1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.115 2005/04/06 14:09:17 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
10 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
11 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
13 TK/02 Fix ACL "control" statement not being available in MIME ACL.
15 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
17 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
18 to test Sieve filters that use "vacation".
20 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
21 that changes the way the GnuTLS parameters are stored in the cache file.
22 The new format can be generated externally. For backward compatibility,
23 if the data in the cache doesn't make sense, Exim assumes it has read an
24 old-format file, and it generates new data and writes a new file. This
25 means that you can't go back to an older release without removing the
28 PH/03 A redirect router that has both "unseen" and "one_time" set does not
29 work if there are any delivery delays because "one_time" forces the
30 parent to be marked "delivered", so its unseen clone is never tried
31 again. For this reason, Exim now forbids the simultaneous setting of
34 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
35 redirected to themselves ("homonym" addresses). Read the long ChangeLog
36 entry if you want to know the details. The fix, however, neglected to
37 consider the case when local delivery batching is involved. The test for
38 "previously delivered" was not happening when checking to see if an
39 address could be batched with a previous (undelivered) one; under
40 certain circumstances this could lead to multiple deliveries to the same
43 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
44 in its include files, and this causes problems building Exim.
46 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
47 header_syntax) but Exim was just ignoring anything given after a slash.
48 In particular, this caused confusion with an attempt to use "verify =
49 reverse_host_lookup/defer_ok". An error is now given when options are
50 supplied for verify items that do not have them. (Maybe reverse_host_
51 lookup should have a defer_ok option, but that's a different point.)
53 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
54 defined by RFC 821) to 2048, because there were problems with some AUTH
55 commands, and RFC 1869 says the size should be increased for extended
56 SMTP commands that take arguments.
58 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
61 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
62 "unknown" error; now it says that the functionality isn't in the binary.
64 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
65 an address' error message when a string expansion fails (syntax or
66 whatever). Otherwise not only does the password appear in the log, it may
67 also be put in a bounce message.
69 PH/11 Installed exipick version 20050225.0 from John Jetmore.
71 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
72 of its addresses was ever tried. (Bugzilla bug #2.)
74 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
75 the result incorrectly in the debug output. (It correctly added a newline
76 to what was transported.)
78 TF/01 Added $received_time.
80 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
81 commented out examples of how to interface to a virus scanner and to
82 SpamAssassin. Also added commented examples of av_scanner and
83 spamd_address settings.
85 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
86 and controls are allowed in which ACLs. There were a couple of minor
87 errors. Some of the entries in the conditions table (which is a table of
88 where they are NOT allowed) were getting very unwieldy; rewrote them as a
89 negation of where the condition IS allowed.
91 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
93 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
94 header file does not have a version number, so I've had to invent a new
95 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
96 API. The code is untested by me (my Linux distribution still has 0.3.2 of
97 radiusclient), but it was contributed by a Radius user.
99 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
100 files or directories, for OpenSSL.
102 PH/19 When an Exim process that is running as root has to create an Exim log
103 file, it does so in a subprocess that runs as exim:exim so as to get the
104 ownership right at creation (otherwise, other Exim processes might see
105 the file with the wrong ownership). There was no test for failure of this
106 fork() call, which would lead to the process getting stuck as it waited
107 for a non-existent subprocess. Forks do occasionally fail when resources
108 run out. I reviewed all the other calls to fork(); they all seem to check
111 PH/20 When checking for unexpected SMTP input at connect time (before writing
112 the banner), Exim was not dealing correctly with a non-positive return
113 from the read() function. If the client had disconnected by this time,
114 the result was a log entry for a synchronization error with an empty
115 string after "input=" when read() returned zero. If read() returned -1
116 (an event I could not check), uninitialized data bytes were printed.
117 There were reports of junk text (parts of files, etc) appearing after
120 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
122 PH/22 Added support for macro redefinition, and (re)definition in between
123 driver and ACL definitions.
125 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
126 forgetting to use the resulting value; it was using the unexpanded value.
128 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
129 hadn't been configured. The fix is from Juergen Kreileder, who
130 understands it better than I do:
132 "Here's what I see happening with three configured cyrus_sasl
133 authenticators configured (plain, login, cram-md5):
135 On startup auth_cyrus_sasl_init() gets called for each of these.
136 This means three calls to sasl_listmech() without a specified mech_list.
137 => SASL tests which mechs of all available mechs actually work
138 => three warnings about OTP not working
139 => the returned list contains: plain, login, cram-md5, digest-md5, ...
141 With the patch, sasl_listmech() also gets called three times. But now
142 SASL's mech_list option is set to the server_mech specified in the the
143 authenticator. Or in other words, the answer from sasl_listmech()
144 gets limited to just the mech you're testing for (which is different
146 => the return list contains just 'plain' or 'login', 'cram-md5' or
147 nothing depending on the value of ob->server_mech.
149 I've just tested the patch: Authentication still works fine,
150 unavailable mechs specified in the exim configuration are still
151 caught, and the auth.log warnings about OTP are gone."
153 PH/25 When debugging is enabled, the contents of the command line are added
154 to the debugging output, even when log_selector=+arguments is not
157 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
158 answer is "GNU", and only if the return is "GNU/something" is the answer
161 PH/27 $acl_verify_message is now set immediately after the failure of a
162 verification in an ACL, and so is available in subsequent modifiers. In
163 particular, the message can be preserved by coding like this:
165 warn !verify = sender
166 set acl_m0 = $acl_verify_message
168 Previously, $acl_verify_message was set only while expanding "message"
169 and "log_message" when a very denied access.
171 PH/28 Modified OS/os.c-Linux with
173 -#ifndef OS_LOAD_AVERAGE
174 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
176 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
177 nomenclature these days.)
181 A note about Exim versions 4.44 and 4.50
182 ----------------------------------------
184 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
185 changes of various kinds. As a consequence, a big documentation update was
186 needed. This delayed the release for rather longer than seemed good, especially
187 in the light of a couple of (minor) security issues. Therefore, the changes
188 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
189 release. So 4.44 and 4.50 are in effect two different branches that both start
192 I have left the 4.50 change log unchanged; it contains all the changes since
193 4.43. The change log for 4.44 is below; many of its items are identical to
194 those for 4.50. This seems to be the most sensible way to preserve the
195 historical information.
201 1. Minor wording change to the doc/README.SIEVE file.
203 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
204 computation of the current number of files was incorrect.
206 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
207 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
210 4. Give more explanation in the error message when the command for a transport
211 filter fails to execute.
213 5. There are several places where Exim runs a non-Exim command in a
214 subprocess. The SIGUSR1 signal should be disabled for these processes. This
215 was being done only for the command run by the queryprogram router. It is
216 now done for all such subprocesses. The other cases are: ${run, transport
217 filters, and the commands run by the lmtp and pipe transports.
219 6. Added CONFIGURE_GROUP build-time option.
221 7. Some older OS have a limit of 256 on the maximum number of file
222 descriptors. Exim was using setrlimit() to set 1000 as a large value
223 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
224 systems. I've change it so that if it can't get 1000, it tries for 256.
226 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
227 was an oversight, and furthermore, ever since the addition of extra
228 controls (e.g. 4.43/32), the checks on when to allow different forms of
229 "control" were broken. There should now be diagnostics for all cases when a
230 control that does not make sense is encountered.
232 9. Added the /retain_sender option to "control=submission".
234 10. $recipients is now available in the predata ACL (oversight).
236 11. Tidy the search cache before the fork to do a delivery from a message
237 received from the command line. Otherwise the child will trigger a lookup
238 failure and thereby defer the delivery if it tries to use (for example) a
239 cached ldap connection that the parent has called unbind on.
241 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
242 of $address_data from the recipient verification was clobbered by the
245 13. The value of address_data from a sender verification is now available in
246 $sender_address_data in subsequent conditions in the ACL statement.
248 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
250 15. Added a new option "connect=<time>" to callout options, to set a different
253 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
254 was its contents. (It was OK if the option was not defined at all.)
256 17. A "Completed" log line is now written for messages that are removed from
257 the spool by the -Mrm option.
259 18. New variables $sender_verify_failure and $recipient_verify_failure contain
260 information about exactly what failed.
262 19. Added -dd to debug only the daemon process.
264 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
265 handles timeouts, both on the server side and network timeouts. Renamed the
266 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
268 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
269 It is now set to "smtps".
271 22. $host_address is now set to the target address during the checking of
274 23. When checking ignore_target_hosts for an ipliteral router, no host name was
275 being passed; this would have caused $sender_host_name to have been used if
276 matching the list had actually called for a host name (not very likely,
277 since this list is usually IP addresses). A host name is now passed as
280 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
281 code that specifies a non-restarting handler (typically sigaction() in
282 modern systems) in an attempt to fix a rare and obscure crash bug.
284 25. Narrowed the window for a race in the daemon that could cause it to ignore
285 SIGCHLD signals. This is not a major problem, because they are used only to
286 wake it up if nothing else does.
288 26. A malformed maildirsize file could cause Exim to calculate negative values
289 for the mailbox size or file count. Odd effects could occur as a result.
290 The maildirsize information is now recalculated if the size or filecount
293 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
294 support for a long time. Removed HAVE_SYS_VFS_H.
296 28. Installed the latest version of exipick from John Jetmore.
298 29. In an address list, if the pattern was not a regular expression, an empty
299 subject address (from a bounce message) matched only if the pattern was an
300 empty string. Non-empty patterns were not even tested. This was the wrong
301 because it is perfectly reasonable to use an empty address as part of a
302 database query. An empty address is now tested by patterns that are
303 lookups. However, all the other forms of pattern expect the subject to
304 contain a local part and a domain, and therefore, for them, an empty
305 address still always fails if the pattern is not itself empty.
307 30. Exim went into a mad DNS loop when attempting to do a callout where the
308 host was specified on an smtp transport, and looking it up yielded more
311 31. Re-factored the code for checking spool and log partition space into a
312 function that finds that data and another that does the check. The former
313 is then used to implement four new variables: $spool_space, $log_space,
314 $spool_inodes, and $log_inodes.
316 32. The RFC2047 encoding function was originally intended for short strings
317 such as real names; it was not keeping to the 75-character limit for
318 encoded words that the RFC imposes. It now respects the limit, and
319 generates multiple encoded words if necessary. To be on the safe side, I
320 have increased the buffer size for the ${rfc2047: expansion operator from
323 33. It is now permitted to omit both strings after an "if" condition; if the
324 condition is true, the result is "true". As before, when the second string
325 is omitted, a false condition yields an empty string. This makes it less
326 cumbersome to write custom ACL and router conditions.
328 34. Failure to deliver a bounce message always caused it to be frozen, even if
329 there was an errors_to setting on the router. The errors_to setting is now
332 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
333 canonical form (fully expanded) before being placed in
334 $sender_host_address.
336 36. The table in the code that translates DNS record types into text (T_A to
337 "A" for instance) was missing entries for NS and CNAME. It is just possible
338 that this could have caused confusion if both these types were looked up
339 for the same domain, because the text type is used as part of Exim's
340 per-process caching. But the chance of anyone hitting this buglet seems
343 37. The dnsdb lookup has been extended in a number of ways.
345 (1) There is a new type, "zns", which walks up the domain tree until it
346 finds some nameserver records. It should be used with care.
348 (2) There is a new type, "mxh", which is like "mx" except that it returns
349 just the host names, not the priorities.
351 (3) It is now possible to give a list of domains (or IP addresses) to be
352 looked up. The behaviour when one of the lookups defers can be
353 controlled by a keyword.
355 (4) It is now possible to specify the separator character for use when
356 multiple records are returned.
358 38. The dnslists ACL condition has been extended: it is now possible to supply
359 a list of IP addresses and/or domains to be looked up in a particular DNS
362 39. Added log_selector=+queue_time_overall.
364 40. When running the queue in the test harness, wait just a tad after forking a
365 delivery process, to get repeatability of debugging output.
367 41. Include certificate and key file names in error message when GnuTLS fails
368 to set them up, because the GnuTLS error message doesn't include the name
369 of the failing file when there is a problem reading it.
371 42. Allow both -bf and -bF in the same test run.
373 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
375 44. The "Exiscan patch" is now merged into the mainline Exim source.
377 45. Sometimes the final signoff response after QUIT could fail to get
378 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
379 < 0 before doing a fflush(). This bug looks as though it goes back to the
380 introduction of TLS in release 3.20, but "sometimes" must have been rare
381 because the tests only now provoked it.
383 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
384 (this can affect the format of dates).
386 47. exim_tidydb, when checking for the continued existence of a message for
387 which it has found a message-specific retry record, was not finding
388 messages that were in split spool directories. Consequently, it was
389 deleting retry records that should have stayed in existence.
391 48. Steve fixed some bugs in eximstats.
393 49. The SPA authentication driver was not abandoning authentication and moving
394 on to the next authenticator when an expansion was forced to fail,
395 contradicting the general specification for all authenticators. Instead it
396 was generating a temporary error. It now behaves as specified.
398 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
399 (the order specifies the preference for clients). The order is now AES256,
400 AES128, 3DES, ARCFOUR128.
402 51. Small patch to Sieve code - explicitly set From: when generating an
405 52. Exim crashed if a remote delivery caused a very long error message to be
406 recorded - for instance if somebody sent an entire SpamAssassin report back
407 as a large number of 550 error lines. This bug was coincidentally fixed by
408 increasing the size of one of Exim's internal buffers (big_buffer) that
409 happened as part of the Exiscan merge. However, to be on the safe side, I
410 have made the code more robust (and fixed the comments that describe what
413 53. Now that there can be additional text after "Completed" in log lines (if
414 the queue_time_overall log selector is set), a one-byte patch to exigrep
415 was needed to allow it to recognize "Completed" as not the last thing in
418 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
419 patch that reportedly fixes this has been added. I am not expert enough to
420 create a test for it. This is what the patch creator wrote:
422 "I found a little strange behaviour of ldap code when working with
423 Windows 2003 AD Domain, where users was placed in more than one
424 Organization Units. When I tried to give exim partial DN, the exit code
425 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
426 But simultaneously result of request was absolutely normal ldap result,
427 so I produce this patch..."
429 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
430 REFERENCE, so I have modified the code to exclude the patch when that macro
433 55. Some experimental protocols are using DNS PTR records for new purposes. The
434 keys for these records are domain names, not reversed IP addresses. The
435 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
436 leaves it alone. Component reversal etc. now happens only for IP addresses.
439 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
441 57. Double the size of the debug message buffer (to 2048) so that more of very
442 long debug lines gets shown.
444 58. The exicyclog utility now does better if the number of log files to keep
445 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
447 59. Two changes related to the smtp_active_hostname option:
449 (1) $smtp_active_hostname is now available as a variable.
450 (2) The default for smtp_banner uses $smtp_active_hostname instead
451 of $primary_hostname.
453 60. The host_aton() function is supposed to be passed a string that is known
454 to be a valid IP address. However, in the case of IPv6 addresses, it was
455 not checking this. This is a hostage to fortune. Exim now panics and dies
456 if the condition is not met. A case was found where this could be provoked
457 from a dnsdb PTR lookup with an IPv6 address that had more than 8
458 components; fortuitously, this particular loophole had already been fixed
459 by change 4.50/55 above.
461 If there are any other similar loopholes, the new check in host_aton()
462 itself should stop them being exploited. The report I received stated that
463 data on the command line could provoke the exploit when Exim was running as
464 exim, but did not say which command line option was involved. All I could
465 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
469 61. There was a buffer overflow vulnerability in the SPA authentication code
470 (which came originally from the Samba project). I have added a test to the
471 spa_base64_to_bits() function which I hope fixes it.
474 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
475 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
477 63. The daemon start-up calls getloadavg() while still root for those OS that
478 need the first call to be done as root, but it missed one case: when
479 deliver_queue_load_max is set with deliver_drop_privilege. This is
480 necessary for the benefit of the queue runner, because there is no re-exec
481 when deliver_drop_privilege is set.
483 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
486 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
487 was in use, was not putting the data itself into the right store pool;
488 consequently, it could be overwritten for a subsequent message in the same
489 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
492 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
494 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
495 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
496 the function were treating the return as a boolean value, which happened to
497 work because 0=false and not-0=true, but is not correct code.
499 68. The host_aton() function was not handling scoped IPv6 addresses (those
500 with, for example, "%eth0" on the end) correctly.
502 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
503 negated items (that is, ~something) in unsigned ints. Some compilers
504 apparently mutter when there is no cast.
506 70. If an address verification called from an ACL failed, and did not produce a
507 user-specific message (i.e. there was only a "system" message), nothing was
508 put in $acl_verify_message. In this situation, it now puts the system
511 71. Change 4.23/11 added synchronization checking at the start of an SMTP
512 session; change 4.31/43 added the unwanted input to the log line - except
513 that it did not do this in the start of session case. It now does.
515 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
516 This is wrong and can cause the other end to generate a synchronization
517 error if it is another Exim or anything else that does the synchronization
518 check. A QUIT command is no longer sent after a timeout.
520 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
523 74. The defer_ok option of callout verification was not working if it was used
524 when verifying addresses in header lines, that is, for this case:
526 verify = header_sender/callout=defer_ok
528 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
529 those file descriptors could be used for SMTP connections. If anything
530 wrote to stderr (the example that came up was "warn" in embedded Perl), it
531 could be sent to the SMTP client, causing chaos. The daemon now opens
532 stdin, stdout, and stderr to /dev/null when it puts itself into the
535 76. Arrange for output from Perl's "warn" command to be written to Exim's main
536 log by default. The user can override this with suitable Perl magic.
538 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
539 the log message when discard triggers, was not working for the DATA ACL or
540 for the non-SMTP ACL.
542 78. Error message wording change in sieve.c.
544 79. If smtp_accept_max_per_host was set, the number of connections could be
545 restricted to fewer than expected, because the daemon was trying to set up
546 a new connection before checking whether the processes handling previous
547 connections had finished. The check for completed processes is now done
548 earlier. On busy systems, this bug wouldn't be noticed because something
549 else would have woken the daemon, and it would have reaped the completed
552 80. If a message was submitted locally by a user whose login name contained one
553 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
554 It caused a spool format error. I have fixed the spool reading code. A
555 related problem was that the "from" clause in the Received: line became
556 illegal because of the space(s). It is now covered by ${quote_local_part.
558 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
561 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
563 83. Patch from Sieve maintainer to fix unterminated string problem in
566 84. Some minor changes to the Linux configuration files to help with other
567 OS variants using glibc.
569 85. One more patch for Sieve to update vacation handling to latest spec.
572 ----------------------------------------------------
573 See the note above about the 4.44 and 4.50 releases.
574 ----------------------------------------------------
580 1. Change 4.43/35 introduced a bug that caused file counts to be
581 incorrectly computed when quota_filecount was set in an appendfile
584 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
585 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
588 3. Give more explanation in the error message when the command for a transport
589 filter fails to execute.
591 4. There are several places where Exim runs a non-Exim command in a
592 subprocess. The SIGUSR1 signal should be disabled for these processes. This
593 was being done only for the command run by the queryprogram router. It is
594 now done for all such subprocesses. The other cases are: ${run, transport
595 filters, and the commands run by the lmtp and pipe transports.
597 5. Some older OS have a limit of 256 on the maximum number of file
598 descriptors. Exim was using setrlimit() to set 1000 as a large value
599 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
600 systems. I've change it so that if it can't get 1000, it tries for 256.
602 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
603 was an oversight, and furthermore, ever since the addition of extra
604 controls (e.g. 4.43/32), the checks on when to allow different forms of
605 "control" were broken. There should now be diagnostics for all cases when a
606 control that does not make sense is encountered.
608 7. $recipients is now available in the predata ACL (oversight).
610 8. Tidy the search cache before the fork to do a delivery from a message
611 received from the command line. Otherwise the child will trigger a lookup
612 failure and thereby defer the delivery if it tries to use (for example) a
613 cached ldap connection that the parent has called unbind on.
615 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
616 of $address_data from the recipient verification was clobbered by the
619 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
620 was its contents. (It was OK if the option was not defined at all.)
622 11. A "Completed" log line is now written for messages that are removed from
623 the spool by the -Mrm option.
625 12. $host_address is now set to the target address during the checking of
628 13. When checking ignore_target_hosts for an ipliteral router, no host name was
629 being passed; this would have caused $sender_host_name to have been used if
630 matching the list had actually called for a host name (not very likely,
631 since this list is usually IP addresses). A host name is now passed as
634 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
635 code that specifies a non-restarting handler (typically sigaction() in
636 modern systems) in an attempt to fix a rare and obscure crash bug.
638 15. Narrowed the window for a race in the daemon that could cause it to ignore
639 SIGCHLD signals. This is not a major problem, because they are used only to
640 wake it up if nothing else does.
642 16. A malformed maildirsize file could cause Exim to calculate negative values
643 for the mailbox size or file count. Odd effects could occur as a result.
644 The maildirsize information is now recalculated if the size or filecount
647 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
648 support for a long time. Removed HAVE_SYS_VFS_H.
650 18. Updated exipick to current release from John Jetmore.
652 19. Allow an empty sender to be matched against a lookup in an address list.
653 Previously the only cases considered were a regular expression, or an
656 20. Exim went into a mad DNS lookup loop when doing a callout where the
657 host was specified on the transport, if the DNS lookup yielded more than
660 21. The RFC2047 encoding function was originally intended for short strings
661 such as real names; it was not keeping to the 75-character limit for
662 encoded words that the RFC imposes. It now respects the limit, and
663 generates multiple encoded words if necessary. To be on the safe side, I
664 have increased the buffer size for the ${rfc2047: expansion operator from
667 22. Failure to deliver a bounce message always caused it to be frozen, even if
668 there was an errors_to setting on the router. The errors_to setting is now
671 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
672 canonical form (fully expanded) before being placed in
673 $sender_host_address.
675 24. Updated eximstats to version 1.33
677 25. Include certificate and key file names in error message when GnuTLS fails
678 to set them up, because the GnuTLS error message doesn't include the name
679 of the failing file when there is a problem reading it.
681 26. Expand error message when OpenSSL has problems setting up cert/key files.
684 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
685 (this can affect the format of dates).
687 28. exim_tidydb, when checking for the continued existence of a message for
688 which it has found a message-specific retry record, was not finding
689 messages that were in split spool directories. Consequently, it was
690 deleting retry records that should have stayed in existence.
692 29. eximstats updated to version 1.35
693 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
694 1.35 - bugfix such that pie charts by volume are generated correctly
696 30. The SPA authentication driver was not abandoning authentication and moving
697 on to the next authenticator when an expansion was forced to fail,
698 contradicting the general specification for all authenticators. Instead it
699 was generating a temporary error. It now behaves as specified.
701 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
702 (the order specifies the preference for clients). The order is now AES256,
703 AES128, 3DES, ARCFOUR128.
705 31. Small patch to Sieve code - explicitly set From: when generating an
708 32. Exim crashed if a remote delivery caused a very long error message to be
709 recorded - for instance if somebody sent an entire SpamAssassin report back
710 as a large number of 550 error lines. This bug was coincidentally fixed by
711 increasing the size of one of Exim's internal buffers (big_buffer) that
712 happened as part of the Exiscan merge. However, to be on the safe side, I
713 have made the code more robust (and fixed the comments that describe what
716 33. Some experimental protocols are using DNS PTR records for new purposes. The
717 keys for these records are domain names, not reversed IP addresses. The
718 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
719 leaves it alone. Component reversal etc. now happens only for IP addresses.
722 34. The host_aton() function is supposed to be passed a string that is known
723 to be a valid IP address. However, in the case of IPv6 addresses, it was
724 not checking this. This is a hostage to fortune. Exim now panics and dies
725 if the condition is not met. A case was found where this could be provoked
726 from a dnsdb PTR lookup with an IPv6 address that had more than 8
727 components; fortuitously, this particular loophole had already been fixed
728 by change 4.50/55 or 4.44/33 above.
730 If there are any other similar loopholes, the new check in host_aton()
731 itself should stop them being exploited. The report I received stated that
732 data on the command line could provoke the exploit when Exim was running as
733 exim, but did not say which command line option was involved. All I could
734 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
738 35. There was a buffer overflow vulnerability in the SPA authentication code
739 (which came originally from the Samba project). I have added a test to the
740 spa_base64_to_bits() function which I hope fixes it.
743 36. The daemon start-up calls getloadavg() while still root for those OS that
744 need the first call to be done as root, but it missed one case: when
745 deliver_queue_load_max is set with deliver_drop_privilege. This is
746 necessary for the benefit of the queue runner, because there is no re-exec
747 when deliver_drop_privilege is set.
749 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
750 was in use, was not putting the data itself into the right store pool;
751 consequently, it could be overwritten for a subsequent message in the same
752 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
755 38. Sometimes the final signoff response after QUIT could fail to get
756 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
757 < 0 before doing a fflush(). This bug looks as though it goes back to the
758 introduction of TLS in release 3.20, but "sometimes" must have been rare
759 because the tests only now provoked it.
765 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
766 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
767 Now it returns an integer. A number of calls were still expecting a T/F
768 return. Fortuitously, in all cases, the tests worked in OK situations,
769 which is the norm. However, things would have gone wrong on any write
770 failures on the smtp file descriptor. This function is used when sending
771 messages over SMTP and also when doing verify callouts.
773 2. When Exim is called to do synchronous delivery of a locally submitted
774 message (the -odf or -odi options), it no longer closes stderr before doing
777 3. Implemented the mua_wrapper option.
779 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
781 5. Implemented the functions header_remove(), header_testname(),
782 header_add_at_position(), and receive_remove_recipient(), and exported them
785 6. If an ACL "warn" statement specified the addition of headers, Exim already
786 inserted X-ACL-Warn: at the start if there was no header name. However, it
787 was not making this test for the second and subsequent header lines if
788 there were newlines in the string. This meant that an invalid header could
789 be inserted if Exim was badly configured.
791 7. Allow an ACL "warn" statement to add header lines at the start or after all
792 the Received: headers, as well as at the end.
794 8. Added the rcpt_4xx retry error code.
796 9. Added postmaster_mailfrom=xxx to callout verification option.
798 10. Added mailfrom=xxxx to the callout verification option, for verify=
801 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
802 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
804 12. Inserted some casts to stop certain compilers warning when using pointer
805 differences as field lengths or precisions in printf-type calls (mostly
806 affecting debugging statements).
808 13. Added optional readline() support for -be (dynamically loaded).
810 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
811 same clock tick as a message's arrival, so that its received time was the
812 same as the "first fail" time on the retry record, and that message
813 remained on the queue past the ultimate address timeout, every queue runner
814 would try a delivery (because it was past the ultimate address timeout) but
815 after another failure, the ultimate address timeout, which should have then
816 bounced the address, did not kick in. This was a "< instead of <=" error;
817 in most cases the first failure would have been in the next clock tick
818 after the received time, and all would be well.
820 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
821 being recognized when the domain list was tested by the match_domain
822 condition in an expansion string.
824 16. Added the ${str2b64: operator.
826 17. Exim was always calling setrlimit() to set a large limit for the number of
827 processes, without checking whether the existing limit was already
828 adequate. (It did check for the limit on file descriptors.) Furthermore,
829 errors from getrlimit() and setrlimit() were being ignored. Now they are
830 logged to the main and panic logs, but Exim does carry on, to try to do its
831 job under whatever limits there are.
833 18. Imported PCRE 5.0.
835 19. Trivial typo in log message " temporarily refused connection" (the leading
838 20. If the log selector return_path_on_delivery was set and an address was
839 redirected to /dev/null, the delivery process crashed because it assumed
840 that a return path would always be set for a "successful" delivery. In this
841 case, the whole delivery is bypassed as an optimization, and therefore no
844 21. Internal re-arrangement: the function for sending a challenge and reading
845 a response while authentication was assuming a zero-terminated challenge
846 string. It's now changed to take a pointer and a length, to allow for
847 binary data in such strings.
849 22. Added the cyrus_sasl authenticator (code supplied by MBM).
851 23. Exim was not respecting finduser_retries when seeking the login of the
852 uid under which it was called; it was always trying 10 times. (The default
853 setting of finduser_retries is zero.) Also, it was sleeping after the final
854 failure, which is pointless.
856 24. Implemented tls_on_connect_ports.
858 25. Implemented acl_smtp_predata.
860 26. If the domain in control=submission is set empty, Exim assumes that the
861 authenticated id is a complete email address when it generates From: or
862 Sender: header lines.
864 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
865 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
866 chown and chgrp in /bin and hostname in /usr/bin.
868 28. Exim was keeping the "process log" file open after each use, just as it
869 does for the main log. This opens the possibility of it remaining open for
870 long periods when the USR1 signal hits a daemon. Occasional processlog
871 errors were reported, that could have been caused by this. Anyway, it seems
872 much more sensible not to leave this file open at all, so that is what now
875 29. The long-running daemon process does not normally write to the log once it
876 has entered its main loop, and it closes the log before doing so. This is
877 so that log files can straightforwardly be renamed and moved. However,
878 there are a couple of unusual error situations where the daemon does write
879 log entries, and I had neglected to close the log afterwards.
881 30. The text of an SMTP error response that was received during a remote
882 delivery was being truncated at 512 bytes. This is too short for some of
883 the long messages that one sometimes sees. I've increased the limit to
886 31. It is now possible to make retry rules that apply only when a message has a
887 specific sender, in particular, an empty sender.
889 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
890 it possible to be selective about when SMTP synchronization is enforced.
892 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
894 32. Implemented hosts_connection_nolog.
896 33. Added an ACL for QUIT.
898 34. Setting "delay_warning=" to disable warnings was not working; it gave a
901 35. Added mailbox_size and mailbox_filecount to appendfile.
903 36. Added control = no_multiline_responses to ACLs.
905 37. There was a bug in the logic of the code that waits for the clock to tick
906 in the case where the clock went backwards by a substantial amount such
907 that the microsecond fraction of "now" was more than the microsecond
908 fraction of "then" (but the whole seconds number was less).
910 38. Added support for the libradius Radius client library this is found on
911 FreeBSD (previously only the radiusclient library was supported).
917 1. When certain lookups returned multiple values in the form name=value, the
918 quoting of the values was not always being done properly. Specifically:
919 (a) If the value started with a double quote, but contained no whitespace,
921 (b) If the value contained whitespace other than a space character (i.e.
922 tabs or newlines or carriage returns) it was not quoted.
923 This fix has been applied to the mysql and pgsql lookups by writing a
924 separate quoting function and calling it from the lookup code. The fix
925 should probably also be applied to nisplus, ibase and oracle lookups, but
926 since I cannot test any of those, I have not disturbed their existing code.
928 2. A hit in the callout cache for a specific address caused a log line with no
929 reason for rejecting RCPT. Now it says "Previous (cached) callout
930 verification failure".
932 3. There was an off-by-one bug in the queryprogram router. An over-long
933 return line was truncated at 256 instead of 255 characters, thereby
934 overflowing its buffer with the terminating zero. As well as fixing this, I
935 have increased the buffer size to 1024 (and made a note to document this).
937 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
938 when Exim is waiting for an SMTP response from a remote server, Exim
939 restarts its select() call on the socket, thereby resetting its timeout.
940 This is not a problem when such interrupts are rare. Somebody set up a cron
941 job to run exiwhat every 2 minutes, which is less than the normal select()
942 timeout (5 or 10 minutes). This meant that the select() timeout never
943 kicked in because it was always reset. I have fixed this by comparing the
944 time when an interrupt arrives with the time at the start of the first call
945 to select(). If more time than the timeout has elapsed, the interrupt is
946 treated as a timeout.
948 5. Some internal re-factoring in preparation for the addition of Sieve
949 extensions (by MH). In particular, the "personal" test is moved to a
950 separate function, and given an option for scanning Cc: and Bcc: (which is
951 not set for Exim filters).
953 6. When Exim created an email address using the login of the caller as the
954 local part (e.g. when creating a From: or Sender: header line), it was not
955 quoting the local part when it contained special characters such as @.
957 7. Installed new OpenBSD configuration files.
959 8. Reworded some messages for syntax errors in "and" and "or" conditions to
960 try to make them clearer.
962 9. Callout options, other than the timeout value, were being ignored when
963 verifying sender addresses in header lines. For example, when using
965 verify = header_sender/callout=no_cache
967 the cache was (incorrectly) being used.
969 10. Added a missing instance of ${EXE} to the exim_install script; this affects
970 only the Cygwin environment.
972 11. When return_path_on_delivery was set as a log selector, if different remote
973 addresses in the same message used different return paths and parallel
974 remote delivery occurred, the wrong values would sometimes be logged.
975 (Whenever a remote delivery process finished, the return path value from
976 the most recently started remote delivery process was logged.)
978 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
979 lines when AUTH and/or TLS are in use. This is the "received protocol"
980 field. Exim used to use "asmtp" for authenticated SMTP, without any
981 indication (in the protocol name) for TLS use. Now it follows the RFC and
982 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
983 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
984 names appear in log lines as well as in Received: header lines.
986 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
987 extensions, and comparison tests, and to fix some bugs.
989 14. Changes to the "personal" filter test:
991 (1) The test was buggy in that it was just doing the equivalent of
992 "contains" tests on header lines. For example, if a user's address was
993 anne@some.where, the "personal" test would incorrectly be true for
995 To: susanne@some.where
997 This test is now done by extracting each address from the header in turn,
998 and checking the entire address. Other tests that are part of "personal"
999 are now done using regular expressions (for example, to check local parts
1000 of addresses in From: header lines).
1002 (2) The list of non-personal local parts in From: addresses has been
1003 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1004 taken from the Sieve specification recommendations.
1006 (3) If the message contains any header line starting with "List-" it is
1007 treated as non-personal.
1009 (4) The test for "circular" in the Subject: header line has been removed
1010 because it now seems ill-conceived.
1012 15. Minor typos in src/EDITME comments corrected.
1014 16. Installed latest exipick from John Jetmore.
1016 17. If headers_add on a router specified a text string that was too long for
1017 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1018 of string_sprintf() is now avoided.
1020 18. $message_body_size was not set (it was always zero) when running the DATA
1021 ACL and the local_scan() function.
1023 19. For the "mail" command in an Exim filter, no default was being set for
1024 the once_repeat time, causing a random time value to be used if "once" was
1025 specified. (If the value happened to be <= 0, no repeat happened.) The
1026 default is now 0s, meaning "never repeat". The "vacation" command was OK
1027 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1028 (I found it when inspecting the code).
1030 20. There is now an overall timeout for performing a callout verification. It
1031 defaults to 4 times the callout timeout, which applies to individual SMTP
1032 commands during the callout. The overall timeout applies when there is more
1033 than one host that can be tried. The timeout is checked before trying the
1034 next host. This prevents very long delays if there are a large number of
1035 hosts and all are timing out (e.g. when the network connections are timing
1036 out). The value of the overall timeout can be changed by specifying an
1037 additional sub-option for "callout", called "maxwait". For example:
1039 verify = sender/callout=5s,maxwait=20s
1041 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1042 to the end before writing, but this should make it even safer).
1044 22. Exim was forgetting that it had advertised PIPELINING for the second and
1045 subsequent messages on an SMTP connection. It was also not resetting its
1046 memory on STARTTLS and an internal HELO.
1048 23. When Exim logs an SMTP synchronization error within a session, it now
1049 records whether PIPELINING has been advertised or not.
1051 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1052 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1053 rather than long int.
1055 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1057 26. Added the never_mail option to autoreply.
1063 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1064 crash if the getsockname() call failed; this can happen if a connection is
1065 closed very soon after it is established. The problem was simply in the
1066 order in which certain operations were done, causing Exim to try to write
1067 to the SMTP stream before it had set up the file descriptor. The bug has
1068 been fixed by making things happen in the correct order.
1074 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1075 before the connection was closed, thus losing the rejection response.
1077 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1078 some early Solaris releases, but causes trouble in current releases where
1079 socklen_t is defined.
1081 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1084 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1087 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1088 that releasing the top of it at the end releases what was used for sub-
1089 expansions (unless the block got too big). However, discard this block if
1090 the first thing is a variable or header, so that we can use its block when
1091 it is dynamic (useful for very large $message_headers, for example).
1093 6. Lookups now cache *every* query, not just the most recent. A new, separate
1094 store pool is used for this. It can be recovered when all lookup caches are
1095 flushed. Lookups now release memory at the end of their result strings.
1096 This has involved some general refactoring of the lookup sources.
1098 7. Some code has been added to the store_xxx() functions to reduce the amount
1099 of flapping under certain conditions.
1101 8. log_incoming_interface used to affect only the <= reception log lines. Now
1102 it causes the local interface and port to be added to several more SMTP log
1103 lines, for example "SMTP connection from", and rejection lines.
1105 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1107 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1109 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1110 could be overwritten at the end of the current message (or the start of a
1111 new message if it was set in a HELO ACL). The value is now preserved for
1112 the duration of the SMTP connection.
1114 12. If a transport had a headers_rewrite setting, and a matching header line
1115 contained an unqualified address, that address was qualified, even if it
1116 did not match any rewriting rules. The underlying bug was that the values
1117 of the flags that permit the existence of unqualified sender and recipient
1118 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1119 non-local messages, and by -bnq for local messages) were not being
1120 preserved with the message after it was received.
1122 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1123 "next input=" as part of the text comprising the host identity instead of
1124 the correct text. The code was using the same buffer for two different
1125 strings. However, depending on which order the printing function evaluated
1126 its arguments, the bug did not always show up. Under Linux, for example, my
1127 test suite worked just fine.
1129 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1130 doesn't work with some older versions of Perl. It has been changed to "my",
1131 which in any case is probably the better facility to use.
1133 15. A really picky compiler found some instances of statements for creating
1134 error messages that either had too many or two few arguments for the format
1137 16. The size of the buffer for calls to the DNS resolver has been increased
1138 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1139 for addresses that have a lot of PTR records. This alleviates a problem; it
1140 does not fully solve it.
1142 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1143 buffer now truncates the list and logs the incident, which is the same
1144 action as happens when Exim is looking up a host name and its aliases.
1145 Previously in this situation something unpredictable would happen;
1146 sometimes it was "internal error: store_reset failed".
1148 18. If a server dropped the connection unexpectedly when an Exim client was
1149 using GnuTLS and trying to read a response, the client delivery process
1150 crashed while trying to generate an error log message.
1152 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1153 string, for example:
1155 warn message = H1: something\nH2: something
1157 the text was added as a single header line from Exim's point of view
1158 though it ended up OK in the delivered message. However, searching for the
1159 second and subsequent header lines using $h_h2: did not work. This has been
1160 fixed. Similarly, if a system filter added multiple headers in this way,
1161 the routers could not see them.
1163 20. Expanded the error message when iplsearch is called with an invalid key to
1164 suggest using net-iplsearch in a host list.
1166 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1167 ACLs are no longer actually imposed (and a message to that effect is
1170 22. If a "gecos" field in a passwd entry contained escaped characters, in
1171 particular, if it contained a \" sequence, Exim got it wrong when building
1172 a From: or a Sender: header from that name. A second bug also caused
1173 incorrect handling when an unquoted " was present following a character
1174 that needed quoting.
1176 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1177 was not being matched caselessly.
1179 24. Arranged for all hyphens in the exim.8 source to be escaped with
1182 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1183 properly. Recipient callout cache records were still being keyed to include
1184 the sender, even when use_sender was set false. This led to far more
1185 callouts that were necessary. The sender is no longer included in the key
1186 when use_sender is false.
1188 26. Added "control = submission" modifier to ACLs.
1190 27. Added the ${base62d: operator to decode base 62 numbers.
1192 28. dnsdb lookups can now access SRV records.
1194 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1195 the configuration file.
1197 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1198 (-v) mode. This makes the output for a verbose queue run more intelligible.
1200 31. Added a use_postmaster feature to recipient callouts.
1202 32. Added the $body_zerocount variable, containing the number of binary zero
1203 bytes in the message body.
1205 33. The time of last modification of the "new" subdirectory is now used as the
1206 "mailbox time last read" when there is a quota error for a maildir
1209 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1211 35. Added +ignore_unknown as a special item in host lists.
1213 36. Code for decoding IPv6 addresses in host lists is now included, even if
1214 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1215 address was recognized as an IP address, but was then not correctly decoded
1216 into binary, causing unexpected and incorrect effects when compared with
1223 1. Very minor rewording of debugging text in manualroute to say "list of
1224 hosts" instead of "hostlist".
1226 2. If verify=header_syntax was set, and a header line with an unqualified
1227 address (no domain) and a large number of spaces between the end of the
1228 name and the colon was received, the reception process suffered a buffer
1229 overflow, and (when I tested it) crashed. This was caused by some obsolete
1230 code that should have been removed. The fix is to remove it!
1232 3. When running in the test harness, delay a bit after writing a bounce
1233 message to get a bit more predictability in the log output.
1235 4. Added a call to search_tidyup() just before forking a reception process. In
1236 theory, someone could use a lookup in the expansion of smtp_accept_max_
1237 per_host which, without the tidyup, could leave open a database connection.
1239 5. Added the variables $recipient_data and $sender_data which get set from a
1240 lookup success in an ACL "recipients" or "senders" condition, or a router
1241 "senders" option, similar to $domain_data and $local_part_data.
1243 6. Moved the writing of debug_print from before to after the "senders" test
1246 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1247 problems for message scanning, either using a data ACL, or using
1248 local_scan() because the Received: header was not generated till after they
1249 were called (in order to set the time as the time of reception completion).
1250 I have revised the way this works. The header is now generated after the
1251 body is received, but before the ACL or local_scan() are called. After they
1252 are run, the timestamp in the header is updated.
1258 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1259 before starting a queue runner without re-exec. This happened only when
1260 deliver_drop_privilege was set or when the Exim user was set to root. The
1261 effect of the bug was that timeouts during subsequent deliveries caused
1262 crashes instead of being properly handled. The handler is now left at its
1263 default (and expected) setting.
1265 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1266 message, again when deliver_drop_privilege is set or Exim is run as root.
1267 The bug described in (1) was not present in this case, but the tidying up
1268 of the other signals was missing. I have made the two cases consistent.
1270 3. The ignore_target_hosts setting on a manualroute router was being ignored
1271 for hosts that were looked up using the /MX notation.
1273 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1276 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1277 operated on the sender address. After changing the $sender_address to <>
1278 for the sender address verify, Exim was re-instated it as the original
1279 (before rewriting) address, but remembering that it had rewritten it, so it
1280 wasn't rewriting it again. This bug also had the effect of breaking the
1281 sender address verification caching when the sender address was rewritten.
1283 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1284 This has been changed so that if the ip literal address matches
1285 ignore_target_hosts, the router declines.
1287 7. Added expansion conditions match_domain, match_address, and match_local_
1288 part (NOT match_host).
1290 8. The placeholder for the Received: header didn't have a length field set.
1292 9. Added code to Exim itself and to exim_lock to test for a specific race
1293 condition that could lead to file corruption when using MBX delivery. The
1294 issue is with the lockfile that is created in /tmp. If this file is removed
1295 after a process has opened it but before that process has acquired a lock,
1296 there is the potential for a second process to recreate the file and also
1297 acquire a lock. This could lead to two Exim processes writing to the file
1298 at the same time. The added code performs the same test as UW imapd; it
1299 checks after acquiring the lock that its file descriptor still refers to
1300 the same named file.
1302 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1303 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1306 11. Added the smtp_active_hostname option. If used, this will typically be made
1307 to depend on the incoming interface address. Because $interface_address is
1308 not set up until the daemon has forked a reception process, error responses
1309 that can happen earlier (such as "too many connections") no longer contain
1312 12. If an expansion in a condition on a "warn" statement fails because a lookup
1313 defers, the "warn" statement is abandoned, and the next ACL statement is
1314 processed. Previously this caused the whole ACL to be aborted.
1316 13. Added the iplsearch lookup type.
1318 14. Added ident_timeout as a log selector.
1320 15. Added tls_certificate_verified as a log selector.
1322 16. Added a global option tls_require_ciphers (compare the smtp transport
1323 option of the same name). This controls incoming TLS connections.
1325 17. I finally figured out how to make tls_require_ciphers do a similar thing
1326 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1327 before starting the TLS session.
1329 18. Tabs are now shown as \t in -bP output.
1331 19. If the log selector return_path_on_delivery was set, Exim crashed when
1332 bouncing a message because it had too many Received: header lines.
1334 20. If two routers both had headers_remove settings, and the first one included
1335 a superfluous trailing colon, the final name in the first list and the
1336 first name in the second list were incorrectly joined into one item (with a
1337 colon in the middle).
1343 1. Added -C and -D options to the exinext utility, mainly to make it easier
1344 to include in the automated testing, but these could be helpful when
1345 multiple configurations are in use.
1347 2. The exinext utility was not formatting the output nicely when there was
1348 an alternate port involved in the retry record key, nor when there was a
1349 message id as well (for retries that were specific to a specific message
1350 and a specific host). It was also confused by IPv6 addresses, because of
1351 the additional colons they contain. I have fixed the IPv4 problem, and
1352 patched it up to do a reasonable job for IPv6.
1354 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1355 delivery, the log line now contains "pipelined" if PIPELINING was used.
1357 4. An SMTP transport process used to panic and die if the bind() call to set
1358 an explicit outgoing interface failed. This has been changed; it is now
1359 treated in the same way as a connect() failure.
1361 5. A reference to $sender_host_name in the part of a conditional expansion
1362 that was being skipped was still causing a DNS lookup. This no longer
1365 6. The def: expansion condition was not recognizing references to header lines
1366 that used bh_ and bheader_.
1368 7. Added the _cache feature to named lists.
1370 8. The code for checking quota_filecount in the appendfile transport was
1371 allowing one more file than it should have been.
1373 9. For compatibility with Sendmail, the command line option
1381 and sets the incoming protocol and host name (for trusted callers). The
1382 host name and its colon can be omitted when only the protocol is to be set.
1383 Note the Exim already has two private options, -pd and -ps, that refer to
1384 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1385 "s", but I don't think that's a major issue.
1387 10. A number of refactoring changes to the code, none of which should affect
1390 (a) The number of logging options was getting close to filling up the
1391 32-bit word that was used as a bit map. I have split them into two classes:
1392 those that are passed in the argument to log_write(), and those that are
1393 only ever tested independently outside of that function. These are now in
1394 separate 32-bit words, so there is plenty of room for expansion again.
1395 There is no change in the user interface or the logging behaviour.
1397 (b) When building, for example, log lines, the code previously used a
1398 macro that called string_cat() twice, in order to add two strings. This is
1399 not really sufficiently general. Furthermore, there was one instance where
1400 it was actually wrong because one of the argument was used twice, and in
1401 one call a function was used. (As it happened, calling the function twice
1402 did not affect the overall behaviour.) The macro has been replaced by a
1403 function that can join an arbitrary number of extra strings onto a growing
1406 (c) The code for expansion conditions now uses a table and a binary chop
1407 instead of a serial search (which was left over from when there were very
1408 few conditions). Also, it now recognizes conditions like "pam" even when
1409 the relevant support is not compiled in: a suitably worded error message is
1410 given if an attempt is made to use such a condition.
1412 11. Added ${time_interval:xxxxx}.
1414 12. A bug was causing one of the ddress fields not to be passed back correctly
1415 from remote delivery subprocesses. The field in question was not being
1416 subsequently used, so this caused to problems in practice.
1418 13. Added new log selectors queue_time and deliver_time.
1420 14. Might have fixed a bug in maildirsizefile handling that threw up
1421 "unexpected character" debug warnings, and recalculated the data
1422 unnecessarily. In any case, I expanded the warning message to give more
1425 15. Added the message "Restricted characters in address" to the statements in
1426 the default ACL that block characters like @ and % in local parts.
1428 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1429 Three changes have been made:
1431 (a) There was a serious bug; a negative response to MAIL caused the whole
1432 recipient domain to be cached as invalid, thereby blocking all messages
1433 to all local parts at the same domain, from all senders. This bug has
1434 been fixed. The domain is no longer cached after a negative response to
1435 MAIL if the sender used is not empty.
1437 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1440 (c) A new callout option, "use_sender" has been added for people who want
1441 the modified behaviour.
1447 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1450 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1451 indeed breaks things for older releases.
1453 3. Added additional logging to the case where there is a problem reading data
1454 from a filter that is running in a subprocess using a pipe, in order to
1455 try to track down a specific problem.
1457 4. Testing facility fudge: when running in the test harness and attempting
1458 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1459 getting "No route to host". Convert this to a timeout.
1461 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1464 6. Some OS don't have socklen_t but use size_t instead. This affects the
1465 fifth argument of getsockopt() amongst other things. This is now
1466 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1467 can be set for individual OS. I have set it for SunOS5, OSF1, and
1468 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1469 some earlier ones do not.
1471 7. Change 4.30/15 was not doing the test caselessly.
1473 8. The standard form for an IPv6 address literal was being rejected by address
1474 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1475 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1476 this, as well as the form without the "IPv6" on the front (but only when
1477 address literals are enabled, of course).
1479 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1481 10. Exim crashed if a message with an empty sender address specified by -f
1482 encountered a router with an errors_to setting. This could be provoked only
1483 by a command such as
1487 where an empty string was supplied; "<>" did not hit this bug.
1489 11. Installed PCRE release 4.5.
1491 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1492 remained set. It is now erased.
1494 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1495 times from message ids (which are base 36 rather than the normal 62).
1497 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1498 were being counted as actual protocol errors, and logged if the log
1499 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1500 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1501 and DATA following a set of rejected RCPTs do not count as protocol errors.
1502 In other words, Exim assumes they were pipelined, though this may not
1503 actually be the case. Of course, in all cases the client gets an
1504 appropriate error code.
1506 15. If a lookup fails in an ACL condition, a message about the failure may
1507 be available; it is used if testing the ACL cannot continue, because most
1508 such messages specify what the cause of the deferral is. However, some
1509 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1510 that caused an old message to be retained and used if a later statement
1511 caused a defer, replacing the real cause of the deferral.
1513 16. If an IP address had so many PTR records that the DNS lookup buffer
1514 was not large enough to hold them, Exim could crash while trying to process
1515 the truncated data. It now detects and logs this case.
1517 17. Further to 4.21/58, another change has been made: if (and only if) the
1518 first line of a message (the first header line) ends with CRLF, a bare LF
1519 in a subsequent header line has a space inserted after it, so as not to
1520 terminate the header.
1522 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1523 unnecessarily, used atoi() instead of strtol(), and didn't check the
1524 termination when getting file sizes from file names by regex.
1526 19. Completely re-implemented the support for maildirsize files, in the light
1527 of a number of problems with the previous contributed implementation
1528 (4.30/29). In particular:
1530 . If the quota is zero, the maildirsize file is maintained, but no quota is
1533 . If the maildir directory does not exist, it is created before any attempt
1534 to write a maildirsize file.
1536 . The quota value in the file is just a cache; if the quota is changed in
1537 the transport, the new value overrides.
1539 . A regular expression is available for excluding directories from the
1542 20. The autoreply transport checks the characters in options that define the
1543 message's headers; it allows continued headers, but it was checking with
1544 isspace() after an embedded newline instead of explicitly looking for a
1547 21. If all the "regular" hosts to which an address was routed had passed their
1548 expiry times, and had not reached their retry times, the address was
1549 bounced, even if fallback hosts were defined. Now Exim should go on to try
1552 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1553 equivalent code in the SMTP transport. Some hosts send humungous responses
1554 to HELO/EHLO, more than 1024 it seems.
1556 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1557 gives compiler warnings in some environments. I've now done it "properly",
1560 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1561 (because of problems with the built-in one) was declared to return uschar *
1562 instead of char *, causing compiler failure.
1564 25. Fixed a file descriptor leak when processing alias/forward files.
1566 26. Fixed a minor format string issue in dbfn.c.
1568 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1570 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1571 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1574 29. When a sender address is verified, it is cached, to save repeating the test
1575 when there is more than one recipient in a message. However, when the
1576 verification involves a callout, it is possible for different callout
1577 options to be set for different recipients. It is too complicated to keep
1578 track of this in the cache, so now Exim always runs a verification when a
1579 callout is required, relying on the callout cache for the optimization.
1580 The overhead is duplication of the address routing, but this should not be
1583 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1584 to be verified with callout=postmaster, and the main callout worked but the
1585 postmaster check failed, the verification correctly failed. However, if a
1586 subsequent RCPT command asked for sender verification *without* the
1587 postmaster check, incorrect caching caused this verification also to fail,
1590 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1591 it was not caching the DNS options (qualify_single, search_parents) that
1592 were used when the lookup failed. A subsequent lookup with different
1593 options therefore always gave the same answer, though there were cases
1594 where it should not have. (Example: a "domains = !$mx_any" option on a
1595 dnslookup router: the "domains" option is always processed without any
1596 widening, but the router might have qualify_single set.) Now Exim uses the
1597 cached value only when the same options are set.
1599 32. Added John Jetmore's "exipick" utility to the distribution.
1601 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1602 than a timeout (e.g. a certificate is required, and is not provided), an
1603 Exim server now closes the connection immediately. Previously it waited for
1604 the client to close - but if the client is SSL, it seems that they each
1605 wait for each other, leading to a delay before one of them times out.
1607 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1608 maintained 0.8.x compatibility because I don't think many are using it, and
1609 it is clearly obsolete.
1611 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1614 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1615 client certificate was expired. A simple patch fixes this, though I don't
1616 understand the full logic of why the verify callback is called multiple
1619 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1620 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1621 which causes problems with some clients (such as the Certicom SSL Plus
1622 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1623 disables the coutermeasure allowing Eudora to connect."
1625 38. Exim was not checking that a write() to a log file succeeded. This could
1626 lead to Bad Things if a log got too big, in particular if it hit a file
1627 size limit. Exim now panics and dies if it cannot write to a log file, just
1628 as it does if it cannot open a log file.
1630 39. Modified OS/Makefile-Linux so that it now contains
1632 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1634 The two -D definitions ensure that Exim is compiled with large file
1635 support, which makes it possible to handle log files that are bigger than
1638 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1639 instance) a domain was checked against a named list that involved a lookup,
1640 causing $domain_data to be set, then another domain was checked against the
1641 same list, then the first domain was re-checked, the value of $domain_data
1642 after the final check could be wrong. In particular, if the second check
1643 failed, it could be set empty. This bug probably also applied to
1646 41. The strip_trailing_dot option was not being applied to the address given
1647 with the -f command-line option.
1649 42. The code for reading a message's header from the spool was incrementing
1650 $received_count, but never initializing it. This meant that the value was
1651 incorrect (doubled) while delivering a message in the same process in which
1652 it was received. In the most common configuration of Exim, this never
1653 happens - a fresh exec is done - but it can happen when
1654 deliver_drop_privilege is set.
1656 43. When Exim logs an SMTP synchronization error - client data sent too soon -
1657 it now includes up to 150 characters of the unexpected data in the log
1660 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1661 and building data strings. The size of both of these buffers was 10 000
1662 bytes - far larger than anybody would *ever* want, thought I. Needless to
1663 say, somebody hit the limit. I have increased the maximum line length to
1664 20 000 and the maximum data length of concatenated lines to 100 000. I have
1665 also fixed two bugs, because there was no checking on these buffers. Tsk,
1666 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1667 buffer is too small.
1669 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1670 lsearch lookups. Now it does.
1672 46. When parsing a route_list item in a manualroute router, a fixed-length
1673 buffer was used for the list of hosts. I made this 1024 bytes long,
1674 thinking that nobody would ever have a list of hosts that long. Wrong.
1675 Somebody had a whole pile of complicated expansion conditions, and the
1676 string was silently truncated, leading to an expansion error. It turns out
1677 that it is easier to change to an unlimited length (owing to other changes
1678 that have happened since this code was originally written) than to build
1679 structure for giving a limitation error. The length of the item that
1680 expands into the list of hosts is now unlimited.
1682 47. The lsearch lookup could not handle data where the length of text line was
1683 more than 4095 characters. Such lines were truncated, leading to shortened
1684 data being returned. It should now handle lines of any length.
1686 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1687 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1690 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
1691 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1692 no longer happens in comment lines. A list of replacements is now placed
1693 at the head of all of the source files, except those whose only change is
1694 to replace PERL_COMMAND in the very first #! line.
1696 50. Replaced the slow insertion sort in queue.c, for sorting the list of
1697 messages on the queue, with a bottom-up merge sort, using code contributed
1698 by Michael Haardt. This should make operations like -bp somewhat faster on
1699 large queues. It won't affect queue runners, except when queue_run_in_order
1702 51. Installed eximstats 1.31 in the distribution.
1704 52. Added support for SRV lookups to the dnslookup router.
1706 53. If an ACL referred to $message_body or $message_body_end, the value was not
1707 reset for any messages that followed in the same SMTP session.
1709 54. The store-handling optimization for building very long strings was not
1710 differentiating between the different store pools. I don't think this
1711 actually made any difference in practice, but I've tidied it.
1713 55. While running the routers to verify a sender address, $sender_address
1714 was still set to the sender address. This is wrong, because when routing to
1715 send a bounce to the sender, it would be empty. Therefore, I have changed
1716 it so that, while verifying a sender address, $sender_address is set to <>.
1717 (There is no change to what happens when verifying a recipient address.)
1719 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1720 target A or AAAA records (if not already returned) without resetting the
1721 qualify_single or search_parents options of the DNS resolver. These are
1722 inappropriate in this case because the targets of MX and SRV records must
1723 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1724 target that, when qualified, matched something in the local domain. These
1725 two options are now turned off when doing these lookups.
1727 57. It seems that at least some releases of Reiserfs (which does not have the
1728 concept of a fixed number of inodes) returns zero and not -1 for the
1729 number of available inodes. This interacted badly with check_spool_inodes,
1730 which assumed that -1 was the "no such thing" setting. What I have done is
1731 to check that the total number of inodes is greater than zero before doing
1732 the test of how many are available.
1734 58. When a "warn" ACL statement has a log_message modifier, the message is
1735 remembered, and not repeated. This is to avoid a lot of repetition when a
1736 message has many recipients that cause the same warning to be written.
1737 Howewer, Exim was preserving the list of already written lines for an
1738 entire SMTP session, which doesn't seem right. The memory is now reset if a
1739 new message is started.
1741 59. The "rewrite" debugging flag was not showing the result of rewriting in the
1742 debugging output unless log_rewrite was also set.
1744 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1745 of (int)(handle) when we know that handle contains (void *)(-1).
1747 61. The Exim daemon panic-logs an error return when it closes the incoming
1748 connection. However "connection reset by peer" seems to be common, and
1749 isn't really an error worthy of noting specially, so that particular error
1752 62. When Exim is trying to find all the local interfaces, it used to panic and
1753 die if the ioctl to get the interface flags failed. However, it seems that
1754 on at least one OS (Solaris 9) it is possible to have an interface that is
1755 included in the list of interfaces, but for which you get a failure error
1756 for this call. This happens when the interface is not "plumbed" into a
1757 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1758 failure of the "get flags" call assumes that the interface is down.
1760 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
1761 makes life easier for people who are doing arithmetic on fields extracted
1762 from dates, where you often get leading zeros that should not be
1763 interpreted as octal.
1765 64. Added qualify_domain to the redirect router, to override the global
1768 65. If a pathologically long header line contained very many addresses (the
1769 report of this problem mentioned 10 000) and each of them was rewritten,
1770 Exim could use up a very large amount of memory. (It kept on making new
1771 copies of the header line as it rewrote, and never released the old ones.)
1772 At the expense of a bit more processing, the header rewriting function has
1773 been changed so that it no longer eats memory in this way.
1775 66. The generation of the Received: header has been moved from the time that a
1776 message starts to be received, to the time that it finishes. The timestamp
1777 in the Received: header should now be very close to that of the <= log
1778 line. There are two side-effects of this change:
1780 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1781 logged header lines no longer include the local Received: line, because
1782 it has not yet been created. The same applies to a copy of the message
1783 that is returned to a non-SMTP sender when a message is rejected.
1785 (b) When a filter file is tested using -bf, no additional Received: header
1786 is added to the test message. After some thought, I decided that this
1789 This change does not affect the value of $received_for. It is still set
1790 after address rewriting, but before local_scan() is called.
1792 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1794 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1795 gave an unhelpful panic error message, and a defer error. I have managed to
1796 change this behaviour so that it now rejects any supplied certificate,
1797 which seems right, as the list of acceptable certificates is empty.
1799 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1800 gave an unhelpful defer error. I have not managed to make this reject any
1801 supplied certificates, but the error message it gives is "no certificate
1802 supplied", which is not helpful.
1804 70. exigrep's output now also includes lines that are not associated with any
1805 message, but which match the given pattern. Implemented by a patch from
1806 Martin Sluka, which also tidied up the Perl a bit.
1808 71. Recipient callout verification, like sender verification, was using <> in
1809 the MAIL FROM command. This isn't really the right thing, since the actual
1810 sender may affect whether the remote host accepts the recipient or not. I
1811 have changed it to use the actual sender in the callout; this means that
1812 the cache record is now keyed on a recipient/sender pair, not just the
1813 recipient address. There doesn't seem to be a real danger of callout loops,
1814 since a callout by the remote host to check the sender would use <>.
1815 [SEE ABOVE: changed after hitting problems.]
1817 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1818 temporary errors. However, in the case of such a code being given after
1819 the end of a data transmission (i.e. after ".") Exim was failing to write
1820 a retry record for the message. (Yes, there was some broken host that was
1821 actually sending 8xx at this point.)
1823 73. An unknown lookup type in a host list could cause Exim to panic-die when
1824 the list was checked. (An example that provoked this was putting <; in the
1825 middle of a list instead of at the start.) If this happened during a DATA
1826 ACL check, a -D file could be left lying around. This kind of configuration
1827 error no longer causes Exim to die; instead it causes a defer errror. The
1828 incident is still logged to the main and panic logs.
1830 74. Buglet left over from Exim 3 conversion. The message "too many messages
1831 in one connection" was written to the rejectlog but not the mainlog, except
1832 when address rewriting (yes!) was being logged.
1834 75. Added write_rejectlog option.
1836 76. When a system filter was run not as root (that is, when system_filter_user
1837 was set), the values of the $n variables were not being returned to the
1838 main process; thus, they were not subsequently available in the $sn
1841 77. Added +return_path_on_delivery log selector.
1843 78. A connection timeout was being treated differently from recipients deferred
1844 when testing hosts_max_try with a message that was older than the host's
1845 retry timeout. (The host should not be counted, thus allowing all hosts to
1846 be tried at least once before bouncing.) This may have been the cause of an
1847 occasionally reported bug whereby a message would remain on the queue
1848 longer than the retry timeout, but would be bounced if a delivery was
1849 forced. I say "may" because I never totally pinned down the problem;
1850 setting up timeout/retry tests is difficult. See also the next item.
1852 79. The ultimate address timeout was not being applied to errors that involved
1853 a combination of host plus message (for example, a timeout on a MAIL
1854 command). When an address resolved to a number of possible hosts, and they
1855 were not all tried for each delivery (e.g. because of hosts_max_try), a
1856 message could remain on the queue longer than the retry timeout.
1858 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1861 81. Fixed an obscure SMTP outgoing bug which required at least the following
1862 conditions: (a) there was another message waiting for the same server;
1863 (b) the server returned 5xx to all RCPT commands in the first message so
1864 that the message was not completed; (c) the server dropped the connection
1865 or gave a negative response to the RSET that Exim sends to abort the
1866 transaction. The observed case was a dropped connection after DATA that had
1867 been sent in pipelining mode. That is, the server had advertised PIPELINING
1868 but was not implementing it correctly. The effect of the bug was incorrect
1869 behaviour, such as trying another host, and this could lead to a crash.
1875 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1876 and daemon.c were passed as pointers to ints; they should have been
1877 pointers to socklen_t variables (which are typically unsigned ints).
1879 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1882 3. Fixed a really odd bug that affected only the testing scheme; patching a
1883 certain fixed string in the binary changed the value of another string that
1884 happened to be identical to the end of the original first string.
1886 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1887 name", it returns that address as the IP address. On some operating
1888 systems (e.g. Solaris), it also passes back the IP address string as the
1889 "host name". However, on others (e.g. Linux), it passes back an empty
1890 string. Exim wasn't checking for this, and was changing the host name to an
1891 empty string, assuming it had been canonicized.
1893 5. Although rare, it is permitted to have more than one PTR record for a given
1894 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1895 all the names associated with an address, because they do in Solaris.
1896 However, it seems that they do not in Linux for data that comes from the
1897 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1898 I found this out when I moved to a new Linux workstation and tried to run
1899 the Exim test suite.
1901 To get round this problem I have changed the code so that it now does its
1902 own call to the DNS to look up PTR records when searching for a host name.
1903 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1904 addresses that are only in /etc/hosts are still found.
1906 This behaviour is, however, controlled by an option called host_lookup_
1907 order, which defaults to "bydns:byaddr". If people want to use the other
1908 order, or indeed, just use one or the other means of lookup, they can
1909 specify it in this variable.
1911 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1912 some operating systems, this comes back from gethostbyaddr() as an empty
1913 string, and this is what Exim used to test for. However, it seems that in
1914 other systems, "." is yielded. Exim now tests for this case too.
1916 7. The values of check_spool_space and check_log_space are now held internally
1917 as a number of kilobytes instead of an absolute number of bytes. If a
1918 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1919 kilobyte. This means that much larger values can be stored.
1921 8. Exim monitor: an attempt to get the action menu when not actually pointing
1922 at a message produces an empty menu entitled "No message selected". This
1923 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1924 no entries in it ("Shell widget menu has zero width and/or height"). So I
1925 have added a single, blank menu entry in this case.
1927 9. Added ${quote_local_part.
1929 10. MIME decoding is now applied to the contents of Subject: header lines when
1932 11. Now that a reference to $sender_host_address automatically causes a reverse
1933 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1934 host lookup before query-style lookups in lists that might use this
1935 variable. This has therefore been abolished, and the "net-" prefix is no
1936 longer necessary for query-style lookups.
1938 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1939 have been a typo for LFLAGS, so it has been changed.
1941 13. The install script calls Exim with "-C /dev/null" in order to find the
1942 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1943 to be output. Howeve, since Exim outputs its version number before the
1944 error, it didn't break the script. It just looked ugly. I fixed this by
1945 always allowing "-C /dev/null" if the caller is root.
1947 14. Ignore overlarge ACL variable number when reading spool file - insurance
1948 against a later release with more variables having written the file.
1950 15. The standard form for an IPv6 address literal was being rejected by EHLO.
1951 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1952 this, as well as the form without the "IPv6" on the front.
1954 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1955 OS/Makefile-Darwin file.
1957 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1958 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1960 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
1961 variable. However, some OS use ints for inodes. I've added cast to long int
1962 to get rid of the compiler warning.
1964 19. I had forgotten to lock out "/../" in configuration file names when
1965 ALT_CONFIG_PREFIX was set.
1967 20. Routers used for verification do not need to specify transports. However,
1968 if such a router generated a host list, and callout was configured, Exim
1969 crashed, because it could not find a port number from the (non-existent)
1970 transport. It now assumes port 25 in this circumstance.
1972 21. Added the -t option to exigrep.
1974 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1975 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1976 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1977 them is set, LOOKUP_LSEARCH is forced.
1979 23. "exim -bV" now outputs a list of lookups that are included in the binary.
1981 24. Added sender and host information to the "rejected by local_scan()" log
1982 line; previously there was no indication of these.
1984 25. Added .include_if_exists.
1986 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
1987 Linux. It turns out that not all file systems support this. Apparently some
1988 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1989 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1990 not supported on the file descriptor, is now ignored when Exim is trying to
1991 sync a directory. This applies only to Linux.
1993 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1995 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1996 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1999 29. Added support for maildirsize files from supplied patch (modified a bit).
2001 30. The use of :fail: followed by an empty string could lead Exim to respond to
2002 sender verification failures with (e.g.):
2004 550 Verification failed for <xxx>
2005 550 Sender verify failed
2007 where the first response line was missing the '-' that indicates it is not
2008 the final line of the response.
2010 31. The loop for finding the name of the user that called Exim had a hardwired
2011 limit of 10; it now uses the value of finduser_retries, which is used for
2012 all other user lookups.
2014 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2017 33. Exim was neglecting to zero errno before one call of strtol() when
2018 expanding a string and expecting an integer value. On some systems this
2019 resulted in spurious "integer overflow" errors. Also, it was casting the
2020 result into an int without checking.
2022 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2023 did not work. The code looks as if it has *never* worked, though it appears
2024 to have been documented since at least releast 1.62. I have made it work.
2026 35. The "timeout_DNS" error in retry rules, also documented since at least
2027 1.62, also never worked. As it isn't clear exactly what this means, and
2028 clearly it isn't a major issue, I have abolished the feature by treating it
2029 as "timeout", and writing a warning to the main and panic logs.
2031 36. The display of retry rules for -brt wasn't always showing the error code
2034 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2035 timeout_connect_A, timeout_connect_MX.
2037 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2038 to the empty sender.
2040 39. The daemon was not analysing the content of -oX till after it had closed
2041 stderr and disconnected from the controlling terminal. This meant that any
2042 syntax errors were only noted on the panic log, and the return code from
2043 the command was 0. By re-arranging the code a little, I've made the
2044 decoding happen first, so such errors now appear on stderr, and the return
2045 code is 1. However, the actual setting up of the sockets still happens in
2046 the disconnected process, so errors there are still only recorded on the
2049 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2050 connections (as happens on some IP stacks) was logged at start up time as
2051 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2052 it from "IPv6 and IPv4", which means that two separate sockets are being
2055 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2056 says whether AF_INET or AF_INET6 was passed as an argument.
2058 42. Exiwhat output was messed up when time zones were included in log
2061 43. Exiwhat now gives more information about the daemon's listening ports,
2062 and whether -tls-on-connect was used.
2064 44. The "port" option of the smtp transport is now expanded.
2066 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2067 silently ignored. Now an error message is written to the main and panic
2070 46. There's a new ACL modifier called "logwrite" which writes to a log file
2071 as soon as it is encountered.
2073 47. Added $local_user_uid and $local_user_gid at routing time.
2075 48. Exim crashed when trying to verify a sender address that was being
2078 49. Exim was recognizing only a space character after ".include". It now also
2079 recognizes a tab character.
2081 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2082 extracting the relevant information from the specification. The man page no
2083 longer contains scrambled data for the -d option, and I've added a section
2084 at the front about calling Exim under different names.
2086 51. Added "extra_headers" argument to the "mail" command in filter files.
2088 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2091 53. Installed eximstats 1.29.
2093 54. Added transport_filter_timeout as a generic transport option.
2095 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2096 Cc: header lines. This was required by RFC 822, but it not required by RFC
2099 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2100 incoming messages that did not have them. Now it does so only if the
2101 message originates locally, that is, if there is no associated remote host
2102 address. When Resent- header lines are present, this applies to the Resent-
2103 lines rather than the non-Resent- lines.
2105 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2106 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2108 58. Messages for configuration errors now include the name of the main
2109 configuration file - useful now that there may be more than one file in a
2110 list (.included file names were always shown).
2112 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2113 for those rare installations that do not start the daemon as root or run it
2114 setuid root. I've cut out the call to initgroups() if the daemon is not
2117 60. The Exim user and group can now be bound into the binary as text strings
2118 that are looked up at the start of Exim's processing.
2120 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2122 62. Added $mailstore_basename variable.
2124 63. Installed patch to sieve.c from Michael Haardt.
2126 64. When Exim failed to open the panic log after failing to open the main log,
2127 the original message it was trying to log was written to stderr and debug
2128 output, but if they were not available (the usual case in production), it
2129 was lost. Now it is written to syslog before the two lines that record the
2130 failures to open the logs.
2132 65. Users' Exim filters run in subprocesses under the user's uid. It is
2133 possible for a "deliver" command or an alias in a "personal" command to
2134 provoke an address rewrite. If logging of address rewriting is configured,
2135 this fails because the process is not running as root or exim. There may be
2136 a better way of dealing with this, but for the moment (because 4.30 needs
2137 to be released), I have disabled address rewrite logging when running a
2138 filter in a non-root, non-exim process.
2144 1. The buildconfig auxiliary program wasn't quoting the value set for
2145 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2146 not defined. This bug was masked in 4.22 by the effect that was fixed in
2149 2. Some messages that were rejected after a message id was allocated were
2150 shown as "incomplete" by exigrep. It no longer does this for messages that
2151 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2153 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2154 have allow_domain_literals set, the ID did not get logged in the <= line.
2155 Domain literals are now always recognized in Message-ID: header lines.
2157 4. The first argument for a ${extract expansion item is the key name or field
2158 number. Leading and trailing spaces in this item were not being ignored,
2159 causing some misleading effects.
2161 5. When deliver_drop_privilege was set, single queue runner processes started
2162 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2163 same command in the process it spins off) were not dropping privilege.
2165 6. When the daemon running as "exim" started a queue runner, it always
2166 re-executed Exim in the spun-off process. This is a waste of effort when
2167 deliver_drop_privilege is set. The new process now just calls the
2168 queue-runner function directly.
2174 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2177 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2178 ignored. Though the use of -oP was forcing the writing of a pid file, it
2179 was always written to the default place.
2181 3. If the message "no IP address found for host xxxx" is generated during
2182 incoming verification, it is now followed by identification of the incoming
2183 connection (so you can more easily find what provoked it).
2185 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2187 5. Added some features to "harden" Exim a bit more against certain attacks:
2189 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2190 be put in Local/Makefile. This is like the never_users runtime option,
2191 but it cannot be overridden. The default setting is "root".
2193 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2194 prefix string with which any file named in a -C command line option
2197 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2198 is retained for -C and -D only if the caller of Exim is root. Without
2199 it, the exim user may also use -C and -D and retain privilege.
2201 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2202 command line option is disabled.
2204 6. Macro names set by the -D option must start with an upper case letter, just
2205 like macro names defined in the configuration file.
2207 7. Added "dereference=" facility to LDAP.
2209 8. Two instances of the typo "uknown" in the source files are fixed.
2211 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2212 the Configure-Makefile script screwed up while processing it.
2214 10. Incorporated PCRE 4.4.
2216 11. The SMTP synchronization check was not operating right at the start of an
2217 SMTP session. For example, it could not catch a HELO sent before the client
2218 waited for the greeting. There is now a check for outstanding input at the
2219 point when the greeting is written. Because of the duplex, asynchronous
2220 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2221 way, but not yet received, when the check is performed.
2223 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2224 on TCP/IP sockets, because this apparently causes some broken clients to
2227 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2228 unchanged) from the Cygwin maintainer.
2230 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2231 was supported instead of testing for mbx. Effectively a typo.
2233 15. The spa authenticator server code was not checking that the input it
2234 received was valid base64.
2236 16. The debug output line for the "set" modifier in ACLs was not showing the
2237 name of the variable that was being set.
2239 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2241 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2242 lookup on its own. Something else was needed to trigger the lookup. For
2243 example, a match in host_lookup or the need for a host name in a host list.
2244 Now, if $sender_host_name is referenced and the host name has not yet been
2245 looked up, a lookup is performed. If the lookup fails, the variable remains
2246 empty, and $host_lookup_failed is set to "1".
2248 19. Added "eqi" as a case-independent comparison operator.
2250 20. The saslauthd authentication condition could segfault if neither service
2251 nor realm was specified.
2253 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2254 error message that was logged was misleading, and incoming SMTP
2255 connections were dropped. The message is now more accurate, and temporary
2256 errors are given to SMTP connections.
2258 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2259 (except RSET) with a 421 error, until QUIT is received. However, it was
2260 failing to send a response to QUIT.
2262 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2263 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2264 ACL is now run after the helo_try_verify_hosts code.
2266 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2267 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2268 case-independent, so other case variants are also recognized). Apparently
2269 some systems use these upper case variants.
2271 25. If more than two messages were waiting for the same host, and a transport
2272 filter was specified for the transport, Exim sent two messages over the
2273 same TCP/IP connection, and then failed with "socket operation on non-
2274 socket" when it tried to send the third.
2276 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2278 27. The extern definition of crypt16() in expand.c was not being excluded when
2279 the OS had its own crypt16() function.
2281 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2282 as a preferred synonym for return_size_limit, both as an option and as an
2285 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2287 30. Changed the default configuration ACL to relax the local part checking rule
2288 for addresses that are not in any local domains. For these addresses,
2289 slashes and pipe symbols are allowed within local parts, but the sequence
2290 /../ is explicitly forbidden.
2292 31. SPA server authentication was not clearing the challenge buffer before
2295 32. log_message in a "warn" ACL statement was writing to the reject log as
2296 well as to the main log, which contradicts the documentation and doesn't
2297 seem right (because no rejection is happening). So I have stopped it.
2299 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2300 However, I am unable to do any testing of this.
2302 34. Fixed an infelicity in the appendfile transport. When checking directories
2303 for a mailbox, to see if any needed to be created, it was accidentally
2304 using path names with one or more superfluous leading slashes; tracing
2305 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2307 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2308 contents are added to the log line that is written for every discarded
2309 recipient. (Previously a log_message setting was ignored.)
2311 36. The ${quote: operator now quotes the string if it is empty.
2313 37. The install script runs exim in order to find its version number. If for
2314 some reason other than non-existence or emptiness, which it checks, it
2315 could not run './exim', it was installing it with an empty version number,
2316 i.e. as "exim-". This error state is now caught, and the installation is
2319 38. An argument was missing from the function that creates an error message
2320 when Exim fails to connect to the socket for saslauthd authentication.
2321 This could cause Exim to crash, or give a corrupted message.
2323 39. Added isip, isip4, and isip6 to ${if conditions.
2325 40. The ACL variables $acl_xx are now saved with the message, and can be
2326 accessed later in routers, transports, and filters.
2328 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2329 strings in the file are not string-expanded.
2331 42. If a MAIL command specified a SIZE value that was too large to fit into an
2332 int variable, the check against message_size_limit failed. Such values are
2333 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2334 day this will have to be increased, but I don't think I want to be around
2335 when emails are that large.
2342 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2343 iconv() is not standard in FreeBSD.
2345 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2346 IPv6 enabled. The observed symptom was a segmentation fault on return from
2347 the function os_common_find_running_interfaces() in src/os.c.
2349 3. In the check_special_case() function in daemon.c I had used "errno" as an
2350 argument name, which causes warnings on some systems. This was basically a
2351 typo, since it was named "eno" in the comments!
2353 4. The code that waits for the clock to tick (at a resolution of some fraction
2354 of a second) so as to ensure message-id uniqueness was always waiting for
2355 at least one whole tick, when it could have waited for less. [This is
2356 almost certainly not relevant at current processor speeds, where it is
2357 unlikely to ever wait at all. But we try to future-proof.]
2359 5. The function that sleeps for a time interval that includes fractions of a
2360 second contained a race. It did not block SIGALRM between setting the
2361 timer, and suspending (a couple of lines later). If the interval was short
2362 and the sigsuspend() was delayed until after it had expired, the suspension
2363 never ended. On busy systems this could lead to processes getting stuck for
2366 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2367 process, before it forks any delivery processes. The open lookup caching
2368 mechanism meant that the open file or database connection was passed into
2369 the delivery process. The problem was that delivery processes always tidy
2370 up cached lookup data. This could cause a problem for the next delivery
2371 process started by the queue runner, because the external queue runner
2372 process does not know about the closure. So the next delivery process
2373 still has data in the lookup cache. In the case of a file lookup, there was
2374 no problem because closing a file descriptor in a subprocess doesn't affect
2375 the parent. However, if the lookup was caching a connection to a database,
2376 the connection was closed, and the second delivery process was likely to
2377 see errors such as "PGSQL: query failed: server closed the connection
2378 unexpectedly". The problem has been fixed by closing all cached lookups
2379 in a queue runner before running a delivery process.
2381 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2382 seem to have the "const" qualifier which it has on other OS. I've
2385 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2386 *of the same type* (client or server) with the same public name that an
2387 error should be diagnosed.
2389 9. When Exim looked up a host name for an IP address, but failed to find the
2390 original IP address when looking up the host name (a safety check), it
2391 output the message "<ip address> does not match any IP for NULL", which was
2392 confusing, to say the least. The bug was that the host name should have
2393 appeared instead of "NULL".
2395 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2396 user that is built into the binary, and the -C or -D options is used, root
2397 privilege is dropped before the configuration file is read. In addition,
2398 logging is switched to stderr instead of the normal log files. If the
2399 configuration then re-defines the Exim user, the unprivileged environment
2400 is probably not what is expected, so Exim logs a panic warning message (but
2403 However, if deliver_drop_privilege is set, the unprivileged state may well
2404 be exactly what is intended, so the warning has been cut out in that case,
2405 and Exim is allowed to try to write to its normal log files.
2411 1. smtp_return_error_details was not giving details for temporary sender
2412 or receiver verification errors.
2414 2. Diagnose a configuration error if two authenticators have the same public
2417 3. Exim used not to create the message log file for a message until the first
2418 delivery attempt. This could be confusing when incoming messages were held
2419 for policy or load reasons. The message log file is now created at the time
2420 the message is received, and an initial "Received" line is written to it.
2422 4. The automatically generated man page for command line options had a minor
2423 bug that caused no ill effects; however, a more serious problem was that
2424 the procedure for building the man page automatically didn't always
2425 operate. Consequently, release 4.20 contains an out-of-date version. This
2426 shouldn't happen again.
2428 5. When building Exim with embedded Perl support, the script that builds the
2429 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2430 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2432 6. The freeze_tell option was not being used for messages that were frozen on
2433 arrival, either by an ACL or by local_scan().
2435 7. Added the smtp_incomplete_transaction log selector.
2437 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2438 was accepting AUTH without a new EHLO.
2440 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2441 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2444 10. Logging of TCP/IP connections (when configured) now happens in the main
2445 daemon process instead of the child process, so that the TCP/IP connection
2446 count is more accurate (but it can never be perfect).
2448 11. The use of "drop" in a nested ACL was not being handled correctly in the
2449 outer ACL. Now, if condition failure induced by the nested "drop" causes
2450 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2451 or "require"), the connection is dropped.
2453 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2454 that yield "discard" can now be used with an "accept" or a "discard" verb,
2455 but an error is generated for any others (because I can't see a useful way
2456 to define what should happen).
2458 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2459 are now processed in the same way as lines in the Exim configuration file.
2460 In particular, continuation lines are supported.
2462 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2464 15. Added -ti meaning -t -i.
2466 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2467 domains, and warn by logging if others are found.
2469 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2470 returned from GIFCONF when Exim is trying to find the list of interfaces on
2471 a host. The code in os.c has been modified to copy each ifreq to an aligned
2472 structure in all cases.
2474 Also, in some cases, the returned ifreq's were being copied to a 'struct
2475 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2476 means the last couple of bytes of an IPv6 address could be chopped if the
2477 ifreq contained only a normal sockaddr (14 bytes storage).
2479 18. Named domain lists were not supported in the hosts_treat_as_local option.
2480 An entry such as +xxxx was not recognized, and was treated as a literal
2483 19. Ensure that header lines added by a DATA ACL are included in the reject log
2484 if the ACL subsequently rejects the message.
2486 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2487 MD5 (which is deprecated).
2489 21. When testing a filter file using -bf, Exim was writing a message when it
2490 took the sender from a "From " line in the message, but it was not doing so
2491 when it took $return_path from a Return-Path: header line. It now does.
2493 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2494 with a valid header line field name (a series of printing characters
2495 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2497 23. Changed "disc" in the source to "disk" to conform to the documentation and
2498 the book and for uniformity.
2500 24. Ignore Sendmail's -Ooption=value command line item.
2502 25. When execve() failed while trying to run a command in a pipe transport,
2503 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2504 could be confused with a return value of 69 from the command itself. This
2505 has been changed to 127, the value the shell returns if it is asked to run
2506 a non-existent command. The wording for the related log line suggests a
2507 non-existent command as the problem.
2509 26. If received_header_text expands to an empty string, do not add a Received:
2510 header line to the message. (Well, it adds a token one on the spool, but
2511 marks it "old" so that it doesn't get used or transmitted.)
2513 27. Installed eximstats 1.28 (addition of -nt option).
2515 28. There was no check for failure on the call to getsockname() in the daemon
2516 code. This can fail if there is a shortage of resources on the system, with
2517 ENOMEM, for example. A temporary error is now given on failure.
2519 29. Contrary to the C standard, it seems that in some environments, the
2520 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2521 program. Exim now does this explicitly; it affects the formatting of
2522 timestamps using strftime().
2524 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2525 complaints. I've now initialized all the variables, to avoid this.
2527 32. Header lines added by a system filter were not being "seen" during
2528 transport-time rewrites.
2530 33. The info_callback() function passed to OpenSSL is set up with type void
2531 (*)(SSL *, int, int), as described somewhere. However, when calling the
2532 function (actually a macro) that sets it up, the type void(*)() is
2533 expected. I've put in a cast to prevent warnings from picky compilers.
2535 34. If a DNS black list lookup found a CNAME record, but there were no A
2536 records associated with the domain it pointed at, Exim crashed.
2538 35. If a DNS black list lookup returned more than one A record, Exim ignored
2539 all but the first. It now scans all returned addresses if a particular IP
2540 value is being sought. In this situation, the contents of the
2541 $dnslist_value variable are a list of all the addresses, separated by a
2544 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2545 to accept a host name and all its aliases if the forward lookup for any of
2546 them yielded the IP address of the incoming connection. Now it accepts only
2547 those names whose forward lookup yields the correct IP address. Any other
2548 names are discarded. This closes a loophole whereby a rogue DNS
2549 administrator could create reverse DNS records to break through a
2550 wildcarded host restriction in an ACL.
2552 37. If a user filter or a system filter that ran in a subprocess used any of
2553 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2554 the wrong values were passed to the pipe command ($thisaddress had the
2555 value of $0, $0 had the value of $1, etc). This bug was introduced by
2556 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2558 38. Improved the line breaking for long SMTP error messages from ACLs.
2559 Previously, if there was no break point between 40 and 75 characters, Exim
2560 left the rest of the message alone. Two changes have been made: (a) I've
2561 reduced the minimum length to 35 characters; (b) if it can't find a break
2562 point between 35 and 75 characters, it looks ahead and uses the first one
2563 that it finds. This may give the occasional overlong line, but at least the
2564 remaining text gets split now.
2566 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2567 file descriptors might be low, and that setting 1000 would always raise it.
2568 It turns out that in some environments, the limit is already over 1000 and
2569 that lowering it causes trouble. So now Exim takes care not to decrease it.
2571 40. When delivering a message, the value of $return_path is set to $sender_
2572 address at the start of routing (routers may change the value). By an
2573 oversight, this default was not being set up when an address was tested by
2574 -bt or -bv, which affected the outcome if any router or filter referred to
2577 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2578 log only when "message" or "log_message" are set. However, if one of the
2579 conditions was an address verification, or a call to a nested ACL, the
2580 messages generated by the underlying test were being passed through. This
2581 no longer happens. The underlying message is available in $acl_verify_
2582 message for both "message" and "log_message" expansions, so it can be
2583 passed through if needed.
2585 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2586 new expansion $bh_ to give the encoded byte string without charset
2587 translation. Translation happens only if iconv() is available; HAVE_ICONV
2588 indicates this at build time. HEADERS_CHARSET gives the charset to
2589 translate to; headers_charset can change it in the configuration, and
2590 "headers charset" can change it in an individual filter file.
2592 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2593 that creates RFC 2047 encoded "words" labels them as that charset instead
2594 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2595 expansion operator; (ii) when Exim creates a From: line for a local
2596 message; (iii) when a header line is rewritten to include a "phrase" part.
2598 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2599 buggy, causing it to skip the first lines of messages whose message ID
2600 ended in 'D'. This would not have bitten before Exim release 4.14, because
2601 message IDs were unlikely to end in 'D' before then. The effect was to have
2602 incorrect size information for certain domains.
2604 45. #include "config.h" was missing at the start of the crypt16.c module. This
2605 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2608 46. If there was a timeout during a "random" callout check, Exim treated it as
2609 a failure of the random address, and carried on sending RSET and the real
2610 address. If the delay was just some slowness somewhere, the response to the
2611 original RCPT would be taken as a response to RSET and so on, causing
2612 mayhem of various kinds.
2614 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2615 when I implemented it. It didn't allow for the fact that some option values
2616 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2617 the right test for positive values.
2619 48. Domain names in DNS records are case-independent. Exim always looks them up
2620 in lower case. Some resolvers return domain names in exactly the case they
2621 appear in the zone file, that is, they may contain uppercase letters. Not
2622 all resolvers do this - some return always lower case. Exim was treating a
2623 change of case by a resolver as a change of domain, similar to a widening
2624 of a domain abbreviation. This triggered its re-routing code and so it was
2625 trying to route what was effectively the same domain again. This normally
2626 caused routing to fail (because the router wouldn't handle the domain
2627 twice). Now Exim checks for this case specially, and just changes the
2628 casing of the domain that it ultimately uses when it transmits the message
2631 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2634 50. If a filter generated a file delivery with a non-absolute name (possible if
2635 no home directory exists for the router), the forbid_file option was not
2638 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2639 the existing equality matching.
2641 52. Exim was using ints instead of ino_t variables in some places where it was
2642 dealing with inode numbers.
2644 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2645 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2646 called TMPDIR, and if it finds it is different, it changes its value.
2648 54. The smtp_printf() function is now made available to local_scan() so
2649 additional output lines can be written before returning. There is also an
2650 smtp_fflush() function to enable the detection of a dropped connection.
2651 The variables smtp_input and smtp_batched_input are exported to
2654 55. Changed the default runtime configuration: the message "Unknown user"
2655 has been removed from the ACL, and instead placed on the localuser router,
2656 using the cannot_route_message feature. This means that any verification
2657 failures that generate their own messages won't get overridden. Similarly,
2658 the "Unrouteable address" message that was in the ACL for unverifiable
2659 relay addresses has also been removed.
2661 56. Added hosts_avoid_esmtp to the smtp transport.
2663 57. The exicyclog script was not checking for the esoteric option
2664 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2665 will work only if exicyclog is run under the appropriate euid.
2667 58. Following a discussion on the list, the rules by which Exim recognises line
2668 endings on incoming messages have been changed. The -dropcr and drop_cr
2669 options are now no-ops, retained only for backwards compatibility. The
2670 following line terminators are recognized: LF CRLF CR. However, special
2671 processing applies to CR:
2673 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2674 nor a local message in the state where . is a terminator.
2676 (ii) If a bare CR is encountered in a header line, an extra space is added
2677 after the line terminator so as not to end the header. The reasoning
2678 behind this is that bare CRs in header lines are most likely either
2679 to be mistakes, or people trying to play silly games.
2681 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
2682 was being incorrectly given as 18 bytes larger than it should have been.
2683 This is a VOB (very old bug).
2685 60. This may never have affected anything current, but just in case it has:
2686 When the local host is found other than at the start of a list of hosts,
2687 the local host, those with the same MX, and any that follow, are discarded.
2688 When the list in question was part of a longer list of hosts, the following
2689 hosts (not currently being processed) were also being discarded. This no
2690 longer happens. I'm not sure if this situation could ever has previously
2693 61. Added the "/MX" feature to lists of hosts in the manualroute and query
2696 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2697 header. This is something that is recommended in a new Internet Draft, and
2698 is something that is documented as being done by Sendmail. There are two
2699 possible values. For messages generated by the autoreply transport, Exim
2702 Auto-Submitted: auto-replied
2704 whereas for all other generated messages (e.g. bounces) it adds
2706 Auto-Submitted: auto-generated
2708 63. The "personal" condition in filters now includes a test for the
2709 Auto-Submitted: header. If it contains the string "auto-" the message it
2710 not considered personal.
2712 64. Added rcpt_include_affixes as a generic transport option.
2714 65. Added queue_only_override (default true).
2716 66. Added the syslog_duplication option.
2718 67. If what should have been the first header line of a message consisted of
2719 a space followed by a colon, Exim was mis-interpreting it as a header line.
2720 It isn't of course - it is syntactically invalid and should therefore be
2721 treated as the start of the message body. The misbehaviour could have
2722 caused a number of strange effects, including loss of data in subsequent
2723 header lines, and spool format errors.
2725 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2726 client host had authenticated. This control can now be exercised by an ACL
2727 for more flexibility.
2729 69. By default, callouts do not happen when testing with -bh. There is now a
2730 variant, -bhc, which does actually run the callout code, including
2731 consulting and updating the callout cache.
2733 70. Added support for saslauthd authentication, courtesy of Alexander
2736 71. If statvfs() failed on the spool or log directories while checking their
2737 size for availability, Exim confusingly gave the error "space shortage".
2738 Furthermore, in debugging mode it crashed with a floating point exception.
2739 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2740 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2741 serious problem, Exim now writes to the main and panic logs when this
2742 happens, with details of the failure. It then refuses to accept the
2743 incoming message, giving the message "spool directory problem" or "log
2744 directory problem" with a 421 code for SMTP messages.
2746 72. When Exim is about to re-exec itself, it ensures that the file descriptors
2747 0, 1, and 2 exist, because some OS complain for execs without them (see
2748 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2749 descriptors. However, the code omitted to check that the open succeeded,
2750 causing mysterious errors if for some reason the permissions on /dev/null
2751 got screwed. Now Exim writes a message to the main and panic logs, and
2752 bombs out if it can't open /dev/null.
2754 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2755 interact so that it is all more flexible. It is supposed to remain
2756 backwards compatible. Also added extra_local_interfaces.
2758 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2759 to bomb out with an assertion failure - to the client this appears as a
2760 connection drop. This problem occurs in the part of the code that was taken
2761 from the Samba project. Fortunately, the assertion is in a very simple
2762 function, so I have fixed this by reproducing the function inline in the
2763 one place where it is called, and arranging for authentication to fail
2764 instead of killing the process with assert().
2766 75. The SPA client code was not working when the server requested OEM rather
2767 than Unicode encoding.
2769 76. Added code to make require_files with a specific uid setting more usable in
2770 the case where statting the file as root fails - usually a non-root-mounted
2771 NFS file system. When this happens and the failure is EACCES, Exim now
2772 forks a subprocess and does the per-uid checking as the relevant uid.
2774 77. Added process_log_path.
2776 78. If log_file_path was not explicitly set, a setting of check_log_space or
2777 check_log_inodes was ignored.
2779 79. If a space check for the spool or log partitions fails, the incident is now
2780 logged. Of course, in the latter case the data may get lost...
2782 80. Added the %p formatting code to string_format() so that it can be used to
2783 print addresses in debug_print(). Adjusted all the address printing in the
2784 debugging in store.c to use %p rather than %d.
2786 81. There was a concern that a line of code in smtp_in.c could overflow a
2787 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2788 initially expressed, the concern was not well-founded, because trailing
2789 spaces are removed early. However, if the trailing spaces were followed by
2790 a NULL, they did not get removed, so the overflow was possible. Two fixes
2793 (a) I re-wrote the offending code in a cleaner fashion.
2794 (b) If an incoming SMTP command contains a NULL character, it is rejected
2797 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
2798 runs initgroups(), so that if the Exim user is in any additional groups,
2799 they will be used during message reception.
2805 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / heiko / exim.git / blob