1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.90 2005/03/15 12:27:54 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
10 TK/01. Added Yahoo DomainKeys support via libdomainkeys. See
11 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
13 TK/02. Fix ACL "control" statment not being available in MIME ACL.
15 TK/03. Fix ACL "regex" condition not being available in MIME ACL.
17 PH/01. Installed a patch from the Sieve maintainer that allows -bf to be used
18 to test Sieve filters that use "vacation".
20 PH/02. Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
21 that changes the way the GnuTLS parameters are stored in the cache file.
22 The new format can be generated externally. For backward compatibility,
23 if the data in the cache doesn't make sense, Exim assumes it has read an
24 old-format file, and it generates new data and writes a new file. This
25 means that you can't go back to an older release without removing the
28 PH/03. A redirect router that has both "unseen" and "one_time" set does not
29 work if there are any delivery delays because "one_time" forces the
30 parent to be marked "delivered", so its unseen clone is never tried
31 again. For this reason, Exim now forbids the simultaneous setting of
34 PH/04. Change 4.11/85 fixed an obscure bug concerned with addresses that are
35 redirected to themselves ("homonym" addresses). Read the long ChangeLog
36 entry if you want to know the details. The fix, however, neglected to
37 consider the case when local delivery batching is involved. The test for
38 "previously delivered" was not happening when checking to see if an
39 address could be batched with a previous (undelivered) one; under
40 certain circumstances this could lead to multiple deliveries to the same
41 address. A one-line patch to add the appropriate test fixes the bug.
44 A note about Exim versions 4.44 and 4.50
45 ----------------------------------------
47 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
48 changes of various kinds. As a consequence, a big documentation update was
49 needed. This delayed the release for rather longer than seemed good, especially
50 in the light of a couple of (minor) security issues. Therefore, the changes
51 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
52 release. So 4.44 and 4.50 are in effect two different branches that both start
55 I have left the 4.50 change log unchanged; it contains all the changes since
56 4.43. The change log for 4.44 is below; many of its items are identical to
57 those for 4.50. This seems to be the most sensible way to preserve the
58 historical information.
64 1. Minor wording change to the doc/README.SIEVE file.
66 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
67 computation of the current number of files was incorrect.
69 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
70 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
73 4. Give more explanation in the error message when the command for a transport
74 filter fails to execute.
76 5. There are several places where Exim runs a non-Exim command in a
77 subprocess. The SIGUSR1 signal should be disabled for these processes. This
78 was being done only for the command run by the queryprogram router. It is
79 now done for all such subprocesses. The other cases are: ${run, transport
80 filters, and the commands run by the lmtp and pipe transports.
82 6. Added CONFIGURE_GROUP build-time option.
84 7. Some older OS have a limit of 256 on the maximum number of file
85 descriptors. Exim was using setrlimit() to set 1000 as a large value
86 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
87 systems. I've change it so that if it can't get 1000, it tries for 256.
89 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
90 was an oversight, and furthermore, ever since the addition of extra
91 controls (e.g. 4.43/32), the checks on when to allow different forms of
92 "control" were broken. There should now be diagnostics for all cases when a
93 control that does not make sense is encountered.
95 9. Added the /retain_sender option to "control=submission".
97 10. $recipients is now available in the predata ACL (oversight).
99 11. Tidy the search cache before the fork to do a delivery from a message
100 received from the command line. Otherwise the child will trigger a lookup
101 failure and thereby defer the delivery if it tries to use (for example) a
102 cached ldap connection that the parent has called unbind on.
104 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
105 of $address_data from the recipient verification was clobbered by the
108 13. The value of address_data from a sender verification is now available in
109 $sender_address_data in subsequent conditions in the ACL statement.
111 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
113 15. Added a new option "connect=<time>" to callout options, to set a different
116 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
117 was its contents. (It was OK if the option was not defined at all.)
119 17. A "Completed" log line is now written for messages that are removed from
120 the spool by the -Mrm option.
122 18. New variables $sender_verify_failure and $recipient_verify_failure contain
123 information about exactly what failed.
125 19. Added -dd to debug only the daemon process.
127 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
128 handles timeouts, both on the server side and network timeouts. Renamed the
129 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
131 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
132 It is now set to "smtps".
134 22. $host_address is now set to the target address during the checking of
137 23. When checking ignore_target_hosts for an ipliteral router, no host name was
138 being passed; this would have caused $sender_host_name to have been used if
139 matching the list had actually called for a host name (not very likely,
140 since this list is usually IP addresses). A host name is now passed as
143 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
144 code that specifies a non-restarting handler (typically sigaction() in
145 modern systems) in an attempt to fix a rare and obscure crash bug.
147 25. Narrowed the window for a race in the daemon that could cause it to ignore
148 SIGCHLD signals. This is not a major problem, because they are used only to
149 wake it up if nothing else does.
151 26. A malformed maildirsize file could cause Exim to calculate negative values
152 for the mailbox size or file count. Odd effects could occur as a result.
153 The maildirsize information is now recalculated if the size or filecount
156 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
157 support for a long time. Removed HAVE_SYS_VFS_H.
159 28. Installed the latest version of exipick from John Jetmore.
161 29. In an address list, if the pattern was not a regular expression, an empty
162 subject address (from a bounce message) matched only if the pattern was an
163 empty string. Non-empty patterns were not even tested. This was the wrong
164 because it is perfectly reasonable to use an empty address as part of a
165 database query. An empty address is now tested by patterns that are
166 lookups. However, all the other forms of pattern expect the subject to
167 contain a local part and a domain, and therefore, for them, an empty
168 address still always fails if the pattern is not itself empty.
170 30. Exim went into a mad DNS loop when attempting to do a callout where the
171 host was specified on an smtp transport, and looking it up yielded more
174 31. Re-factored the code for checking spool and log partition space into a
175 function that finds that data and another that does the check. The former
176 is then used to implement four new variables: $spool_space, $log_space,
177 $spool_inodes, and $log_inodes.
179 32. The RFC2047 encoding function was originally intended for short strings
180 such as real names; it was not keeping to the 75-character limit for
181 encoded words that the RFC imposes. It now respects the limit, and
182 generates multiple encoded words if necessary. To be on the safe side, I
183 have increased the buffer size for the ${rfc2047: expansion operator from
186 33. It is now permitted to omit both strings after an "if" condition; if the
187 condition is true, the result is "true". As before, when the second string
188 is omitted, a false condition yields an empty string. This makes it less
189 cumbersome to write custom ACL and router conditions.
191 34. Failure to deliver a bounce message always caused it to be frozen, even if
192 there was an errors_to setting on the router. The errors_to setting is now
195 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
196 canonical form (fully expanded) before being placed in
197 $sender_host_address.
199 36. The table in the code that translates DNS record types into text (T_A to
200 "A" for instance) was missing entries for NS and CNAME. It is just possible
201 that this could have caused confusion if both these types were looked up
202 for the same domain, because the text type is used as part of Exim's
203 per-process caching. But the chance of anyone hitting this buglet seems
206 37. The dnsdb lookup has been extended in a number of ways.
208 (1) There is a new type, "zns", which walks up the domain tree until it
209 finds some nameserver records. It should be used with care.
211 (2) There is a new type, "mxh", which is like "mx" except that it returns
212 just the host names, not the priorities.
214 (3) It is now possible to give a list of domains (or IP addresses) to be
215 looked up. The behaviour when one of the lookups defers can be
216 controlled by a keyword.
218 (4) It is now possible to specify the separator character for use when
219 multiple records are returned.
221 38. The dnslists ACL condition has been extended: it is now possible to supply
222 a list of IP addresses and/or domains to be looked up in a particular DNS
225 39. Added log_selector=+queue_time_overall.
227 40. When running the queue in the test harness, wait just a tad after forking a
228 delivery process, to get repeatability of debugging output.
230 41. Include certificate and key file names in error message when GnuTLS fails
231 to set them up, because the GnuTLS error message doesn't include the name
232 of the failing file when there is a problem reading it.
234 42. Allow both -bf and -bF in the same test run.
236 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
238 44. The "Exiscan patch" is now merged into the mainline Exim source.
240 45. Sometimes the final signoff response after QUIT could fail to get
241 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
242 < 0 before doing a fflush(). This bug looks as though it goes back to the
243 introduction of TLS in release 3.20, but "sometimes" must have been rare
244 because the tests only now provoked it.
246 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
247 (this can affect the format of dates).
249 47. exim_tidydb, when checking for the continued existence of a message for
250 which it has found a message-specific retry record, was not finding
251 messages that were in split spool directories. Consequently, it was
252 deleting retry records that should have stayed in existence.
254 48. Steve fixed some bugs in eximstats.
256 49. The SPA authentication driver was not abandoning authentication and moving
257 on to the next authenticator when an expansion was forced to fail,
258 contradicting the general specification for all authenticators. Instead it
259 was generating a temporary error. It now behaves as specified.
261 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
262 (the order specifies the preference for clients). The order is now AES256,
263 AES128, 3DES, ARCFOUR128.
265 51. Small patch to Sieve code - explicitly set From: when generating an
268 52. Exim crashed if a remote delivery caused a very long error message to be
269 recorded - for instance if somebody sent an entire SpamAssassin report back
270 as a large number of 550 error lines. This bug was coincidentally fixed by
271 increasing the size of one of Exim's internal buffers (big_buffer) that
272 happened as part of the Exiscan merge. However, to be on the safe side, I
273 have made the code more robust (and fixed the comments that describe what
276 53. Now that there can be additional text after "Completed" in log lines (if
277 the queue_time_overall log selector is set), a one-byte patch to exigrep
278 was needed to allow it to recognize "Completed" as not the last thing in
281 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
282 patch that reportedly fixes this has been added. I am not expert enough to
283 create a test for it. This is what the patch creator wrote:
285 "I found a little strange behaviour of ldap code when working with
286 Windows 2003 AD Domain, where users was placed in more than one
287 Organization Units. When I tried to give exim partial DN, the exit code
288 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
289 But simultaneously result of request was absolutely normal ldap result,
290 so I produce this patch..."
292 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
293 REFERENCE, so I have modified the code to exclude the patch when that macro
296 55. Some experimental protocols are using DNS PTR records for new purposes. The
297 keys for these records are domain names, not reversed IP addresses. The
298 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
299 leaves it alone. Component reversal etc. now happens only for IP addresses.
302 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
304 57. Double the size of the debug message buffer (to 2048) so that more of very
305 long debug lines gets shown.
307 58. The exicyclog utility now does better if the number of log files to keep
308 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
310 59. Two changes related to the smtp_active_hostname option:
312 (1) $smtp_active_hostname is now available as a variable.
313 (2) The default for smtp_banner uses $smtp_active_hostname instead
314 of $primary_hostname.
316 60. The host_aton() function is supposed to be passed a string that is known
317 to be a valid IP address. However, in the case of IPv6 addresses, it was
318 not checking this. This is a hostage to fortune. Exim now panics and dies
319 if the condition is not met. A case was found where this could be provoked
320 from a dnsdb PTR lookup with an IPv6 address that had more than 8
321 components; fortuitously, this particular loophole had already been fixed
322 by change 4.50/55 above.
324 If there are any other similar loopholes, the new check in host_aton()
325 itself should stop them being exploited. The report I received stated that
326 data on the command line could provoke the exploit when Exim was running as
327 exim, but did not say which command line option was involved. All I could
328 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
332 61. There was a buffer overflow vulnerability in the SPA authentication code
333 (which came originally from the Samba project). I have added a test to the
334 spa_base64_to_bits() function which I hope fixes it.
337 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
338 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
340 63. The daemon start-up calls getloadavg() while still root for those OS that
341 need the first call to be done as root, but it missed one case: when
342 deliver_queue_load_max is set with deliver_drop_privilege. This is
343 necessary for the benefit of the queue runner, because there is no re-exec
344 when deliver_drop_privilege is set.
346 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
349 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
350 was in use, was not putting the data itself into the right store pool;
351 consequently, it could be overwritten for a subsequent message in the same
352 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
355 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
357 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
358 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
359 the function were treating the return as a boolean value, which happened to
360 work because 0=false and not-0=true, but is not correct code.
362 68. The host_aton() function was not handling scoped IPv6 addresses (those
363 with, for example, "%eth0" on the end) correctly.
365 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
366 negated items (that is, ~something) in unsigned ints. Some compilers
367 apparently mutter when there is no cast.
369 70. If an address verification called from an ACL failed, and did not produce a
370 user-specific message (i.e. there was only a "system" message), nothing was
371 put in $acl_verify_message. In this situation, it now puts the system
374 71. Change 4.23/11 added synchronization checking at the start of an SMTP
375 session; change 4.31/43 added the unwanted input to the log line - except
376 that it did not do this in the start of session case. It now does.
378 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
379 This is wrong and can cause the other end to generate a synchronization
380 error if it is another Exim or anything else that does the synchronization
381 check. A QUIT command is no longer sent after a timeout.
383 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
386 74. The defer_ok option of callout verification was not working if it was used
387 when verifying addresses in header lines, that is, for this case:
389 verify = header_sender/callout=defer_ok
391 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
392 those file descriptors could be used for SMTP connections. If anything
393 wrote to stderr (the example that came up was "warn" in embedded Perl), it
394 could be sent to the SMTP client, causing chaos. The daemon now opens
395 stdin, stdout, and stderr to /dev/null when it puts itself into the
398 76. Arrange for output from Perl's "warn" command to be written to Exim's main
399 log by default. The user can override this with suitable Perl magic.
401 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
402 the log message when discard triggers, was not working for the DATA ACL or
403 for the non-SMTP ACL.
405 78. Error message wording change in sieve.c.
407 79. If smtp_accept_max_per_host was set, the number of connections could be
408 restricted to fewer than expected, because the daemon was trying to set up
409 a new connection before checking whether the processes handling previous
410 connections had finished. The check for completed processes is now done
411 earlier. On busy systems, this bug wouldn't be noticed because something
412 else would have woken the daemon, and it would have reaped the completed
415 80. If a message was submitted locally by a user whose login name contained one
416 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
417 It caused a spool format error. I have fixed the spool reading code. A
418 related problem was that the "from" clause in the Received: line became
419 illegal because of the space(s). It is now covered by ${quote_local_part.
421 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
424 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
426 83. Patch from Sieve maintainer to fix unterminated string problem in
429 84. Some minor changes to the Linux configuration files to help with other
430 OS variants using glibc.
432 85. One more patch for Sieve to update vacation handling to latest spec.
435 ----------------------------------------------------
436 See the note above about the 4.44 and 4.50 releases.
437 ----------------------------------------------------
443 1. Change 4.43/35 introduced a bug that caused file counts to be
444 incorrectly computed when quota_filecount was set in an appendfile
447 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
448 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
451 3. Give more explanation in the error message when the command for a transport
452 filter fails to execute.
454 4. There are several places where Exim runs a non-Exim command in a
455 subprocess. The SIGUSR1 signal should be disabled for these processes. This
456 was being done only for the command run by the queryprogram router. It is
457 now done for all such subprocesses. The other cases are: ${run, transport
458 filters, and the commands run by the lmtp and pipe transports.
460 5. Some older OS have a limit of 256 on the maximum number of file
461 descriptors. Exim was using setrlimit() to set 1000 as a large value
462 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
463 systems. I've change it so that if it can't get 1000, it tries for 256.
465 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
466 was an oversight, and furthermore, ever since the addition of extra
467 controls (e.g. 4.43/32), the checks on when to allow different forms of
468 "control" were broken. There should now be diagnostics for all cases when a
469 control that does not make sense is encountered.
471 7. $recipients is now available in the predata ACL (oversight).
473 8. Tidy the search cache before the fork to do a delivery from a message
474 received from the command line. Otherwise the child will trigger a lookup
475 failure and thereby defer the delivery if it tries to use (for example) a
476 cached ldap connection that the parent has called unbind on.
478 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
479 of $address_data from the recipient verification was clobbered by the
482 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
483 was its contents. (It was OK if the option was not defined at all.)
485 11. A "Completed" log line is now written for messages that are removed from
486 the spool by the -Mrm option.
488 12. $host_address is now set to the target address during the checking of
491 13. When checking ignore_target_hosts for an ipliteral router, no host name was
492 being passed; this would have caused $sender_host_name to have been used if
493 matching the list had actually called for a host name (not very likely,
494 since this list is usually IP addresses). A host name is now passed as
497 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
498 code that specifies a non-restarting handler (typically sigaction() in
499 modern systems) in an attempt to fix a rare and obscure crash bug.
501 15. Narrowed the window for a race in the daemon that could cause it to ignore
502 SIGCHLD signals. This is not a major problem, because they are used only to
503 wake it up if nothing else does.
505 16. A malformed maildirsize file could cause Exim to calculate negative values
506 for the mailbox size or file count. Odd effects could occur as a result.
507 The maildirsize information is now recalculated if the size or filecount
510 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
511 support for a long time. Removed HAVE_SYS_VFS_H.
513 18. Updated exipick to current release from John Jetmore.
515 19. Allow an empty sender to be matched against a lookup in an address list.
516 Previously the only cases considered were a regular expression, or an
519 20. Exim went into a mad DNS lookup loop when doing a callout where the
520 host was specified on the transport, if the DNS lookup yielded more than
523 21. The RFC2047 encoding function was originally intended for short strings
524 such as real names; it was not keeping to the 75-character limit for
525 encoded words that the RFC imposes. It now respects the limit, and
526 generates multiple encoded words if necessary. To be on the safe side, I
527 have increased the buffer size for the ${rfc2047: expansion operator from
530 22. Failure to deliver a bounce message always caused it to be frozen, even if
531 there was an errors_to setting on the router. The errors_to setting is now
534 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
535 canonical form (fully expanded) before being placed in
536 $sender_host_address.
538 24. Updated eximstats to version 1.33
540 25. Include certificate and key file names in error message when GnuTLS fails
541 to set them up, because the GnuTLS error message doesn't include the name
542 of the failing file when there is a problem reading it.
544 26. Expand error message when OpenSSL has problems setting up cert/key files.
547 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
548 (this can affect the format of dates).
550 28. exim_tidydb, when checking for the continued existence of a message for
551 which it has found a message-specific retry record, was not finding
552 messages that were in split spool directories. Consequently, it was
553 deleting retry records that should have stayed in existence.
555 29. eximstats updated to version 1.35
556 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
557 1.35 - bugfix such that pie charts by volume are generated correctly
559 30. The SPA authentication driver was not abandoning authentication and moving
560 on to the next authenticator when an expansion was forced to fail,
561 contradicting the general specification for all authenticators. Instead it
562 was generating a temporary error. It now behaves as specified.
564 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
565 (the order specifies the preference for clients). The order is now AES256,
566 AES128, 3DES, ARCFOUR128.
568 31. Small patch to Sieve code - explicitly set From: when generating an
571 32. Exim crashed if a remote delivery caused a very long error message to be
572 recorded - for instance if somebody sent an entire SpamAssassin report back
573 as a large number of 550 error lines. This bug was coincidentally fixed by
574 increasing the size of one of Exim's internal buffers (big_buffer) that
575 happened as part of the Exiscan merge. However, to be on the safe side, I
576 have made the code more robust (and fixed the comments that describe what
579 33. Some experimental protocols are using DNS PTR records for new purposes. The
580 keys for these records are domain names, not reversed IP addresses. The
581 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
582 leaves it alone. Component reversal etc. now happens only for IP addresses.
585 34. The host_aton() function is supposed to be passed a string that is known
586 to be a valid IP address. However, in the case of IPv6 addresses, it was
587 not checking this. This is a hostage to fortune. Exim now panics and dies
588 if the condition is not met. A case was found where this could be provoked
589 from a dnsdb PTR lookup with an IPv6 address that had more than 8
590 components; fortuitously, this particular loophole had already been fixed
591 by change 4.50/55 or 4.44/33 above.
593 If there are any other similar loopholes, the new check in host_aton()
594 itself should stop them being exploited. The report I received stated that
595 data on the command line could provoke the exploit when Exim was running as
596 exim, but did not say which command line option was involved. All I could
597 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
601 35. There was a buffer overflow vulnerability in the SPA authentication code
602 (which came originally from the Samba project). I have added a test to the
603 spa_base64_to_bits() function which I hope fixes it.
606 36. The daemon start-up calls getloadavg() while still root for those OS that
607 need the first call to be done as root, but it missed one case: when
608 deliver_queue_load_max is set with deliver_drop_privilege. This is
609 necessary for the benefit of the queue runner, because there is no re-exec
610 when deliver_drop_privilege is set.
612 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
613 was in use, was not putting the data itself into the right store pool;
614 consequently, it could be overwritten for a subsequent message in the same
615 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
618 38. Sometimes the final signoff response after QUIT could fail to get
619 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
620 < 0 before doing a fflush(). This bug looks as though it goes back to the
621 introduction of TLS in release 3.20, but "sometimes" must have been rare
622 because the tests only now provoked it.
628 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
629 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
630 Now it returns an integer. A number of calls were still expecting a T/F
631 return. Fortuitously, in all cases, the tests worked in OK situations,
632 which is the norm. However, things would have gone wrong on any write
633 failures on the smtp file descriptor. This function is used when sending
634 messages over SMTP and also when doing verify callouts.
636 2. When Exim is called to do synchronous delivery of a locally submitted
637 message (the -odf or -odi options), it no longer closes stderr before doing
640 3. Implemented the mua_wrapper option.
642 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
644 5. Implemented the functions header_remove(), header_testname(),
645 header_add_at_position(), and receive_remove_recipient(), and exported them
648 6. If an ACL "warn" statement specified the addition of headers, Exim already
649 inserted X-ACL-Warn: at the start if there was no header name. However, it
650 was not making this test for the second and subsequent header lines if
651 there were newlines in the string. This meant that an invalid header could
652 be inserted if Exim was badly configured.
654 7. Allow an ACL "warn" statement to add header lines at the start or after all
655 the Received: headers, as well as at the end.
657 8. Added the rcpt_4xx retry error code.
659 9. Added postmaster_mailfrom=xxx to callout verification option.
661 10. Added mailfrom=xxxx to the callout verification option, for verify=
664 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
665 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
667 12. Inserted some casts to stop certain compilers warning when using pointer
668 differences as field lengths or precisions in printf-type calls (mostly
669 affecting debugging statements).
671 13. Added optional readline() support for -be (dynamically loaded).
673 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
674 same clock tick as a message's arrival, so that its received time was the
675 same as the "first fail" time on the retry record, and that message
676 remained on the queue past the ultimate address timeout, every queue runner
677 would try a delivery (because it was past the ultimate address timeout) but
678 after another failure, the ultimate address timeout, which should have then
679 bounced the address, did not kick in. This was a "< instead of <=" error;
680 in most cases the first failure would have been in the next clock tick
681 after the received time, and all would be well.
683 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
684 being recognized when the domain list was tested by the match_domain
685 condition in an expansion string.
687 16. Added the ${str2b64: operator.
689 17. Exim was always calling setrlimit() to set a large limit for the number of
690 processes, without checking whether the existing limit was already
691 adequate. (It did check for the limit on file descriptors.) Furthermore,
692 errors from getrlimit() and setrlimit() were being ignored. Now they are
693 logged to the main and panic logs, but Exim does carry on, to try to do its
694 job under whatever limits there are.
696 18. Imported PCRE 5.0.
698 19. Trivial typo in log message " temporarily refused connection" (the leading
701 20. If the log selector return_path_on_delivery was set and an address was
702 redirected to /dev/null, the delivery process crashed because it assumed
703 that a return path would always be set for a "successful" delivery. In this
704 case, the whole delivery is bypassed as an optimization, and therefore no
707 21. Internal re-arrangement: the function for sending a challenge and reading
708 a response while authentication was assuming a zero-terminated challenge
709 string. It's now changed to take a pointer and a length, to allow for
710 binary data in such strings.
712 22. Added the cyrus_sasl authenticator (code supplied by MBM).
714 23. Exim was not respecting finduser_retries when seeking the login of the
715 uid under which it was called; it was always trying 10 times. (The default
716 setting of finduser_retries is zero.) Also, it was sleeping after the final
717 failure, which is pointless.
719 24. Implemented tls_on_connect_ports.
721 25. Implemented acl_smtp_predata.
723 26. If the domain in control=submission is set empty, Exim assumes that the
724 authenticated id is a complete email address when it generates From: or
725 Sender: header lines.
727 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
728 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
729 chown and chgrp in /bin and hostname in /usr/bin.
731 28. Exim was keeping the "process log" file open after each use, just as it
732 does for the main log. This opens the possibility of it remaining open for
733 long periods when the USR1 signal hits a daemon. Occasional processlog
734 errors were reported, that could have been caused by this. Anyway, it seems
735 much more sensible not to leave this file open at all, so that is what now
738 29. The long-running daemon process does not normally write to the log once it
739 has entered its main loop, and it closes the log before doing so. This is
740 so that log files can straightforwardly be renamed and moved. However,
741 there are a couple of unusual error situations where the daemon does write
742 log entries, and I had neglected to close the log afterwards.
744 30. The text of an SMTP error response that was received during a remote
745 delivery was being truncated at 512 bytes. This is too short for some of
746 the long messages that one sometimes sees. I've increased the limit to
749 31. It is now possible to make retry rules that apply only when a message has a
750 specific sender, in particular, an empty sender.
752 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
753 it possible to be selective about when SMTP synchronization is enforced.
755 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
757 32. Implemented hosts_connection_nolog.
759 33. Added an ACL for QUIT.
761 34. Setting "delay_warning=" to disable warnings was not working; it gave a
764 35. Added mailbox_size and mailbox_filecount to appendfile.
766 36. Added control = no_multiline_responses to ACLs.
768 37. There was a bug in the logic of the code that waits for the clock to tick
769 in the case where the clock went backwards by a substantial amount such
770 that the microsecond fraction of "now" was more than the microsecond
771 fraction of "then" (but the whole seconds number was less).
773 38. Added support for the libradius Radius client library this is found on
774 FreeBSD (previously only the radiusclient library was supported).
780 1. When certain lookups returned multiple values in the form name=value, the
781 quoting of the values was not always being done properly. Specifically:
782 (a) If the value started with a double quote, but contained no whitespace,
784 (b) If the value contained whitespace other than a space character (i.e.
785 tabs or newlines or carriage returns) it was not quoted.
786 This fix has been applied to the mysql and pgsql lookups by writing a
787 separate quoting function and calling it from the lookup code. The fix
788 should probably also be applied to nisplus, ibase and oracle lookups, but
789 since I cannot test any of those, I have not disturbed their existing code.
791 2. A hit in the callout cache for a specific address caused a log line with no
792 reason for rejecting RCPT. Now it says "Previous (cached) callout
793 verification failure".
795 3. There was an off-by-one bug in the queryprogram router. An over-long
796 return line was truncated at 256 instead of 255 characters, thereby
797 overflowing its buffer with the terminating zero. As well as fixing this, I
798 have increased the buffer size to 1024 (and made a note to document this).
800 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
801 when Exim is waiting for an SMTP response from a remote server, Exim
802 restarts its select() call on the socket, thereby resetting its timeout.
803 This is not a problem when such interrupts are rare. Somebody set up a cron
804 job to run exiwhat every 2 minutes, which is less than the normal select()
805 timeout (5 or 10 minutes). This meant that the select() timeout never
806 kicked in because it was always reset. I have fixed this by comparing the
807 time when an interrupt arrives with the time at the start of the first call
808 to select(). If more time than the timeout has elapsed, the interrupt is
809 treated as a timeout.
811 5. Some internal re-factoring in preparation for the addition of Sieve
812 extensions (by MH). In particular, the "personal" test is moved to a
813 separate function, and given an option for scanning Cc: and Bcc: (which is
814 not set for Exim filters).
816 6. When Exim created an email address using the login of the caller as the
817 local part (e.g. when creating a From: or Sender: header line), it was not
818 quoting the local part when it contained special characters such as @.
820 7. Installed new OpenBSD configuration files.
822 8. Reworded some messages for syntax errors in "and" and "or" conditions to
823 try to make them clearer.
825 9. Callout options, other than the timeout value, were being ignored when
826 verifying sender addresses in header lines. For example, when using
828 verify = header_sender/callout=no_cache
830 the cache was (incorrectly) being used.
832 10. Added a missing instance of ${EXE} to the exim_install script; this affects
833 only the Cygwin environment.
835 11. When return_path_on_delivery was set as a log selector, if different remote
836 addresses in the same message used different return paths and parallel
837 remote delivery occurred, the wrong values would sometimes be logged.
838 (Whenever a remote delivery process finished, the return path value from
839 the most recently started remote delivery process was logged.)
841 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
842 lines when AUTH and/or TLS are in use. This is the "received protocol"
843 field. Exim used to use "asmtp" for authenticated SMTP, without any
844 indication (in the protocol name) for TLS use. Now it follows the RFC and
845 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
846 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
847 names appear in log lines as well as in Received: header lines.
849 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
850 extensions, and comparison tests, and to fix some bugs.
852 14. Changes to the "personal" filter test:
854 (1) The test was buggy in that it was just doing the equivalent of
855 "contains" tests on header lines. For example, if a user's address was
856 anne@some.where, the "personal" test would incorrectly be true for
858 To: susanne@some.where
860 This test is now done by extracting each address from the header in turn,
861 and checking the entire address. Other tests that are part of "personal"
862 are now done using regular expressions (for example, to check local parts
863 of addresses in From: header lines).
865 (2) The list of non-personal local parts in From: addresses has been
866 extended to include "listserv", "majordomo", "*-request", and "owner-*",
867 taken from the Sieve specification recommendations.
869 (3) If the message contains any header line starting with "List-" it is
870 treated as non-personal.
872 (4) The test for "circular" in the Subject: header line has been removed
873 because it now seems ill-conceived.
875 15. Minor typos in src/EDITME comments corrected.
877 16. Installed latest exipick from John Jetmore.
879 17. If headers_add on a router specified a text string that was too long for
880 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
881 of string_sprintf() is now avoided.
883 18. $message_body_size was not set (it was always zero) when running the DATA
884 ACL and the local_scan() function.
886 19. For the "mail" command in an Exim filter, no default was being set for
887 the once_repeat time, causing a random time value to be used if "once" was
888 specified. (If the value happened to be <= 0, no repeat happened.) The
889 default is now 0s, meaning "never repeat". The "vacation" command was OK
890 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
891 (I found it when inspecting the code).
893 20. There is now an overall timeout for performing a callout verification. It
894 defaults to 4 times the callout timeout, which applies to individual SMTP
895 commands during the callout. The overall timeout applies when there is more
896 than one host that can be tried. The timeout is checked before trying the
897 next host. This prevents very long delays if there are a large number of
898 hosts and all are timing out (e.g. when the network connections are timing
899 out). The value of the overall timeout can be changed by specifying an
900 additional sub-option for "callout", called "maxwait". For example:
902 verify = sender/callout=5s,maxwait=20s
904 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
905 to the end before writing, but this should make it even safer).
907 22. Exim was forgetting that it had advertised PIPELINING for the second and
908 subsequent messages on an SMTP connection. It was also not resetting its
909 memory on STARTTLS and an internal HELO.
911 23. When Exim logs an SMTP synchronization error within a session, it now
912 records whether PIPELINING has been advertised or not.
914 24. Added 3 instances of "(long int)" casts to time_t variables that were being
915 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
916 rather than long int.
918 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
920 26. Added the never_mail option to autoreply.
926 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
927 crash if the getsockname() call failed; this can happen if a connection is
928 closed very soon after it is established. The problem was simply in the
929 order in which certain operations were done, causing Exim to try to write
930 to the SMTP stream before it had set up the file descriptor. The bug has
931 been fixed by making things happen in the correct order.
937 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
938 before the connection was closed, thus losing the rejection response.
940 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
941 some early Solaris releases, but causes trouble in current releases where
942 socklen_t is defined.
944 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
947 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
950 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
951 that releasing the top of it at the end releases what was used for sub-
952 expansions (unless the block got too big). However, discard this block if
953 the first thing is a variable or header, so that we can use its block when
954 it is dynamic (useful for very large $message_headers, for example).
956 6. Lookups now cache *every* query, not just the most recent. A new, separate
957 store pool is used for this. It can be recovered when all lookup caches are
958 flushed. Lookups now release memory at the end of their result strings.
959 This has involved some general refactoring of the lookup sources.
961 7. Some code has been added to the store_xxx() functions to reduce the amount
962 of flapping under certain conditions.
964 8. log_incoming_interface used to affect only the <= reception log lines. Now
965 it causes the local interface and port to be added to several more SMTP log
966 lines, for example "SMTP connection from", and rejection lines.
968 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
970 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
972 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
973 could be overwritten at the end of the current message (or the start of a
974 new message if it was set in a HELO ACL). The value is now preserved for
975 the duration of the SMTP connection.
977 12. If a transport had a headers_rewrite setting, and a matching header line
978 contained an unqualified address, that address was qualified, even if it
979 did not match any rewriting rules. The underlying bug was that the values
980 of the flags that permit the existence of unqualified sender and recipient
981 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
982 non-local messages, and by -bnq for local messages) were not being
983 preserved with the message after it was received.
985 13. When Exim was logging an SMTP synchronization error, it could sometimes log
986 "next input=" as part of the text comprising the host identity instead of
987 the correct text. The code was using the same buffer for two different
988 strings. However, depending on which order the printing function evaluated
989 its arguments, the bug did not always show up. Under Linux, for example, my
990 test suite worked just fine.
992 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
993 doesn't work with some older versions of Perl. It has been changed to "my",
994 which in any case is probably the better facility to use.
996 15. A really picky compiler found some instances of statements for creating
997 error messages that either had too many or two few arguments for the format
1000 16. The size of the buffer for calls to the DNS resolver has been increased
1001 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1002 for addresses that have a lot of PTR records. This alleviates a problem; it
1003 does not fully solve it.
1005 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1006 buffer now truncates the list and logs the incident, which is the same
1007 action as happens when Exim is looking up a host name and its aliases.
1008 Previously in this situation something unpredictable would happen;
1009 sometimes it was "internal error: store_reset failed".
1011 18. If a server dropped the connection unexpectedly when an Exim client was
1012 using GnuTLS and trying to read a response, the client delivery process
1013 crashed while trying to generate an error log message.
1015 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1016 string, for example:
1018 warn message = H1: something\nH2: something
1020 the text was added as a single header line from Exim's point of view
1021 though it ended up OK in the delivered message. However, searching for the
1022 second and subsequent header lines using $h_h2: did not work. This has been
1023 fixed. Similarly, if a system filter added multiple headers in this way,
1024 the routers could not see them.
1026 20. Expanded the error message when iplsearch is called with an invalid key to
1027 suggest using net-iplsearch in a host list.
1029 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1030 ACLs are no longer actually imposed (and a message to that effect is
1033 22. If a "gecos" field in a passwd entry contained escaped characters, in
1034 particular, if it contained a \" sequence, Exim got it wrong when building
1035 a From: or a Sender: header from that name. A second bug also caused
1036 incorrect handling when an unquoted " was present following a character
1037 that needed quoting.
1039 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1040 was not being matched caselessly.
1042 24. Arranged for all hyphens in the exim.8 source to be escaped with
1045 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1046 properly. Recipient callout cache records were still being keyed to include
1047 the sender, even when use_sender was set false. This led to far more
1048 callouts that were necessary. The sender is no longer included in the key
1049 when use_sender is false.
1051 26. Added "control = submission" modifier to ACLs.
1053 27. Added the ${base62d: operator to decode base 62 numbers.
1055 28. dnsdb lookups can now access SRV records.
1057 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1058 the configuration file.
1060 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1061 (-v) mode. This makes the output for a verbose queue run more intelligible.
1063 31. Added a use_postmaster feature to recipient callouts.
1065 32. Added the $body_zerocount variable, containing the number of binary zero
1066 bytes in the message body.
1068 33. The time of last modification of the "new" subdirectory is now used as the
1069 "mailbox time last read" when there is a quota error for a maildir
1072 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1074 35. Added +ignore_unknown as a special item in host lists.
1076 36. Code for decoding IPv6 addresses in host lists is now included, even if
1077 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1078 address was recognized as an IP address, but was then not correctly decoded
1079 into binary, causing unexpected and incorrect effects when compared with
1086 1. Very minor rewording of debugging text in manualroute to say "list of
1087 hosts" instead of "hostlist".
1089 2. If verify=header_syntax was set, and a header line with an unqualified
1090 address (no domain) and a large number of spaces between the end of the
1091 name and the colon was received, the reception process suffered a buffer
1092 overflow, and (when I tested it) crashed. This was caused by some obsolete
1093 code that should have been removed. The fix is to remove it!
1095 3. When running in the test harness, delay a bit after writing a bounce
1096 message to get a bit more predictability in the log output.
1098 4. Added a call to search_tidyup() just before forking a reception process. In
1099 theory, someone could use a lookup in the expansion of smtp_accept_max_
1100 per_host which, without the tidyup, could leave open a database connection.
1102 5. Added the variables $recipient_data and $sender_data which get set from a
1103 lookup success in an ACL "recipients" or "senders" condition, or a router
1104 "senders" option, similar to $domain_data and $local_part_data.
1106 6. Moved the writing of debug_print from before to after the "senders" test
1109 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1110 problems for message scanning, either using a data ACL, or using
1111 local_scan() because the Received: header was not generated till after they
1112 were called (in order to set the time as the time of reception completion).
1113 I have revised the way this works. The header is now generated after the
1114 body is received, but before the ACL or local_scan() are called. After they
1115 are run, the timestamp in the header is updated.
1121 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1122 before starting a queue runner without re-exec. This happened only when
1123 deliver_drop_privilege was set or when the Exim user was set to root. The
1124 effect of the bug was that timeouts during subsequent deliveries caused
1125 crashes instead of being properly handled. The handler is now left at its
1126 default (and expected) setting.
1128 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1129 message, again when deliver_drop_privilege is set or Exim is run as root.
1130 The bug described in (1) was not present in this case, but the tidying up
1131 of the other signals was missing. I have made the two cases consistent.
1133 3. The ignore_target_hosts setting on a manualroute router was being ignored
1134 for hosts that were looked up using the /MX notation.
1136 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1139 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1140 operated on the sender address. After changing the $sender_address to <>
1141 for the sender address verify, Exim was re-instated it as the original
1142 (before rewriting) address, but remembering that it had rewritten it, so it
1143 wasn't rewriting it again. This bug also had the effect of breaking the
1144 sender address verification caching when the sender address was rewritten.
1146 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1147 This has been changed so that if the ip literal address matches
1148 ignore_target_hosts, the router declines.
1150 7. Added expansion conditions match_domain, match_address, and match_local_
1151 part (NOT match_host).
1153 8. The placeholder for the Received: header didn't have a length field set.
1155 9. Added code to Exim itself and to exim_lock to test for a specific race
1156 condition that could lead to file corruption when using MBX delivery. The
1157 issue is with the lockfile that is created in /tmp. If this file is removed
1158 after a process has opened it but before that process has acquired a lock,
1159 there is the potential for a second process to recreate the file and also
1160 acquire a lock. This could lead to two Exim processes writing to the file
1161 at the same time. The added code performs the same test as UW imapd; it
1162 checks after acquiring the lock that its file descriptor still refers to
1163 the same named file.
1165 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1166 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1169 11. Added the smtp_active_hostname option. If used, this will typically be made
1170 to depend on the incoming interface address. Because $interface_address is
1171 not set up until the daemon has forked a reception process, error responses
1172 that can happen earlier (such as "too many connections") no longer contain
1175 12. If an expansion in a condition on a "warn" statement fails because a lookup
1176 defers, the "warn" statement is abandoned, and the next ACL statement is
1177 processed. Previously this caused the whole ACL to be aborted.
1179 13. Added the iplsearch lookup type.
1181 14. Added ident_timeout as a log selector.
1183 15. Added tls_certificate_verified as a log selector.
1185 16. Added a global option tls_require_ciphers (compare the smtp transport
1186 option of the same name). This controls incoming TLS connections.
1188 17. I finally figured out how to make tls_require_ciphers do a similar thing
1189 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1190 before starting the TLS session.
1192 18. Tabs are now shown as \t in -bP output.
1194 19. If the log selector return_path_on_delivery was set, Exim crashed when
1195 bouncing a message because it had too many Received: header lines.
1197 20. If two routers both had headers_remove settings, and the first one included
1198 a superfluous trailing colon, the final name in the first list and the
1199 first name in the second list were incorrectly joined into one item (with a
1200 colon in the middle).
1206 1. Added -C and -D options to the exinext utility, mainly to make it easier
1207 to include in the automated testing, but these could be helpful when
1208 multiple configurations are in use.
1210 2. The exinext utility was not formatting the output nicely when there was
1211 an alternate port involved in the retry record key, nor when there was a
1212 message id as well (for retries that were specific to a specific message
1213 and a specific host). It was also confused by IPv6 addresses, because of
1214 the additional colons they contain. I have fixed the IPv4 problem, and
1215 patched it up to do a reasonable job for IPv6.
1217 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1218 delivery, the log line now contains "pipelined" if PIPELINING was used.
1220 4. An SMTP transport process used to panic and die if the bind() call to set
1221 an explicit outgoing interface failed. This has been changed; it is now
1222 treated in the same way as a connect() failure.
1224 5. A reference to $sender_host_name in the part of a conditional expansion
1225 that was being skipped was still causing a DNS lookup. This no longer
1228 6. The def: expansion condition was not recognizing references to header lines
1229 that used bh_ and bheader_.
1231 7. Added the _cache feature to named lists.
1233 8. The code for checking quota_filecount in the appendfile transport was
1234 allowing one more file than it should have been.
1236 9. For compatibility with Sendmail, the command line option
1244 and sets the incoming protocol and host name (for trusted callers). The
1245 host name and its colon can be omitted when only the protocol is to be set.
1246 Note the Exim already has two private options, -pd and -ps, that refer to
1247 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1248 "s", but I don't think that's a major issue.
1250 10. A number of refactoring changes to the code, none of which should affect
1253 (a) The number of logging options was getting close to filling up the
1254 32-bit word that was used as a bit map. I have split them into two classes:
1255 those that are passed in the argument to log_write(), and those that are
1256 only ever tested independently outside of that function. These are now in
1257 separate 32-bit words, so there is plenty of room for expansion again.
1258 There is no change in the user interface or the logging behaviour.
1260 (b) When building, for example, log lines, the code previously used a
1261 macro that called string_cat() twice, in order to add two strings. This is
1262 not really sufficiently general. Furthermore, there was one instance where
1263 it was actually wrong because one of the argument was used twice, and in
1264 one call a function was used. (As it happened, calling the function twice
1265 did not affect the overall behaviour.) The macro has been replaced by a
1266 function that can join an arbitrary number of extra strings onto a growing
1269 (c) The code for expansion conditions now uses a table and a binary chop
1270 instead of a serial search (which was left over from when there were very
1271 few conditions). Also, it now recognizes conditions like "pam" even when
1272 the relevant support is not compiled in: a suitably worded error message is
1273 given if an attempt is made to use such a condition.
1275 11. Added ${time_interval:xxxxx}.
1277 12. A bug was causing one of the ddress fields not to be passed back correctly
1278 from remote delivery subprocesses. The field in question was not being
1279 subsequently used, so this caused to problems in practice.
1281 13. Added new log selectors queue_time and deliver_time.
1283 14. Might have fixed a bug in maildirsizefile handling that threw up
1284 "unexpected character" debug warnings, and recalculated the data
1285 unnecessarily. In any case, I expanded the warning message to give more
1288 15. Added the message "Restricted characters in address" to the statements in
1289 the default ACL that block characters like @ and % in local parts.
1291 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1292 Three changes have been made:
1294 (a) There was a serious bug; a negative response to MAIL caused the whole
1295 recipient domain to be cached as invalid, thereby blocking all messages
1296 to all local parts at the same domain, from all senders. This bug has
1297 been fixed. The domain is no longer cached after a negative response to
1298 MAIL if the sender used is not empty.
1300 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1303 (c) A new callout option, "use_sender" has been added for people who want
1304 the modified behaviour.
1310 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1313 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1314 indeed breaks things for older releases.
1316 3. Added additional logging to the case where there is a problem reading data
1317 from a filter that is running in a subprocess using a pipe, in order to
1318 try to track down a specific problem.
1320 4. Testing facility fudge: when running in the test harness and attempting
1321 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1322 getting "No route to host". Convert this to a timeout.
1324 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1327 6. Some OS don't have socklen_t but use size_t instead. This affects the
1328 fifth argument of getsockopt() amongst other things. This is now
1329 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1330 can be set for individual OS. I have set it for SunOS5, OSF1, and
1331 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1332 some earlier ones do not.
1334 7. Change 4.30/15 was not doing the test caselessly.
1336 8. The standard form for an IPv6 address literal was being rejected by address
1337 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1338 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1339 this, as well as the form without the "IPv6" on the front (but only when
1340 address literals are enabled, of course).
1342 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1344 10. Exim crashed if a message with an empty sender address specified by -f
1345 encountered a router with an errors_to setting. This could be provoked only
1346 by a command such as
1350 where an empty string was supplied; "<>" did not hit this bug.
1352 11. Installed PCRE release 4.5.
1354 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1355 remained set. It is now erased.
1357 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1358 times from message ids (which are base 36 rather than the normal 62).
1360 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1361 were being counted as actual protocol errors, and logged if the log
1362 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1363 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1364 and DATA following a set of rejected RCPTs do not count as protocol errors.
1365 In other words, Exim assumes they were pipelined, though this may not
1366 actually be the case. Of course, in all cases the client gets an
1367 appropriate error code.
1369 15. If a lookup fails in an ACL condition, a message about the failure may
1370 be available; it is used if testing the ACL cannot continue, because most
1371 such messages specify what the cause of the deferral is. However, some
1372 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1373 that caused an old message to be retained and used if a later statement
1374 caused a defer, replacing the real cause of the deferral.
1376 16. If an IP address had so many PTR records that the DNS lookup buffer
1377 was not large enough to hold them, Exim could crash while trying to process
1378 the truncated data. It now detects and logs this case.
1380 17. Further to 4.21/58, another change has been made: if (and only if) the
1381 first line of a message (the first header line) ends with CRLF, a bare LF
1382 in a subsequent header line has a space inserted after it, so as not to
1383 terminate the header.
1385 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1386 unnecessarily, used atoi() instead of strtol(), and didn't check the
1387 termination when getting file sizes from file names by regex.
1389 19. Completely re-implemented the support for maildirsize files, in the light
1390 of a number of problems with the previous contributed implementation
1391 (4.30/29). In particular:
1393 . If the quota is zero, the maildirsize file is maintained, but no quota is
1396 . If the maildir directory does not exist, it is created before any attempt
1397 to write a maildirsize file.
1399 . The quota value in the file is just a cache; if the quota is changed in
1400 the transport, the new value overrides.
1402 . A regular expression is available for excluding directories from the
1405 20. The autoreply transport checks the characters in options that define the
1406 message's headers; it allows continued headers, but it was checking with
1407 isspace() after an embedded newline instead of explicitly looking for a
1410 21. If all the "regular" hosts to which an address was routed had passed their
1411 expiry times, and had not reached their retry times, the address was
1412 bounced, even if fallback hosts were defined. Now Exim should go on to try
1415 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1416 equivalent code in the SMTP transport. Some hosts send humungous responses
1417 to HELO/EHLO, more than 1024 it seems.
1419 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1420 gives compiler warnings in some environments. I've now done it "properly",
1423 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1424 (because of problems with the built-in one) was declared to return uschar *
1425 instead of char *, causing compiler failure.
1427 25. Fixed a file descriptor leak when processing alias/forward files.
1429 26. Fixed a minor format string issue in dbfn.c.
1431 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1433 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1434 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1437 29. When a sender address is verified, it is cached, to save repeating the test
1438 when there is more than one recipient in a message. However, when the
1439 verification involves a callout, it is possible for different callout
1440 options to be set for different recipients. It is too complicated to keep
1441 track of this in the cache, so now Exim always runs a verification when a
1442 callout is required, relying on the callout cache for the optimization.
1443 The overhead is duplication of the address routing, but this should not be
1446 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1447 to be verified with callout=postmaster, and the main callout worked but the
1448 postmaster check failed, the verification correctly failed. However, if a
1449 subsequent RCPT command asked for sender verification *without* the
1450 postmaster check, incorrect caching caused this verification also to fail,
1453 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1454 it was not caching the DNS options (qualify_single, search_parents) that
1455 were used when the lookup failed. A subsequent lookup with different
1456 options therefore always gave the same answer, though there were cases
1457 where it should not have. (Example: a "domains = !$mx_any" option on a
1458 dnslookup router: the "domains" option is always processed without any
1459 widening, but the router might have qualify_single set.) Now Exim uses the
1460 cached value only when the same options are set.
1462 32. Added John Jetmore's "exipick" utility to the distribution.
1464 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1465 than a timeout (e.g. a certificate is required, and is not provided), an
1466 Exim server now closes the connection immediately. Previously it waited for
1467 the client to close - but if the client is SSL, it seems that they each
1468 wait for each other, leading to a delay before one of them times out.
1470 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1471 maintained 0.8.x compatibility because I don't think many are using it, and
1472 it is clearly obsolete.
1474 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1477 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1478 client certificate was expired. A simple patch fixes this, though I don't
1479 understand the full logic of why the verify callback is called multiple
1482 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1483 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1484 which causes problems with some clients (such as the Certicom SSL Plus
1485 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1486 disables the coutermeasure allowing Eudora to connect."
1488 38. Exim was not checking that a write() to a log file succeeded. This could
1489 lead to Bad Things if a log got too big, in particular if it hit a file
1490 size limit. Exim now panics and dies if it cannot write to a log file, just
1491 as it does if it cannot open a log file.
1493 39. Modified OS/Makefile-Linux so that it now contains
1495 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1497 The two -D definitions ensure that Exim is compiled with large file
1498 support, which makes it possible to handle log files that are bigger than
1501 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1502 instance) a domain was checked against a named list that involved a lookup,
1503 causing $domain_data to be set, then another domain was checked against the
1504 same list, then the first domain was re-checked, the value of $domain_data
1505 after the final check could be wrong. In particular, if the second check
1506 failed, it could be set empty. This bug probably also applied to
1509 41. The strip_trailing_dot option was not being applied to the address given
1510 with the -f command-line option.
1512 42. The code for reading a message's header from the spool was incrementing
1513 $received_count, but never initializing it. This meant that the value was
1514 incorrect (doubled) while delivering a message in the same process in which
1515 it was received. In the most common configuration of Exim, this never
1516 happens - a fresh exec is done - but it can happen when
1517 deliver_drop_privilege is set.
1519 43. When Exim logs an SMTP synchronization error - client data sent too soon -
1520 it now includes up to 150 characters of the unexpected data in the log
1523 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1524 and building data strings. The size of both of these buffers was 10 000
1525 bytes - far larger than anybody would *ever* want, thought I. Needless to
1526 say, somebody hit the limit. I have increased the maximum line length to
1527 20 000 and the maximum data length of concatenated lines to 100 000. I have
1528 also fixed two bugs, because there was no checking on these buffers. Tsk,
1529 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1530 buffer is too small.
1532 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1533 lsearch lookups. Now it does.
1535 46. When parsing a route_list item in a manualroute router, a fixed-length
1536 buffer was used for the list of hosts. I made this 1024 bytes long,
1537 thinking that nobody would ever have a list of hosts that long. Wrong.
1538 Somebody had a whole pile of complicated expansion conditions, and the
1539 string was silently truncated, leading to an expansion error. It turns out
1540 that it is easier to change to an unlimited length (owing to other changes
1541 that have happened since this code was originally written) than to build
1542 structure for giving a limitation error. The length of the item that
1543 expands into the list of hosts is now unlimited.
1545 47. The lsearch lookup could not handle data where the length of text line was
1546 more than 4095 characters. Such lines were truncated, leading to shortened
1547 data being returned. It should now handle lines of any length.
1549 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1550 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1553 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
1554 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1555 no longer happens in comment lines. A list of replacements is now placed
1556 at the head of all of the source files, except those whose only change is
1557 to replace PERL_COMMAND in the very first #! line.
1559 50. Replaced the slow insertion sort in queue.c, for sorting the list of
1560 messages on the queue, with a bottom-up merge sort, using code contributed
1561 by Michael Haardt. This should make operations like -bp somewhat faster on
1562 large queues. It won't affect queue runners, except when queue_run_in_order
1565 51. Installed eximstats 1.31 in the distribution.
1567 52. Added support for SRV lookups to the dnslookup router.
1569 53. If an ACL referred to $message_body or $message_body_end, the value was not
1570 reset for any messages that followed in the same SMTP session.
1572 54. The store-handling optimization for building very long strings was not
1573 differentiating between the different store pools. I don't think this
1574 actually made any difference in practice, but I've tidied it.
1576 55. While running the routers to verify a sender address, $sender_address
1577 was still set to the sender address. This is wrong, because when routing to
1578 send a bounce to the sender, it would be empty. Therefore, I have changed
1579 it so that, while verifying a sender address, $sender_address is set to <>.
1580 (There is no change to what happens when verifying a recipient address.)
1582 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1583 target A or AAAA records (if not already returned) without resetting the
1584 qualify_single or search_parents options of the DNS resolver. These are
1585 inappropriate in this case because the targets of MX and SRV records must
1586 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1587 target that, when qualified, matched something in the local domain. These
1588 two options are now turned off when doing these lookups.
1590 57. It seems that at least some releases of Reiserfs (which does not have the
1591 concept of a fixed number of inodes) returns zero and not -1 for the
1592 number of available inodes. This interacted badly with check_spool_inodes,
1593 which assumed that -1 was the "no such thing" setting. What I have done is
1594 to check that the total number of inodes is greater than zero before doing
1595 the test of how many are available.
1597 58. When a "warn" ACL statement has a log_message modifier, the message is
1598 remembered, and not repeated. This is to avoid a lot of repetition when a
1599 message has many recipients that cause the same warning to be written.
1600 Howewer, Exim was preserving the list of already written lines for an
1601 entire SMTP session, which doesn't seem right. The memory is now reset if a
1602 new message is started.
1604 59. The "rewrite" debugging flag was not showing the result of rewriting in the
1605 debugging output unless log_rewrite was also set.
1607 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1608 of (int)(handle) when we know that handle contains (void *)(-1).
1610 61. The Exim daemon panic-logs an error return when it closes the incoming
1611 connection. However "connection reset by peer" seems to be common, and
1612 isn't really an error worthy of noting specially, so that particular error
1615 62. When Exim is trying to find all the local interfaces, it used to panic and
1616 die if the ioctl to get the interface flags failed. However, it seems that
1617 on at least one OS (Solaris 9) it is possible to have an interface that is
1618 included in the list of interfaces, but for which you get a failure error
1619 for this call. This happens when the interface is not "plumbed" into a
1620 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1621 failure of the "get flags" call assumes that the interface is down.
1623 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
1624 makes life easier for people who are doing arithmetic on fields extracted
1625 from dates, where you often get leading zeros that should not be
1626 interpreted as octal.
1628 64. Added qualify_domain to the redirect router, to override the global
1631 65. If a pathologically long header line contained very many addresses (the
1632 report of this problem mentioned 10 000) and each of them was rewritten,
1633 Exim could use up a very large amount of memory. (It kept on making new
1634 copies of the header line as it rewrote, and never released the old ones.)
1635 At the expense of a bit more processing, the header rewriting function has
1636 been changed so that it no longer eats memory in this way.
1638 66. The generation of the Received: header has been moved from the time that a
1639 message starts to be received, to the time that it finishes. The timestamp
1640 in the Received: header should now be very close to that of the <= log
1641 line. There are two side-effects of this change:
1643 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1644 logged header lines no longer include the local Received: line, because
1645 it has not yet been created. The same applies to a copy of the message
1646 that is returned to a non-SMTP sender when a message is rejected.
1648 (b) When a filter file is tested using -bf, no additional Received: header
1649 is added to the test message. After some thought, I decided that this
1652 This change does not affect the value of $received_for. It is still set
1653 after address rewriting, but before local_scan() is called.
1655 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1657 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1658 gave an unhelpful panic error message, and a defer error. I have managed to
1659 change this behaviour so that it now rejects any supplied certificate,
1660 which seems right, as the list of acceptable certificates is empty.
1662 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1663 gave an unhelpful defer error. I have not managed to make this reject any
1664 supplied certificates, but the error message it gives is "no certificate
1665 supplied", which is not helpful.
1667 70. exigrep's output now also includes lines that are not associated with any
1668 message, but which match the given pattern. Implemented by a patch from
1669 Martin Sluka, which also tidied up the Perl a bit.
1671 71. Recipient callout verification, like sender verification, was using <> in
1672 the MAIL FROM command. This isn't really the right thing, since the actual
1673 sender may affect whether the remote host accepts the recipient or not. I
1674 have changed it to use the actual sender in the callout; this means that
1675 the cache record is now keyed on a recipient/sender pair, not just the
1676 recipient address. There doesn't seem to be a real danger of callout loops,
1677 since a callout by the remote host to check the sender would use <>.
1678 [SEE ABOVE: changed after hitting problems.]
1680 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1681 temporary errors. However, in the case of such a code being given after
1682 the end of a data transmission (i.e. after ".") Exim was failing to write
1683 a retry record for the message. (Yes, there was some broken host that was
1684 actually sending 8xx at this point.)
1686 73. An unknown lookup type in a host list could cause Exim to panic-die when
1687 the list was checked. (An example that provoked this was putting <; in the
1688 middle of a list instead of at the start.) If this happened during a DATA
1689 ACL check, a -D file could be left lying around. This kind of configuration
1690 error no longer causes Exim to die; instead it causes a defer errror. The
1691 incident is still logged to the main and panic logs.
1693 74. Buglet left over from Exim 3 conversion. The message "too many messages
1694 in one connection" was written to the rejectlog but not the mainlog, except
1695 when address rewriting (yes!) was being logged.
1697 75. Added write_rejectlog option.
1699 76. When a system filter was run not as root (that is, when system_filter_user
1700 was set), the values of the $n variables were not being returned to the
1701 main process; thus, they were not subsequently available in the $sn
1704 77. Added +return_path_on_delivery log selector.
1706 78. A connection timeout was being treated differently from recipients deferred
1707 when testing hosts_max_try with a message that was older than the host's
1708 retry timeout. (The host should not be counted, thus allowing all hosts to
1709 be tried at least once before bouncing.) This may have been the cause of an
1710 occasionally reported bug whereby a message would remain on the queue
1711 longer than the retry timeout, but would be bounced if a delivery was
1712 forced. I say "may" because I never totally pinned down the problem;
1713 setting up timeout/retry tests is difficult. See also the next item.
1715 79. The ultimate address timeout was not being applied to errors that involved
1716 a combination of host plus message (for example, a timeout on a MAIL
1717 command). When an address resolved to a number of possible hosts, and they
1718 were not all tried for each delivery (e.g. because of hosts_max_try), a
1719 message could remain on the queue longer than the retry timeout.
1721 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1724 81. Fixed an obscure SMTP outgoing bug which required at least the following
1725 conditions: (a) there was another message waiting for the same server;
1726 (b) the server returned 5xx to all RCPT commands in the first message so
1727 that the message was not completed; (c) the server dropped the connection
1728 or gave a negative response to the RSET that Exim sends to abort the
1729 transaction. The observed case was a dropped connection after DATA that had
1730 been sent in pipelining mode. That is, the server had advertised PIPELINING
1731 but was not implementing it correctly. The effect of the bug was incorrect
1732 behaviour, such as trying another host, and this could lead to a crash.
1738 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1739 and daemon.c were passed as pointers to ints; they should have been
1740 pointers to socklen_t variables (which are typically unsigned ints).
1742 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1745 3. Fixed a really odd bug that affected only the testing scheme; patching a
1746 certain fixed string in the binary changed the value of another string that
1747 happened to be identical to the end of the original first string.
1749 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1750 name", it returns that address as the IP address. On some operating
1751 systems (e.g. Solaris), it also passes back the IP address string as the
1752 "host name". However, on others (e.g. Linux), it passes back an empty
1753 string. Exim wasn't checking for this, and was changing the host name to an
1754 empty string, assuming it had been canonicized.
1756 5. Although rare, it is permitted to have more than one PTR record for a given
1757 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1758 all the names associated with an address, because they do in Solaris.
1759 However, it seems that they do not in Linux for data that comes from the
1760 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1761 I found this out when I moved to a new Linux workstation and tried to run
1762 the Exim test suite.
1764 To get round this problem I have changed the code so that it now does its
1765 own call to the DNS to look up PTR records when searching for a host name.
1766 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1767 addresses that are only in /etc/hosts are still found.
1769 This behaviour is, however, controlled by an option called host_lookup_
1770 order, which defaults to "bydns:byaddr". If people want to use the other
1771 order, or indeed, just use one or the other means of lookup, they can
1772 specify it in this variable.
1774 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1775 some operating systems, this comes back from gethostbyaddr() as an empty
1776 string, and this is what Exim used to test for. However, it seems that in
1777 other systems, "." is yielded. Exim now tests for this case too.
1779 7. The values of check_spool_space and check_log_space are now held internally
1780 as a number of kilobytes instead of an absolute number of bytes. If a
1781 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1782 kilobyte. This means that much larger values can be stored.
1784 8. Exim monitor: an attempt to get the action menu when not actually pointing
1785 at a message produces an empty menu entitled "No message selected". This
1786 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1787 no entries in it ("Shell widget menu has zero width and/or height"). So I
1788 have added a single, blank menu entry in this case.
1790 9. Added ${quote_local_part.
1792 10. MIME decoding is now applied to the contents of Subject: header lines when
1795 11. Now that a reference to $sender_host_address automatically causes a reverse
1796 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1797 host lookup before query-style lookups in lists that might use this
1798 variable. This has therefore been abolished, and the "net-" prefix is no
1799 longer necessary for query-style lookups.
1801 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1802 have been a typo for LFLAGS, so it has been changed.
1804 13. The install script calls Exim with "-C /dev/null" in order to find the
1805 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1806 to be output. Howeve, since Exim outputs its version number before the
1807 error, it didn't break the script. It just looked ugly. I fixed this by
1808 always allowing "-C /dev/null" if the caller is root.
1810 14. Ignore overlarge ACL variable number when reading spool file - insurance
1811 against a later release with more variables having written the file.
1813 15. The standard form for an IPv6 address literal was being rejected by EHLO.
1814 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1815 this, as well as the form without the "IPv6" on the front.
1817 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1818 OS/Makefile-Darwin file.
1820 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1821 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1823 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
1824 variable. However, some OS use ints for inodes. I've added cast to long int
1825 to get rid of the compiler warning.
1827 19. I had forgotten to lock out "/../" in configuration file names when
1828 ALT_CONFIG_PREFIX was set.
1830 20. Routers used for verification do not need to specify transports. However,
1831 if such a router generated a host list, and callout was configured, Exim
1832 crashed, because it could not find a port number from the (non-existent)
1833 transport. It now assumes port 25 in this circumstance.
1835 21. Added the -t option to exigrep.
1837 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1838 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1839 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1840 them is set, LOOKUP_LSEARCH is forced.
1842 23. "exim -bV" now outputs a list of lookups that are included in the binary.
1844 24. Added sender and host information to the "rejected by local_scan()" log
1845 line; previously there was no indication of these.
1847 25. Added .include_if_exists.
1849 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
1850 Linux. It turns out that not all file systems support this. Apparently some
1851 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1852 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1853 not supported on the file descriptor, is now ignored when Exim is trying to
1854 sync a directory. This applies only to Linux.
1856 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1858 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1859 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1862 29. Added support for maildirsize files from supplied patch (modified a bit).
1864 30. The use of :fail: followed by an empty string could lead Exim to respond to
1865 sender verification failures with (e.g.):
1867 550 Verification failed for <xxx>
1868 550 Sender verify failed
1870 where the first response line was missing the '-' that indicates it is not
1871 the final line of the response.
1873 31. The loop for finding the name of the user that called Exim had a hardwired
1874 limit of 10; it now uses the value of finduser_retries, which is used for
1875 all other user lookups.
1877 32. Added $received_count variable, available in data and not_smtp ACLs, and at
1880 33. Exim was neglecting to zero errno before one call of strtol() when
1881 expanding a string and expecting an integer value. On some systems this
1882 resulted in spurious "integer overflow" errors. Also, it was casting the
1883 result into an int without checking.
1885 34. Testing for a connection timeout using "timeout_connect" in the retry rules
1886 did not work. The code looks as if it has *never* worked, though it appears
1887 to have been documented since at least releast 1.62. I have made it work.
1889 35. The "timeout_DNS" error in retry rules, also documented since at least
1890 1.62, also never worked. As it isn't clear exactly what this means, and
1891 clearly it isn't a major issue, I have abolished the feature by treating it
1892 as "timeout", and writing a warning to the main and panic logs.
1894 36. The display of retry rules for -brt wasn't always showing the error code
1897 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
1898 timeout_connect_A, timeout_connect_MX.
1900 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
1901 to the empty sender.
1903 39. The daemon was not analysing the content of -oX till after it had closed
1904 stderr and disconnected from the controlling terminal. This meant that any
1905 syntax errors were only noted on the panic log, and the return code from
1906 the command was 0. By re-arranging the code a little, I've made the
1907 decoding happen first, so such errors now appear on stderr, and the return
1908 code is 1. However, the actual setting up of the sockets still happens in
1909 the disconnected process, so errors there are still only recorded on the
1912 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
1913 connections (as happens on some IP stacks) was logged at start up time as
1914 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
1915 it from "IPv6 and IPv4", which means that two separate sockets are being
1918 41. The debug output for gethostbyname2() or getipnodebyname() failures now
1919 says whether AF_INET or AF_INET6 was passed as an argument.
1921 42. Exiwhat output was messed up when time zones were included in log
1924 43. Exiwhat now gives more information about the daemon's listening ports,
1925 and whether -tls-on-connect was used.
1927 44. The "port" option of the smtp transport is now expanded.
1929 45. A "message" modifier in a "warn" statement in a non-message ACL was being
1930 silently ignored. Now an error message is written to the main and panic
1933 46. There's a new ACL modifier called "logwrite" which writes to a log file
1934 as soon as it is encountered.
1936 47. Added $local_user_uid and $local_user_gid at routing time.
1938 48. Exim crashed when trying to verify a sender address that was being
1941 49. Exim was recognizing only a space character after ".include". It now also
1942 recognizes a tab character.
1944 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
1945 extracting the relevant information from the specification. The man page no
1946 longer contains scrambled data for the -d option, and I've added a section
1947 at the front about calling Exim under different names.
1949 51. Added "extra_headers" argument to the "mail" command in filter files.
1951 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
1954 53. Installed eximstats 1.29.
1956 54. Added transport_filter_timeout as a generic transport option.
1958 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
1959 Cc: header lines. This was required by RFC 822, but it not required by RFC
1962 56. Exim used to add From:, Date:, and Message-Id: header lines to any
1963 incoming messages that did not have them. Now it does so only if the
1964 message originates locally, that is, if there is no associated remote host
1965 address. When Resent- header lines are present, this applies to the Resent-
1966 lines rather than the non-Resent- lines.
1968 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
1969 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
1971 58. Messages for configuration errors now include the name of the main
1972 configuration file - useful now that there may be more than one file in a
1973 list (.included file names were always shown).
1975 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
1976 for those rare installations that do not start the daemon as root or run it
1977 setuid root. I've cut out the call to initgroups() if the daemon is not
1980 60. The Exim user and group can now be bound into the binary as text strings
1981 that are looked up at the start of Exim's processing.
1983 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
1985 62. Added $mailstore_basename variable.
1987 63. Installed patch to sieve.c from Michael Haardt.
1989 64. When Exim failed to open the panic log after failing to open the main log,
1990 the original message it was trying to log was written to stderr and debug
1991 output, but if they were not available (the usual case in production), it
1992 was lost. Now it is written to syslog before the two lines that record the
1993 failures to open the logs.
1995 65. Users' Exim filters run in subprocesses under the user's uid. It is
1996 possible for a "deliver" command or an alias in a "personal" command to
1997 provoke an address rewrite. If logging of address rewriting is configured,
1998 this fails because the process is not running as root or exim. There may be
1999 a better way of dealing with this, but for the moment (because 4.30 needs
2000 to be released), I have disabled address rewrite logging when running a
2001 filter in a non-root, non-exim process.
2007 1. The buildconfig auxiliary program wasn't quoting the value set for
2008 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2009 not defined. This bug was masked in 4.22 by the effect that was fixed in
2012 2. Some messages that were rejected after a message id was allocated were
2013 shown as "incomplete" by exigrep. It no longer does this for messages that
2014 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2016 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2017 have allow_domain_literals set, the ID did not get logged in the <= line.
2018 Domain literals are now always recognized in Message-ID: header lines.
2020 4. The first argument for a ${extract expansion item is the key name or field
2021 number. Leading and trailing spaces in this item were not being ignored,
2022 causing some misleading effects.
2024 5. When deliver_drop_privilege was set, single queue runner processes started
2025 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2026 same command in the process it spins off) were not dropping privilege.
2028 6. When the daemon running as "exim" started a queue runner, it always
2029 re-executed Exim in the spun-off process. This is a waste of effort when
2030 deliver_drop_privilege is set. The new process now just calls the
2031 queue-runner function directly.
2037 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2040 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2041 ignored. Though the use of -oP was forcing the writing of a pid file, it
2042 was always written to the default place.
2044 3. If the message "no IP address found for host xxxx" is generated during
2045 incoming verification, it is now followed by identification of the incoming
2046 connection (so you can more easily find what provoked it).
2048 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2050 5. Added some features to "harden" Exim a bit more against certain attacks:
2052 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2053 be put in Local/Makefile. This is like the never_users runtime option,
2054 but it cannot be overridden. The default setting is "root".
2056 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2057 prefix string with which any file named in a -C command line option
2060 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2061 is retained for -C and -D only if the caller of Exim is root. Without
2062 it, the exim user may also use -C and -D and retain privilege.
2064 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2065 command line option is disabled.
2067 6. Macro names set by the -D option must start with an upper case letter, just
2068 like macro names defined in the configuration file.
2070 7. Added "dereference=" facility to LDAP.
2072 8. Two instances of the typo "uknown" in the source files are fixed.
2074 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2075 the Configure-Makefile script screwed up while processing it.
2077 10. Incorporated PCRE 4.4.
2079 11. The SMTP synchronization check was not operating right at the start of an
2080 SMTP session. For example, it could not catch a HELO sent before the client
2081 waited for the greeting. There is now a check for outstanding input at the
2082 point when the greeting is written. Because of the duplex, asynchronous
2083 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2084 way, but not yet received, when the check is performed.
2086 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2087 on TCP/IP sockets, because this apparently causes some broken clients to
2090 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2091 unchanged) from the Cygwin maintainer.
2093 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2094 was supported instead of testing for mbx. Effectively a typo.
2096 15. The spa authenticator server code was not checking that the input it
2097 received was valid base64.
2099 16. The debug output line for the "set" modifier in ACLs was not showing the
2100 name of the variable that was being set.
2102 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2104 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2105 lookup on its own. Something else was needed to trigger the lookup. For
2106 example, a match in host_lookup or the need for a host name in a host list.
2107 Now, if $sender_host_name is referenced and the host name has not yet been
2108 looked up, a lookup is performed. If the lookup fails, the variable remains
2109 empty, and $host_lookup_failed is set to "1".
2111 19. Added "eqi" as a case-independent comparison operator.
2113 20. The saslauthd authentication condition could segfault if neither service
2114 nor realm was specified.
2116 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2117 error message that was logged was misleading, and incoming SMTP
2118 connections were dropped. The message is now more accurate, and temporary
2119 errors are given to SMTP connections.
2121 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2122 (except RSET) with a 421 error, until QUIT is received. However, it was
2123 failing to send a response to QUIT.
2125 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2126 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2127 ACL is now run after the helo_try_verify_hosts code.
2129 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2130 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2131 case-independent, so other case variants are also recognized). Apparently
2132 some systems use these upper case variants.
2134 25. If more than two messages were waiting for the same host, and a transport
2135 filter was specified for the transport, Exim sent two messages over the
2136 same TCP/IP connection, and then failed with "socket operation on non-
2137 socket" when it tried to send the third.
2139 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2141 27. The extern definition of crypt16() in expand.c was not being excluded when
2142 the OS had its own crypt16() function.
2144 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2145 as a preferred synonym for return_size_limit, both as an option and as an
2148 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2150 30. Changed the default configuration ACL to relax the local part checking rule
2151 for addresses that are not in any local domains. For these addresses,
2152 slashes and pipe symbols are allowed within local parts, but the sequence
2153 /../ is explicitly forbidden.
2155 31. SPA server authentication was not clearing the challenge buffer before
2158 32. log_message in a "warn" ACL statement was writing to the reject log as
2159 well as to the main log, which contradicts the documentation and doesn't
2160 seem right (because no rejection is happening). So I have stopped it.
2162 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2163 However, I am unable to do any testing of this.
2165 34. Fixed an infelicity in the appendfile transport. When checking directories
2166 for a mailbox, to see if any needed to be created, it was accidentally
2167 using path names with one or more superfluous leading slashes; tracing
2168 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2170 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2171 contents are added to the log line that is written for every discarded
2172 recipient. (Previously a log_message setting was ignored.)
2174 36. The ${quote: operator now quotes the string if it is empty.
2176 37. The install script runs exim in order to find its version number. If for
2177 some reason other than non-existence or emptiness, which it checks, it
2178 could not run './exim', it was installing it with an empty version number,
2179 i.e. as "exim-". This error state is now caught, and the installation is
2182 38. An argument was missing from the function that creates an error message
2183 when Exim fails to connect to the socket for saslauthd authentication.
2184 This could cause Exim to crash, or give a corrupted message.
2186 39. Added isip, isip4, and isip6 to ${if conditions.
2188 40. The ACL variables $acl_xx are now saved with the message, and can be
2189 accessed later in routers, transports, and filters.
2191 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2192 strings in the file are not string-expanded.
2194 42. If a MAIL command specified a SIZE value that was too large to fit into an
2195 int variable, the check against message_size_limit failed. Such values are
2196 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2197 day this will have to be increased, but I don't think I want to be around
2198 when emails are that large.
2205 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2206 iconv() is not standard in FreeBSD.
2208 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2209 IPv6 enabled. The observed symptom was a segmentation fault on return from
2210 the function os_common_find_running_interfaces() in src/os.c.
2212 3. In the check_special_case() function in daemon.c I had used "errno" as an
2213 argument name, which causes warnings on some systems. This was basically a
2214 typo, since it was named "eno" in the comments!
2216 4. The code that waits for the clock to tick (at a resolution of some fraction
2217 of a second) so as to ensure message-id uniqueness was always waiting for
2218 at least one whole tick, when it could have waited for less. [This is
2219 almost certainly not relevant at current processor speeds, where it is
2220 unlikely to ever wait at all. But we try to future-proof.]
2222 5. The function that sleeps for a time interval that includes fractions of a
2223 second contained a race. It did not block SIGALRM between setting the
2224 timer, and suspending (a couple of lines later). If the interval was short
2225 and the sigsuspend() was delayed until after it had expired, the suspension
2226 never ended. On busy systems this could lead to processes getting stuck for
2229 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2230 process, before it forks any delivery processes. The open lookup caching
2231 mechanism meant that the open file or database connection was passed into
2232 the delivery process. The problem was that delivery processes always tidy
2233 up cached lookup data. This could cause a problem for the next delivery
2234 process started by the queue runner, because the external queue runner
2235 process does not know about the closure. So the next delivery process
2236 still has data in the lookup cache. In the case of a file lookup, there was
2237 no problem because closing a file descriptor in a subprocess doesn't affect
2238 the parent. However, if the lookup was caching a connection to a database,
2239 the connection was closed, and the second delivery process was likely to
2240 see errors such as "PGSQL: query failed: server closed the connection
2241 unexpectedly". The problem has been fixed by closing all cached lookups
2242 in a queue runner before running a delivery process.
2244 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2245 seem to have the "const" qualifier which it has on other OS. I've
2248 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2249 *of the same type* (client or server) with the same public name that an
2250 error should be diagnosed.
2252 9. When Exim looked up a host name for an IP address, but failed to find the
2253 original IP address when looking up the host name (a safety check), it
2254 output the message "<ip address> does not match any IP for NULL", which was
2255 confusing, to say the least. The bug was that the host name should have
2256 appeared instead of "NULL".
2258 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2259 user that is built into the binary, and the -C or -D options is used, root
2260 privilege is dropped before the configuration file is read. In addition,
2261 logging is switched to stderr instead of the normal log files. If the
2262 configuration then re-defines the Exim user, the unprivileged environment
2263 is probably not what is expected, so Exim logs a panic warning message (but
2266 However, if deliver_drop_privilege is set, the unprivileged state may well
2267 be exactly what is intended, so the warning has been cut out in that case,
2268 and Exim is allowed to try to write to its normal log files.
2274 1. smtp_return_error_details was not giving details for temporary sender
2275 or receiver verification errors.
2277 2. Diagnose a configuration error if two authenticators have the same public
2280 3. Exim used not to create the message log file for a message until the first
2281 delivery attempt. This could be confusing when incoming messages were held
2282 for policy or load reasons. The message log file is now created at the time
2283 the message is received, and an initial "Received" line is written to it.
2285 4. The automatically generated man page for command line options had a minor
2286 bug that caused no ill effects; however, a more serious problem was that
2287 the procedure for building the man page automatically didn't always
2288 operate. Consequently, release 4.20 contains an out-of-date version. This
2289 shouldn't happen again.
2291 5. When building Exim with embedded Perl support, the script that builds the
2292 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2293 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2295 6. The freeze_tell option was not being used for messages that were frozen on
2296 arrival, either by an ACL or by local_scan().
2298 7. Added the smtp_incomplete_transaction log selector.
2300 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2301 was accepting AUTH without a new EHLO.
2303 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2304 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2307 10. Logging of TCP/IP connections (when configured) now happens in the main
2308 daemon process instead of the child process, so that the TCP/IP connection
2309 count is more accurate (but it can never be perfect).
2311 11. The use of "drop" in a nested ACL was not being handled correctly in the
2312 outer ACL. Now, if condition failure induced by the nested "drop" causes
2313 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2314 or "require"), the connection is dropped.
2316 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2317 that yield "discard" can now be used with an "accept" or a "discard" verb,
2318 but an error is generated for any others (because I can't see a useful way
2319 to define what should happen).
2321 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2322 are now processed in the same way as lines in the Exim configuration file.
2323 In particular, continuation lines are supported.
2325 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2327 15. Added -ti meaning -t -i.
2329 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2330 domains, and warn by logging if others are found.
2332 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2333 returned from GIFCONF when Exim is trying to find the list of interfaces on
2334 a host. The code in os.c has been modified to copy each ifreq to an aligned
2335 structure in all cases.
2337 Also, in some cases, the returned ifreq's were being copied to a 'struct
2338 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2339 means the last couple of bytes of an IPv6 address could be chopped if the
2340 ifreq contained only a normal sockaddr (14 bytes storage).
2342 18. Named domain lists were not supported in the hosts_treat_as_local option.
2343 An entry such as +xxxx was not recognized, and was treated as a literal
2346 19. Ensure that header lines added by a DATA ACL are included in the reject log
2347 if the ACL subsequently rejects the message.
2349 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2350 MD5 (which is deprecated).
2352 21. When testing a filter file using -bf, Exim was writing a message when it
2353 took the sender from a "From " line in the message, but it was not doing so
2354 when it took $return_path from a Return-Path: header line. It now does.
2356 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2357 with a valid header line field name (a series of printing characters
2358 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2360 23. Changed "disc" in the source to "disk" to conform to the documentation and
2361 the book and for uniformity.
2363 24. Ignore Sendmail's -Ooption=value command line item.
2365 25. When execve() failed while trying to run a command in a pipe transport,
2366 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2367 could be confused with a return value of 69 from the command itself. This
2368 has been changed to 127, the value the shell returns if it is asked to run
2369 a non-existent command. The wording for the related log line suggests a
2370 non-existent command as the problem.
2372 26. If received_header_text expands to an empty string, do not add a Received:
2373 header line to the message. (Well, it adds a token one on the spool, but
2374 marks it "old" so that it doesn't get used or transmitted.)
2376 27. Installed eximstats 1.28 (addition of -nt option).
2378 28. There was no check for failure on the call to getsockname() in the daemon
2379 code. This can fail if there is a shortage of resources on the system, with
2380 ENOMEM, for example. A temporary error is now given on failure.
2382 29. Contrary to the C standard, it seems that in some environments, the
2383 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2384 program. Exim now does this explicitly; it affects the formatting of
2385 timestamps using strftime().
2387 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2388 complaints. I've now initialized all the variables, to avoid this.
2390 32. Header lines added by a system filter were not being "seen" during
2391 transport-time rewrites.
2393 33. The info_callback() function passed to OpenSSL is set up with type void
2394 (*)(SSL *, int, int), as described somewhere. However, when calling the
2395 function (actually a macro) that sets it up, the type void(*)() is
2396 expected. I've put in a cast to prevent warnings from picky compilers.
2398 34. If a DNS black list lookup found a CNAME record, but there were no A
2399 records associated with the domain it pointed at, Exim crashed.
2401 35. If a DNS black list lookup returned more than one A record, Exim ignored
2402 all but the first. It now scans all returned addresses if a particular IP
2403 value is being sought. In this situation, the contents of the
2404 $dnslist_value variable are a list of all the addresses, separated by a
2407 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2408 to accept a host name and all its aliases if the forward lookup for any of
2409 them yielded the IP address of the incoming connection. Now it accepts only
2410 those names whose forward lookup yields the correct IP address. Any other
2411 names are discarded. This closes a loophole whereby a rogue DNS
2412 administrator could create reverse DNS records to break through a
2413 wildcarded host restriction in an ACL.
2415 37. If a user filter or a system filter that ran in a subprocess used any of
2416 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2417 the wrong values were passed to the pipe command ($thisaddress had the
2418 value of $0, $0 had the value of $1, etc). This bug was introduced by
2419 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2421 38. Improved the line breaking for long SMTP error messages from ACLs.
2422 Previously, if there was no break point between 40 and 75 characters, Exim
2423 left the rest of the message alone. Two changes have been made: (a) I've
2424 reduced the minimum length to 35 characters; (b) if it can't find a break
2425 point between 35 and 75 characters, it looks ahead and uses the first one
2426 that it finds. This may give the occasional overlong line, but at least the
2427 remaining text gets split now.
2429 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2430 file descriptors might be low, and that setting 1000 would always raise it.
2431 It turns out that in some environments, the limit is already over 1000 and
2432 that lowering it causes trouble. So now Exim takes care not to decrease it.
2434 40. When delivering a message, the value of $return_path is set to $sender_
2435 address at the start of routing (routers may change the value). By an
2436 oversight, this default was not being set up when an address was tested by
2437 -bt or -bv, which affected the outcome if any router or filter referred to
2440 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2441 log only when "message" or "log_message" are set. However, if one of the
2442 conditions was an address verification, or a call to a nested ACL, the
2443 messages generated by the underlying test were being passed through. This
2444 no longer happens. The underlying message is available in $acl_verify_
2445 message for both "message" and "log_message" expansions, so it can be
2446 passed through if needed.
2448 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2449 new expansion $bh_ to give the encoded byte string without charset
2450 translation. Translation happens only if iconv() is available; HAVE_ICONV
2451 indicates this at build time. HEADERS_CHARSET gives the charset to
2452 translate to; headers_charset can change it in the configuration, and
2453 "headers charset" can change it in an individual filter file.
2455 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2456 that creates RFC 2047 encoded "words" labels them as that charset instead
2457 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2458 expansion operator; (ii) when Exim creates a From: line for a local
2459 message; (iii) when a header line is rewritten to include a "phrase" part.
2461 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2462 buggy, causing it to skip the first lines of messages whose message ID
2463 ended in 'D'. This would not have bitten before Exim release 4.14, because
2464 message IDs were unlikely to end in 'D' before then. The effect was to have
2465 incorrect size information for certain domains.
2467 45. #include "config.h" was missing at the start of the crypt16.c module. This
2468 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2471 46. If there was a timeout during a "random" callout check, Exim treated it as
2472 a failure of the random address, and carried on sending RSET and the real
2473 address. If the delay was just some slowness somewhere, the response to the
2474 original RCPT would be taken as a response to RSET and so on, causing
2475 mayhem of various kinds.
2477 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2478 when I implemented it. It didn't allow for the fact that some option values
2479 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2480 the right test for positive values.
2482 48. Domain names in DNS records are case-independent. Exim always looks them up
2483 in lower case. Some resolvers return domain names in exactly the case they
2484 appear in the zone file, that is, they may contain uppercase letters. Not
2485 all resolvers do this - some return always lower case. Exim was treating a
2486 change of case by a resolver as a change of domain, similar to a widening
2487 of a domain abbreviation. This triggered its re-routing code and so it was
2488 trying to route what was effectively the same domain again. This normally
2489 caused routing to fail (because the router wouldn't handle the domain
2490 twice). Now Exim checks for this case specially, and just changes the
2491 casing of the domain that it ultimately uses when it transmits the message
2494 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2497 50. If a filter generated a file delivery with a non-absolute name (possible if
2498 no home directory exists for the router), the forbid_file option was not
2501 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2502 the existing equality matching.
2504 52. Exim was using ints instead of ino_t variables in some places where it was
2505 dealing with inode numbers.
2507 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2508 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2509 called TMPDIR, and if it finds it is different, it changes its value.
2511 54. The smtp_printf() function is now made available to local_scan() so
2512 additional output lines can be written before returning. There is also an
2513 smtp_fflush() function to enable the detection of a dropped connection.
2514 The variables smtp_input and smtp_batched_input are exported to
2517 55. Changed the default runtime configuration: the message "Unknown user"
2518 has been removed from the ACL, and instead placed on the localuser router,
2519 using the cannot_route_message feature. This means that any verification
2520 failures that generate their own messages won't get overridden. Similarly,
2521 the "Unrouteable address" message that was in the ACL for unverifiable
2522 relay addresses has also been removed.
2524 56. Added hosts_avoid_esmtp to the smtp transport.
2526 57. The exicyclog script was not checking for the esoteric option
2527 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2528 will work only if exicyclog is run under the appropriate euid.
2530 58. Following a discussion on the list, the rules by which Exim recognises line
2531 endings on incoming messages have been changed. The -dropcr and drop_cr
2532 options are now no-ops, retained only for backwards compatibility. The
2533 following line terminators are recognized: LF CRLF CR. However, special
2534 processing applies to CR:
2536 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2537 nor a local message in the state where . is a terminator.
2539 (ii) If a bare CR is encountered in a header line, an extra space is added
2540 after the line terminator so as not to end the header. The reasoning
2541 behind this is that bare CRs in header lines are most likely either
2542 to be mistakes, or people trying to play silly games.
2544 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
2545 was being incorrectly given as 18 bytes larger than it should have been.
2546 This is a VOB (very old bug).
2548 60. This may never have affected anything current, but just in case it has:
2549 When the local host is found other than at the start of a list of hosts,
2550 the local host, those with the same MX, and any that follow, are discarded.
2551 When the list in question was part of a longer list of hosts, the following
2552 hosts (not currently being processed) were also being discarded. This no
2553 longer happens. I'm not sure if this situation could ever has previously
2556 61. Added the "/MX" feature to lists of hosts in the manualroute and query
2559 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2560 header. This is something that is recommended in a new Internet Draft, and
2561 is something that is documented as being done by Sendmail. There are two
2562 possible values. For messages generated by the autoreply transport, Exim
2565 Auto-Submitted: auto-replied
2567 whereas for all other generated messages (e.g. bounces) it adds
2569 Auto-Submitted: auto-generated
2571 63. The "personal" condition in filters now includes a test for the
2572 Auto-Submitted: header. If it contains the string "auto-" the message it
2573 not considered personal.
2575 64. Added rcpt_include_affixes as a generic transport option.
2577 65. Added queue_only_override (default true).
2579 66. Added the syslog_duplication option.
2581 67. If what should have been the first header line of a message consisted of
2582 a space followed by a colon, Exim was mis-interpreting it as a header line.
2583 It isn't of course - it is syntactically invalid and should therefore be
2584 treated as the start of the message body. The misbehaviour could have
2585 caused a number of strange effects, including loss of data in subsequent
2586 header lines, and spool format errors.
2588 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2589 client host had authenticated. This control can now be exercised by an ACL
2590 for more flexibility.
2592 69. By default, callouts do not happen when testing with -bh. There is now a
2593 variant, -bhc, which does actually run the callout code, including
2594 consulting and updating the callout cache.
2596 70. Added support for saslauthd authentication, courtesy of Alexander
2599 71. If statvfs() failed on the spool or log directories while checking their
2600 size for availability, Exim confusingly gave the error "space shortage".
2601 Furthermore, in debugging mode it crashed with a floating point exception.
2602 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2603 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2604 serious problem, Exim now writes to the main and panic logs when this
2605 happens, with details of the failure. It then refuses to accept the
2606 incoming message, giving the message "spool directory problem" or "log
2607 directory problem" with a 421 code for SMTP messages.
2609 72. When Exim is about to re-exec itself, it ensures that the file descriptors
2610 0, 1, and 2 exist, because some OS complain for execs without them (see
2611 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2612 descriptors. However, the code omitted to check that the open succeeded,
2613 causing mysterious errors if for some reason the permissions on /dev/null
2614 got screwed. Now Exim writes a message to the main and panic logs, and
2615 bombs out if it can't open /dev/null.
2617 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2618 interact so that it is all more flexible. It is supposed to remain
2619 backwards compatible. Also added extra_local_interfaces.
2621 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2622 to bomb out with an assertion failure - to the client this appears as a
2623 connection drop. This problem occurs in the part of the code that was taken
2624 from the Samba project. Fortunately, the assertion is in a very simple
2625 function, so I have fixed this by reproducing the function inline in the
2626 one place where it is called, and arranging for authentication to fail
2627 instead of killing the process with assert().
2629 75. The SPA client code was not working when the server requested OEM rather
2630 than Unicode encoding.
2632 76. Added code to make require_files with a specific uid setting more usable in
2633 the case where statting the file as root fails - usually a non-root-mounted
2634 NFS file system. When this happens and the failure is EACCES, Exim now
2635 forks a subprocess and does the per-uid checking as the relevant uid.
2637 77. Added process_log_path.
2639 78. If log_file_path was not explicitly set, a setting of check_log_space or
2640 check_log_inodes was ignored.
2642 79. If a space check for the spool or log partitions fails, the incident is now
2643 logged. Of course, in the latter case the data may get lost...
2645 80. Added the %p formatting code to string_format() so that it can be used to
2646 print addresses in debug_print(). Adjusted all the address printing in the
2647 debugging in store.c to use %p rather than %d.
2649 81. There was a concern that a line of code in smtp_in.c could overflow a
2650 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2651 initially expressed, the concern was not well-founded, because trailing
2652 spaces are removed early. However, if the trailing spaces were followed by
2653 a NULL, they did not get removed, so the overflow was possible. Two fixes
2656 (a) I re-wrote the offending code in a cleaner fashion.
2657 (b) If an incoming SMTP command contains a NULL character, it is rejected
2660 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
2661 runs initgroups(), so that if the Exim user is in any additional groups,
2662 they will be used during message reception.
2668 The change log for 4.20 and earlier releases has been archived.
projects / exim.git / blob