1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.86 2005/03/08 11:38:21 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
10 PH/01. Installed a patch from the Sieve maintainer that allows -bf to be used
11 to test Sieve filters that use "vacation".
13 PH/02. Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
14 that changes the way the GnuTLS parameters are stored in the cache file.
15 The new format can be generated externally. For backward compatibility,
16 if the data in the cache doesn't make sense, Exim assumes it has read an
17 old-format file, and it generates new data and writes a new file. This
18 means that you can't go back to an older release without removing the
22 A note about Exim versions 4.44 and 4.50
23 ----------------------------------------
25 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
26 changes of various kinds. As a consequence, a big documentation update was
27 needed. This delayed the release for rather longer than seemed good, especially
28 in the light of a couple of (minor) security issues. Therefore, the changes
29 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
30 release. So 4.44 and 4.50 are in effect two different branches that both start
33 I have left the 4.50 change log unchanged; it contains all the changes since
34 4.43. The change log for 4.44 is below; many of its items are identical to
35 those for 4.50. This seems to be the most sensible way to preserve the
36 historical information.
42 1. Minor wording change to the doc/README.SIEVE file.
44 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
45 computation of the current number of files was incorrect.
47 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
48 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
51 4. Give more explanation in the error message when the command for a transport
52 filter fails to execute.
54 5. There are several places where Exim runs a non-Exim command in a
55 subprocess. The SIGUSR1 signal should be disabled for these processes. This
56 was being done only for the command run by the queryprogram router. It is
57 now done for all such subprocesses. The other cases are: ${run, transport
58 filters, and the commands run by the lmtp and pipe transports.
60 6. Added CONFIGURE_GROUP build-time option.
62 7. Some older OS have a limit of 256 on the maximum number of file
63 descriptors. Exim was using setrlimit() to set 1000 as a large value
64 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
65 systems. I've change it so that if it can't get 1000, it tries for 256.
67 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
68 was an oversight, and furthermore, ever since the addition of extra
69 controls (e.g. 4.43/32), the checks on when to allow different forms of
70 "control" were broken. There should now be diagnostics for all cases when a
71 control that does not make sense is encountered.
73 9. Added the /retain_sender option to "control=submission".
75 10. $recipients is now available in the predata ACL (oversight).
77 11. Tidy the search cache before the fork to do a delivery from a message
78 received from the command line. Otherwise the child will trigger a lookup
79 failure and thereby defer the delivery if it tries to use (for example) a
80 cached ldap connection that the parent has called unbind on.
82 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
83 of $address_data from the recipient verification was clobbered by the
86 13. The value of address_data from a sender verification is now available in
87 $sender_address_data in subsequent conditions in the ACL statement.
89 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
91 15. Added a new option "connect=<time>" to callout options, to set a different
94 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
95 was its contents. (It was OK if the option was not defined at all.)
97 17. A "Completed" log line is now written for messages that are removed from
98 the spool by the -Mrm option.
100 18. New variables $sender_verify_failure and $recipient_verify_failure contain
101 information about exactly what failed.
103 19. Added -dd to debug only the daemon process.
105 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
106 handles timeouts, both on the server side and network timeouts. Renamed the
107 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
109 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
110 It is now set to "smtps".
112 22. $host_address is now set to the target address during the checking of
115 23. When checking ignore_target_hosts for an ipliteral router, no host name was
116 being passed; this would have caused $sender_host_name to have been used if
117 matching the list had actually called for a host name (not very likely,
118 since this list is usually IP addresses). A host name is now passed as
121 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
122 code that specifies a non-restarting handler (typically sigaction() in
123 modern systems) in an attempt to fix a rare and obscure crash bug.
125 25. Narrowed the window for a race in the daemon that could cause it to ignore
126 SIGCHLD signals. This is not a major problem, because they are used only to
127 wake it up if nothing else does.
129 26. A malformed maildirsize file could cause Exim to calculate negative values
130 for the mailbox size or file count. Odd effects could occur as a result.
131 The maildirsize information is now recalculated if the size or filecount
134 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
135 support for a long time. Removed HAVE_SYS_VFS_H.
137 28. Installed the latest version of exipick from John Jetmore.
139 29. In an address list, if the pattern was not a regular expression, an empty
140 subject address (from a bounce message) matched only if the pattern was an
141 empty string. Non-empty patterns were not even tested. This was the wrong
142 because it is perfectly reasonable to use an empty address as part of a
143 database query. An empty address is now tested by patterns that are
144 lookups. However, all the other forms of pattern expect the subject to
145 contain a local part and a domain, and therefore, for them, an empty
146 address still always fails if the pattern is not itself empty.
148 30. Exim went into a mad DNS loop when attempting to do a callout where the
149 host was specified on an smtp transport, and looking it up yielded more
152 31. Re-factored the code for checking spool and log partition space into a
153 function that finds that data and another that does the check. The former
154 is then used to implement four new variables: $spool_space, $log_space,
155 $spool_inodes, and $log_inodes.
157 32. The RFC2047 encoding function was originally intended for short strings
158 such as real names; it was not keeping to the 75-character limit for
159 encoded words that the RFC imposes. It now respects the limit, and
160 generates multiple encoded words if necessary. To be on the safe side, I
161 have increased the buffer size for the ${rfc2047: expansion operator from
164 33. It is now permitted to omit both strings after an "if" condition; if the
165 condition is true, the result is "true". As before, when the second string
166 is omitted, a false condition yields an empty string. This makes it less
167 cumbersome to write custom ACL and router conditions.
169 34. Failure to deliver a bounce message always caused it to be frozen, even if
170 there was an errors_to setting on the router. The errors_to setting is now
173 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
174 canonical form (fully expanded) before being placed in
175 $sender_host_address.
177 36. The table in the code that translates DNS record types into text (T_A to
178 "A" for instance) was missing entries for NS and CNAME. It is just possible
179 that this could have caused confusion if both these types were looked up
180 for the same domain, because the text type is used as part of Exim's
181 per-process caching. But the chance of anyone hitting this buglet seems
184 37. The dnsdb lookup has been extended in a number of ways.
186 (1) There is a new type, "zns", which walks up the domain tree until it
187 finds some nameserver records. It should be used with care.
189 (2) There is a new type, "mxh", which is like "mx" except that it returns
190 just the host names, not the priorities.
192 (3) It is now possible to give a list of domains (or IP addresses) to be
193 looked up. The behaviour when one of the lookups defers can be
194 controlled by a keyword.
196 (4) It is now possible to specify the separator character for use when
197 multiple records are returned.
199 38. The dnslists ACL condition has been extended: it is now possible to supply
200 a list of IP addresses and/or domains to be looked up in a particular DNS
203 39. Added log_selector=+queue_time_overall.
205 40. When running the queue in the test harness, wait just a tad after forking a
206 delivery process, to get repeatability of debugging output.
208 41. Include certificate and key file names in error message when GnuTLS fails
209 to set them up, because the GnuTLS error message doesn't include the name
210 of the failing file when there is a problem reading it.
212 42. Allow both -bf and -bF in the same test run.
214 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
216 44. The "Exiscan patch" is now merged into the mainline Exim source.
218 45. Sometimes the final signoff response after QUIT could fail to get
219 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
220 < 0 before doing a fflush(). This bug looks as though it goes back to the
221 introduction of TLS in release 3.20, but "sometimes" must have been rare
222 because the tests only now provoked it.
224 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
225 (this can affect the format of dates).
227 47. exim_tidydb, when checking for the continued existence of a message for
228 which it has found a message-specific retry record, was not finding
229 messages that were in split spool directories. Consequently, it was
230 deleting retry records that should have stayed in existence.
232 48. Steve fixed some bugs in eximstats.
234 49. The SPA authentication driver was not abandoning authentication and moving
235 on to the next authenticator when an expansion was forced to fail,
236 contradicting the general specification for all authenticators. Instead it
237 was generating a temporary error. It now behaves as specified.
239 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
240 (the order specifies the preference for clients). The order is now AES256,
241 AES128, 3DES, ARCFOUR128.
243 51. Small patch to Sieve code - explicitly set From: when generating an
246 52. Exim crashed if a remote delivery caused a very long error message to be
247 recorded - for instance if somebody sent an entire SpamAssassin report back
248 as a large number of 550 error lines. This bug was coincidentally fixed by
249 increasing the size of one of Exim's internal buffers (big_buffer) that
250 happened as part of the Exiscan merge. However, to be on the safe side, I
251 have made the code more robust (and fixed the comments that describe what
254 53. Now that there can be additional text after "Completed" in log lines (if
255 the queue_time_overall log selector is set), a one-byte patch to exigrep
256 was needed to allow it to recognize "Completed" as not the last thing in
259 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
260 patch that reportedly fixes this has been added. I am not expert enough to
261 create a test for it. This is what the patch creator wrote:
263 "I found a little strange behaviour of ldap code when working with
264 Windows 2003 AD Domain, where users was placed in more than one
265 Organization Units. When I tried to give exim partial DN, the exit code
266 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
267 But simultaneously result of request was absolutely normal ldap result,
268 so I produce this patch..."
270 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
271 REFERENCE, so I have modified the code to exclude the patch when that macro
274 55. Some experimental protocols are using DNS PTR records for new purposes. The
275 keys for these records are domain names, not reversed IP addresses. The
276 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
277 leaves it alone. Component reversal etc. now happens only for IP addresses.
280 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
282 57. Double the size of the debug message buffer (to 2048) so that more of very
283 long debug lines gets shown.
285 58. The exicyclog utility now does better if the number of log files to keep
286 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
288 59. Two changes related to the smtp_active_hostname option:
290 (1) $smtp_active_hostname is now available as a variable.
291 (2) The default for smtp_banner uses $smtp_active_hostname instead
292 of $primary_hostname.
294 60. The host_aton() function is supposed to be passed a string that is known
295 to be a valid IP address. However, in the case of IPv6 addresses, it was
296 not checking this. This is a hostage to fortune. Exim now panics and dies
297 if the condition is not met. A case was found where this could be provoked
298 from a dnsdb PTR lookup with an IPv6 address that had more than 8
299 components; fortuitously, this particular loophole had already been fixed
300 by change 4.50/55 above.
302 If there are any other similar loopholes, the new check in host_aton()
303 itself should stop them being exploited. The report I received stated that
304 data on the command line could provoke the exploit when Exim was running as
305 exim, but did not say which command line option was involved. All I could
306 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
310 61. There was a buffer overflow vulnerability in the SPA authentication code
311 (which came originally from the Samba project). I have added a test to the
312 spa_base64_to_bits() function which I hope fixes it.
315 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
316 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
318 63. The daemon start-up calls getloadavg() while still root for those OS that
319 need the first call to be done as root, but it missed one case: when
320 deliver_queue_load_max is set with deliver_drop_privilege. This is
321 necessary for the benefit of the queue runner, because there is no re-exec
322 when deliver_drop_privilege is set.
324 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
327 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
328 was in use, was not putting the data itself into the right store pool;
329 consequently, it could be overwritten for a subsequent message in the same
330 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
333 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
335 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
336 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
337 the function were treating the return as a boolean value, which happened to
338 work because 0=false and not-0=true, but is not correct code.
340 68. The host_aton() function was not handling scoped IPv6 addresses (those
341 with, for example, "%eth0" on the end) correctly.
343 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
344 negated items (that is, ~something) in unsigned ints. Some compilers
345 apparently mutter when there is no cast.
347 70. If an address verification called from an ACL failed, and did not produce a
348 user-specific message (i.e. there was only a "system" message), nothing was
349 put in $acl_verify_message. In this situation, it now puts the system
352 71. Change 4.23/11 added synchronization checking at the start of an SMTP
353 session; change 4.31/43 added the unwanted input to the log line - except
354 that it did not do this in the start of session case. It now does.
356 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
357 This is wrong and can cause the other end to generate a synchronization
358 error if it is another Exim or anything else that does the synchronization
359 check. A QUIT command is no longer sent after a timeout.
361 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
364 74. The defer_ok option of callout verification was not working if it was used
365 when verifying addresses in header lines, that is, for this case:
367 verify = header_sender/callout=defer_ok
369 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
370 those file descriptors could be used for SMTP connections. If anything
371 wrote to stderr (the example that came up was "warn" in embedded Perl), it
372 could be sent to the SMTP client, causing chaos. The daemon now opens
373 stdin, stdout, and stderr to /dev/null when it puts itself into the
376 76. Arrange for output from Perl's "warn" command to be written to Exim's main
377 log by default. The user can override this with suitable Perl magic.
379 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
380 the log message when discard triggers, was not working for the DATA ACL or
381 for the non-SMTP ACL.
383 78. Error message wording change in sieve.c.
385 79. If smtp_accept_max_per_host was set, the number of connections could be
386 restricted to fewer than expected, because the daemon was trying to set up
387 a new connection before checking whether the processes handling previous
388 connections had finished. The check for completed processes is now done
389 earlier. On busy systems, this bug wouldn't be noticed because something
390 else would have woken the daemon, and it would have reaped the completed
393 80. If a message was submitted locally by a user whose login name contained one
394 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
395 It caused a spool format error. I have fixed the spool reading code. A
396 related problem was that the "from" clause in the Received: line became
397 illegal because of the space(s). It is now covered by ${quote_local_part.
399 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
402 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
404 83. Patch from Sieve maintainer to fix unterminated string problem in
407 84. Some minor changes to the Linux configuration files to help with other
408 OS variants using glibc.
410 85. One more patch for Sieve to update vacation handling to latest spec.
413 ----------------------------------------------------
414 See the note above about the 4.44 and 4.50 releases.
415 ----------------------------------------------------
421 1. Change 4.43/35 introduced a bug that caused file counts to be
422 incorrectly computed when quota_filecount was set in an appendfile
425 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
426 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
429 3. Give more explanation in the error message when the command for a transport
430 filter fails to execute.
432 4. There are several places where Exim runs a non-Exim command in a
433 subprocess. The SIGUSR1 signal should be disabled for these processes. This
434 was being done only for the command run by the queryprogram router. It is
435 now done for all such subprocesses. The other cases are: ${run, transport
436 filters, and the commands run by the lmtp and pipe transports.
438 5. Some older OS have a limit of 256 on the maximum number of file
439 descriptors. Exim was using setrlimit() to set 1000 as a large value
440 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
441 systems. I've change it so that if it can't get 1000, it tries for 256.
443 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
444 was an oversight, and furthermore, ever since the addition of extra
445 controls (e.g. 4.43/32), the checks on when to allow different forms of
446 "control" were broken. There should now be diagnostics for all cases when a
447 control that does not make sense is encountered.
449 7. $recipients is now available in the predata ACL (oversight).
451 8. Tidy the search cache before the fork to do a delivery from a message
452 received from the command line. Otherwise the child will trigger a lookup
453 failure and thereby defer the delivery if it tries to use (for example) a
454 cached ldap connection that the parent has called unbind on.
456 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
457 of $address_data from the recipient verification was clobbered by the
460 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
461 was its contents. (It was OK if the option was not defined at all.)
463 11. A "Completed" log line is now written for messages that are removed from
464 the spool by the -Mrm option.
466 12. $host_address is now set to the target address during the checking of
469 13. When checking ignore_target_hosts for an ipliteral router, no host name was
470 being passed; this would have caused $sender_host_name to have been used if
471 matching the list had actually called for a host name (not very likely,
472 since this list is usually IP addresses). A host name is now passed as
475 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
476 code that specifies a non-restarting handler (typically sigaction() in
477 modern systems) in an attempt to fix a rare and obscure crash bug.
479 15. Narrowed the window for a race in the daemon that could cause it to ignore
480 SIGCHLD signals. This is not a major problem, because they are used only to
481 wake it up if nothing else does.
483 16. A malformed maildirsize file could cause Exim to calculate negative values
484 for the mailbox size or file count. Odd effects could occur as a result.
485 The maildirsize information is now recalculated if the size or filecount
488 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
489 support for a long time. Removed HAVE_SYS_VFS_H.
491 18. Updated exipick to current release from John Jetmore.
493 19. Allow an empty sender to be matched against a lookup in an address list.
494 Previously the only cases considered were a regular expression, or an
497 20. Exim went into a mad DNS lookup loop when doing a callout where the
498 host was specified on the transport, if the DNS lookup yielded more than
501 21. The RFC2047 encoding function was originally intended for short strings
502 such as real names; it was not keeping to the 75-character limit for
503 encoded words that the RFC imposes. It now respects the limit, and
504 generates multiple encoded words if necessary. To be on the safe side, I
505 have increased the buffer size for the ${rfc2047: expansion operator from
508 22. Failure to deliver a bounce message always caused it to be frozen, even if
509 there was an errors_to setting on the router. The errors_to setting is now
512 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
513 canonical form (fully expanded) before being placed in
514 $sender_host_address.
516 24. Updated eximstats to version 1.33
518 25. Include certificate and key file names in error message when GnuTLS fails
519 to set them up, because the GnuTLS error message doesn't include the name
520 of the failing file when there is a problem reading it.
522 26. Expand error message when OpenSSL has problems setting up cert/key files.
525 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
526 (this can affect the format of dates).
528 28. exim_tidydb, when checking for the continued existence of a message for
529 which it has found a message-specific retry record, was not finding
530 messages that were in split spool directories. Consequently, it was
531 deleting retry records that should have stayed in existence.
533 29. eximstats updated to version 1.35
534 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
535 1.35 - bugfix such that pie charts by volume are generated correctly
537 30. The SPA authentication driver was not abandoning authentication and moving
538 on to the next authenticator when an expansion was forced to fail,
539 contradicting the general specification for all authenticators. Instead it
540 was generating a temporary error. It now behaves as specified.
542 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
543 (the order specifies the preference for clients). The order is now AES256,
544 AES128, 3DES, ARCFOUR128.
546 31. Small patch to Sieve code - explicitly set From: when generating an
549 32. Exim crashed if a remote delivery caused a very long error message to be
550 recorded - for instance if somebody sent an entire SpamAssassin report back
551 as a large number of 550 error lines. This bug was coincidentally fixed by
552 increasing the size of one of Exim's internal buffers (big_buffer) that
553 happened as part of the Exiscan merge. However, to be on the safe side, I
554 have made the code more robust (and fixed the comments that describe what
557 33. Some experimental protocols are using DNS PTR records for new purposes. The
558 keys for these records are domain names, not reversed IP addresses. The
559 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
560 leaves it alone. Component reversal etc. now happens only for IP addresses.
563 34. The host_aton() function is supposed to be passed a string that is known
564 to be a valid IP address. However, in the case of IPv6 addresses, it was
565 not checking this. This is a hostage to fortune. Exim now panics and dies
566 if the condition is not met. A case was found where this could be provoked
567 from a dnsdb PTR lookup with an IPv6 address that had more than 8
568 components; fortuitously, this particular loophole had already been fixed
569 by change 4.50/55 or 4.44/33 above.
571 If there are any other similar loopholes, the new check in host_aton()
572 itself should stop them being exploited. The report I received stated that
573 data on the command line could provoke the exploit when Exim was running as
574 exim, but did not say which command line option was involved. All I could
575 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
579 35. There was a buffer overflow vulnerability in the SPA authentication code
580 (which came originally from the Samba project). I have added a test to the
581 spa_base64_to_bits() function which I hope fixes it.
584 36. The daemon start-up calls getloadavg() while still root for those OS that
585 need the first call to be done as root, but it missed one case: when
586 deliver_queue_load_max is set with deliver_drop_privilege. This is
587 necessary for the benefit of the queue runner, because there is no re-exec
588 when deliver_drop_privilege is set.
590 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
591 was in use, was not putting the data itself into the right store pool;
592 consequently, it could be overwritten for a subsequent message in the same
593 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
596 38. Sometimes the final signoff response after QUIT could fail to get
597 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
598 < 0 before doing a fflush(). This bug looks as though it goes back to the
599 introduction of TLS in release 3.20, but "sometimes" must have been rare
600 because the tests only now provoked it.
606 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
607 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
608 Now it returns an integer. A number of calls were still expecting a T/F
609 return. Fortuitously, in all cases, the tests worked in OK situations,
610 which is the norm. However, things would have gone wrong on any write
611 failures on the smtp file descriptor. This function is used when sending
612 messages over SMTP and also when doing verify callouts.
614 2. When Exim is called to do synchronous delivery of a locally submitted
615 message (the -odf or -odi options), it no longer closes stderr before doing
618 3. Implemented the mua_wrapper option.
620 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
622 5. Implemented the functions header_remove(), header_testname(),
623 header_add_at_position(), and receive_remove_recipient(), and exported them
626 6. If an ACL "warn" statement specified the addition of headers, Exim already
627 inserted X-ACL-Warn: at the start if there was no header name. However, it
628 was not making this test for the second and subsequent header lines if
629 there were newlines in the string. This meant that an invalid header could
630 be inserted if Exim was badly configured.
632 7. Allow an ACL "warn" statement to add header lines at the start or after all
633 the Received: headers, as well as at the end.
635 8. Added the rcpt_4xx retry error code.
637 9. Added postmaster_mailfrom=xxx to callout verification option.
639 10. Added mailfrom=xxxx to the callout verification option, for verify=
642 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
643 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
645 12. Inserted some casts to stop certain compilers warning when using pointer
646 differences as field lengths or precisions in printf-type calls (mostly
647 affecting debugging statements).
649 13. Added optional readline() support for -be (dynamically loaded).
651 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
652 same clock tick as a message's arrival, so that its received time was the
653 same as the "first fail" time on the retry record, and that message
654 remained on the queue past the ultimate address timeout, every queue runner
655 would try a delivery (because it was past the ultimate address timeout) but
656 after another failure, the ultimate address timeout, which should have then
657 bounced the address, did not kick in. This was a "< instead of <=" error;
658 in most cases the first failure would have been in the next clock tick
659 after the received time, and all would be well.
661 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
662 being recognized when the domain list was tested by the match_domain
663 condition in an expansion string.
665 16. Added the ${str2b64: operator.
667 17. Exim was always calling setrlimit() to set a large limit for the number of
668 processes, without checking whether the existing limit was already
669 adequate. (It did check for the limit on file descriptors.) Furthermore,
670 errors from getrlimit() and setrlimit() were being ignored. Now they are
671 logged to the main and panic logs, but Exim does carry on, to try to do its
672 job under whatever limits there are.
674 18. Imported PCRE 5.0.
676 19. Trivial typo in log message " temporarily refused connection" (the leading
679 20. If the log selector return_path_on_delivery was set and an address was
680 redirected to /dev/null, the delivery process crashed because it assumed
681 that a return path would always be set for a "successful" delivery. In this
682 case, the whole delivery is bypassed as an optimization, and therefore no
685 21. Internal re-arrangement: the function for sending a challenge and reading
686 a response while authentication was assuming a zero-terminated challenge
687 string. It's now changed to take a pointer and a length, to allow for
688 binary data in such strings.
690 22. Added the cyrus_sasl authenticator (code supplied by MBM).
692 23. Exim was not respecting finduser_retries when seeking the login of the
693 uid under which it was called; it was always trying 10 times. (The default
694 setting of finduser_retries is zero.) Also, it was sleeping after the final
695 failure, which is pointless.
697 24. Implemented tls_on_connect_ports.
699 25. Implemented acl_smtp_predata.
701 26. If the domain in control=submission is set empty, Exim assumes that the
702 authenticated id is a complete email address when it generates From: or
703 Sender: header lines.
705 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
706 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
707 chown and chgrp in /bin and hostname in /usr/bin.
709 28. Exim was keeping the "process log" file open after each use, just as it
710 does for the main log. This opens the possibility of it remaining open for
711 long periods when the USR1 signal hits a daemon. Occasional processlog
712 errors were reported, that could have been caused by this. Anyway, it seems
713 much more sensible not to leave this file open at all, so that is what now
716 29. The long-running daemon process does not normally write to the log once it
717 has entered its main loop, and it closes the log before doing so. This is
718 so that log files can straightforwardly be renamed and moved. However,
719 there are a couple of unusual error situations where the daemon does write
720 log entries, and I had neglected to close the log afterwards.
722 30. The text of an SMTP error response that was received during a remote
723 delivery was being truncated at 512 bytes. This is too short for some of
724 the long messages that one sometimes sees. I've increased the limit to
727 31. It is now possible to make retry rules that apply only when a message has a
728 specific sender, in particular, an empty sender.
730 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
731 it possible to be selective about when SMTP synchronization is enforced.
733 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
735 32. Implemented hosts_connection_nolog.
737 33. Added an ACL for QUIT.
739 34. Setting "delay_warning=" to disable warnings was not working; it gave a
742 35. Added mailbox_size and mailbox_filecount to appendfile.
744 36. Added control = no_multiline_responses to ACLs.
746 37. There was a bug in the logic of the code that waits for the clock to tick
747 in the case where the clock went backwards by a substantial amount such
748 that the microsecond fraction of "now" was more than the microsecond
749 fraction of "then" (but the whole seconds number was less).
751 38. Added support for the libradius Radius client library this is found on
752 FreeBSD (previously only the radiusclient library was supported).
758 1. When certain lookups returned multiple values in the form name=value, the
759 quoting of the values was not always being done properly. Specifically:
760 (a) If the value started with a double quote, but contained no whitespace,
762 (b) If the value contained whitespace other than a space character (i.e.
763 tabs or newlines or carriage returns) it was not quoted.
764 This fix has been applied to the mysql and pgsql lookups by writing a
765 separate quoting function and calling it from the lookup code. The fix
766 should probably also be applied to nisplus, ibase and oracle lookups, but
767 since I cannot test any of those, I have not disturbed their existing code.
769 2. A hit in the callout cache for a specific address caused a log line with no
770 reason for rejecting RCPT. Now it says "Previous (cached) callout
771 verification failure".
773 3. There was an off-by-one bug in the queryprogram router. An over-long
774 return line was truncated at 256 instead of 255 characters, thereby
775 overflowing its buffer with the terminating zero. As well as fixing this, I
776 have increased the buffer size to 1024 (and made a note to document this).
778 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
779 when Exim is waiting for an SMTP response from a remote server, Exim
780 restarts its select() call on the socket, thereby resetting its timeout.
781 This is not a problem when such interrupts are rare. Somebody set up a cron
782 job to run exiwhat every 2 minutes, which is less than the normal select()
783 timeout (5 or 10 minutes). This meant that the select() timeout never
784 kicked in because it was always reset. I have fixed this by comparing the
785 time when an interrupt arrives with the time at the start of the first call
786 to select(). If more time than the timeout has elapsed, the interrupt is
787 treated as a timeout.
789 5. Some internal re-factoring in preparation for the addition of Sieve
790 extensions (by MH). In particular, the "personal" test is moved to a
791 separate function, and given an option for scanning Cc: and Bcc: (which is
792 not set for Exim filters).
794 6. When Exim created an email address using the login of the caller as the
795 local part (e.g. when creating a From: or Sender: header line), it was not
796 quoting the local part when it contained special characters such as @.
798 7. Installed new OpenBSD configuration files.
800 8. Reworded some messages for syntax errors in "and" and "or" conditions to
801 try to make them clearer.
803 9. Callout options, other than the timeout value, were being ignored when
804 verifying sender addresses in header lines. For example, when using
806 verify = header_sender/callout=no_cache
808 the cache was (incorrectly) being used.
810 10. Added a missing instance of ${EXE} to the exim_install script; this affects
811 only the Cygwin environment.
813 11. When return_path_on_delivery was set as a log selector, if different remote
814 addresses in the same message used different return paths and parallel
815 remote delivery occurred, the wrong values would sometimes be logged.
816 (Whenever a remote delivery process finished, the return path value from
817 the most recently started remote delivery process was logged.)
819 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
820 lines when AUTH and/or TLS are in use. This is the "received protocol"
821 field. Exim used to use "asmtp" for authenticated SMTP, without any
822 indication (in the protocol name) for TLS use. Now it follows the RFC and
823 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
824 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
825 names appear in log lines as well as in Received: header lines.
827 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
828 extensions, and comparison tests, and to fix some bugs.
830 14. Changes to the "personal" filter test:
832 (1) The test was buggy in that it was just doing the equivalent of
833 "contains" tests on header lines. For example, if a user's address was
834 anne@some.where, the "personal" test would incorrectly be true for
836 To: susanne@some.where
838 This test is now done by extracting each address from the header in turn,
839 and checking the entire address. Other tests that are part of "personal"
840 are now done using regular expressions (for example, to check local parts
841 of addresses in From: header lines).
843 (2) The list of non-personal local parts in From: addresses has been
844 extended to include "listserv", "majordomo", "*-request", and "owner-*",
845 taken from the Sieve specification recommendations.
847 (3) If the message contains any header line starting with "List-" it is
848 treated as non-personal.
850 (4) The test for "circular" in the Subject: header line has been removed
851 because it now seems ill-conceived.
853 15. Minor typos in src/EDITME comments corrected.
855 16. Installed latest exipick from John Jetmore.
857 17. If headers_add on a router specified a text string that was too long for
858 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
859 of string_sprintf() is now avoided.
861 18. $message_body_size was not set (it was always zero) when running the DATA
862 ACL and the local_scan() function.
864 19. For the "mail" command in an Exim filter, no default was being set for
865 the once_repeat time, causing a random time value to be used if "once" was
866 specified. (If the value happened to be <= 0, no repeat happened.) The
867 default is now 0s, meaning "never repeat". The "vacation" command was OK
868 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
869 (I found it when inspecting the code).
871 20. There is now an overall timeout for performing a callout verification. It
872 defaults to 4 times the callout timeout, which applies to individual SMTP
873 commands during the callout. The overall timeout applies when there is more
874 than one host that can be tried. The timeout is checked before trying the
875 next host. This prevents very long delays if there are a large number of
876 hosts and all are timing out (e.g. when the network connections are timing
877 out). The value of the overall timeout can be changed by specifying an
878 additional sub-option for "callout", called "maxwait". For example:
880 verify = sender/callout=5s,maxwait=20s
882 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
883 to the end before writing, but this should make it even safer).
885 22. Exim was forgetting that it had advertised PIPELINING for the second and
886 subsequent messages on an SMTP connection. It was also not resetting its
887 memory on STARTTLS and an internal HELO.
889 23. When Exim logs an SMTP synchronization error within a session, it now
890 records whether PIPELINING has been advertised or not.
892 24. Added 3 instances of "(long int)" casts to time_t variables that were being
893 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
894 rather than long int.
896 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
898 26. Added the never_mail option to autoreply.
904 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
905 crash if the getsockname() call failed; this can happen if a connection is
906 closed very soon after it is established. The problem was simply in the
907 order in which certain operations were done, causing Exim to try to write
908 to the SMTP stream before it had set up the file descriptor. The bug has
909 been fixed by making things happen in the correct order.
915 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
916 before the connection was closed, thus losing the rejection response.
918 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
919 some early Solaris releases, but causes trouble in current releases where
920 socklen_t is defined.
922 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
925 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
928 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
929 that releasing the top of it at the end releases what was used for sub-
930 expansions (unless the block got too big). However, discard this block if
931 the first thing is a variable or header, so that we can use its block when
932 it is dynamic (useful for very large $message_headers, for example).
934 6. Lookups now cache *every* query, not just the most recent. A new, separate
935 store pool is used for this. It can be recovered when all lookup caches are
936 flushed. Lookups now release memory at the end of their result strings.
937 This has involved some general refactoring of the lookup sources.
939 7. Some code has been added to the store_xxx() functions to reduce the amount
940 of flapping under certain conditions.
942 8. log_incoming_interface used to affect only the <= reception log lines. Now
943 it causes the local interface and port to be added to several more SMTP log
944 lines, for example "SMTP connection from", and rejection lines.
946 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
948 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
950 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
951 could be overwritten at the end of the current message (or the start of a
952 new message if it was set in a HELO ACL). The value is now preserved for
953 the duration of the SMTP connection.
955 12. If a transport had a headers_rewrite setting, and a matching header line
956 contained an unqualified address, that address was qualified, even if it
957 did not match any rewriting rules. The underlying bug was that the values
958 of the flags that permit the existence of unqualified sender and recipient
959 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
960 non-local messages, and by -bnq for local messages) were not being
961 preserved with the message after it was received.
963 13. When Exim was logging an SMTP synchronization error, it could sometimes log
964 "next input=" as part of the text comprising the host identity instead of
965 the correct text. The code was using the same buffer for two different
966 strings. However, depending on which order the printing function evaluated
967 its arguments, the bug did not always show up. Under Linux, for example, my
968 test suite worked just fine.
970 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
971 doesn't work with some older versions of Perl. It has been changed to "my",
972 which in any case is probably the better facility to use.
974 15. A really picky compiler found some instances of statements for creating
975 error messages that either had too many or two few arguments for the format
978 16. The size of the buffer for calls to the DNS resolver has been increased
979 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
980 for addresses that have a lot of PTR records. This alleviates a problem; it
981 does not fully solve it.
983 17. A dnsdb lookup for PTR records that receives more data than will fit in the
984 buffer now truncates the list and logs the incident, which is the same
985 action as happens when Exim is looking up a host name and its aliases.
986 Previously in this situation something unpredictable would happen;
987 sometimes it was "internal error: store_reset failed".
989 18. If a server dropped the connection unexpectedly when an Exim client was
990 using GnuTLS and trying to read a response, the client delivery process
991 crashed while trying to generate an error log message.
993 19. If a "warn" verb in an ACL added multiple headers to a message in a single
996 warn message = H1: something\nH2: something
998 the text was added as a single header line from Exim's point of view
999 though it ended up OK in the delivered message. However, searching for the
1000 second and subsequent header lines using $h_h2: did not work. This has been
1001 fixed. Similarly, if a system filter added multiple headers in this way,
1002 the routers could not see them.
1004 20. Expanded the error message when iplsearch is called with an invalid key to
1005 suggest using net-iplsearch in a host list.
1007 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1008 ACLs are no longer actually imposed (and a message to that effect is
1011 22. If a "gecos" field in a passwd entry contained escaped characters, in
1012 particular, if it contained a \" sequence, Exim got it wrong when building
1013 a From: or a Sender: header from that name. A second bug also caused
1014 incorrect handling when an unquoted " was present following a character
1015 that needed quoting.
1017 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1018 was not being matched caselessly.
1020 24. Arranged for all hyphens in the exim.8 source to be escaped with
1023 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1024 properly. Recipient callout cache records were still being keyed to include
1025 the sender, even when use_sender was set false. This led to far more
1026 callouts that were necessary. The sender is no longer included in the key
1027 when use_sender is false.
1029 26. Added "control = submission" modifier to ACLs.
1031 27. Added the ${base62d: operator to decode base 62 numbers.
1033 28. dnsdb lookups can now access SRV records.
1035 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1036 the configuration file.
1038 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1039 (-v) mode. This makes the output for a verbose queue run more intelligible.
1041 31. Added a use_postmaster feature to recipient callouts.
1043 32. Added the $body_zerocount variable, containing the number of binary zero
1044 bytes in the message body.
1046 33. The time of last modification of the "new" subdirectory is now used as the
1047 "mailbox time last read" when there is a quota error for a maildir
1050 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1052 35. Added +ignore_unknown as a special item in host lists.
1054 36. Code for decoding IPv6 addresses in host lists is now included, even if
1055 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1056 address was recognized as an IP address, but was then not correctly decoded
1057 into binary, causing unexpected and incorrect effects when compared with
1064 1. Very minor rewording of debugging text in manualroute to say "list of
1065 hosts" instead of "hostlist".
1067 2. If verify=header_syntax was set, and a header line with an unqualified
1068 address (no domain) and a large number of spaces between the end of the
1069 name and the colon was received, the reception process suffered a buffer
1070 overflow, and (when I tested it) crashed. This was caused by some obsolete
1071 code that should have been removed. The fix is to remove it!
1073 3. When running in the test harness, delay a bit after writing a bounce
1074 message to get a bit more predictability in the log output.
1076 4. Added a call to search_tidyup() just before forking a reception process. In
1077 theory, someone could use a lookup in the expansion of smtp_accept_max_
1078 per_host which, without the tidyup, could leave open a database connection.
1080 5. Added the variables $recipient_data and $sender_data which get set from a
1081 lookup success in an ACL "recipients" or "senders" condition, or a router
1082 "senders" option, similar to $domain_data and $local_part_data.
1084 6. Moved the writing of debug_print from before to after the "senders" test
1087 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1088 problems for message scanning, either using a data ACL, or using
1089 local_scan() because the Received: header was not generated till after they
1090 were called (in order to set the time as the time of reception completion).
1091 I have revised the way this works. The header is now generated after the
1092 body is received, but before the ACL or local_scan() are called. After they
1093 are run, the timestamp in the header is updated.
1099 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1100 before starting a queue runner without re-exec. This happened only when
1101 deliver_drop_privilege was set or when the Exim user was set to root. The
1102 effect of the bug was that timeouts during subsequent deliveries caused
1103 crashes instead of being properly handled. The handler is now left at its
1104 default (and expected) setting.
1106 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1107 message, again when deliver_drop_privilege is set or Exim is run as root.
1108 The bug described in (1) was not present in this case, but the tidying up
1109 of the other signals was missing. I have made the two cases consistent.
1111 3. The ignore_target_hosts setting on a manualroute router was being ignored
1112 for hosts that were looked up using the /MX notation.
1114 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1117 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1118 operated on the sender address. After changing the $sender_address to <>
1119 for the sender address verify, Exim was re-instated it as the original
1120 (before rewriting) address, but remembering that it had rewritten it, so it
1121 wasn't rewriting it again. This bug also had the effect of breaking the
1122 sender address verification caching when the sender address was rewritten.
1124 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1125 This has been changed so that if the ip literal address matches
1126 ignore_target_hosts, the router declines.
1128 7. Added expansion conditions match_domain, match_address, and match_local_
1129 part (NOT match_host).
1131 8. The placeholder for the Received: header didn't have a length field set.
1133 9. Added code to Exim itself and to exim_lock to test for a specific race
1134 condition that could lead to file corruption when using MBX delivery. The
1135 issue is with the lockfile that is created in /tmp. If this file is removed
1136 after a process has opened it but before that process has acquired a lock,
1137 there is the potential for a second process to recreate the file and also
1138 acquire a lock. This could lead to two Exim processes writing to the file
1139 at the same time. The added code performs the same test as UW imapd; it
1140 checks after acquiring the lock that its file descriptor still refers to
1141 the same named file.
1143 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1144 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1147 11. Added the smtp_active_hostname option. If used, this will typically be made
1148 to depend on the incoming interface address. Because $interface_address is
1149 not set up until the daemon has forked a reception process, error responses
1150 that can happen earlier (such as "too many connections") no longer contain
1153 12. If an expansion in a condition on a "warn" statement fails because a lookup
1154 defers, the "warn" statement is abandoned, and the next ACL statement is
1155 processed. Previously this caused the whole ACL to be aborted.
1157 13. Added the iplsearch lookup type.
1159 14. Added ident_timeout as a log selector.
1161 15. Added tls_certificate_verified as a log selector.
1163 16. Added a global option tls_require_ciphers (compare the smtp transport
1164 option of the same name). This controls incoming TLS connections.
1166 17. I finally figured out how to make tls_require_ciphers do a similar thing
1167 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1168 before starting the TLS session.
1170 18. Tabs are now shown as \t in -bP output.
1172 19. If the log selector return_path_on_delivery was set, Exim crashed when
1173 bouncing a message because it had too many Received: header lines.
1175 20. If two routers both had headers_remove settings, and the first one included
1176 a superfluous trailing colon, the final name in the first list and the
1177 first name in the second list were incorrectly joined into one item (with a
1178 colon in the middle).
1184 1. Added -C and -D options to the exinext utility, mainly to make it easier
1185 to include in the automated testing, but these could be helpful when
1186 multiple configurations are in use.
1188 2. The exinext utility was not formatting the output nicely when there was
1189 an alternate port involved in the retry record key, nor when there was a
1190 message id as well (for retries that were specific to a specific message
1191 and a specific host). It was also confused by IPv6 addresses, because of
1192 the additional colons they contain. I have fixed the IPv4 problem, and
1193 patched it up to do a reasonable job for IPv6.
1195 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1196 delivery, the log line now contains "pipelined" if PIPELINING was used.
1198 4. An SMTP transport process used to panic and die if the bind() call to set
1199 an explicit outgoing interface failed. This has been changed; it is now
1200 treated in the same way as a connect() failure.
1202 5. A reference to $sender_host_name in the part of a conditional expansion
1203 that was being skipped was still causing a DNS lookup. This no longer
1206 6. The def: expansion condition was not recognizing references to header lines
1207 that used bh_ and bheader_.
1209 7. Added the _cache feature to named lists.
1211 8. The code for checking quota_filecount in the appendfile transport was
1212 allowing one more file than it should have been.
1214 9. For compatibility with Sendmail, the command line option
1222 and sets the incoming protocol and host name (for trusted callers). The
1223 host name and its colon can be omitted when only the protocol is to be set.
1224 Note the Exim already has two private options, -pd and -ps, that refer to
1225 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1226 "s", but I don't think that's a major issue.
1228 10. A number of refactoring changes to the code, none of which should affect
1231 (a) The number of logging options was getting close to filling up the
1232 32-bit word that was used as a bit map. I have split them into two classes:
1233 those that are passed in the argument to log_write(), and those that are
1234 only ever tested independently outside of that function. These are now in
1235 separate 32-bit words, so there is plenty of room for expansion again.
1236 There is no change in the user interface or the logging behaviour.
1238 (b) When building, for example, log lines, the code previously used a
1239 macro that called string_cat() twice, in order to add two strings. This is
1240 not really sufficiently general. Furthermore, there was one instance where
1241 it was actually wrong because one of the argument was used twice, and in
1242 one call a function was used. (As it happened, calling the function twice
1243 did not affect the overall behaviour.) The macro has been replaced by a
1244 function that can join an arbitrary number of extra strings onto a growing
1247 (c) The code for expansion conditions now uses a table and a binary chop
1248 instead of a serial search (which was left over from when there were very
1249 few conditions). Also, it now recognizes conditions like "pam" even when
1250 the relevant support is not compiled in: a suitably worded error message is
1251 given if an attempt is made to use such a condition.
1253 11. Added ${time_interval:xxxxx}.
1255 12. A bug was causing one of the ddress fields not to be passed back correctly
1256 from remote delivery subprocesses. The field in question was not being
1257 subsequently used, so this caused to problems in practice.
1259 13. Added new log selectors queue_time and deliver_time.
1261 14. Might have fixed a bug in maildirsizefile handling that threw up
1262 "unexpected character" debug warnings, and recalculated the data
1263 unnecessarily. In any case, I expanded the warning message to give more
1266 15. Added the message "Restricted characters in address" to the statements in
1267 the default ACL that block characters like @ and % in local parts.
1269 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1270 Three changes have been made:
1272 (a) There was a serious bug; a negative response to MAIL caused the whole
1273 recipient domain to be cached as invalid, thereby blocking all messages
1274 to all local parts at the same domain, from all senders. This bug has
1275 been fixed. The domain is no longer cached after a negative response to
1276 MAIL if the sender used is not empty.
1278 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1281 (c) A new callout option, "use_sender" has been added for people who want
1282 the modified behaviour.
1288 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1291 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1292 indeed breaks things for older releases.
1294 3. Added additional logging to the case where there is a problem reading data
1295 from a filter that is running in a subprocess using a pipe, in order to
1296 try to track down a specific problem.
1298 4. Testing facility fudge: when running in the test harness and attempting
1299 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1300 getting "No route to host". Convert this to a timeout.
1302 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1305 6. Some OS don't have socklen_t but use size_t instead. This affects the
1306 fifth argument of getsockopt() amongst other things. This is now
1307 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1308 can be set for individual OS. I have set it for SunOS5, OSF1, and
1309 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1310 some earlier ones do not.
1312 7. Change 4.30/15 was not doing the test caselessly.
1314 8. The standard form for an IPv6 address literal was being rejected by address
1315 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1316 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1317 this, as well as the form without the "IPv6" on the front (but only when
1318 address literals are enabled, of course).
1320 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1322 10. Exim crashed if a message with an empty sender address specified by -f
1323 encountered a router with an errors_to setting. This could be provoked only
1324 by a command such as
1328 where an empty string was supplied; "<>" did not hit this bug.
1330 11. Installed PCRE release 4.5.
1332 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1333 remained set. It is now erased.
1335 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1336 times from message ids (which are base 36 rather than the normal 62).
1338 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1339 were being counted as actual protocol errors, and logged if the log
1340 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1341 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1342 and DATA following a set of rejected RCPTs do not count as protocol errors.
1343 In other words, Exim assumes they were pipelined, though this may not
1344 actually be the case. Of course, in all cases the client gets an
1345 appropriate error code.
1347 15. If a lookup fails in an ACL condition, a message about the failure may
1348 be available; it is used if testing the ACL cannot continue, because most
1349 such messages specify what the cause of the deferral is. However, some
1350 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1351 that caused an old message to be retained and used if a later statement
1352 caused a defer, replacing the real cause of the deferral.
1354 16. If an IP address had so many PTR records that the DNS lookup buffer
1355 was not large enough to hold them, Exim could crash while trying to process
1356 the truncated data. It now detects and logs this case.
1358 17. Further to 4.21/58, another change has been made: if (and only if) the
1359 first line of a message (the first header line) ends with CRLF, a bare LF
1360 in a subsequent header line has a space inserted after it, so as not to
1361 terminate the header.
1363 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1364 unnecessarily, used atoi() instead of strtol(), and didn't check the
1365 termination when getting file sizes from file names by regex.
1367 19. Completely re-implemented the support for maildirsize files, in the light
1368 of a number of problems with the previous contributed implementation
1369 (4.30/29). In particular:
1371 . If the quota is zero, the maildirsize file is maintained, but no quota is
1374 . If the maildir directory does not exist, it is created before any attempt
1375 to write a maildirsize file.
1377 . The quota value in the file is just a cache; if the quota is changed in
1378 the transport, the new value overrides.
1380 . A regular expression is available for excluding directories from the
1383 20. The autoreply transport checks the characters in options that define the
1384 message's headers; it allows continued headers, but it was checking with
1385 isspace() after an embedded newline instead of explicitly looking for a
1388 21. If all the "regular" hosts to which an address was routed had passed their
1389 expiry times, and had not reached their retry times, the address was
1390 bounced, even if fallback hosts were defined. Now Exim should go on to try
1393 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1394 equivalent code in the SMTP transport. Some hosts send humungous responses
1395 to HELO/EHLO, more than 1024 it seems.
1397 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1398 gives compiler warnings in some environments. I've now done it "properly",
1401 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1402 (because of problems with the built-in one) was declared to return uschar *
1403 instead of char *, causing compiler failure.
1405 25. Fixed a file descriptor leak when processing alias/forward files.
1407 26. Fixed a minor format string issue in dbfn.c.
1409 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1411 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1412 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1415 29. When a sender address is verified, it is cached, to save repeating the test
1416 when there is more than one recipient in a message. However, when the
1417 verification involves a callout, it is possible for different callout
1418 options to be set for different recipients. It is too complicated to keep
1419 track of this in the cache, so now Exim always runs a verification when a
1420 callout is required, relying on the callout cache for the optimization.
1421 The overhead is duplication of the address routing, but this should not be
1424 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1425 to be verified with callout=postmaster, and the main callout worked but the
1426 postmaster check failed, the verification correctly failed. However, if a
1427 subsequent RCPT command asked for sender verification *without* the
1428 postmaster check, incorrect caching caused this verification also to fail,
1431 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1432 it was not caching the DNS options (qualify_single, search_parents) that
1433 were used when the lookup failed. A subsequent lookup with different
1434 options therefore always gave the same answer, though there were cases
1435 where it should not have. (Example: a "domains = !$mx_any" option on a
1436 dnslookup router: the "domains" option is always processed without any
1437 widening, but the router might have qualify_single set.) Now Exim uses the
1438 cached value only when the same options are set.
1440 32. Added John Jetmore's "exipick" utility to the distribution.
1442 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1443 than a timeout (e.g. a certificate is required, and is not provided), an
1444 Exim server now closes the connection immediately. Previously it waited for
1445 the client to close - but if the client is SSL, it seems that they each
1446 wait for each other, leading to a delay before one of them times out.
1448 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1449 maintained 0.8.x compatibility because I don't think many are using it, and
1450 it is clearly obsolete.
1452 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1455 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1456 client certificate was expired. A simple patch fixes this, though I don't
1457 understand the full logic of why the verify callback is called multiple
1460 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1461 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1462 which causes problems with some clients (such as the Certicom SSL Plus
1463 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1464 disables the coutermeasure allowing Eudora to connect."
1466 38. Exim was not checking that a write() to a log file succeeded. This could
1467 lead to Bad Things if a log got too big, in particular if it hit a file
1468 size limit. Exim now panics and dies if it cannot write to a log file, just
1469 as it does if it cannot open a log file.
1471 39. Modified OS/Makefile-Linux so that it now contains
1473 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1475 The two -D definitions ensure that Exim is compiled with large file
1476 support, which makes it possible to handle log files that are bigger than
1479 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1480 instance) a domain was checked against a named list that involved a lookup,
1481 causing $domain_data to be set, then another domain was checked against the
1482 same list, then the first domain was re-checked, the value of $domain_data
1483 after the final check could be wrong. In particular, if the second check
1484 failed, it could be set empty. This bug probably also applied to
1487 41. The strip_trailing_dot option was not being applied to the address given
1488 with the -f command-line option.
1490 42. The code for reading a message's header from the spool was incrementing
1491 $received_count, but never initializing it. This meant that the value was
1492 incorrect (doubled) while delivering a message in the same process in which
1493 it was received. In the most common configuration of Exim, this never
1494 happens - a fresh exec is done - but it can happen when
1495 deliver_drop_privilege is set.
1497 43. When Exim logs an SMTP synchronization error - client data sent too soon -
1498 it now includes up to 150 characters of the unexpected data in the log
1501 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1502 and building data strings. The size of both of these buffers was 10 000
1503 bytes - far larger than anybody would *ever* want, thought I. Needless to
1504 say, somebody hit the limit. I have increased the maximum line length to
1505 20 000 and the maximum data length of concatenated lines to 100 000. I have
1506 also fixed two bugs, because there was no checking on these buffers. Tsk,
1507 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1508 buffer is too small.
1510 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1511 lsearch lookups. Now it does.
1513 46. When parsing a route_list item in a manualroute router, a fixed-length
1514 buffer was used for the list of hosts. I made this 1024 bytes long,
1515 thinking that nobody would ever have a list of hosts that long. Wrong.
1516 Somebody had a whole pile of complicated expansion conditions, and the
1517 string was silently truncated, leading to an expansion error. It turns out
1518 that it is easier to change to an unlimited length (owing to other changes
1519 that have happened since this code was originally written) than to build
1520 structure for giving a limitation error. The length of the item that
1521 expands into the list of hosts is now unlimited.
1523 47. The lsearch lookup could not handle data where the length of text line was
1524 more than 4095 characters. Such lines were truncated, leading to shortened
1525 data being returned. It should now handle lines of any length.
1527 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1528 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1531 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
1532 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1533 no longer happens in comment lines. A list of replacements is now placed
1534 at the head of all of the source files, except those whose only change is
1535 to replace PERL_COMMAND in the very first #! line.
1537 50. Replaced the slow insertion sort in queue.c, for sorting the list of
1538 messages on the queue, with a bottom-up merge sort, using code contributed
1539 by Michael Haardt. This should make operations like -bp somewhat faster on
1540 large queues. It won't affect queue runners, except when queue_run_in_order
1543 51. Installed eximstats 1.31 in the distribution.
1545 52. Added support for SRV lookups to the dnslookup router.
1547 53. If an ACL referred to $message_body or $message_body_end, the value was not
1548 reset for any messages that followed in the same SMTP session.
1550 54. The store-handling optimization for building very long strings was not
1551 differentiating between the different store pools. I don't think this
1552 actually made any difference in practice, but I've tidied it.
1554 55. While running the routers to verify a sender address, $sender_address
1555 was still set to the sender address. This is wrong, because when routing to
1556 send a bounce to the sender, it would be empty. Therefore, I have changed
1557 it so that, while verifying a sender address, $sender_address is set to <>.
1558 (There is no change to what happens when verifying a recipient address.)
1560 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1561 target A or AAAA records (if not already returned) without resetting the
1562 qualify_single or search_parents options of the DNS resolver. These are
1563 inappropriate in this case because the targets of MX and SRV records must
1564 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1565 target that, when qualified, matched something in the local domain. These
1566 two options are now turned off when doing these lookups.
1568 57. It seems that at least some releases of Reiserfs (which does not have the
1569 concept of a fixed number of inodes) returns zero and not -1 for the
1570 number of available inodes. This interacted badly with check_spool_inodes,
1571 which assumed that -1 was the "no such thing" setting. What I have done is
1572 to check that the total number of inodes is greater than zero before doing
1573 the test of how many are available.
1575 58. When a "warn" ACL statement has a log_message modifier, the message is
1576 remembered, and not repeated. This is to avoid a lot of repetition when a
1577 message has many recipients that cause the same warning to be written.
1578 Howewer, Exim was preserving the list of already written lines for an
1579 entire SMTP session, which doesn't seem right. The memory is now reset if a
1580 new message is started.
1582 59. The "rewrite" debugging flag was not showing the result of rewriting in the
1583 debugging output unless log_rewrite was also set.
1585 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1586 of (int)(handle) when we know that handle contains (void *)(-1).
1588 61. The Exim daemon panic-logs an error return when it closes the incoming
1589 connection. However "connection reset by peer" seems to be common, and
1590 isn't really an error worthy of noting specially, so that particular error
1593 62. When Exim is trying to find all the local interfaces, it used to panic and
1594 die if the ioctl to get the interface flags failed. However, it seems that
1595 on at least one OS (Solaris 9) it is possible to have an interface that is
1596 included in the list of interfaces, but for which you get a failure error
1597 for this call. This happens when the interface is not "plumbed" into a
1598 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1599 failure of the "get flags" call assumes that the interface is down.
1601 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
1602 makes life easier for people who are doing arithmetic on fields extracted
1603 from dates, where you often get leading zeros that should not be
1604 interpreted as octal.
1606 64. Added qualify_domain to the redirect router, to override the global
1609 65. If a pathologically long header line contained very many addresses (the
1610 report of this problem mentioned 10 000) and each of them was rewritten,
1611 Exim could use up a very large amount of memory. (It kept on making new
1612 copies of the header line as it rewrote, and never released the old ones.)
1613 At the expense of a bit more processing, the header rewriting function has
1614 been changed so that it no longer eats memory in this way.
1616 66. The generation of the Received: header has been moved from the time that a
1617 message starts to be received, to the time that it finishes. The timestamp
1618 in the Received: header should now be very close to that of the <= log
1619 line. There are two side-effects of this change:
1621 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1622 logged header lines no longer include the local Received: line, because
1623 it has not yet been created. The same applies to a copy of the message
1624 that is returned to a non-SMTP sender when a message is rejected.
1626 (b) When a filter file is tested using -bf, no additional Received: header
1627 is added to the test message. After some thought, I decided that this
1630 This change does not affect the value of $received_for. It is still set
1631 after address rewriting, but before local_scan() is called.
1633 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1635 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1636 gave an unhelpful panic error message, and a defer error. I have managed to
1637 change this behaviour so that it now rejects any supplied certificate,
1638 which seems right, as the list of acceptable certificates is empty.
1640 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1641 gave an unhelpful defer error. I have not managed to make this reject any
1642 supplied certificates, but the error message it gives is "no certificate
1643 supplied", which is not helpful.
1645 70. exigrep's output now also includes lines that are not associated with any
1646 message, but which match the given pattern. Implemented by a patch from
1647 Martin Sluka, which also tidied up the Perl a bit.
1649 71. Recipient callout verification, like sender verification, was using <> in
1650 the MAIL FROM command. This isn't really the right thing, since the actual
1651 sender may affect whether the remote host accepts the recipient or not. I
1652 have changed it to use the actual sender in the callout; this means that
1653 the cache record is now keyed on a recipient/sender pair, not just the
1654 recipient address. There doesn't seem to be a real danger of callout loops,
1655 since a callout by the remote host to check the sender would use <>.
1656 [SEE ABOVE: changed after hitting problems.]
1658 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1659 temporary errors. However, in the case of such a code being given after
1660 the end of a data transmission (i.e. after ".") Exim was failing to write
1661 a retry record for the message. (Yes, there was some broken host that was
1662 actually sending 8xx at this point.)
1664 73. An unknown lookup type in a host list could cause Exim to panic-die when
1665 the list was checked. (An example that provoked this was putting <; in the
1666 middle of a list instead of at the start.) If this happened during a DATA
1667 ACL check, a -D file could be left lying around. This kind of configuration
1668 error no longer causes Exim to die; instead it causes a defer errror. The
1669 incident is still logged to the main and panic logs.
1671 74. Buglet left over from Exim 3 conversion. The message "too many messages
1672 in one connection" was written to the rejectlog but not the mainlog, except
1673 when address rewriting (yes!) was being logged.
1675 75. Added write_rejectlog option.
1677 76. When a system filter was run not as root (that is, when system_filter_user
1678 was set), the values of the $n variables were not being returned to the
1679 main process; thus, they were not subsequently available in the $sn
1682 77. Added +return_path_on_delivery log selector.
1684 78. A connection timeout was being treated differently from recipients deferred
1685 when testing hosts_max_try with a message that was older than the host's
1686 retry timeout. (The host should not be counted, thus allowing all hosts to
1687 be tried at least once before bouncing.) This may have been the cause of an
1688 occasionally reported bug whereby a message would remain on the queue
1689 longer than the retry timeout, but would be bounced if a delivery was
1690 forced. I say "may" because I never totally pinned down the problem;
1691 setting up timeout/retry tests is difficult. See also the next item.
1693 79. The ultimate address timeout was not being applied to errors that involved
1694 a combination of host plus message (for example, a timeout on a MAIL
1695 command). When an address resolved to a number of possible hosts, and they
1696 were not all tried for each delivery (e.g. because of hosts_max_try), a
1697 message could remain on the queue longer than the retry timeout.
1699 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1702 81. Fixed an obscure SMTP outgoing bug which required at least the following
1703 conditions: (a) there was another message waiting for the same server;
1704 (b) the server returned 5xx to all RCPT commands in the first message so
1705 that the message was not completed; (c) the server dropped the connection
1706 or gave a negative response to the RSET that Exim sends to abort the
1707 transaction. The observed case was a dropped connection after DATA that had
1708 been sent in pipelining mode. That is, the server had advertised PIPELINING
1709 but was not implementing it correctly. The effect of the bug was incorrect
1710 behaviour, such as trying another host, and this could lead to a crash.
1716 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1717 and daemon.c were passed as pointers to ints; they should have been
1718 pointers to socklen_t variables (which are typically unsigned ints).
1720 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1723 3. Fixed a really odd bug that affected only the testing scheme; patching a
1724 certain fixed string in the binary changed the value of another string that
1725 happened to be identical to the end of the original first string.
1727 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1728 name", it returns that address as the IP address. On some operating
1729 systems (e.g. Solaris), it also passes back the IP address string as the
1730 "host name". However, on others (e.g. Linux), it passes back an empty
1731 string. Exim wasn't checking for this, and was changing the host name to an
1732 empty string, assuming it had been canonicized.
1734 5. Although rare, it is permitted to have more than one PTR record for a given
1735 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1736 all the names associated with an address, because they do in Solaris.
1737 However, it seems that they do not in Linux for data that comes from the
1738 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1739 I found this out when I moved to a new Linux workstation and tried to run
1740 the Exim test suite.
1742 To get round this problem I have changed the code so that it now does its
1743 own call to the DNS to look up PTR records when searching for a host name.
1744 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1745 addresses that are only in /etc/hosts are still found.
1747 This behaviour is, however, controlled by an option called host_lookup_
1748 order, which defaults to "bydns:byaddr". If people want to use the other
1749 order, or indeed, just use one or the other means of lookup, they can
1750 specify it in this variable.
1752 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1753 some operating systems, this comes back from gethostbyaddr() as an empty
1754 string, and this is what Exim used to test for. However, it seems that in
1755 other systems, "." is yielded. Exim now tests for this case too.
1757 7. The values of check_spool_space and check_log_space are now held internally
1758 as a number of kilobytes instead of an absolute number of bytes. If a
1759 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1760 kilobyte. This means that much larger values can be stored.
1762 8. Exim monitor: an attempt to get the action menu when not actually pointing
1763 at a message produces an empty menu entitled "No message selected". This
1764 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1765 no entries in it ("Shell widget menu has zero width and/or height"). So I
1766 have added a single, blank menu entry in this case.
1768 9. Added ${quote_local_part.
1770 10. MIME decoding is now applied to the contents of Subject: header lines when
1773 11. Now that a reference to $sender_host_address automatically causes a reverse
1774 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1775 host lookup before query-style lookups in lists that might use this
1776 variable. This has therefore been abolished, and the "net-" prefix is no
1777 longer necessary for query-style lookups.
1779 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1780 have been a typo for LFLAGS, so it has been changed.
1782 13. The install script calls Exim with "-C /dev/null" in order to find the
1783 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1784 to be output. Howeve, since Exim outputs its version number before the
1785 error, it didn't break the script. It just looked ugly. I fixed this by
1786 always allowing "-C /dev/null" if the caller is root.
1788 14. Ignore overlarge ACL variable number when reading spool file - insurance
1789 against a later release with more variables having written the file.
1791 15. The standard form for an IPv6 address literal was being rejected by EHLO.
1792 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1793 this, as well as the form without the "IPv6" on the front.
1795 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1796 OS/Makefile-Darwin file.
1798 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1799 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1801 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
1802 variable. However, some OS use ints for inodes. I've added cast to long int
1803 to get rid of the compiler warning.
1805 19. I had forgotten to lock out "/../" in configuration file names when
1806 ALT_CONFIG_PREFIX was set.
1808 20. Routers used for verification do not need to specify transports. However,
1809 if such a router generated a host list, and callout was configured, Exim
1810 crashed, because it could not find a port number from the (non-existent)
1811 transport. It now assumes port 25 in this circumstance.
1813 21. Added the -t option to exigrep.
1815 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1816 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1817 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1818 them is set, LOOKUP_LSEARCH is forced.
1820 23. "exim -bV" now outputs a list of lookups that are included in the binary.
1822 24. Added sender and host information to the "rejected by local_scan()" log
1823 line; previously there was no indication of these.
1825 25. Added .include_if_exists.
1827 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
1828 Linux. It turns out that not all file systems support this. Apparently some
1829 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1830 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1831 not supported on the file descriptor, is now ignored when Exim is trying to
1832 sync a directory. This applies only to Linux.
1834 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1836 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1837 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1840 29. Added support for maildirsize files from supplied patch (modified a bit).
1842 30. The use of :fail: followed by an empty string could lead Exim to respond to
1843 sender verification failures with (e.g.):
1845 550 Verification failed for <xxx>
1846 550 Sender verify failed
1848 where the first response line was missing the '-' that indicates it is not
1849 the final line of the response.
1851 31. The loop for finding the name of the user that called Exim had a hardwired
1852 limit of 10; it now uses the value of finduser_retries, which is used for
1853 all other user lookups.
1855 32. Added $received_count variable, available in data and not_smtp ACLs, and at
1858 33. Exim was neglecting to zero errno before one call of strtol() when
1859 expanding a string and expecting an integer value. On some systems this
1860 resulted in spurious "integer overflow" errors. Also, it was casting the
1861 result into an int without checking.
1863 34. Testing for a connection timeout using "timeout_connect" in the retry rules
1864 did not work. The code looks as if it has *never* worked, though it appears
1865 to have been documented since at least releast 1.62. I have made it work.
1867 35. The "timeout_DNS" error in retry rules, also documented since at least
1868 1.62, also never worked. As it isn't clear exactly what this means, and
1869 clearly it isn't a major issue, I have abolished the feature by treating it
1870 as "timeout", and writing a warning to the main and panic logs.
1872 36. The display of retry rules for -brt wasn't always showing the error code
1875 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
1876 timeout_connect_A, timeout_connect_MX.
1878 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
1879 to the empty sender.
1881 39. The daemon was not analysing the content of -oX till after it had closed
1882 stderr and disconnected from the controlling terminal. This meant that any
1883 syntax errors were only noted on the panic log, and the return code from
1884 the command was 0. By re-arranging the code a little, I've made the
1885 decoding happen first, so such errors now appear on stderr, and the return
1886 code is 1. However, the actual setting up of the sockets still happens in
1887 the disconnected process, so errors there are still only recorded on the
1890 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
1891 connections (as happens on some IP stacks) was logged at start up time as
1892 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
1893 it from "IPv6 and IPv4", which means that two separate sockets are being
1896 41. The debug output for gethostbyname2() or getipnodebyname() failures now
1897 says whether AF_INET or AF_INET6 was passed as an argument.
1899 42. Exiwhat output was messed up when time zones were included in log
1902 43. Exiwhat now gives more information about the daemon's listening ports,
1903 and whether -tls-on-connect was used.
1905 44. The "port" option of the smtp transport is now expanded.
1907 45. A "message" modifier in a "warn" statement in a non-message ACL was being
1908 silently ignored. Now an error message is written to the main and panic
1911 46. There's a new ACL modifier called "logwrite" which writes to a log file
1912 as soon as it is encountered.
1914 47. Added $local_user_uid and $local_user_gid at routing time.
1916 48. Exim crashed when trying to verify a sender address that was being
1919 49. Exim was recognizing only a space character after ".include". It now also
1920 recognizes a tab character.
1922 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
1923 extracting the relevant information from the specification. The man page no
1924 longer contains scrambled data for the -d option, and I've added a section
1925 at the front about calling Exim under different names.
1927 51. Added "extra_headers" argument to the "mail" command in filter files.
1929 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
1932 53. Installed eximstats 1.29.
1934 54. Added transport_filter_timeout as a generic transport option.
1936 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
1937 Cc: header lines. This was required by RFC 822, but it not required by RFC
1940 56. Exim used to add From:, Date:, and Message-Id: header lines to any
1941 incoming messages that did not have them. Now it does so only if the
1942 message originates locally, that is, if there is no associated remote host
1943 address. When Resent- header lines are present, this applies to the Resent-
1944 lines rather than the non-Resent- lines.
1946 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
1947 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
1949 58. Messages for configuration errors now include the name of the main
1950 configuration file - useful now that there may be more than one file in a
1951 list (.included file names were always shown).
1953 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
1954 for those rare installations that do not start the daemon as root or run it
1955 setuid root. I've cut out the call to initgroups() if the daemon is not
1958 60. The Exim user and group can now be bound into the binary as text strings
1959 that are looked up at the start of Exim's processing.
1961 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
1963 62. Added $mailstore_basename variable.
1965 63. Installed patch to sieve.c from Michael Haardt.
1967 64. When Exim failed to open the panic log after failing to open the main log,
1968 the original message it was trying to log was written to stderr and debug
1969 output, but if they were not available (the usual case in production), it
1970 was lost. Now it is written to syslog before the two lines that record the
1971 failures to open the logs.
1973 65. Users' Exim filters run in subprocesses under the user's uid. It is
1974 possible for a "deliver" command or an alias in a "personal" command to
1975 provoke an address rewrite. If logging of address rewriting is configured,
1976 this fails because the process is not running as root or exim. There may be
1977 a better way of dealing with this, but for the moment (because 4.30 needs
1978 to be released), I have disabled address rewrite logging when running a
1979 filter in a non-root, non-exim process.
1985 1. The buildconfig auxiliary program wasn't quoting the value set for
1986 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
1987 not defined. This bug was masked in 4.22 by the effect that was fixed in
1990 2. Some messages that were rejected after a message id was allocated were
1991 shown as "incomplete" by exigrep. It no longer does this for messages that
1992 are rejected by local_scan() or the DATA or non-SMTP ACLs.
1994 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
1995 have allow_domain_literals set, the ID did not get logged in the <= line.
1996 Domain literals are now always recognized in Message-ID: header lines.
1998 4. The first argument for a ${extract expansion item is the key name or field
1999 number. Leading and trailing spaces in this item were not being ignored,
2000 causing some misleading effects.
2002 5. When deliver_drop_privilege was set, single queue runner processes started
2003 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2004 same command in the process it spins off) were not dropping privilege.
2006 6. When the daemon running as "exim" started a queue runner, it always
2007 re-executed Exim in the spun-off process. This is a waste of effort when
2008 deliver_drop_privilege is set. The new process now just calls the
2009 queue-runner function directly.
2015 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2018 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2019 ignored. Though the use of -oP was forcing the writing of a pid file, it
2020 was always written to the default place.
2022 3. If the message "no IP address found for host xxxx" is generated during
2023 incoming verification, it is now followed by identification of the incoming
2024 connection (so you can more easily find what provoked it).
2026 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2028 5. Added some features to "harden" Exim a bit more against certain attacks:
2030 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2031 be put in Local/Makefile. This is like the never_users runtime option,
2032 but it cannot be overridden. The default setting is "root".
2034 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2035 prefix string with which any file named in a -C command line option
2038 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2039 is retained for -C and -D only if the caller of Exim is root. Without
2040 it, the exim user may also use -C and -D and retain privilege.
2042 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2043 command line option is disabled.
2045 6. Macro names set by the -D option must start with an upper case letter, just
2046 like macro names defined in the configuration file.
2048 7. Added "dereference=" facility to LDAP.
2050 8. Two instances of the typo "uknown" in the source files are fixed.
2052 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2053 the Configure-Makefile script screwed up while processing it.
2055 10. Incorporated PCRE 4.4.
2057 11. The SMTP synchronization check was not operating right at the start of an
2058 SMTP session. For example, it could not catch a HELO sent before the client
2059 waited for the greeting. There is now a check for outstanding input at the
2060 point when the greeting is written. Because of the duplex, asynchronous
2061 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2062 way, but not yet received, when the check is performed.
2064 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2065 on TCP/IP sockets, because this apparently causes some broken clients to
2068 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2069 unchanged) from the Cygwin maintainer.
2071 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2072 was supported instead of testing for mbx. Effectively a typo.
2074 15. The spa authenticator server code was not checking that the input it
2075 received was valid base64.
2077 16. The debug output line for the "set" modifier in ACLs was not showing the
2078 name of the variable that was being set.
2080 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2082 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2083 lookup on its own. Something else was needed to trigger the lookup. For
2084 example, a match in host_lookup or the need for a host name in a host list.
2085 Now, if $sender_host_name is referenced and the host name has not yet been
2086 looked up, a lookup is performed. If the lookup fails, the variable remains
2087 empty, and $host_lookup_failed is set to "1".
2089 19. Added "eqi" as a case-independent comparison operator.
2091 20. The saslauthd authentication condition could segfault if neither service
2092 nor realm was specified.
2094 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2095 error message that was logged was misleading, and incoming SMTP
2096 connections were dropped. The message is now more accurate, and temporary
2097 errors are given to SMTP connections.
2099 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2100 (except RSET) with a 421 error, until QUIT is received. However, it was
2101 failing to send a response to QUIT.
2103 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2104 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2105 ACL is now run after the helo_try_verify_hosts code.
2107 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2108 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2109 case-independent, so other case variants are also recognized). Apparently
2110 some systems use these upper case variants.
2112 25. If more than two messages were waiting for the same host, and a transport
2113 filter was specified for the transport, Exim sent two messages over the
2114 same TCP/IP connection, and then failed with "socket operation on non-
2115 socket" when it tried to send the third.
2117 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2119 27. The extern definition of crypt16() in expand.c was not being excluded when
2120 the OS had its own crypt16() function.
2122 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2123 as a preferred synonym for return_size_limit, both as an option and as an
2126 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2128 30. Changed the default configuration ACL to relax the local part checking rule
2129 for addresses that are not in any local domains. For these addresses,
2130 slashes and pipe symbols are allowed within local parts, but the sequence
2131 /../ is explicitly forbidden.
2133 31. SPA server authentication was not clearing the challenge buffer before
2136 32. log_message in a "warn" ACL statement was writing to the reject log as
2137 well as to the main log, which contradicts the documentation and doesn't
2138 seem right (because no rejection is happening). So I have stopped it.
2140 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2141 However, I am unable to do any testing of this.
2143 34. Fixed an infelicity in the appendfile transport. When checking directories
2144 for a mailbox, to see if any needed to be created, it was accidentally
2145 using path names with one or more superfluous leading slashes; tracing
2146 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2148 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2149 contents are added to the log line that is written for every discarded
2150 recipient. (Previously a log_message setting was ignored.)
2152 36. The ${quote: operator now quotes the string if it is empty.
2154 37. The install script runs exim in order to find its version number. If for
2155 some reason other than non-existence or emptiness, which it checks, it
2156 could not run './exim', it was installing it with an empty version number,
2157 i.e. as "exim-". This error state is now caught, and the installation is
2160 38. An argument was missing from the function that creates an error message
2161 when Exim fails to connect to the socket for saslauthd authentication.
2162 This could cause Exim to crash, or give a corrupted message.
2164 39. Added isip, isip4, and isip6 to ${if conditions.
2166 40. The ACL variables $acl_xx are now saved with the message, and can be
2167 accessed later in routers, transports, and filters.
2169 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2170 strings in the file are not string-expanded.
2172 42. If a MAIL command specified a SIZE value that was too large to fit into an
2173 int variable, the check against message_size_limit failed. Such values are
2174 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2175 day this will have to be increased, but I don't think I want to be around
2176 when emails are that large.
2183 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2184 iconv() is not standard in FreeBSD.
2186 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2187 IPv6 enabled. The observed symptom was a segmentation fault on return from
2188 the function os_common_find_running_interfaces() in src/os.c.
2190 3. In the check_special_case() function in daemon.c I had used "errno" as an
2191 argument name, which causes warnings on some systems. This was basically a
2192 typo, since it was named "eno" in the comments!
2194 4. The code that waits for the clock to tick (at a resolution of some fraction
2195 of a second) so as to ensure message-id uniqueness was always waiting for
2196 at least one whole tick, when it could have waited for less. [This is
2197 almost certainly not relevant at current processor speeds, where it is
2198 unlikely to ever wait at all. But we try to future-proof.]
2200 5. The function that sleeps for a time interval that includes fractions of a
2201 second contained a race. It did not block SIGALRM between setting the
2202 timer, and suspending (a couple of lines later). If the interval was short
2203 and the sigsuspend() was delayed until after it had expired, the suspension
2204 never ended. On busy systems this could lead to processes getting stuck for
2207 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2208 process, before it forks any delivery processes. The open lookup caching
2209 mechanism meant that the open file or database connection was passed into
2210 the delivery process. The problem was that delivery processes always tidy
2211 up cached lookup data. This could cause a problem for the next delivery
2212 process started by the queue runner, because the external queue runner
2213 process does not know about the closure. So the next delivery process
2214 still has data in the lookup cache. In the case of a file lookup, there was
2215 no problem because closing a file descriptor in a subprocess doesn't affect
2216 the parent. However, if the lookup was caching a connection to a database,
2217 the connection was closed, and the second delivery process was likely to
2218 see errors such as "PGSQL: query failed: server closed the connection
2219 unexpectedly". The problem has been fixed by closing all cached lookups
2220 in a queue runner before running a delivery process.
2222 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2223 seem to have the "const" qualifier which it has on other OS. I've
2226 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2227 *of the same type* (client or server) with the same public name that an
2228 error should be diagnosed.
2230 9. When Exim looked up a host name for an IP address, but failed to find the
2231 original IP address when looking up the host name (a safety check), it
2232 output the message "<ip address> does not match any IP for NULL", which was
2233 confusing, to say the least. The bug was that the host name should have
2234 appeared instead of "NULL".
2236 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2237 user that is built into the binary, and the -C or -D options is used, root
2238 privilege is dropped before the configuration file is read. In addition,
2239 logging is switched to stderr instead of the normal log files. If the
2240 configuration then re-defines the Exim user, the unprivileged environment
2241 is probably not what is expected, so Exim logs a panic warning message (but
2244 However, if deliver_drop_privilege is set, the unprivileged state may well
2245 be exactly what is intended, so the warning has been cut out in that case,
2246 and Exim is allowed to try to write to its normal log files.
2252 1. smtp_return_error_details was not giving details for temporary sender
2253 or receiver verification errors.
2255 2. Diagnose a configuration error if two authenticators have the same public
2258 3. Exim used not to create the message log file for a message until the first
2259 delivery attempt. This could be confusing when incoming messages were held
2260 for policy or load reasons. The message log file is now created at the time
2261 the message is received, and an initial "Received" line is written to it.
2263 4. The automatically generated man page for command line options had a minor
2264 bug that caused no ill effects; however, a more serious problem was that
2265 the procedure for building the man page automatically didn't always
2266 operate. Consequently, release 4.20 contains an out-of-date version. This
2267 shouldn't happen again.
2269 5. When building Exim with embedded Perl support, the script that builds the
2270 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2271 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2273 6. The freeze_tell option was not being used for messages that were frozen on
2274 arrival, either by an ACL or by local_scan().
2276 7. Added the smtp_incomplete_transaction log selector.
2278 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2279 was accepting AUTH without a new EHLO.
2281 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2282 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2285 10. Logging of TCP/IP connections (when configured) now happens in the main
2286 daemon process instead of the child process, so that the TCP/IP connection
2287 count is more accurate (but it can never be perfect).
2289 11. The use of "drop" in a nested ACL was not being handled correctly in the
2290 outer ACL. Now, if condition failure induced by the nested "drop" causes
2291 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2292 or "require"), the connection is dropped.
2294 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2295 that yield "discard" can now be used with an "accept" or a "discard" verb,
2296 but an error is generated for any others (because I can't see a useful way
2297 to define what should happen).
2299 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2300 are now processed in the same way as lines in the Exim configuration file.
2301 In particular, continuation lines are supported.
2303 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2305 15. Added -ti meaning -t -i.
2307 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2308 domains, and warn by logging if others are found.
2310 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2311 returned from GIFCONF when Exim is trying to find the list of interfaces on
2312 a host. The code in os.c has been modified to copy each ifreq to an aligned
2313 structure in all cases.
2315 Also, in some cases, the returned ifreq's were being copied to a 'struct
2316 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2317 means the last couple of bytes of an IPv6 address could be chopped if the
2318 ifreq contained only a normal sockaddr (14 bytes storage).
2320 18. Named domain lists were not supported in the hosts_treat_as_local option.
2321 An entry such as +xxxx was not recognized, and was treated as a literal
2324 19. Ensure that header lines added by a DATA ACL are included in the reject log
2325 if the ACL subsequently rejects the message.
2327 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2328 MD5 (which is deprecated).
2330 21. When testing a filter file using -bf, Exim was writing a message when it
2331 took the sender from a "From " line in the message, but it was not doing so
2332 when it took $return_path from a Return-Path: header line. It now does.
2334 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2335 with a valid header line field name (a series of printing characters
2336 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2338 23. Changed "disc" in the source to "disk" to conform to the documentation and
2339 the book and for uniformity.
2341 24. Ignore Sendmail's -Ooption=value command line item.
2343 25. When execve() failed while trying to run a command in a pipe transport,
2344 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2345 could be confused with a return value of 69 from the command itself. This
2346 has been changed to 127, the value the shell returns if it is asked to run
2347 a non-existent command. The wording for the related log line suggests a
2348 non-existent command as the problem.
2350 26. If received_header_text expands to an empty string, do not add a Received:
2351 header line to the message. (Well, it adds a token one on the spool, but
2352 marks it "old" so that it doesn't get used or transmitted.)
2354 27. Installed eximstats 1.28 (addition of -nt option).
2356 28. There was no check for failure on the call to getsockname() in the daemon
2357 code. This can fail if there is a shortage of resources on the system, with
2358 ENOMEM, for example. A temporary error is now given on failure.
2360 29. Contrary to the C standard, it seems that in some environments, the
2361 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2362 program. Exim now does this explicitly; it affects the formatting of
2363 timestamps using strftime().
2365 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2366 complaints. I've now initialized all the variables, to avoid this.
2368 32. Header lines added by a system filter were not being "seen" during
2369 transport-time rewrites.
2371 33. The info_callback() function passed to OpenSSL is set up with type void
2372 (*)(SSL *, int, int), as described somewhere. However, when calling the
2373 function (actually a macro) that sets it up, the type void(*)() is
2374 expected. I've put in a cast to prevent warnings from picky compilers.
2376 34. If a DNS black list lookup found a CNAME record, but there were no A
2377 records associated with the domain it pointed at, Exim crashed.
2379 35. If a DNS black list lookup returned more than one A record, Exim ignored
2380 all but the first. It now scans all returned addresses if a particular IP
2381 value is being sought. In this situation, the contents of the
2382 $dnslist_value variable are a list of all the addresses, separated by a
2385 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2386 to accept a host name and all its aliases if the forward lookup for any of
2387 them yielded the IP address of the incoming connection. Now it accepts only
2388 those names whose forward lookup yields the correct IP address. Any other
2389 names are discarded. This closes a loophole whereby a rogue DNS
2390 administrator could create reverse DNS records to break through a
2391 wildcarded host restriction in an ACL.
2393 37. If a user filter or a system filter that ran in a subprocess used any of
2394 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2395 the wrong values were passed to the pipe command ($thisaddress had the
2396 value of $0, $0 had the value of $1, etc). This bug was introduced by
2397 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2399 38. Improved the line breaking for long SMTP error messages from ACLs.
2400 Previously, if there was no break point between 40 and 75 characters, Exim
2401 left the rest of the message alone. Two changes have been made: (a) I've
2402 reduced the minimum length to 35 characters; (b) if it can't find a break
2403 point between 35 and 75 characters, it looks ahead and uses the first one
2404 that it finds. This may give the occasional overlong line, but at least the
2405 remaining text gets split now.
2407 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2408 file descriptors might be low, and that setting 1000 would always raise it.
2409 It turns out that in some environments, the limit is already over 1000 and
2410 that lowering it causes trouble. So now Exim takes care not to decrease it.
2412 40. When delivering a message, the value of $return_path is set to $sender_
2413 address at the start of routing (routers may change the value). By an
2414 oversight, this default was not being set up when an address was tested by
2415 -bt or -bv, which affected the outcome if any router or filter referred to
2418 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2419 log only when "message" or "log_message" are set. However, if one of the
2420 conditions was an address verification, or a call to a nested ACL, the
2421 messages generated by the underlying test were being passed through. This
2422 no longer happens. The underlying message is available in $acl_verify_
2423 message for both "message" and "log_message" expansions, so it can be
2424 passed through if needed.
2426 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2427 new expansion $bh_ to give the encoded byte string without charset
2428 translation. Translation happens only if iconv() is available; HAVE_ICONV
2429 indicates this at build time. HEADERS_CHARSET gives the charset to
2430 translate to; headers_charset can change it in the configuration, and
2431 "headers charset" can change it in an individual filter file.
2433 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2434 that creates RFC 2047 encoded "words" labels them as that charset instead
2435 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2436 expansion operator; (ii) when Exim creates a From: line for a local
2437 message; (iii) when a header line is rewritten to include a "phrase" part.
2439 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2440 buggy, causing it to skip the first lines of messages whose message ID
2441 ended in 'D'. This would not have bitten before Exim release 4.14, because
2442 message IDs were unlikely to end in 'D' before then. The effect was to have
2443 incorrect size information for certain domains.
2445 45. #include "config.h" was missing at the start of the crypt16.c module. This
2446 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2449 46. If there was a timeout during a "random" callout check, Exim treated it as
2450 a failure of the random address, and carried on sending RSET and the real
2451 address. If the delay was just some slowness somewhere, the response to the
2452 original RCPT would be taken as a response to RSET and so on, causing
2453 mayhem of various kinds.
2455 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2456 when I implemented it. It didn't allow for the fact that some option values
2457 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2458 the right test for positive values.
2460 48. Domain names in DNS records are case-independent. Exim always looks them up
2461 in lower case. Some resolvers return domain names in exactly the case they
2462 appear in the zone file, that is, they may contain uppercase letters. Not
2463 all resolvers do this - some return always lower case. Exim was treating a
2464 change of case by a resolver as a change of domain, similar to a widening
2465 of a domain abbreviation. This triggered its re-routing code and so it was
2466 trying to route what was effectively the same domain again. This normally
2467 caused routing to fail (because the router wouldn't handle the domain
2468 twice). Now Exim checks for this case specially, and just changes the
2469 casing of the domain that it ultimately uses when it transmits the message
2472 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2475 50. If a filter generated a file delivery with a non-absolute name (possible if
2476 no home directory exists for the router), the forbid_file option was not
2479 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2480 the existing equality matching.
2482 52. Exim was using ints instead of ino_t variables in some places where it was
2483 dealing with inode numbers.
2485 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2486 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2487 called TMPDIR, and if it finds it is different, it changes its value.
2489 54. The smtp_printf() function is now made available to local_scan() so
2490 additional output lines can be written before returning. There is also an
2491 smtp_fflush() function to enable the detection of a dropped connection.
2492 The variables smtp_input and smtp_batched_input are exported to
2495 55. Changed the default runtime configuration: the message "Unknown user"
2496 has been removed from the ACL, and instead placed on the localuser router,
2497 using the cannot_route_message feature. This means that any verification
2498 failures that generate their own messages won't get overridden. Similarly,
2499 the "Unrouteable address" message that was in the ACL for unverifiable
2500 relay addresses has also been removed.
2502 56. Added hosts_avoid_esmtp to the smtp transport.
2504 57. The exicyclog script was not checking for the esoteric option
2505 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2506 will work only if exicyclog is run under the appropriate euid.
2508 58. Following a discussion on the list, the rules by which Exim recognises line
2509 endings on incoming messages have been changed. The -dropcr and drop_cr
2510 options are now no-ops, retained only for backwards compatibility. The
2511 following line terminators are recognized: LF CRLF CR. However, special
2512 processing applies to CR:
2514 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2515 nor a local message in the state where . is a terminator.
2517 (ii) If a bare CR is encountered in a header line, an extra space is added
2518 after the line terminator so as not to end the header. The reasoning
2519 behind this is that bare CRs in header lines are most likely either
2520 to be mistakes, or people trying to play silly games.
2522 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
2523 was being incorrectly given as 18 bytes larger than it should have been.
2524 This is a VOB (very old bug).
2526 60. This may never have affected anything current, but just in case it has:
2527 When the local host is found other than at the start of a list of hosts,
2528 the local host, those with the same MX, and any that follow, are discarded.
2529 When the list in question was part of a longer list of hosts, the following
2530 hosts (not currently being processed) were also being discarded. This no
2531 longer happens. I'm not sure if this situation could ever has previously
2534 61. Added the "/MX" feature to lists of hosts in the manualroute and query
2537 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2538 header. This is something that is recommended in a new Internet Draft, and
2539 is something that is documented as being done by Sendmail. There are two
2540 possible values. For messages generated by the autoreply transport, Exim
2543 Auto-Submitted: auto-replied
2545 whereas for all other generated messages (e.g. bounces) it adds
2547 Auto-Submitted: auto-generated
2549 63. The "personal" condition in filters now includes a test for the
2550 Auto-Submitted: header. If it contains the string "auto-" the message it
2551 not considered personal.
2553 64. Added rcpt_include_affixes as a generic transport option.
2555 65. Added queue_only_override (default true).
2557 66. Added the syslog_duplication option.
2559 67. If what should have been the first header line of a message consisted of
2560 a space followed by a colon, Exim was mis-interpreting it as a header line.
2561 It isn't of course - it is syntactically invalid and should therefore be
2562 treated as the start of the message body. The misbehaviour could have
2563 caused a number of strange effects, including loss of data in subsequent
2564 header lines, and spool format errors.
2566 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2567 client host had authenticated. This control can now be exercised by an ACL
2568 for more flexibility.
2570 69. By default, callouts do not happen when testing with -bh. There is now a
2571 variant, -bhc, which does actually run the callout code, including
2572 consulting and updating the callout cache.
2574 70. Added support for saslauthd authentication, courtesy of Alexander
2577 71. If statvfs() failed on the spool or log directories while checking their
2578 size for availability, Exim confusingly gave the error "space shortage".
2579 Furthermore, in debugging mode it crashed with a floating point exception.
2580 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2581 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2582 serious problem, Exim now writes to the main and panic logs when this
2583 happens, with details of the failure. It then refuses to accept the
2584 incoming message, giving the message "spool directory problem" or "log
2585 directory problem" with a 421 code for SMTP messages.
2587 72. When Exim is about to re-exec itself, it ensures that the file descriptors
2588 0, 1, and 2 exist, because some OS complain for execs without them (see
2589 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2590 descriptors. However, the code omitted to check that the open succeeded,
2591 causing mysterious errors if for some reason the permissions on /dev/null
2592 got screwed. Now Exim writes a message to the main and panic logs, and
2593 bombs out if it can't open /dev/null.
2595 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2596 interact so that it is all more flexible. It is supposed to remain
2597 backwards compatible. Also added extra_local_interfaces.
2599 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2600 to bomb out with an assertion failure - to the client this appears as a
2601 connection drop. This problem occurs in the part of the code that was taken
2602 from the Samba project. Fortunately, the assertion is in a very simple
2603 function, so I have fixed this by reproducing the function inline in the
2604 one place where it is called, and arranging for authentication to fail
2605 instead of killing the process with assert().
2607 75. The SPA client code was not working when the server requested OEM rather
2608 than Unicode encoding.
2610 76. Added code to make require_files with a specific uid setting more usable in
2611 the case where statting the file as root fails - usually a non-root-mounted
2612 NFS file system. When this happens and the failure is EACCES, Exim now
2613 forks a subprocess and does the per-uid checking as the relevant uid.
2615 77. Added process_log_path.
2617 78. If log_file_path was not explicitly set, a setting of check_log_space or
2618 check_log_inodes was ignored.
2620 79. If a space check for the spool or log partitions fails, the incident is now
2621 logged. Of course, in the latter case the data may get lost...
2623 80. Added the %p formatting code to string_format() so that it can be used to
2624 print addresses in debug_print(). Adjusted all the address printing in the
2625 debugging in store.c to use %p rather than %d.
2627 81. There was a concern that a line of code in smtp_in.c could overflow a
2628 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2629 initially expressed, the concern was not well-founded, because trailing
2630 spaces are removed early. However, if the trailing spaces were followed by
2631 a NULL, they did not get removed, so the overflow was possible. Two fixes
2634 (a) I re-wrote the offending code in a cleaner fashion.
2635 (b) If an incoming SMTP command contains a NULL character, it is rejected
2638 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
2639 runs initgroups(), so that if the Exim user is in any additional groups,
2640 they will be used during message reception.
2646 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / heiko / exim.git / blob