now
892 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
893 canonical form (fully expanded) before being placed in
894 $sender_host_address.
896 36. The table in the code that translates DNS record types into text (T_A to
897 "A" for instance) was missing entries for NS and CNAME. It is just possible
898 that this could have caused confusion if both these types were looked up
899 for the same domain, because the text type is used as part of Exim's
900 per-process caching. But the chance of anyone hitting this buglet seems
903 37. The dnsdb lookup has been extended in a number of ways.
905 (1) There is a new type, "zns", which walks up the domain tree until it
906 finds some nameserver records. It should be used with care.
908 (2) There is a new type, "mxh", which is like "mx" except that it returns
909 just the host names, not the priorities.
911 (3) It is now possible to give a list of domains (or IP addresses) to be
912 looked up. The behaviour when one of the lookups defers can be
913 controlled by a keyword.
915 (4) It is now possible to specify the separator character for use when
916 multiple records are returned.
918 38. The dnslists ACL condition has been extended: it is now possible to supply
919 a list of IP addresses and/or domains to be looked up in a particular DNS
922 39. Added log_selector=+queue_time_overall.
924 40. When running the queue in the test harness, wait just a tad after forking a
925 delivery process, to get repeatability of debugging output.
927 41. Include certificate and key file names in error message when GnuTLS fails
928 to set them up, because the GnuTLS error message doesn't include the name
929 of the failing file when there is a problem reading it.
931 42. Allow both -bf and -bF in the same test run.
933 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
935 44. The "Exiscan patch" is now merged into the mainline Exim source.
937 45. Sometimes the final signoff response after QUIT could fail to get
938 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
939 < 0 before doing a fflush(). This bug looks as though it goes back to the
940 introduction of TLS in release 3.20, but "sometimes" must have been rare
941 because the tests only now provoked it.
943 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
944 (this can affect the format of dates).
946 47. exim_tidydb, when checking for the continued existence of a message for
947 which it has found a message-specific retry record, was not finding
948 messages that were in split spool directories. Consequently, it was
949 deleting retry records that should have stayed in existence.
951 48. Steve fixed some bugs in eximstats.
953 49. The SPA authentication driver was not abandoning authentication and moving
954 on to the next authenticator when an expansion was forced to fail,
955 contradicting the general specification for all authenticators. Instead it
956 was generating a temporary error. It now behaves as specified.
958 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
959 (the order specifies the preference for clients). The order is now AES256,
960 AES128, 3DES, ARCFOUR128.
962 51. Small patch to Sieve code - explicitly set From: when generating an
965 52. Exim crashed if a remote delivery caused a very long error message to be
966 recorded - for instance if somebody sent an entire SpamAssassin report back
967 as a large number of 550 error lines. This bug was coincidentally fixed by
968 increasing the size of one of Exim's internal buffers (big_buffer) that
969 happened as part of the Exiscan merge. However, to be on the safe side, I
970 have made the code more robust (and fixed the comments that describe what
973 53. Now that there can be additional text after "Completed" in log lines (if
974 the queue_time_overall log selector is set), a one-byte patch to exigrep
975 was needed to allow it to recognize "Completed" as not the last thing in
978 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
979 patch that reportedly fixes this has been added. I am not expert enough to
980 create a test for it. This is what the patch creator wrote:
982 "I found a little strange behaviour of ldap code when working with
983 Windows 2003 AD Domain, where users was placed in more than one
984 Organization Units. When I tried to give exim partial DN, the exit code
985 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
986 But simultaneously result of request was absolutely normal ldap result,
987 so I produce this patch..."
989 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
990 REFERENCE, so I have modified the code to exclude the patch when that macro
993 55. Some experimental protocols are using DNS PTR records for new purposes. The
994 keys for these records are domain names, not reversed IP addresses. The
995 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
996 leaves it alone. Component reversal etc. now happens only for IP addresses.
999 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1001 57. Double the size of the debug message buffer (to 2048) so that more of very
1002 long debug lines gets shown.
1004 58. The exicyclog utility now does better if the number of log files to keep
1005 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1007 59. Two changes related to the smtp_active_hostname option:
1009 (1) $smtp_active_hostname is now available as a variable.
1010 (2) The default for smtp_banner uses $smtp_active_hostname instead
1011 of $primary_hostname.
1013 60. The host_aton() function is supposed to be passed a string that is known
1014 to be a valid IP address. However, in the case of IPv6 addresses, it was
1015 not checking this. This is a hostage to fortune. Exim now panics and dies
1016 if the condition is not met. A case was found where this could be provoked
1017 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1018 components; fortuitously, this particular loophole had already been fixed
1019 by change 4.50/55 above.
1021 If there are any other similar loopholes, the new check in host_aton()
1022 itself should stop them being exploited. The report I received stated that
1023 data on the command line could provoke the exploit when Exim was running as
1024 exim, but did not say which command line option was involved. All I could
1025 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1026 running as the user.
1029 61. There was a buffer overflow vulnerability in the SPA authentication code
1030 (which came originally from the Samba project). I have added a test to the
1031 spa_base64_to_bits() function which I hope fixes it.
1034 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1035 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1037 63. The daemon start-up calls getloadavg() while still root for those OS that
1038 need the first call to be done as root, but it missed one case: when
1039 deliver_queue_load_max is set with deliver_drop_privilege. This is
1040 necessary for the benefit of the queue runner, because there is no re-exec
1041 when deliver_drop_privilege is set.
1043 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1044 This has been fixed.
1046 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1047 was in use, was not putting the data itself into the right store pool;
1048 consequently, it could be overwritten for a subsequent message in the same
1049 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1052 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1054 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1055 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1056 the function were treating the return as a boolean value, which happened to
1057 work because 0=false and not-0=true, but is not correct code.
1059 68. The host_aton() function was not handling scoped IPv6 addresses (those
1060 with, for example, "%eth0" on the end) correctly.
1062 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1063 negated items (that is, ~something) in unsigned ints. Some compilers
1064 apparently mutter when there is no cast.
1066 70. If an address verification called from an ACL failed, and did not produce a
1067 user-specific message (i.e. there was only a "system" message), nothing was
1068 put in $acl_verify_message. In this situation, it now puts the system
1071 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1072 session; change 4.31/43 added the unwanted input to the log line - except
1073 that it did not do this in the start of session case. It now does.
1075 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1076 This is wrong and can cause the other end to generate a synchronization
1077 error if it is another Exim or anything else that does the synchronization
1078 check. A QUIT command is no longer sent after a timeout.
1080 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1081 during host lookups.
1083 74. The defer_ok option of callout verification was not working if it was used
1084 when verifying addresses in header lines, that is, for this case:
1086 verify = header_sender/callout=defer_ok
1088 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1089 those file descriptors could be used for SMTP connections. If anything
1090 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1091 could be sent to the SMTP client, causing chaos. The daemon now opens
1092 stdin, stdout, and stderr to /dev/null when it puts itself into the
1095 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1096 log by default. The user can override this with suitable Perl magic.
1098 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1099 the log message when discard triggers, was not working for the DATA ACL or
1100 for the non-SMTP ACL.
1102 78. Error message wording change in sieve.c.
1104 79. If smtp_accept_max_per_host was set, the number of connections could be
1105 restricted to fewer than expected, because the daemon was trying to set up
1106 a new connection before checking whether the processes handling previous
1107 connections had finished. The check for completed processes is now done
1108 earlier. On busy systems, this bug wouldn't be noticed because something
1109 else would have woken the daemon, and it would have reaped the completed
1112 80. If a message was submitted locally by a user whose login name contained one
1113 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1114 It caused a spool format error. I have fixed the spool reading code. A
1115 related problem was that the "from" clause in the Received: line became
1116 illegal because of the space(s). It is now covered by ${quote_local_part.
1118 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1121 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1123 83. Patch from Sieve maintainer to fix unterminated string problem in
1124 "vacation" handling.
1126 84. Some minor changes to the Linux configuration files to help with other
1127 OS variants using glibc.
1129 85. One more patch for Sieve to update vacation handling to latest spec.
1132 ----------------------------------------------------
1133 See the note above about the 4.44 and 4.50 releases.
1134 ----------------------------------------------------
1140 1. Change 4.43/35 introduced a bug that caused file counts to be
1141 incorrectly computed when quota_filecount was set in an appendfile
1144 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1145 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1148 3. Give more explanation in the error message when the command for a transport
1149 filter fails to execute.
1151 4. There are several places where Exim runs a non-Exim command in a
1152 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1153 was being done only for the command run by the queryprogram router. It is
1154 now done for all such subprocesses. The other cases are: ${run, transport
1155 filters, and the commands run by the lmtp and pipe transports.
1157 5. Some older OS have a limit of 256 on the maximum number of file
1158 descriptors. Exim was using setrlimit() to set 1000 as a large value
1159 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1160 systems. I've change it so that if it can't get 1000, it tries for 256.
1162 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1163 was an oversight, and furthermore, ever since the addition of extra
1164 controls (e.g. 4.43/32), the checks on when to allow different forms of
1165 "control" were broken. There should now be diagnostics for all cases when a
1166 control that does not make sense is encountered.
1168 7. $recipients is now available in the predata ACL (oversight).
1170 8. Tidy the search cache before the fork to do a delivery from a message
1171 received from the command line. Otherwise the child will trigger a lookup
1172 failure and thereby defer the delivery if it tries to use (for example) a
1173 cached ldap connection that the parent has called unbind on.
1175 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1176 of $address_data from the recipient verification was clobbered by the
1177 sender verification.
1179 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1180 was its contents. (It was OK if the option was not defined at all.)
1182 11. A "Completed" log line is now written for messages that are removed from
1183 the spool by the -Mrm option.
1185 12. $host_address is now set to the target address during the checking of
1186 ignore_target_hosts.
1188 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1189 being passed; this would have caused $sender_host_name to have been used if
1190 matching the list had actually called for a host name (not very likely,
1191 since this list is usually IP addresses). A host name is now passed as
1194 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1195 code that specifies a non-restarting handler (typically sigaction() in
1196 modern systems) in an attempt to fix a rare and obscure crash bug.
1198 15. Narrowed the window for a race in the daemon that could cause it to ignore
1199 SIGCHLD signals. This is not a major problem, because they are used only to
1200 wake it up if nothing else does.
1202 16. A malformed maildirsize file could cause Exim to calculate negative values
1203 for the mailbox size or file count. Odd effects could occur as a result.
1204 The maildirsize information is now recalculated if the size or filecount
1207 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1208 support for a long time. Removed HAVE_SYS_VFS_H.
1210 18. Updated exipick to current release from John Jetmore.
1212 19. Allow an empty sender to be matched against a lookup in an address list.
1213 Previously the only cases considered were a regular expression, or an
1216 20. Exim went into a mad DNS lookup loop when doing a callout where the
1217 host was specified on the transport, if the DNS lookup yielded more than
1220 21. The RFC2047 encoding function was originally intended for short strings
1221 such as real names; it was not keeping to the 75-character limit for
1222 encoded words that the RFC imposes. It now respects the limit, and
1223 generates multiple encoded words if necessary. To be on the safe side, I
1224 have increased the buffer size for the ${rfc2047: expansion operator from
1227 22. Failure to deliver a bounce message always caused it to be frozen, even if
1228 there was an errors_to setting on the router. The errors_to setting is now
1231 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1232 canonical form (fully expanded) before being placed in
1233 $sender_host_address.
1235 24. Updated eximstats to version 1.33
1237 25. Include certificate and key file names in error message when GnuTLS fails
1238 to set them up, because the GnuTLS error message doesn't include the name
1239 of the failing file when there is a problem reading it.
1241 26. Expand error message when OpenSSL has problems setting up cert/key files.
1244 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1245 (this can affect the format of dates).
1247 28. exim_tidydb, when checking for the continued existence of a message for
1248 which it has found a message-specific retry record, was not finding
1249 messages that were in split spool directories. Consequently, it was
1250 deleting retry records that should have stayed in existence.
1252 29. eximstats updated to version 1.35
1253 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1254 1.35 - bugfix such that pie charts by volume are generated correctly
1256 30. The SPA authentication driver was not abandoning authentication and moving
1257 on to the next authenticator when an expansion was forced to fail,
1258 contradicting the general specification for all authenticators. Instead it
1259 was generating a temporary error. It now behaves as specified.
1261 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1262 (the order specifies the preference for clients). The order is now AES256,
1263 AES128, 3DES, ARCFOUR128.
1265 31. Small patch to Sieve code - explicitly set From: when generating an
1268 32. Exim crashed if a remote delivery caused a very long error message to be
1269 recorded - for instance if somebody sent an entire SpamAssassin report back
1270 as a large number of 550 error lines. This bug was coincidentally fixed by
1271 increasing the size of one of Exim's internal buffers (big_buffer) that
1272 happened as part of the Exiscan merge. However, to be on the safe side, I
1273 have made the code more robust (and fixed the comments that describe what
1276 33. Some experimental protocols are using DNS PTR records for new purposes. The
1277 keys for these records are domain names, not reversed IP addresses. The
1278 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1279 leaves it alone. Component reversal etc. now happens only for IP addresses.
1282 34. The host_aton() function is supposed to be passed a string that is known
1283 to be a valid IP address. However, in the case of IPv6 addresses, it was
1284 not checking this. This is a hostage to fortune. Exim now panics and dies
1285 if the condition is not met. A case was found where this could be provoked
1286 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1287 components; fortuitously, this particular loophole had already been fixed
1288 by change 4.50/55 or 4.44/33 above.
1290 If there are any other similar loopholes, the new check in host_aton()
1291 itself should stop them being exploited. The report I received stated that
1292 data on the command line could provoke the exploit when Exim was running as
1293 exim, but did not say which command line option was involved. All I could
1294 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1295 running as the user.
1298 35. There was a buffer overflow vulnerability in the SPA authentication code
1299 (which came originally from the Samba project). I have added a test to the
1300 spa_base64_to_bits() function which I hope fixes it.
1303 36. The daemon start-up calls getloadavg() while still root for those OS that
1304 need the first call to be done as root, but it missed one case: when
1305 deliver_queue_load_max is set with deliver_drop_privilege. This is
1306 necessary for the benefit of the queue runner, because there is no re-exec
1307 when deliver_drop_privilege is set.
1309 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1310 was in use, was not putting the data itself into the right store pool;
1311 consequently, it could be overwritten for a subsequent message in the same
1312 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1315 38. Sometimes the final signoff response after QUIT could fail to get
1316 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1317 < 0 before doing a fflush(). This bug looks as though it goes back to the
1318 introduction of TLS in release 3.20, but "sometimes" must have been rare
1319 because the tests only now provoked it.
1325 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1326 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1327 Now it returns an integer. A number of calls were still expecting a T/F
1328 return. Fortuitously, in all cases, the tests worked in OK situations,
1329 which is the norm. However, things would have gone wrong on any write
1330 failures on the smtp file descriptor. This function is used when sending
1331 messages over SMTP and also when doing verify callouts.
1333 2. When Exim is called to do synchronous delivery of a locally submitted
1334 message (the -odf or -odi options), it no longer closes stderr before doing
1337 3. Implemented the mua_wrapper option.
1339 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1341 5. Implemented the functions header_remove(), header_testname(),
1342 header_add_at_position(), and receive_remove_recipient(), and exported them
1345 6. If an ACL "warn" statement specified the addition of headers, Exim already
1346 inserted X-ACL-Warn: at the start if there was no header name. However, it
1347 was not making this test for the second and subsequent header lines if
1348 there were newlines in the string. This meant that an invalid header could
1349 be inserted if Exim was badly configured.
1351 7. Allow an ACL "warn" statement to add header lines at the start or after all
1352 the Received: headers, as well as at the end.
1354 8. Added the rcpt_4xx retry error code.
1356 9. Added postmaster_mailfrom=xxx to callout verification option.
1358 10. Added mailfrom=xxxx to the callout verification option, for verify=
1361 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1362 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1364 12. Inserted some casts to stop certain compilers warning when using pointer
1365 differences as field lengths or precisions in printf-type calls (mostly
1366 affecting debugging statements).
1368 13. Added optional readline() support for -be (dynamically loaded).
1370 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1371 same clock tick as a message's arrival, so that its received time was the
1372 same as the "first fail" time on the retry record, and that message
1373 remained on the queue past the ultimate address timeout, every queue runner
1374 would try a delivery (because it was past the ultimate address timeout) but
1375 after another failure, the ultimate address timeout, which should have then
1376 bounced the address, did not kick in. This was a "< instead of <=" error;
1377 in most cases the first failure would have been in the next clock tick
1378 after the received time, and all would be well.
1380 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1381 being recognized when the domain list was tested by the match_domain
1382 condition in an expansion string.
1384 16. Added the ${str2b64: operator.
1386 17. Exim was always calling setrlimit() to set a large limit for the number of
1387 processes, without checking whether the existing limit was already
1388 adequate. (It did check for the limit on file descriptors.) Furthermore,
1389 errors from getrlimit() and setrlimit() were being ignored. Now they are
1390 logged to the main and panic logs, but Exim does carry on, to try to do its
1391 job under whatever limits there are.
1393 18. Imported PCRE 5.0.
1395 19. Trivial typo in log message " temporarily refused connection" (the leading
1398 20. If the log selector return_path_on_delivery was set and an address was
1399 redirected to /dev/null, the delivery process crashed because it assumed
1400 that a return path would always be set for a "successful" delivery. In this
1401 case, the whole delivery is bypassed as an optimization, and therefore no
1404 21. Internal re-arrangement: the function for sending a challenge and reading
1405 a response while authentication was assuming a zero-terminated challenge
1406 string. It's now changed to take a pointer and a length, to allow for
1407 binary data in such strings.
1409 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1411 23. Exim was not respecting finduser_retries when seeking the login of the
1412 uid under which it was called; it was always trying 10 times. (The default
1413 setting of finduser_retries is zero.) Also, it was sleeping after the final
1414 failure, which is pointless.
1416 24. Implemented tls_on_connect_ports.
1418 25. Implemented acl_smtp_predata.
1420 26. If the domain in control=submission is set empty, Exim assumes that the
1421 authenticated id is a complete email address when it generates From: or
1422 Sender: header lines.
1424 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1425 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1426 chown and chgrp in /bin and hostname in /usr/bin.
1428 28. Exim was keeping the "process log" file open after each use, just as it
1429 does for the main log. This opens the possibility of it remaining open for
1430 long periods when the USR1 signal hits a daemon. Occasional processlog
1431 errors were reported, that could have been caused by this. Anyway, it seems
1432 much more sensible not to leave this file open at all, so that is what now
1435 29. The long-running daemon process does not normally write to the log once it
1436 has entered its main loop, and it closes the log before doing so. This is
1437 so that log files can straightforwardly be renamed and moved. However,
1438 there are a couple of unusual error situations where the daemon does write
1439 log entries, and I had neglected to close the log afterwards.
1441 30. The text of an SMTP error response that was received during a remote
1442 delivery was being truncated at 512 bytes. This is too short for some of
1443 the long messages that one sometimes sees. I've increased the limit to
1446 31. It is now possible to make retry rules that apply only when a message has a
1447 specific sender, in particular, an empty sender.
1449 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1450 it possible to be selective about when SMTP synchronization is enforced.
1452 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1454 32. Implemented hosts_connection_nolog.
1456 33. Added an ACL for QUIT.
1458 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1461 35. Added mailbox_size and mailbox_filecount to appendfile.
1463 36. Added control = no_multiline_responses to ACLs.
1465 37. There was a bug in the logic of the code that waits for the clock to tick
1466 in the case where the clock went backwards by a substantial amount such
1467 that the microsecond fraction of "now" was more than the microsecond
1468 fraction of "then" (but the whole seconds number was less).
1470 38. Added support for the libradius Radius client library this is found on
1471 FreeBSD (previously only the radiusclient library was supported).
1477 1. When certain lookups returned multiple values in the form name=value, the
1478 quoting of the values was not always being done properly. Specifically:
1479 (a) If the value started with a double quote, but contained no whitespace,
1481 (b) If the value contained whitespace other than a space character (i.e.
1482 tabs or newlines or carriage returns) it was not quoted.
1483 This fix has been applied to the mysql and pgsql lookups by writing a
1484 separate quoting function and calling it from the lookup code. The fix
1485 should probably also be applied to nisplus, ibase and oracle lookups, but
1486 since I cannot test any of those, I have not disturbed their existing code.
1488 2. A hit in the callout cache for a specific address caused a log line with no
1489 reason for rejecting RCPT. Now it says "Previous (cached) callout
1490 verification failure".
1492 3. There was an off-by-one bug in the queryprogram router. An over-long
1493 return line was truncated at 256 instead of 255 characters, thereby
1494 overflowing its buffer with the terminating zero. As well as fixing this, I
1495 have increased the buffer size to 1024 (and made a note to document this).
1497 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1498 when Exim is waiting for an SMTP response from a remote server, Exim
1499 restarts its select() call on the socket, thereby resetting its timeout.
1500 This is not a problem when such interrupts are rare. Somebody set up a cron
1501 job to run exiwhat every 2 minutes, which is less than the normal select()
1502 timeout (5 or 10 minutes). This meant that the select() timeout never
1503 kicked in because it was always reset. I have fixed this by comparing the
1504 time when an interrupt arrives with the time at the start of the first call
1505 to select(). If more time than the timeout has elapsed, the interrupt is
1506 treated as a timeout.
1508 5. Some internal re-factoring in preparation for the addition of Sieve
1509 extensions (by MH). In particular, the "personal" test is moved to a
1510 separate function, and given an option for scanning Cc: and Bcc: (which is
1511 not set for Exim filters).
1513 6. When Exim created an email address using the login of the caller as the
1514 local part (e.g. when creating a From: or Sender: header line), it was not
1515 quoting the local part when it contained special characters such as @.
1517 7. Installed new OpenBSD configuration files.
1519 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1520 try to make them clearer.
1522 9. Callout options, other than the timeout value, were being ignored when
1523 verifying sender addresses in header lines. For example, when using
1525 verify = header_sender/callout=no_cache
1527 the cache was (incorrectly) being used.
1529 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1530 only the Cygwin environment.
1532 11. When return_path_on_delivery was set as a log selector, if different remote
1533 addresses in the same message used different return paths and parallel
1534 remote delivery occurred, the wrong values would sometimes be logged.
1535 (Whenever a remote delivery process finished, the return path value from
1536 the most recently started remote delivery process was logged.)
1538 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1539 lines when AUTH and/or TLS are in use. This is the "received protocol"
1540 field. Exim used to use "asmtp" for authenticated SMTP, without any
1541 indication (in the protocol name) for TLS use. Now it follows the RFC and
1542 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1543 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1544 names appear in log lines as well as in Received: header lines.
1546 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1547 extensions, and comparison tests, and to fix some bugs.
1549 14. Changes to the "personal" filter test:
1551 (1) The test was buggy in that it was just doing the equivalent of
1552 "contains" tests on header lines. For example, if a user's address was
1553 anne@some.where, the "personal" test would incorrectly be true for
1555 To: susanne@some.where
1557 This test is now done by extracting each address from the header in turn,
1558 and checking the entire address. Other tests that are part of "personal"
1559 are now done using regular expressions (for example, to check local parts
1560 of addresses in From: header lines).
1562 (2) The list of non-personal local parts in From: addresses has been
1563 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1564 taken from the Sieve specification recommendations.
1566 (3) If the message contains any header line starting with "List-" it is
1567 treated as non-personal.
1569 (4) The test for "circular" in the Subject: header line has been removed
1570 because it now seems ill-conceived.
1572 15. Minor typos in src/EDITME comments corrected.
1574 16. Installed latest exipick from John Jetmore.
1576 17. If headers_add on a router specified a text string that was too long for
1577 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1578 of string_sprintf() is now avoided.
1580 18. $message_body_size was not set (it was always zero) when running the DATA
1581 ACL and the local_scan() function.
1583 19. For the "mail" command in an Exim filter, no default was being set for
1584 the once_repeat time, causing a random time value to be used if "once" was
1585 specified. (If the value happened to be <= 0, no repeat happened.) The
1586 default is now 0s, meaning "never repeat". The "vacation" command was OK
1587 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1588 (I found it when inspecting the code).
1590 20. There is now an overall timeout for performing a callout verification. It
1591 defaults to 4 times the callout timeout, which applies to individual SMTP
1592 commands during the callout. The overall timeout applies when there is more
1593 than one host that can be tried. The timeout is checked before trying the
1594 next host. This prevents very long delays if there are a large number of
1595 hosts and all are timing out (e.g. when the network connections are timing
1596 out). The value of the overall timeout can be changed by specifying an
1597 additional sub-option for "callout", called "maxwait". For example:
1599 verify = sender/callout=5s,maxwait=20s
1601 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1602 to the end before writing, but this should make it even safer).
1604 22. Exim was forgetting that it had advertised PIPELINING for the second and
1605 subsequent messages on an SMTP connection. It was also not resetting its
1606 memory on STARTTLS and an internal HELO.
1608 23. When Exim logs an SMTP synchronization error within a session, it now
1609 records whether PIPELINING has been advertised or not.
1611 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1612 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1613 rather than long int.
1615 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1617 26. Added the never_mail option to autoreply.
1623 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1624 crash if the getsockname() call failed; this can happen if a connection is
1625 closed very soon after it is established. The problem was simply in the
1626 order in which certain operations were done, causing Exim to try to write
1627 to the SMTP stream before it had set up the file descriptor. The bug has
1628 been fixed by making things happen in the correct order.
1634 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1635 before the connection was closed, thus losing the rejection response.
1637 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1638 some early Solaris releases, but causes trouble in current releases where
1639 socklen_t is defined.
1641 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1644 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1647 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1648 that releasing the top of it at the end releases what was used for sub-
1649 expansions (unless the block got too big). However, discard this block if
1650 the first thing is a variable or header, so that we can use its block when
1651 it is dynamic (useful for very large $message_headers, for example).
1653 6. Lookups now cache *every* query, not just the most recent. A new, separate
1654 store pool is used for this. It can be recovered when all lookup caches are
1655 flushed. Lookups now release memory at the end of their result strings.
1656 This has involved some general refactoring of the lookup sources.
1658 7. Some code has been added to the store_xxx() functions to reduce the amount
1659 of flapping under certain conditions.
1661 8. log_incoming_interface used to affect only the <= reception log lines. Now
1662 it causes the local interface and port to be added to several more SMTP log
1663 lines, for example "SMTP connection from", and rejection lines.
1665 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1667 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1669 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1670 could be overwritten at the end of the current message (or the start of a
1671 new message if it was set in a HELO ACL). The value is now preserved for
1672 the duration of the SMTP connection.
1674 12. If a transport had a headers_rewrite setting, and a matching header line
1675 contained an unqualified address, that address was qualified, even if it
1676 did not match any rewriting rules. The underlying bug was that the values
1677 of the flags that permit the existence of unqualified sender and recipient
1678 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1679 non-local messages, and by -bnq for local messages) were not being
1680 preserved with the message after it was received.
1682 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1683 "next input=" as part of the text comprising the host identity instead of
1684 the correct text. The code was using the same buffer for two different
1685 strings. However, depending on which order the printing function evaluated
1686 its arguments, the bug did not always show up. Under Linux, for example, my
1687 test suite worked just fine.
1689 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1690 doesn't work with some older versions of Perl. It has been changed to "my",
1691 which in any case is probably the better facility to use.
1693 15. A really picky compiler found some instances of statements for creating
1694 error messages that either had too many or two few arguments for the format
1697 16. The size of the buffer for calls to the DNS resolver has been increased
1698 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1699 for addresses that have a lot of PTR records. This alleviates a problem; it
1700 does not fully solve it.
1702 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1703 buffer now truncates the list and logs the incident, which is the same
1704 action as happens when Exim is looking up a host name and its aliases.
1705 Previously in this situation something unpredictable would happen;
1706 sometimes it was "internal error: store_reset failed".
1708 18. If a server dropped the connection unexpectedly when an Exim client was
1709 using GnuTLS and trying to read a response, the client delivery process
1710 crashed while trying to generate an error log message.
1712 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1713 string, for example:
1715 warn message = H1: something\nH2: something
1717 the text was added as a single header line from Exim's point of view
1718 though it ended up OK in the delivered message. However, searching for the
1719 second and subsequent header lines using $h_h2: did not work. This has been
1720 fixed. Similarly, if a system filter added multiple headers in this way,
1721 the routers could not see them.
1723 20. Expanded the error message when iplsearch is called with an invalid key to
1724 suggest using net-iplsearch in a host list.
1726 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1727 ACLs are no longer actually imposed (and a message to that effect is
1730 22. If a "gecos" field in a passwd entry contained escaped characters, in
1731 particular, if it contained a \" sequence, Exim got it wrong when building
1732 a From: or a Sender: header from that name. A second bug also caused
1733 incorrect handling when an unquoted " was present following a character
1734 that needed quoting.
1736 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1737 was not being matched caselessly.
1739 24. Arranged for all hyphens in the exim.8 source to be escaped with
1742 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1743 properly. Recipient callout cache records were still being keyed to include
1744 the sender, even when use_sender was set false. This led to far more
1745 callouts that were necessary. The sender is no longer included in the key
1746 when use_sender is false.
1748 26. Added "control = submission" modifier to ACLs.
1750 27. Added the ${base62d: operator to decode base 62 numbers.
1752 28. dnsdb lookups can now access SRV records.
1754 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1755 the configuration file.
1757 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1758 (-v) mode. This makes the output for a verbose queue run more intelligible.
1760 31. Added a use_postmaster feature to recipient callouts.
1762 32. Added the $body_zerocount variable, containing the number of binary zero
1763 bytes in the message body.
1765 33. The time of last modification of the "new" subdirectory is now used as the
1766 "mailbox time last read" when there is a quota error for a maildir
1769 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1771 35. Added +ignore_unknown as a special item in host lists.
1773 36. Code for decoding IPv6 addresses in host lists is now included, even if
1774 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1775 address was recognized as an IP address, but was then not correctly decoded
1776 into binary, causing unexpected and incorrect effects when compared with
1783 1. Very minor rewording of debugging text in manualroute to say "list of
1784 hosts" instead of "hostlist".
1786 2. If verify=header_syntax was set, and a header line with an unqualified
1787 address (no domain) and a large number of spaces between the end of the
1788 name and the colon was received, the reception process suffered a buffer
1789 overflow, and (when I tested it) crashed. This was caused by some obsolete
1790 code that should have been removed. The fix is to remove it!
1792 3. When running in the test harness, delay a bit after writing a bounce
1793 message to get a bit more predictability in the log output.
1795 4. Added a call to search_tidyup() just before forking a reception process. In
1796 theory, someone could use a lookup in the expansion of smtp_accept_max_
1797 per_host which, without the tidyup, could leave open a database connection.
1799 5. Added the variables $recipient_data and $sender_data which get set from a
1800 lookup success in an ACL "recipients" or "senders" condition, or a router
1801 "senders" option, similar to $domain_data and $local_part_data.
1803 6. Moved the writing of debug_print from before to after the "senders" test
1806 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1807 problems for message scanning, either using a data ACL, or using
1808 local_scan() because the Received: header was not generated till after they
1809 were called (in order to set the time as the time of reception completion).
1810 I have revised the way this works. The header is now generated after the
1811 body is received, but before the ACL or local_scan() are called. After they
1812 are run, the timestamp in the header is updated.
1818 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1819 before starting a queue runner without re-exec. This happened only when
1820 deliver_drop_privilege was set or when the Exim user was set to root. The
1821 effect of the bug was that timeouts during subsequent deliveries caused
1822 crashes instead of being properly handled. The handler is now left at its
1823 default (and expected) setting.
1825 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1826 message, again when deliver_drop_privilege is set or Exim is run as root.
1827 The bug described in (1) was not present in this case, but the tidying up
1828 of the other signals was missing. I have made the two cases consistent.
1830 3. The ignore_target_hosts setting on a manualroute router was being ignored
1831 for hosts that were looked up using the /MX notation.
1833 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1836 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1837 operated on the sender address. After changing the $sender_address to <>
1838 for the sender address verify, Exim was re-instated it as the original
1839 (before rewriting) address, but remembering that it had rewritten it, so it
1840 wasn't rewriting it again. This bug also had the effect of breaking the
1841 sender address verification caching when the sender address was rewritten.
1843 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1844 This has been changed so that if the ip literal address matches
1845 ignore_target_hosts, the router declines.
1847 7. Added expansion conditions match_domain, match_address, and match_local_
1848 part (NOT match_host).
1850 8. The placeholder for the Received: header didn't have a length field set.
1852 9. Added code to Exim itself and to exim_lock to test for a specific race
1853 condition that could lead to file corruption when using MBX delivery. The
1854 issue is with the lockfile that is created in /tmp. If this file is removed
1855 after a process has opened it but before that process has acquired a lock,
1856 there is the potential for a second process to recreate the file and also
1857 acquire a lock. This could lead to two Exim processes writing to the file
1858 at the same time. The added code performs the same test as UW imapd; it
1859 checks after acquiring the lock that its file descriptor still refers to
1860 the same named file.
1862 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1863 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1866 11. Added the smtp_active_hostname option. If used, this will typically be made
1867 to depend on the incoming interface address. Because $interface_address is
1868 not set up until the daemon has forked a reception process, error responses
1869 that can happen earlier (such as "too many connections") no longer contain
1872 12. If an expansion in a condition on a "warn" statement fails because a lookup
1873 defers, the "warn" statement is abandoned, and the next ACL statement is
1874 processed. Previously this caused the whole ACL to be aborted.
1876 13. Added the iplsearch lookup type.
1878 14. Added ident_timeout as a log selector.
1880 15. Added tls_certificate_verified as a log selector.
1882 16. Added a global option tls_require_ciphers (compare the smtp transport
1883 option of the same name). This controls incoming TLS connections.
1885 17. I finally figured out how to make tls_require_ciphers do a similar thing
1886 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1887 before starting the TLS session.
1889 18. Tabs are now shown as \t in -bP output.
1891 19. If the log selector return_path_on_delivery was set, Exim crashed when
1892 bouncing a message because it had too many Received: header lines.
1894 20. If two routers both had headers_remove settings, and the first one included
1895 a superfluous trailing colon, the final name in the first list and the
1896 first name in the second list were incorrectly joined into one item (with a
1897 colon in the middle).
1903 1. Added -C and -D options to the exinext utility, mainly to make it easier
1904 to include in the automated testing, but these could be helpful when
1905 multiple configurations are in use.
1907 2. The exinext utility was not formatting the output nicely when there was
1908 an alternate port involved in the retry record key, nor when there was a
1909 message id as well (for retries that were specific to a specific message
1910 and a specific host). It was also confused by IPv6 addresses, because of
1911 the additional colons they contain. I have fixed the IPv4 problem, and
1912 patched it up to do a reasonable job for IPv6.
1914 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1915 delivery, the log line now contains "pipelined" if PIPELINING was used.
1917 4. An SMTP transport process used to panic and die if the bind() call to set
1918 an explicit outgoing interface failed. This has been changed; it is now
1919 treated in the same way as a connect() failure.
1921 5. A reference to $sender_host_name in the part of a conditional expansion
1922 that was being skipped was still causing a DNS lookup. This no longer
1925 6. The def: expansion condition was not recognizing references to header lines
1926 that used bh_ and bheader_.
1928 7. Added the _cache feature to named lists.
1930 8. The code for checking quota_filecount in the appendfile transport was
1931 allowing one more file than it should have been.
1933 9. For compatibility with Sendmail, the command line option
1941 and sets the incoming protocol and host name (for trusted callers). The
1942 host name and its colon can be omitted when only the protocol is to be set.
1943 Note the Exim already has two private options, -pd and -ps, that refer to
1944 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1945 "s", but I don't think that's a major issue.
1947 10. A number of refactoring changes to the code, none of which should affect
1950 (a) The number of logging options was getting close to filling up the
1951 32-bit word that was used as a bit map. I have split them into two classes:
1952 those that are passed in the argument to log_write(), and those that are
1953 only ever tested independently outside of that function. These are now in
1954 separate 32-bit words, so there is plenty of room for expansion again.
1955 There is no change in the user interface or the logging behaviour.
1957 (b) When building, for example, log lines, the code previously used a
1958 macro that called string_cat() twice, in order to add two strings. This is
1959 not really sufficiently general. Furthermore, there was one instance where
1960 it was actually wrong because one of the argument was used twice, and in
1961 one call a function was used. (As it happened, calling the function twice
1962 did not affect the overall behaviour.) The macro has been replaced by a
1963 function that can join an arbitrary number of extra strings onto a growing
1966 (c) The code for expansion conditions now uses a table and a binary chop
1967 instead of a serial search (which was left over from when there were very
1968 few conditions). Also, it now recognizes conditions like "pam" even when
1969 the relevant support is not compiled in: a suitably worded error message is
1970 given if an attempt is made to use such a condition.
1972 11. Added ${time_interval:xxxxx}.
1974 12. A bug was causing one of the ddress fields not to be passed back correctly
1975 from remote delivery subprocesses. The field in question was not being
1976 subsequently used, so this caused to problems in practice.
1978 13. Added new log selectors queue_time and deliver_time.
1980 14. Might have fixed a bug in maildirsizefile handling that threw up
1981 "unexpected character" debug warnings, and recalculated the data
1982 unnecessarily. In any case, I expanded the warning message to give more
1985 15. Added the message "Restricted characters in address" to the statements in
1986 the default ACL that block characters like @ and % in local parts.
1988 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1989 Three changes have been made:
1991 (a) There was a serious bug; a negative response to MAIL caused the whole
1992 recipient domain to be cached as invalid, thereby blocking all messages
1993 to all local parts at the same domain, from all senders. This bug has
1994 been fixed. The domain is no longer cached after a negative response to
1995 MAIL if the sender used is not empty.
1997 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2000 (c) A new callout option, "use_sender" has been added for people who want
2001 the modified behaviour.
2007 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2010 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2011 indeed breaks things for older releases.
2013 3. Added additional logging to the case where there is a problem reading data
2014 from a filter that is running in a subprocess using a pipe, in order to
2015 try to track down a specific problem.
2017 4. Testing facility fudge: when running in the test harness and attempting
2018 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2019 getting "No route to host". Convert this to a timeout.
2021 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2024 6. Some OS don't have socklen_t but use size_t instead. This affects the
2025 fifth argument of getsockopt() amongst other things. This is now
2026 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2027 can be set for individual OS. I have set it for SunOS5, OSF1, and
2028 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2029 some earlier ones do not.
2031 7. Change 4.30/15 was not doing the test caselessly.
2033 8. The standard form for an IPv6 address literal was being rejected by address
2034 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2035 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2036 this, as well as the form without the "IPv6" on the front (but only when
2037 address literals are enabled, of course).
2039 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2041 10. Exim crashed if a message with an empty sender address specified by -f
2042 encountered a router with an errors_to setting. This could be provoked only
2043 by a command such as
2047 where an empty string was supplied; "<>" did not hit this bug.
2049 11. Installed PCRE release 4.5.
2051 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2052 remained set. It is now erased.
2054 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2055 times from message ids (which are base 36 rather than the normal 62).
2057 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2058 were being counted as actual protocol errors, and logged if the log
2059 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2060 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2061 and DATA following a set of rejected RCPTs do not count as protocol errors.
2062 In other words, Exim assumes they were pipelined, though this may not
2063 actually be the case. Of course, in all cases the client gets an
2064 appropriate error code.
2066 15. If a lookup fails in an ACL condition, a message about the failure may
2067 be available; it is used if testing the ACL cannot continue, because most
2068 such messages specify what the cause of the deferral is. However, some
2069 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2070 that caused an old message to be retained and used if a later statement
2071 caused a defer, replacing the real cause of the deferral.
2073 16. If an IP address had so many PTR records that the DNS lookup buffer
2074 was not large enough to hold them, Exim could crash while trying to process
2075 the truncated data. It now detects and logs this case.
2077 17. Further to 4.21/58, another change has been made: if (and only if) the
2078 first line of a message (the first header line) ends with CRLF, a bare LF
2079 in a subsequent header line has a space inserted after it, so as not to
2080 terminate the header.
2082 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2083 unnecessarily, used atoi() instead of strtol(), and didn't check the
2084 termination when getting file sizes from file names by regex.
2086 19. Completely re-implemented the support for maildirsize files, in the light
2087 of a number of problems with the previous contributed implementation
2088 (4.30/29). In particular:
2090 . If the quota is zero, the maildirsize file is maintained, but no quota is
2093 . If the maildir directory does not exist, it is created before any attempt
2094 to write a maildirsize file.
2096 . The quota value in the file is just a cache; if the quota is changed in
2097 the transport, the new value overrides.
2099 . A regular expression is available for excluding directories from the
2102 20. The autoreply transport checks the characters in options that define the
2103 message's headers; it allows continued headers, but it was checking with
2104 isspace() after an embedded newline instead of explicitly looking for a
2107 21. If all the "regular" hosts to which an address was routed had passed their
2108 expiry times, and had not reached their retry times, the address was
2109 bounced, even if fallback hosts were defined. Now Exim should go on to try
2112 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2113 equivalent code in the SMTP transport. Some hosts send humungous responses
2114 to HELO/EHLO, more than 1024 it seems.
2116 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2117 gives compiler warnings in some environments. I've now done it "properly",
2120 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2121 (because of problems with the built-in one) was declared to return uschar *
2122 instead of char *, causing compiler failure.
2124 25. Fixed a file descriptor leak when processing alias/forward files.
2126 26. Fixed a minor format string issue in dbfn.c.
2128 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2130 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2131 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2134 29. When a sender address is verified, it is cached, to save repeating the test
2135 when there is more than one recipient in a message. However, when the
2136 verification involves a callout, it is possible for different callout
2137 options to be set for different recipients. It is too complicated to keep
2138 track of this in the cache, so now Exim always runs a verification when a
2139 callout is required, relying on the callout cache for the optimization.
2140 The overhead is duplication of the address routing, but this should not be
2143 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2144 to be verified with callout=postmaster, and the main callout worked but the
2145 postmaster check failed, the verification correctly failed. However, if a
2146 subsequent RCPT command asked for sender verification *without* the
2147 postmaster check, incorrect caching caused this verification also to fail,
2150 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2151 it was not caching the DNS options (qualify_single, search_parents) that
2152 were used when the lookup failed. A subsequent lookup with different
2153 options therefore always gave the same answer, though there were cases
2154 where it should not have. (Example: a "domains = !$mx_any" option on a
2155 dnslookup router: the "domains" option is always processed without any
2156 widening, but the router might have qualify_single set.) Now Exim uses the
2157 cached value only when the same options are set.
2159 32. Added John Jetmore's "exipick" utility to the distribution.
2161 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2162 than a timeout (e.g. a certificate is required, and is not provided), an
2163 Exim server now closes the connection immediately. Previously it waited for
2164 the client to close - but if the client is SSL, it seems that they each
2165 wait for each other, leading to a delay before one of them times out.
2167 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2168 maintained 0.8.x compatibility because I don't think many are using it, and
2169 it is clearly obsolete.
2171 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2174 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2175 client certificate was expired. A simple patch fixes this, though I don't
2176 understand the full logic of why the verify callback is called multiple
2179 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2180 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2181 which causes problems with some clients (such as the Certicom SSL Plus
2182 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2183 disables the coutermeasure allowing Eudora to connect."
2185 38. Exim was not checking that a write() to a log file succeeded. This could
2186 lead to Bad Things if a log got too big, in particular if it hit a file
2187 size limit. Exim now panics and dies if it cannot write to a log file, just
2188 as it does if it cannot open a log file.
2190 39. Modified OS/Makefile-Linux so that it now contains
2192 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2194 The two -D definitions ensure that Exim is compiled with large file
2195 support, which makes it possible to handle log files that are bigger than
2198 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2199 instance) a domain was checked against a named list that involved a lookup,
2200 causing $domain_data to be set, then another domain was checked against the
2201 same list, then the first domain was re-checked, the value of $domain_data
2202 after the final check could be wrong. In particular, if the second check
2203 failed, it could be set empty. This bug probably also applied to
2206 41. The strip_trailing_dot option was not being applied to the address given
2207 with the -f command-line option.
2209 42. The code for reading a message's header from the spool was incrementing
2210 $received_count, but never initializing it. This meant that the value was
2211 incorrect (doubled) while delivering a message in the same process in which
2212 it was received. In the most common configuration of Exim, this never
2213 happens - a fresh exec is done - but it can happen when
2214 deliver_drop_privilege is set.
2216 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2217 it now includes up to 150 characters of the unexpected data in the log
2220 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2221 and building data strings. The size of both of these buffers was 10 000
2222 bytes - far larger than anybody would *ever* want, thought I. Needless to
2223 say, somebody hit the limit. I have increased the maximum line length to
2224 20 000 and the maximum data length of concatenated lines to 100 000. I have
2225 also fixed two bugs, because there was no checking on these buffers. Tsk,
2226 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2227 buffer is too small.
2229 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2230 lsearch lookups. Now it does.
2232 46. When parsing a route_list item in a manualroute router, a fixed-length
2233 buffer was used for the list of hosts. I made this 1024 bytes long,
2234 thinking that nobody would ever have a list of hosts that long. Wrong.
2235 Somebody had a whole pile of complicated expansion conditions, and the
2236 string was silently truncated, leading to an expansion error. It turns out
2237 that it is easier to change to an unlimited length (owing to other changes
2238 that have happened since this code was originally written) than to build
2239 structure for giving a limitation error. The length of the item that
2240 expands into the list of hosts is now unlimited.
2242 47. The lsearch lookup could not handle data where the length of text line was
2243 more than 4095 characters. Such lines were truncated, leading to shortened
2244 data being returned. It should now handle lines of any length.
2246 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2247 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2250 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2251 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2252 no longer happens in comment lines. A list of replacements is now placed
2253 at the head of all of the source files, except those whose only change is
2254 to replace PERL_COMMAND in the very first #! line.
2256 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2257 messages on the queue, with a bottom-up merge sort, using code contributed
2258 by Michael Haardt. This should make operations like -bp somewhat faster on
2259 large queues. It won't affect queue runners, except when queue_run_in_order
2262 51. Installed eximstats 1.31 in the distribution.
2264 52. Added support for SRV lookups to the dnslookup router.
2266 53. If an ACL referred to $message_body or $message_body_end, the value was not
2267 reset for any messages that followed in the same SMTP session.
2269 54. The store-handling optimization for building very long strings was not
2270 differentiating between the different store pools. I don't think this
2271 actually made any difference in practice, but I've tidied it.
2273 55. While running the routers to verify a sender address, $sender_address
2274 was still set to the sender address. This is wrong, because when routing to
2275 send a bounce to the sender, it would be empty. Therefore, I have changed
2276 it so that, while verifying a sender address, $sender_address is set to <>.
2277 (There is no change to what happens when verifying a recipient address.)
2279 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2280 target A or AAAA records (if not already returned) without resetting the
2281 qualify_single or search_parents options of the DNS resolver. These are
2282 inappropriate in this case because the targets of MX and SRV records must
2283 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2284 target that, when qualified, matched something in the local domain. These
2285 two options are now turned off when doing these lookups.
2287 57. It seems that at least some releases of Reiserfs (which does not have the
2288 concept of a fixed number of inodes) returns zero and not -1 for the
2289 number of available inodes. This interacted badly with check_spool_inodes,
2290 which assumed that -1 was the "no such thing" setting. What I have done is
2291 to check that the total number of inodes is greater than zero before doing
2292 the test of how many are available.
2294 58. When a "warn" ACL statement has a log_message modifier, the message is
2295 remembered, and not repeated. This is to avoid a lot of repetition when a
2296 message has many recipients that cause the same warning to be written.
2297 Howewer, Exim was preserving the list of already written lines for an
2298 entire SMTP session, which doesn't seem right. The memory is now reset if a
2299 new message is started.
2301 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2302 debugging output unless log_rewrite was also set.
2304 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2305 of (int)(handle) when we know that handle contains (void *)(-1).
2307 61. The Exim daemon panic-logs an error return when it closes the incoming
2308 connection. However "connection reset by peer" seems to be common, and
2309 isn't really an error worthy of noting specially, so that particular error
2312 62. When Exim is trying to find all the local interfaces, it used to panic and
2313 die if the ioctl to get the interface flags failed. However, it seems that
2314 on at least one OS (Solaris 9) it is possible to have an interface that is
2315 included in the list of interfaces, but for which you get a failure error
2316 for this call. This happens when the interface is not "plumbed" into a
2317 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2318 failure of the "get flags" call assumes that the interface is down.
2320 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2321 makes life easier for people who are doing arithmetic on fields extracted
2322 from dates, where you often get leading zeros that should not be
2323 interpreted as octal.
2325 64. Added qualify_domain to the redirect router, to override the global
2328 65. If a pathologically long header line contained very many addresses (the
2329 report of this problem mentioned 10 000) and each of them was rewritten,
2330 Exim could use up a very large amount of memory. (It kept on making new
2331 copies of the header line as it rewrote, and never released the old ones.)
2332 At the expense of a bit more processing, the header rewriting function has
2333 been changed so that it no longer eats memory in this way.
2335 66. The generation of the Received: header has been moved from the time that a
2336 message starts to be received, to the time that it finishes. The timestamp
2337 in the Received: header should now be very close to that of the <= log
2338 line. There are two side-effects of this change:
2340 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2341 logged header lines no longer include the local Received: line, because
2342 it has not yet been created. The same applies to a copy of the message
2343 that is returned to a non-SMTP sender when a message is rejected.
2345 (b) When a filter file is tested using -bf, no additional Received: header
2346 is added to the test message. After some thought, I decided that this
2349 This change does not affect the value of $received_for. It is still set
2350 after address rewriting, but before local_scan() is called.
2352 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2354 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2355 gave an unhelpful panic error message, and a defer error. I have managed to
2356 change this behaviour so that it now rejects any supplied certificate,
2357 which seems right, as the list of acceptable certificates is empty.
2359 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2360 gave an unhelpful defer error. I have not managed to make this reject any
2361 supplied certificates, but the error message it gives is "no certificate
2362 supplied", which is not helpful.
2364 70. exigrep's output now also includes lines that are not associated with any
2365 message, but which match the given pattern. Implemented by a patch from
2366 Martin Sluka, which also tidied up the Perl a bit.
2368 71. Recipient callout verification, like sender verification, was using <> in
2369 the MAIL FROM command. This isn't really the right thing, since the actual
2370 sender may affect whether the remote host accepts the recipient or not. I
2371 have changed it to use the actual sender in the callout; this means that
2372 the cache record is now keyed on a recipient/sender pair, not just the
2373 recipient address. There doesn't seem to be a real danger of callout loops,
2374 since a callout by the remote host to check the sender would use <>.
2375 [SEE ABOVE: changed after hitting problems.]
2377 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2378 temporary errors. However, in the case of such a code being given after
2379 the end of a data transmission (i.e. after ".") Exim was failing to write
2380 a retry record for the message. (Yes, there was some broken host that was
2381 actually sending 8xx at this point.)
2383 73. An unknown lookup type in a host list could cause Exim to panic-die when
2384 the list was checked. (An example that provoked this was putting <; in the
2385 middle of a list instead of at the start.) If this happened during a DATA
2386 ACL check, a -D file could be left lying around. This kind of configuration
2387 error no longer causes Exim to die; instead it causes a defer errror. The
2388 incident is still logged to the main and panic logs.
2390 74. Buglet left over from Exim 3 conversion. The message "too many messages
2391 in one connection" was written to the rejectlog but not the mainlog, except
2392 when address rewriting (yes!) was being logged.
2394 75. Added write_rejectlog option.
2396 76. When a system filter was run not as root (that is, when system_filter_user
2397 was set), the values of the $n variables were not being returned to the
2398 main process; thus, they were not subsequently available in the $sn
2401 77. Added +return_path_on_delivery log selector.
2403 78. A connection timeout was being treated differently from recipients deferred
2404 when testing hosts_max_try with a message that was older than the host's
2405 retry timeout. (The host should not be counted, thus allowing all hosts to
2406 be tried at least once before bouncing.) This may have been the cause of an
2407 occasionally reported bug whereby a message would remain on the queue
2408 longer than the retry timeout, but would be bounced if a delivery was
2409 forced. I say "may" because I never totally pinned down the problem;
2410 setting up timeout/retry tests is difficult. See also the next item.
2412 79. The ultimate address timeout was not being applied to errors that involved
2413 a combination of host plus message (for example, a timeout on a MAIL
2414 command). When an address resolved to a number of possible hosts, and they
2415 were not all tried for each delivery (e.g. because of hosts_max_try), a
2416 message could remain on the queue longer than the retry timeout.
2418 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2421 81. Fixed an obscure SMTP outgoing bug which required at least the following
2422 conditions: (a) there was another message waiting for the same server;
2423 (b) the server returned 5xx to all RCPT commands in the first message so
2424 that the message was not completed; (c) the server dropped the connection
2425 or gave a negative response to the RSET that Exim sends to abort the
2426 transaction. The observed case was a dropped connection after DATA that had
2427 been sent in pipelining mode. That is, the server had advertised PIPELINING
2428 but was not implementing it correctly. The effect of the bug was incorrect
2429 behaviour, such as trying another host, and this could lead to a crash.
2435 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2436 and daemon.c were passed as pointers to ints; they should have been
2437 pointers to socklen_t variables (which are typically unsigned ints).
2439 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2442 3. Fixed a really odd bug that affected only the testing scheme; patching a
2443 certain fixed string in the binary changed the value of another string that
2444 happened to be identical to the end of the original first string.
2446 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2447 name", it returns that address as the IP address. On some operating
2448 systems (e.g. Solaris), it also passes back the IP address string as the
2449 "host name". However, on others (e.g. Linux), it passes back an empty
2450 string. Exim wasn't checking for this, and was changing the host name to an
2451 empty string, assuming it had been canonicized.
2453 5. Although rare, it is permitted to have more than one PTR record for a given
2454 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2455 all the names associated with an address, because they do in Solaris.
2456 However, it seems that they do not in Linux for data that comes from the
2457 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2458 I found this out when I moved to a new Linux workstation and tried to run
2459 the Exim test suite.
2461 To get round this problem I have changed the code so that it now does its
2462 own call to the DNS to look up PTR records when searching for a host name.
2463 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2464 addresses that are only in /etc/hosts are still found.
2466 This behaviour is, however, controlled by an option called host_lookup_
2467 order, which defaults to "bydns:byaddr". If people want to use the other
2468 order, or indeed, just use one or the other means of lookup, they can
2469 specify it in this variable.
2471 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2472 some operating systems, this comes back from gethostbyaddr() as an empty
2473 string, and this is what Exim used to test for. However, it seems that in
2474 other systems, "." is yielded. Exim now tests for this case too.
2476 7. The values of check_spool_space and check_log_space are now held internally
2477 as a number of kilobytes instead of an absolute number of bytes. If a
2478 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2479 kilobyte. This means that much larger values can be stored.
2481 8. Exim monitor: an attempt to get the action menu when not actually pointing
2482 at a message produces an empty menu entitled "No message selected". This
2483 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2484 no entries in it ("Shell widget menu has zero width and/or height"). So I
2485 have added a single, blank menu entry in this case.
2487 9. Added ${quote_local_part.
2489 10. MIME decoding is now applied to the contents of Subject: header lines when
2492 11. Now that a reference to $sender_host_address automatically causes a reverse
2493 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2494 host lookup before query-style lookups in lists that might use this
2495 variable. This has therefore been abolished, and the "net-" prefix is no
2496 longer necessary for query-style lookups.
2498 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2499 have been a typo for LFLAGS, so it has been changed.
2501 13. The install script calls Exim with "-C /dev/null" in order to find the
2502 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2503 to be output. Howeve, since Exim outputs its version number before the
2504 error, it didn't break the script. It just looked ugly. I fixed this by
2505 always allowing "-C /dev/null" if the caller is root.
2507 14. Ignore overlarge ACL variable number when reading spool file - insurance
2508 against a later release with more variables having written the file.
2510 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2511 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2512 this, as well as the form without the "IPv6" on the front.
2514 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2515 OS/Makefile-Darwin file.
2517 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2518 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2520 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2521 variable. However, some OS use ints for inodes. I've added cast to long int
2522 to get rid of the compiler warning.
2524 19. I had forgotten to lock out "/../" in configuration file names when
2525 ALT_CONFIG_PREFIX was set.
2527 20. Routers used for verification do not need to specify transports. However,
2528 if such a router generated a host list, and callout was configured, Exim
2529 crashed, because it could not find a port number from the (non-existent)
2530 transport. It now assumes port 25 in this circumstance.
2532 21. Added the -t option to exigrep.
2534 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2535 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2536 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2537 them is set, LOOKUP_LSEARCH is forced.
2539 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2541 24. Added sender and host information to the "rejected by local_scan()" log
2542 line; previously there was no indication of these.
2544 25. Added .include_if_exists.
2546 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2547 Linux. It turns out that not all file systems support this. Apparently some
2548 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2549 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2550 not supported on the file descriptor, is now ignored when Exim is trying to
2551 sync a directory. This applies only to Linux.
2553 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2555 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2556 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2559 29. Added support for maildirsize files from supplied patch (modified a bit).
2561 30. The use of :fail: followed by an empty string could lead Exim to respond to
2562 sender verification failures with (e.g.):
2564 550 Verification failed for <xxx>
2565 550 Sender verify failed
2567 where the first response line was missing the '-' that indicates it is not
2568 the final line of the response.
2570 31. The loop for finding the name of the user that called Exim had a hardwired
2571 limit of 10; it now uses the value of finduser_retries, which is used for
2572 all other user lookups.
2574 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2577 33. Exim was neglecting to zero errno before one call of strtol() when
2578 expanding a string and expecting an integer value. On some systems this
2579 resulted in spurious "integer overflow" errors. Also, it was casting the
2580 result into an int without checking.
2582 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2583 did not work. The code looks as if it has *never* worked, though it appears
2584 to have been documented since at least releast 1.62. I have made it work.
2586 35. The "timeout_DNS" error in retry rules, also documented since at least
2587 1.62, also never worked. As it isn't clear exactly what this means, and
2588 clearly it isn't a major issue, I have abolished the feature by treating it
2589 as "timeout", and writing a warning to the main and panic logs.
2591 36. The display of retry rules for -brt wasn't always showing the error code
2594 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2595 timeout_connect_A, timeout_connect_MX.
2597 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2598 to the empty sender.
2600 39. The daemon was not analysing the content of -oX till after it had closed
2601 stderr and disconnected from the controlling terminal. This meant that any
2602 syntax errors were only noted on the panic log, and the return code from
2603 the command was 0. By re-arranging the code a little, I've made the
2604 decoding happen first, so such errors now appear on stderr, and the return
2605 code is 1. However, the actual setting up of the sockets still happens in
2606 the disconnected process, so errors there are still only recorded on the
2609 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2610 connections (as happens on some IP stacks) was logged at start up time as
2611 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2612 it from "IPv6 and IPv4", which means that two separate sockets are being
2615 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2616 says whether AF_INET or AF_INET6 was passed as an argument.
2618 42. Exiwhat output was messed up when time zones were included in log
2621 43. Exiwhat now gives more information about the daemon's listening ports,
2622 and whether -tls-on-connect was used.
2624 44. The "port" option of the smtp transport is now expanded.
2626 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2627 silently ignored. Now an error message is written to the main and panic
2630 46. There's a new ACL modifier called "logwrite" which writes to a log file
2631 as soon as it is encountered.
2633 47. Added $local_user_uid and $local_user_gid at routing time.
2635 48. Exim crashed when trying to verify a sender address that was being
2638 49. Exim was recognizing only a space character after ".include". It now also
2639 recognizes a tab character.
2641 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2642 extracting the relevant information from the specification. The man page no
2643 longer contains scrambled data for the -d option, and I've added a section
2644 at the front about calling Exim under different names.
2646 51. Added "extra_headers" argument to the "mail" command in filter files.
2648 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2651 53. Installed eximstats 1.29.
2653 54. Added transport_filter_timeout as a generic transport option.
2655 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2656 Cc: header lines. This was required by RFC 822, but it not required by RFC
2659 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2660 incoming messages that did not have them. Now it does so only if the
2661 message originates locally, that is, if there is no associated remote host
2662 address. When Resent- header lines are present, this applies to the Resent-
2663 lines rather than the non-Resent- lines.
2665 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2666 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2668 58. Messages for configuration errors now include the name of the main
2669 configuration file - useful now that there may be more than one file in a
2670 list (.included file names were always shown).
2672 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2673 for those rare installations that do not start the daemon as root or run it
2674 setuid root. I've cut out the call to initgroups() if the daemon is not
2677 60. The Exim user and group can now be bound into the binary as text strings
2678 that are looked up at the start of Exim's processing.
2680 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2682 62. Added $mailstore_basename variable.
2684 63. Installed patch to sieve.c from Michael Haardt.
2686 64. When Exim failed to open the panic log after failing to open the main log,
2687 the original message it was trying to log was written to stderr and debug
2688 output, but if they were not available (the usual case in production), it
2689 was lost. Now it is written to syslog before the two lines that record the
2690 failures to open the logs.
2692 65. Users' Exim filters run in subprocesses under the user's uid. It is
2693 possible for a "deliver" command or an alias in a "personal" command to
2694 provoke an address rewrite. If logging of address rewriting is configured,
2695 this fails because the process is not running as root or exim. There may be
2696 a better way of dealing with this, but for the moment (because 4.30 needs
2697 to be released), I have disabled address rewrite logging when running a
2698 filter in a non-root, non-exim process.
2704 1. The buildconfig auxiliary program wasn't quoting the value set for
2705 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2706 not defined. This bug was masked in 4.22 by the effect that was fixed in
2709 2. Some messages that were rejected after a message id was allocated were
2710 shown as "incomplete" by exigrep. It no longer does this for messages that
2711 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2713 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2714 have allow_domain_literals set, the ID did not get logged in the <= line.
2715 Domain literals are now always recognized in Message-ID: header lines.
2717 4. The first argument for a ${extract expansion item is the key name or field
2718 number. Leading and trailing spaces in this item were not being ignored,
2719 causing some misleading effects.
2721 5. When deliver_drop_privilege was set, single queue runner processes started
2722 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2723 same command in the process it spins off) were not dropping privilege.
2725 6. When the daemon running as "exim" started a queue runner, it always
2726 re-executed Exim in the spun-off process. This is a waste of effort when
2727 deliver_drop_privilege is set. The new process now just calls the
2728 queue-runner function directly.
2734 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2737 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2738 ignored. Though the use of -oP was forcing the writing of a pid file, it
2739 was always written to the default place.
2741 3. If the message "no IP address found for host xxxx" is generated during
2742 incoming verification, it is now followed by identification of the incoming
2743 connection (so you can more easily find what provoked it).
2745 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2747 5. Added some features to "harden" Exim a bit more against certain attacks:
2749 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2750 be put in Local/Makefile. This is like the never_users runtime option,
2751 but it cannot be overridden. The default setting is "root".
2753 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2754 prefix string with which any file named in a -C command line option
2757 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2758 is retained for -C and -D only if the caller of Exim is root. Without
2759 it, the exim user may also use -C and -D and retain privilege.
2761 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2762 command line option is disabled.
2764 6. Macro names set by the -D option must start with an upper case letter, just
2765 like macro names defined in the configuration file.
2767 7. Added "dereference=" facility to LDAP.
2769 8. Two instances of the typo "uknown" in the source files are fixed.
2771 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2772 the Configure-Makefile script screwed up while processing it.
2774 10. Incorporated PCRE 4.4.
2776 11. The SMTP synchronization check was not operating right at the start of an
2777 SMTP session. For example, it could not catch a HELO sent before the client
2778 waited for the greeting. There is now a check for outstanding input at the
2779 point when the greeting is written. Because of the duplex, asynchronous
2780 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2781 way, but not yet received, when the check is performed.
2783 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2784 on TCP/IP sockets, because this apparently causes some broken clients to
2787 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2788 unchanged) from the Cygwin maintainer.
2790 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2791 was supported instead of testing for mbx. Effectively a typo.
2793 15. The spa authenticator server code was not checking that the input it
2794 received was valid base64.
2796 16. The debug output line for the "set" modifier in ACLs was not showing the
2797 name of the variable that was being set.
2799 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2801 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2802 lookup on its own. Something else was needed to trigger the lookup. For
2803 example, a match in host_lookup or the need for a host name in a host list.
2804 Now, if $sender_host_name is referenced and the host name has not yet been
2805 looked up, a lookup is performed. If the lookup fails, the variable remains
2806 empty, and $host_lookup_failed is set to "1".
2808 19. Added "eqi" as a case-independent comparison operator.
2810 20. The saslauthd authentication condition could segfault if neither service
2811 nor realm was specified.
2813 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2814 error message that was logged was misleading, and incoming SMTP
2815 connections were dropped. The message is now more accurate, and temporary
2816 errors are given to SMTP connections.
2818 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2819 (except RSET) with a 421 error, until QUIT is received. However, it was
2820 failing to send a response to QUIT.
2822 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2823 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2824 ACL is now run after the helo_try_verify_hosts code.
2826 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2827 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2828 case-independent, so other case variants are also recognized). Apparently
2829 some systems use these upper case variants.
2831 25. If more than two messages were waiting for the same host, and a transport
2832 filter was specified for the transport, Exim sent two messages over the
2833 same TCP/IP connection, and then failed with "socket operation on non-
2834 socket" when it tried to send the third.
2836 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2838 27. The extern definition of crypt16() in expand.c was not being excluded when
2839 the OS had its own crypt16() function.
2841 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2842 as a preferred synonym for return_size_limit, both as an option and as an
2845 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2847 30. Changed the default configuration ACL to relax the local part checking rule
2848 for addresses that are not in any local domains. For these addresses,
2849 slashes and pipe symbols are allowed within local parts, but the sequence
2850 /../ is explicitly forbidden.
2852 31. SPA server authentication was not clearing the challenge buffer before
2855 32. log_message in a "warn" ACL statement was writing to the reject log as
2856 well as to the main log, which contradicts the documentation and doesn't
2857 seem right (because no rejection is happening). So I have stopped it.
2859 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2860 However, I am unable to do any testing of this.
2862 34. Fixed an infelicity in the appendfile transport. When checking directories
2863 for a mailbox, to see if any needed to be created, it was accidentally
2864 using path names with one or more superfluous leading slashes; tracing
2865 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2867 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2868 contents are added to the log line that is written for every discarded
2869 recipient. (Previously a log_message setting was ignored.)
2871 36. The ${quote: operator now quotes the string if it is empty.
2873 37. The install script runs exim in order to find its version number. If for
2874 some reason other than non-existence or emptiness, which it checks, it
2875 could not run './exim', it was installing it with an empty version number,
2876 i.e. as "exim-". This error state is now caught, and the installation is
2879 38. An argument was missing from the function that creates an error message
2880 when Exim fails to connect to the socket for saslauthd authentication.
2881 This could cause Exim to crash, or give a corrupted message.
2883 39. Added isip, isip4, and isip6 to ${if conditions.
2885 40. The ACL variables $acl_xx are now saved with the message, and can be
2886 accessed later in routers, transports, and filters.
2888 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2889 strings in the file are not string-expanded.
2891 42. If a MAIL command specified a SIZE value that was too large to fit into an
2892 int variable, the check against message_size_limit failed. Such values are
2893 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2894 day this will have to be increased, but I don't think I want to be around
2895 when emails are that large.
2902 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2903 iconv() is not standard in FreeBSD.
2905 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2906 IPv6 enabled. The observed symptom was a segmentation fault on return from
2907 the function os_common_find_running_interfaces() in src/os.c.
2909 3. In the check_special_case() function in daemon.c I had used "errno" as an
2910 argument name, which causes warnings on some systems. This was basically a
2911 typo, since it was named "eno" in the comments!
2913 4. The code that waits for the clock to tick (at a resolution of some fraction
2914 of a second) so as to ensure message-id uniqueness was always waiting for
2915 at least one whole tick, when it could have waited for less. [This is
2916 almost certainly not relevant at current processor speeds, where it is
2917 unlikely to ever wait at all. But we try to future-proof.]
2919 5. The function that sleeps for a time interval that includes fractions of a
2920 second contained a race. It did not block SIGALRM between setting the
2921 timer, and suspending (a couple of lines later). If the interval was short
2922 and the sigsuspend() was delayed until after it had expired, the suspension
2923 never ended. On busy systems this could lead to processes getting stuck for
2926 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2927 process, before it forks any delivery processes. The open lookup caching
2928 mechanism meant that the open file or database connection was passed into
2929 the delivery process. The problem was that delivery processes always tidy
2930 up cached lookup data. This could cause a problem for the next delivery
2931 process started by the queue runner, because the external queue runner
2932 process does not know about the closure. So the next delivery process
2933 still has data in the lookup cache. In the case of a file lookup, there was
2934 no problem because closing a file descriptor in a subprocess doesn't affect
2935 the parent. However, if the lookup was caching a connection to a database,
2936 the connection was closed, and the second delivery process was likely to
2937 see errors such as "PGSQL: query failed: server closed the connection
2938 unexpectedly". The problem has been fixed by closing all cached lookups
2939 in a queue runner before running a delivery process.
2941 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2942 seem to have the "const" qualifier which it has on other OS. I've
2945 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2946 *of the same type* (client or server) with the same public name that an
2947 error should be diagnosed.
2949 9. When Exim looked up a host name for an IP address, but failed to find the
2950 original IP address when looking up the host name (a safety check), it
2951 output the message "<ip address> does not match any IP for NULL", which was
2952 confusing, to say the least. The bug was that the host name should have
2953 appeared instead of "NULL".
2955 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2956 user that is built into the binary, and the -C or -D options is used, root
2957 privilege is dropped before the configuration file is read. In addition,
2958 logging is switched to stderr instead of the normal log files. If the
2959 configuration then re-defines the Exim user, the unprivileged environment
2960 is probably not what is expected, so Exim logs a panic warning message (but
2963 However, if deliver_drop_privilege is set, the unprivileged state may well
2964 be exactly what is intended, so the warning has been cut out in that case,
2965 and Exim is allowed to try to write to its normal log files.
2971 1. smtp_return_error_details was not giving details for temporary sender
2972 or receiver verification errors.
2974 2. Diagnose a configuration error if two authenticators have the same public
2977 3. Exim used not to create the message log file for a message until the first
2978 delivery attempt. This could be confusing when incoming messages were held
2979 for policy or load reasons. The message log file is now created at the time
2980 the message is received, and an initial "Received" line is written to it.
2982 4. The automatically generated man page for command line options had a minor
2983 bug that caused no ill effects; however, a more serious problem was that
2984 the procedure for building the man page automatically didn't always
2985 operate. Consequently, release 4.20 contains an out-of-date version. This
2986 shouldn't happen again.
2988 5. When building Exim with embedded Perl support, the script that builds the
2989 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2990 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2992 6. The freeze_tell option was not being used for messages that were frozen on
2993 arrival, either by an ACL or by local_scan().
2995 7. Added the smtp_incomplete_transaction log selector.
2997 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2998 was accepting AUTH without a new EHLO.
3000 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3001 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3004 10. Logging of TCP/IP connections (when configured) now happens in the main
3005 daemon process instead of the child process, so that the TCP/IP connection
3006 count is more accurate (but it can never be perfect).
3008 11. The use of "drop" in a nested ACL was not being handled correctly in the
3009 outer ACL. Now, if condition failure induced by the nested "drop" causes
3010 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3011 or "require"), the connection is dropped.
3013 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3014 that yield "discard" can now be used with an "accept" or a "discard" verb,
3015 but an error is generated for any others (because I can't see a useful way
3016 to define what should happen).
3018 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3019 are now processed in the same way as lines in the Exim configuration file.
3020 In particular, continuation lines are supported.
3022 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3024 15. Added -ti meaning -t -i.
3026 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3027 domains, and warn by logging if others are found.
3029 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3030 returned from GIFCONF when Exim is trying to find the list of interfaces on
3031 a host. The code in os.c has been modified to copy each ifreq to an aligned
3032 structure in all cases.
3034 Also, in some cases, the returned ifreq's were being copied to a 'struct
3035 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3036 means the last couple of bytes of an IPv6 address could be chopped if the
3037 ifreq contained only a normal sockaddr (14 bytes storage).
3039 18. Named domain lists were not supported in the hosts_treat_as_local option.
3040 An entry such as +xxxx was not recognized, and was treated as a literal
3043 19. Ensure that header lines added by a DATA ACL are included in the reject log
3044 if the ACL subsequently rejects the message.
3046 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3047 MD5 (which is deprecated).
3049 21. When testing a filter file using -bf, Exim was writing a message when it
3050 took the sender from a "From " line in the message, but it was not doing so
3051 when it took $return_path from a Return-Path: header line. It now does.
3053 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3054 with a valid header line field name (a series of printing characters
3055 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3057 23. Changed "disc" in the source to "disk" to conform to the documentation and
3058 the book and for uniformity.
3060 24. Ignore Sendmail's -Ooption=value command line item.
3062 25. When execve() failed while trying to run a command in a pipe transport,
3063 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3064 could be confused with a return value of 69 from the command itself. This
3065 has been changed to 127, the value the shell returns if it is asked to run
3066 a non-existent command. The wording for the related log line suggests a
3067 non-existent command as the problem.
3069 26. If received_header_text expands to an empty string, do not add a Received:
3070 header line to the message. (Well, it adds a token one on the spool, but
3071 marks it "old" so that it doesn't get used or transmitted.)
3073 27. Installed eximstats 1.28 (addition of -nt option).
3075 28. There was no check for failure on the call to getsockname() in the daemon
3076 code. This can fail if there is a shortage of resources on the system, with
3077 ENOMEM, for example. A temporary error is now given on failure.
3079 29. Contrary to the C standard, it seems that in some environments, the
3080 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3081 program. Exim now does this explicitly; it affects the formatting of
3082 timestamps using strftime().
3084 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3085 complaints. I've now initialized all the variables, to avoid this.
3087 32. Header lines added by a system filter were not being "seen" during
3088 transport-time rewrites.
3090 33. The info_callback() function passed to OpenSSL is set up with type void
3091 (*)(SSL *, int, int), as described somewhere. However, when calling the
3092 function (actually a macro) that sets it up, the type void(*)() is
3093 expected. I've put in a cast to prevent warnings from picky compilers.
3095 34. If a DNS black list lookup found a CNAME record, but there were no A
3096 records associated with the domain it pointed at, Exim crashed.
3098 35. If a DNS black list lookup returned more than one A record, Exim ignored
3099 all but the first. It now scans all returned addresses if a particular IP
3100 value is being sought. In this situation, the contents of the
3101 $dnslist_value variable are a list of all the addresses, separated by a
3104 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3105 to accept a host name and all its aliases if the forward lookup for any of
3106 them yielded the IP address of the incoming connection. Now it accepts only
3107 those names whose forward lookup yields the correct IP address. Any other
3108 names are discarded. This closes a loophole whereby a rogue DNS
3109 administrator could create reverse DNS records to break through a
3110 wildcarded host restriction in an ACL.
3112 37. If a user filter or a system filter that ran in a subprocess used any of
3113 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3114 the wrong values were passed to the pipe command ($thisaddress had the
3115 value of $0, $0 had the value of $1, etc). This bug was introduced by
3116 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3118 38. Improved the line breaking for long SMTP error messages from ACLs.
3119 Previously, if there was no break point between 40 and 75 characters, Exim
3120 left the rest of the message alone. Two changes have been made: (a) I've
3121 reduced the minimum length to 35 characters; (b) if it can't find a break
3122 point between 35 and 75 characters, it looks ahead and uses the first one
3123 that it finds. This may give the occasional overlong line, but at least the
3124 remaining text gets split now.
3126 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3127 file descriptors might be low, and that setting 1000 would always raise it.
3128 It turns out that in some environments, the limit is already over 1000 and
3129 that lowering it causes trouble. So now Exim takes care not to decrease it.
3131 40. When delivering a message, the value of $return_path is set to $sender_
3132 address at the start of routing (routers may change the value). By an
3133 oversight, this default was not being set up when an address was tested by
3134 -bt or -bv, which affected the outcome if any router or filter referred to
3137 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3138 log only when "message" or "log_message" are set. However, if one of the
3139 conditions was an address verification, or a call to a nested ACL, the
3140 messages generated by the underlying test were being passed through. This
3141 no longer happens. The underlying message is available in $acl_verify_
3142 message for both "message" and "log_message" expansions, so it can be
3143 passed through if needed.
3145 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3146 new expansion $bh_ to give the encoded byte string without charset
3147 translation. Translation happens only if iconv() is available; HAVE_ICONV
3148 indicates this at build time. HEADERS_CHARSET gives the charset to
3149 translate to; headers_charset can change it in the configuration, and
3150 "headers charset" can change it in an individual filter file.
3152 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3153 that creates RFC 2047 encoded "words" labels them as that charset instead
3154 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3155 expansion operator; (ii) when Exim creates a From: line for a local
3156 message; (iii) when a header line is rewritten to include a "phrase" part.
3158 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3159 buggy, causing it to skip the first lines of messages whose message ID
3160 ended in 'D'. This would not have bitten before Exim release 4.14, because
3161 message IDs were unlikely to end in 'D' before then. The effect was to have
3162 incorrect size information for certain domains.
3164 45. #include "config.h" was missing at the start of the crypt16.c module. This
3165 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3168 46. If there was a timeout during a "random" callout check, Exim treated it as
3169 a failure of the random address, and carried on sending RSET and the real
3170 address. If the delay was just some slowness somewhere, the response to the
3171 original RCPT would be taken as a response to RSET and so on, causing
3172 mayhem of various kinds.
3174 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3175 when I implemented it. It didn't allow for the fact that some option values
3176 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3177 the right test for positive values.
3179 48. Domain names in DNS records are case-independent. Exim always looks them up
3180 in lower case. Some resolvers return domain names in exactly the case they
3181 appear in the zone file, that is, they may contain uppercase letters. Not
3182 all resolvers do this - some return always lower case. Exim was treating a
3183 change of case by a resolver as a change of domain, similar to a widening
3184 of a domain abbreviation. This triggered its re-routing code and so it was
3185 trying to route what was effectively the same domain again. This normally
3186 caused routing to fail (because the router wouldn't handle the domain
3187 twice). Now Exim checks for this case specially, and just changes the
3188 casing of the domain that it ultimately uses when it transmits the message
3191 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3194 50. If a filter generated a file delivery with a non-absolute name (possible if
3195 no home directory exists for the router), the forbid_file option was not
3198 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3199 the existing equality matching.
3201 52. Exim was using ints instead of ino_t variables in some places where it was
3202 dealing with inode numbers.
3204 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3205 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3206 called TMPDIR, and if it finds it is different, it changes its value.
3208 54. The smtp_printf() function is now made available to local_scan() so
3209 additional output lines can be written before returning. There is also an
3210 smtp_fflush() function to enable the detection of a dropped connection.
3211 The variables smtp_input and smtp_batched_input are exported to
3214 55. Changed the default runtime configuration: the message "Unknown user"
3215 has been removed from the ACL, and instead placed on the localuser router,
3216 using the cannot_route_message feature. This means that any verification
3217 failures that generate their own messages won't get overridden. Similarly,
3218 the "Unrouteable address" message that was in the ACL for unverifiable
3219 relay addresses has also been removed.
3221 56. Added hosts_avoid_esmtp to the smtp transport.
3223 57. The exicyclog script was not checking for the esoteric option
3224 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3225 will work only if exicyclog is run under the appropriate euid.
3227 58. Following a discussion on the list, the rules by which Exim recognises line
3228 endings on incoming messages have been changed. The -dropcr and drop_cr
3229 options are now no-ops, retained only for backwards compatibility. The
3230 following line terminators are recognized: LF CRLF CR. However, special
3231 processing applies to CR:
3233 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3234 nor a local message in the state where . is a terminator.
3236 (ii) If a bare CR is encountered in a header line, an extra space is added
3237 after the line terminator so as not to end the header. The reasoning
3238 behind this is that bare CRs in header lines are most likely either
3239 to be mistakes, or people trying to play silly games.
3241 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3242 was being incorrectly given as 18 bytes larger than it should have been.
3243 This is a VOB (very old bug).
3245 60. This may never have affected anything current, but just in case it has:
3246 When the local host is found other than at the start of a list of hosts,
3247 the local host, those with the same MX, and any that follow, are discarded.
3248 When the list in question was part of a longer list of hosts, the following
3249 hosts (not currently being processed) were also being discarded. This no
3250 longer happens. I'm not sure if this situation could ever has previously
3253 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3256 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3257 header. This is something that is recommended in a new Internet Draft, and
3258 is something that is documented as being done by Sendmail. There are two
3259 possible values. For messages generated by the autoreply transport, Exim
3262 Auto-Submitted: auto-replied
3264 whereas for all other generated messages (e.g. bounces) it adds
3266 Auto-Submitted: auto-generated
3268 63. The "personal" condition in filters now includes a test for the
3269 Auto-Submitted: header. If it contains the string "auto-" the message it
3270 not considered personal.
3272 64. Added rcpt_include_affixes as a generic transport option.
3274 65. Added queue_only_override (default true).
3276 66. Added the syslog_duplication option.
3278 67. If what should have been the first header line of a message consisted of
3279 a space followed by a colon, Exim was mis-interpreting it as a header line.
3280 It isn't of course - it is syntactically invalid and should therefore be
3281 treated as the start of the message body. The misbehaviour could have
3282 caused a number of strange effects, including loss of data in subsequent
3283 header lines, and spool format errors.
3285 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3286 client host had authenticated. This control can now be exercised by an ACL
3287 for more flexibility.
3289 69. By default, callouts do not happen when testing with -bh. There is now a
3290 variant, -bhc, which does actually run the callout code, including
3291 consulting and updating the callout cache.
3293 70. Added support for saslauthd authentication, courtesy of Alexander
3296 71. If statvfs() failed on the spool or log directories while checking their
3297 size for availability, Exim confusingly gave the error "space shortage".
3298 Furthermore, in debugging mode it crashed with a floating point exception.
3299 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3300 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3301 serious problem, Exim now writes to the main and panic logs when this
3302 happens, with details of the failure. It then refuses to accept the
3303 incoming message, giving the message "spool directory problem" or "log
3304 directory problem" with a 421 code for SMTP messages.
3306 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3307 0, 1, and 2 exist, because some OS complain for execs without them (see
3308 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3309 descriptors. However, the code omitted to check that the open succeeded,
3310 causing mysterious errors if for some reason the permissions on /dev/null
3311 got screwed. Now Exim writes a message to the main and panic logs, and
3312 bombs out if it can't open /dev/null.
3314 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3315 interact so that it is all more flexible. It is supposed to remain
3316 backwards compatible. Also added extra_local_interfaces.
3318 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3319 to bomb out with an assertion failure - to the client this appears as a
3320 connection drop. This problem occurs in the part of the code that was taken
3321 from the Samba project. Fortunately, the assertion is in a very simple
3322 function, so I have fixed this by reproducing the function inline in the
3323 one place where it is called, and arranging for authentication to fail
3324 instead of killing the process with assert().
3326 75. The SPA client code was not working when the server requested OEM rather
3327 than Unicode encoding.
3329 76. Added code to make require_files with a specific uid setting more usable in
3330 the case where statting the file as root fails - usually a non-root-mounted
3331 NFS file system. When this happens and the failure is EACCES, Exim now
3332 forks a subprocess and does the per-uid checking as the relevant uid.
3334 77. Added process_log_path.
3336 78. If log_file_path was not explicitly set, a setting of check_log_space or
3337 check_log_inodes was ignored.
3339 79. If a space check for the spool or log partitions fails, the incident is now
3340 logged. Of course, in the latter case the data may get lost...
3342 80. Added the %p formatting code to string_format() so that it can be used to
3343 print addresses in debug_print(). Adjusted all the address printing in the
3344 debugging in store.c to use %p rather than %d.
3346 81. There was a concern that a line of code in smtp_in.c could overflow a
3347 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3348 initially expressed, the concern was not well-founded, because trailing
3349 spaces are removed early. However, if the trailing spaces were followed by
3350 a NULL, they did not get removed, so the overflow was possible. Two fixes
3353 (a) I re-wrote the offending code in a cleaner fashion.
3354 (b) If an incoming SMTP command contains a NULL character, it is rejected
3357 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3358 runs initgroups(), so that if the Exim user is in any additional groups,
3359 they will be used during message reception.
3365 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob