1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.87 2005/03/08 15:35:46 tom 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 PH/01. Installed a patch from the Sieve maintainer that allows -bf to be used
14 to test Sieve filters that use "vacation".
16 PH/02. Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
17 that changes the way the GnuTLS parameters are stored in the cache file.
18 The new format can be generated externally. For backward compatibility,
19 if the data in the cache doesn't make sense, Exim assumes it has read an
20 old-format file, and it generates new data and writes a new file. This
21 means that you can't go back to an older release without removing the
25 A note about Exim versions 4.44 and 4.50
26 ----------------------------------------
28 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
29 changes of various kinds. As a consequence, a big documentation update was
30 needed. This delayed the release for rather longer than seemed good, especially
31 in the light of a couple of (minor) security issues. Therefore, the changes
32 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
33 release. So 4.44 and 4.50 are in effect two different branches that both start
36 I have left the 4.50 change log unchanged; it contains all the changes since
37 4.43. The change log for 4.44 is below; many of its items are identical to
38 those for 4.50. This seems to be the most sensible way to preserve the
39 historical information.
45 1. Minor wording change to the doc/README.SIEVE file.
47 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
48 computation of the current number of files was incorrect.
50 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
51 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
54 4. Give more explanation in the error message when the command for a transport
55 filter fails to execute.
57 5. There are several places where Exim runs a non-Exim command in a
58 subprocess. The SIGUSR1 signal should be disabled for these processes. This
59 was being done only for the command run by the queryprogram router. It is
60 now done for all such subprocesses. The other cases are: ${run, transport
61 filters, and the commands run by the lmtp and pipe transports.
63 6. Added CONFIGURE_GROUP build-time option.
65 7. Some older OS have a limit of 256 on the maximum number of file
66 descriptors. Exim was using setrlimit() to set 1000 as a large value
67 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
68 systems. I've change it so that if it can't get 1000, it tries for 256.
70 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
71 was an oversight, and furthermore, ever since the addition of extra
72 controls (e.g. 4.43/32), the checks on when to allow different forms of
73 "control" were broken. There should now be diagnostics for all cases when a
74 control that does not make sense is encountered.
76 9. Added the /retain_sender option to "control=submission".
78 10. $recipients is now available in the predata ACL (oversight).
80 11. Tidy the search cache before the fork to do a delivery from a message
81 received from the command line. Otherwise the child will trigger a lookup
82 failure and thereby defer the delivery if it tries to use (for example) a
83 cached ldap connection that the parent has called unbind on.
85 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
86 of $address_data from the recipient verification was clobbered by the
89 13. The value of address_data from a sender verification is now available in
90 $sender_address_data in subsequent conditions in the ACL statement.
92 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
94 15. Added a new option "connect=<time>" to callout options, to set a different
97 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
98 was its contents. (It was OK if the option was not defined at all.)
100 17. A "Completed" log line is now written for messages that are removed from
101 the spool by the -Mrm option.
103 18. New variables $sender_verify_failure and $recipient_verify_failure contain
104 information about exactly what failed.
106 19. Added -dd to debug only the daemon process.
108 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
109 handles timeouts, both on the server side and network timeouts. Renamed the
110 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
112 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
113 It is now set to "smtps".
115 22. $host_address is now set to the target address during the checking of
118 23. When checking ignore_target_hosts for an ipliteral router, no host name was
119 being passed; this would have caused $sender_host_name to have been used if
120 matching the list had actually called for a host name (not very likely,
121 since this list is usually IP addresses). A host name is now passed as
124 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
125 code that specifies a non-restarting handler (typically sigaction() in
126 modern systems) in an attempt to fix a rare and obscure crash bug.
128 25. Narrowed the window for a race in the daemon that could cause it to ignore
129 SIGCHLD signals. This is not a major problem, because they are used only to
130 wake it up if nothing else does.
132 26. A malformed maildirsize file could cause Exim to calculate negative values
133 for the mailbox size or file count. Odd effects could occur as a result.
134 The maildirsize information is now recalculated if the size or filecount
137 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
138 support for a long time. Removed HAVE_SYS_VFS_H.
140 28. Installed the latest version of exipick from John Jetmore.
142 29. In an address list, if the pattern was not a regular expression, an empty
143 subject address (from a bounce message) matched only if the pattern was an
144 empty string. Non-empty patterns were not even tested. This was the wrong
145 because it is perfectly reasonable to use an empty address as part of a
146 database query. An empty address is now tested by patterns that are
147 lookups. However, all the other forms of pattern expect the subject to
148 contain a local part and a domain, and therefore, for them, an empty
149 address still always fails if the pattern is not itself empty.
151 30. Exim went into a mad DNS loop when attempting to do a callout where the
152 host was specified on an smtp transport, and looking it up yielded more
155 31. Re-factored the code for checking spool and log partition space into a
156 function that finds that data and another that does the check. The former
157 is then used to implement four new variables: $spool_space, $log_space,
158 $spool_inodes, and $log_inodes.
160 32. The RFC2047 encoding function was originally intended for short strings
161 such as real names; it was not keeping to the 75-character limit for
162 encoded words that the RFC imposes. It now respects the limit, and
163 generates multiple encoded words if necessary. To be on the safe side, I
164 have increased the buffer size for the ${rfc2047: expansion operator from
167 33. It is now permitted to omit both strings after an "if" condition; if the
168 condition is true, the result is "true". As before, when the second string
169 is omitted, a false condition yields an empty string. This makes it less
170 cumbersome to write custom ACL and router conditions.
172 34. Failure to deliver a bounce message always caused it to be frozen, even if
173 there was an errors_to setting on the router. The errors_to setting is now
176 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
177 canonical form (fully expanded) before being placed in
178 $sender_host_address.
180 36. The table in the code that translates DNS record types into text (T_A to
181 "A" for instance) was missing entries for NS and CNAME. It is just possible
182 that this could have caused confusion if both these types were looked up
183 for the same domain, because the text type is used as part of Exim's
184 per-process caching. But the chance of anyone hitting this buglet seems
187 37. The dnsdb lookup has been extended in a number of ways.
189 (1) There is a new type, "zns", which walks up the domain tree until it
190 finds some nameserver records. It should be used with care.
192 (2) There is a new type, "mxh", which is like "mx" except that it returns
193 just the host names, not the priorities.
195 (3) It is now possible to give a list of domains (or IP addresses) to be
196 looked up. The behaviour when one of the lookups defers can be
197 controlled by a keyword.
199 (4) It is now possible to specify the separator character for use when
200 multiple records are returned.
202 38. The dnslists ACL condition has been extended: it is now possible to supply
203 a list of IP addresses and/or domains to be looked up in a particular DNS
206 39. Added log_selector=+queue_time_overall.
208 40. When running the queue in the test harness, wait just a tad after forking a
209 delivery process, to get repeatability of debugging output.
211 41. Include certificate and key file names in error message when GnuTLS fails
212 to set them up, because the GnuTLS error message doesn't include the name
213 of the failing file when there is a problem reading it.
215 42. Allow both -bf and -bF in the same test run.
217 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
219 44. The "Exiscan patch" is now merged into the mainline Exim source.
221 45. Sometimes the final signoff response after QUIT could fail to get
222 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
223 < 0 before doing a fflush(). This bug looks as though it goes back to the
224 introduction of TLS in release 3.20, but "sometimes" must have been rare
225 because the tests only now provoked it.
227 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
228 (this can affect the format of dates).
230 47. exim_tidydb, when checking for the continued existence of a message for
231 which it has found a message-specific retry record, was not finding
232 messages that were in split spool directories. Consequently, it was
233 deleting retry records that should have stayed in existence.
235 48. Steve fixed some bugs in eximstats.
237 49. The SPA authentication driver was not abandoning authentication and moving
238 on to the next authenticator when an expansion was forced to fail,
239 contradicting the general specification for all authenticators. Instead it
240 was generating a temporary error. It now behaves as specified.
242 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
243 (the order specifies the preference for clients). The order is now AES256,
244 AES128, 3DES, ARCFOUR128.
246 51. Small patch to Sieve code - explicitly set From: when generating an
249 52. Exim crashed if a remote delivery caused a very long error message to be
250 recorded - for instance if somebody sent an entire SpamAssassin report back
251 as a large number of 550 error lines. This bug was coincidentally fixed by
252 increasing the size of one of Exim's internal buffers (big_buffer) that
253 happened as part of the Exiscan merge. However, to be on the safe side, I
254 have made the code more robust (and fixed the comments that describe what
257 53. Now that there can be additional text after "Completed" in log lines (if
258 the queue_time_overall log selector is set), a one-byte patch to exigrep
259 was needed to allow it to recognize "Completed" as not the last thing in
262 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
263 patch that reportedly fixes this has been added. I am not expert enough to
264 create a test for it. This is what the patch creator wrote:
266 "I found a little strange behaviour of ldap code when working with
267 Windows 2003 AD Domain, where users was placed in more than one
268 Organization Units. When I tried to give exim partial DN, the exit code
269 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
270 But simultaneously result of request was absolutely normal ldap result,
271 so I produce this patch..."
273 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
274 REFERENCE, so I have modified the code to exclude the patch when that macro
277 55. Some experimental protocols are using DNS PTR records for new purposes. The
278 keys for these records are domain names, not reversed IP addresses. The
279 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
280 leaves it alone. Component reversal etc. now happens only for IP addresses.
283 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
285 57. Double the size of the debug message buffer (to 2048) so that more of very
286 long debug lines gets shown.
288 58. The exicyclog utility now does better if the number of log files to keep
289 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
291 59. Two changes related to the smtp_active_hostname option:
293 (1) $smtp_active_hostname is now available as a variable.
294 (2) The default for smtp_banner uses $smtp_active_hostname instead
295 of $primary_hostname.
297 60. The host_aton() function is supposed to be passed a string that is known
298 to be a valid IP address. However, in the case of IPv6 addresses, it was
299 not checking this. This is a hostage to fortune. Exim now panics and dies
300 if the condition is not met. A case was found where this could be provoked
301 from a dnsdb PTR lookup with an IPv6 address that had more than 8
302 components; fortuitously, this particular loophole had already been fixed
303 by change 4.50/55 above.
305 If there are any other similar loopholes, the new check in host_aton()
306 itself should stop them being exploited. The report I received stated that
307 data on the command line could provoke the exploit when Exim was running as
308 exim, but did not say which command line option was involved. All I could
309 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
313 61. There was a buffer overflow vulnerability in the SPA authentication code
314 (which came originally from the Samba project). I have added a test to the
315 spa_base64_to_bits() function which I hope fixes it.
318 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
319 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
321 63. The daemon start-up calls getloadavg() while still root for those OS that
322 need the first call to be done as root, but it missed one case: when
323 deliver_queue_load_max is set with deliver_drop_privilege. This is
324 necessary for the benefit of the queue runner, because there is no re-exec
325 when deliver_drop_privilege is set.
327 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
330 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
331 was in use, was not putting the data itself into the right store pool;
332 consequently, it could be overwritten for a subsequent message in the same
333 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
336 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
338 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
339 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
340 the function were treating the return as a boolean value, which happened to
341 work because 0=false and not-0=true, but is not correct code.
343 68. The host_aton() function was not handling scoped IPv6 addresses (those
344 with, for example, "%eth0" on the end) correctly.
346 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
347 negated items (that is, ~something) in unsigned ints. Some compilers
348 apparently mutter when there is no cast.
350 70. If an address verification called from an ACL failed, and did not produce a
351 user-specific message (i.e. there was only a "system" message), nothing was
352 put in $acl_verify_message. In this situation, it now puts the system
355 71. Change 4.23/11 added synchronization checking at the start of an SMTP
356 session; change 4.31/43 added the unwanted input to the log line - except
357 that it did not do this in the start of session case. It now does.
359 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
360 This is wrong and can cause the other end to generate a synchronization
361 error if it is another Exim or anything else that does the synchronization
362 check. A QUIT command is no longer sent after a timeout.
364 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
367 74. The defer_ok option of callout verification was not working if it was used
368 when verifying addresses in header lines, that is, for this case:
370 verify = header_sender/callout=defer_ok
372 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
373 those file descriptors could be used for SMTP connections. If anything
374 wrote to stderr (the example that came up was "warn" in embedded Perl), it
375 could be sent to the SMTP client, causing chaos. The daemon now opens
376 stdin, stdout, and stderr to /dev/null when it puts itself into the
379 76. Arrange for output from Perl's "warn" command to be written to Exim's main
380 log by default. The user can override this with suitable Perl magic.
382 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
383 the log message when discard triggers, was not working for the DATA ACL or
384 for the non-SMTP ACL.
386 78. Error message wording change in sieve.c.
388 79. If smtp_accept_max_per_host was set, the number of connections could be
389 restricted to fewer than expected, because the daemon was trying to set up
390 a new connection before checking whether the processes handling previous
391 connections had finished. The check for completed processes is now done
392 earlier. On busy systems, this bug wouldn't be noticed because something
393 else would have woken the daemon, and it would have reaped the completed
396 80. If a message was submitted locally by a user whose login name contained one
397 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
398 It caused a spool format error. I have fixed the spool reading code. A
399 related problem was that the "from" clause in the Received: line became
400 illegal because of the space(s). It is now covered by ${quote_local_part.
402 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
405 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
407 83. Patch from Sieve maintainer to fix unterminated string problem in
410 84. Some minor changes to the Linux configuration files to help with other
411 OS variants using glibc.
413 85. One more patch for Sieve to update vacation handling to latest spec.
416 ----------------------------------------------------
417 See the note above about the 4.44 and 4.50 releases.
418 ----------------------------------------------------
424 1. Change 4.43/35 introduced a bug that caused file counts to be
425 incorrectly computed when quota_filecount was set in an appendfile
428 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
429 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
432 3. Give more explanation in the error message when the command for a transport
433 filter fails to execute.
435 4. There are several places where Exim runs a non-Exim command in a
436 subprocess. The SIGUSR1 signal should be disabled for these processes. This
437 was being done only for the command run by the queryprogram router. It is
438 now done for all such subprocesses. The other cases are: ${run, transport
439 filters, and the commands run by the lmtp and pipe transports.
441 5. Some older OS have a limit of 256 on the maximum number of file
442 descriptors. Exim was using setrlimit() to set 1000 as a large value
443 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
444 systems. I've change it so that if it can't get 1000, it tries for 256.
446 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
447 was an oversight, and furthermore, ever since the addition of extra
448 controls (e.g. 4.43/32), the checks on when to allow different forms of
449 "control" were broken. There should now be diagnostics for all cases when a
450 control that does not make sense is encountered.
452 7. $recipients is now available in the predata ACL (oversight).
454 8. Tidy the search cache before the fork to do a delivery from a message
455 received from the command line. Otherwise the child will trigger a lookup
456 failure and thereby defer the delivery if it tries to use (for example) a
457 cached ldap connection that the parent has called unbind on.
459 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
460 of $address_data from the recipient verification was clobbered by the
463 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
464 was its contents. (It was OK if the option was not defined at all.)
466 11. A "Completed" log line is now written for messages that are removed from
467 the spool by the -Mrm option.
469 12. $host_address is now set to the target address during the checking of
472 13. When checking ignore_target_hosts for an ipliteral router, no host name was
473 being passed; this would have caused $sender_host_name to have been used if
474 matching the list had actually called for a host name (not very likely,
475 since this list is usually IP addresses). A host name is now passed as
478 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
479 code that specifies a non-restarting handler (typically sigaction() in
480 modern systems) in an attempt to fix a rare and obscure crash bug.
482 15. Narrowed the window for a race in the daemon that could cause it to ignore
483 SIGCHLD signals. This is not a major problem, because they are used only to
484 wake it up if nothing else does.
486 16. A malformed maildirsize file could cause Exim to calculate negative values
487 for the mailbox size or file count. Odd effects could occur as a result.
488 The maildirsize information is now recalculated if the size or filecount
491 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
492 support for a long time. Removed HAVE_SYS_VFS_H.
494 18. Updated exipick to current release from John Jetmore.
496 19. Allow an empty sender to be matched against a lookup in an address list.
497 Previously the only cases considered were a regular expression, or an
500 20. Exim went into a mad DNS lookup loop when doing a callout where the
501 host was specified on the transport, if the DNS lookup yielded more than
504 21. The RFC2047 encoding function was originally intended for short strings
505 such as real names; it was not keeping to the 75-character limit for
506 encoded words that the RFC imposes. It now respects the limit, and
507 generates multiple encoded words if necessary. To be on the safe side, I
508 have increased the buffer size for the ${rfc2047: expansion operator from
511 22. Failure to deliver a bounce message always caused it to be frozen, even if
512 there was an errors_to setting on the router. The errors_to setting is now
515 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
516 canonical form (fully expanded) before being placed in
517 $sender_host_address.
519 24. Updated eximstats to version 1.33
521 25. Include certificate and key file names in error message when GnuTLS fails
522 to set them up, because the GnuTLS error message doesn't include the name
523 of the failing file when there is a problem reading it.
525 26. Expand error message when OpenSSL has problems setting up cert/key files.
528 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
529 (this can affect the format of dates).
531 28. exim_tidydb, when checking for the continued existence of a message for
532 which it has found a message-specific retry record, was not finding
533 messages that were in split spool directories. Consequently, it was
534 deleting retry records that should have stayed in existence.
536 29. eximstats updated to version 1.35
537 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
538 1.35 - bugfix such that pie charts by volume are generated correctly
540 30. The SPA authentication driver was not abandoning authentication and moving
541 on to the next authenticator when an expansion was forced to fail,
542 contradicting the general specification for all authenticators. Instead it
543 was generating a temporary error. It now behaves as specified.
545 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
546 (the order specifies the preference for clients). The order is now AES256,
547 AES128, 3DES, ARCFOUR128.
549 31. Small patch to Sieve code - explicitly set From: when generating an
552 32. Exim crashed if a remote delivery caused a very long error message to be
553 recorded - for instance if somebody sent an entire SpamAssassin report back
554 as a large number of 550 error lines. This bug was coincidentally fixed by
555 increasing the size of one of Exim's internal buffers (big_buffer) that
556 happened as part of the Exiscan merge. However, to be on the safe side, I
557 have made the code more robust (and fixed the comments that describe what
560 33. Some experimental protocols are using DNS PTR records for new purposes. The
561 keys for these records are domain names, not reversed IP addresses. The
562 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
563 leaves it alone. Component reversal etc. now happens only for IP addresses.
566 34. The host_aton() function is supposed to be passed a string that is known
567 to be a valid IP address. However, in the case of IPv6 addresses, it was
568 not checking this. This is a hostage to fortune. Exim now panics and dies
569 if the condition is not met. A case was found where this could be provoked
570 from a dnsdb PTR lookup with an IPv6 address that had more than 8
571 components; fortuitously, this particular loophole had already been fixed
572 by change 4.50/55 or 4.44/33 above.
574 If there are any other similar loopholes, the new check in host_aton()
575 itself should stop them being exploited. The report I received stated that
576 data on the command line could provoke the exploit when Exim was running as
577 exim, but did not say which command line option was involved. All I could
578 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
582 35. There was a buffer overflow vulnerability in the SPA authentication code
583 (which came originally from the Samba project). I have added a test to the
584 spa_base64_to_bits() function which I hope fixes it.
587 36. The daemon start-up calls getloadavg() while still root for those OS that
588 need the first call to be done as root, but it missed one case: when
589 deliver_queue_load_max is set with deliver_drop_privilege. This is
590 necessary for the benefit of the queue runner, because there is no re-exec
591 when deliver_drop_privilege is set.
593 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
594 was in use, was not putting the data itself into the right store pool;
595 consequently, it could be overwritten for a subsequent message in the same
596 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
599 38. Sometimes the final signoff response after QUIT could fail to get
600 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
601 < 0 before doing a fflush(). This bug looks as though it goes back to the
602 introduction of TLS in release 3.20, but "sometimes" must have been rare
603 because the tests only now provoked it.
609 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
610 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
611 Now it returns an integer. A number of calls were still expecting a T/F
612 return. Fortuitously, in all cases, the tests worked in OK situations,
613 which is the norm. However, things would have gone wrong on any write
614 failures on the smtp file descriptor. This function is used when sending
615 messages over SMTP and also when doing verify callouts.
617 2. When Exim is called to do synchronous delivery of a locally submitted
618 message (the -odf or -odi options), it no longer closes stderr before doing
621 3. Implemented the mua_wrapper option.
623 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
625 5. Implemented the functions header_remove(), header_testname(),
626 header_add_at_position(), and receive_remove_recipient(), and exported them
629 6. If an ACL "warn" statement specified the addition of headers, Exim already
630 inserted X-ACL-Warn: at the start if there was no header name. However, it
631 was not making this test for the second and subsequent header lines if
632 there were newlines in the string. This meant that an invalid header could
633 be inserted if Exim was badly configured.
635 7. Allow an ACL "warn" statement to add header lines at the start or after all
636 the Received: headers, as well as at the end.
638 8. Added the rcpt_4xx retry error code.
640 9. Added postmaster_mailfrom=xxx to callout verification option.
642 10. Added mailfrom=xxxx to the callout verification option, for verify=
645 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
646 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
648 12. Inserted some casts to stop certain compilers warning when using pointer
649 differences as field lengths or precisions in printf-type calls (mostly
650 affecting debugging statements).
652 13. Added optional readline() support for -be (dynamically loaded).
654 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
655 same clock tick as a message's arrival, so that its received time was the
656 same as the "first fail" time on the retry record, and that message
657 remained on the queue past the ultimate address timeout, every queue runner
658 would try a delivery (because it was past the ultimate address timeout) but
659 after another failure, the ultimate address timeout, which should have then
660 bounced the address, did not kick in. This was a "< instead of <=" error;
661 in most cases the first failure would have been in the next clock tick
662 after the received time, and all would be well.
664 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
665 being recognized when the domain list was tested by the match_domain
666 condition in an expansion string.
668 16. Added the ${str2b64: operator.
670 17. Exim was always calling setrlimit() to set a large limit for the number of
671 processes, without checking whether the existing limit was already
672 adequate. (It did check for the limit on file descriptors.) Furthermore,
673 errors from getrlimit() and setrlimit() were being ignored. Now they are
674 logged to the main and panic logs, but Exim does carry on, to try to do its
675 job under whatever limits there are.
677 18. Imported PCRE 5.0.
679 19. Trivial typo in log message " temporarily refused connection" (the leading
682 20. If the log selector return_path_on_delivery was set and an address was
683 redirected to /dev/null, the delivery process crashed because it assumed
684 that a return path would always be set for a "successful" delivery. In this
685 case, the whole delivery is bypassed as an optimization, and therefore no
688 21. Internal re-arrangement: the function for sending a challenge and reading
689 a response while authentication was assuming a zero-terminated challenge
690 string. It's now changed to take a pointer and a length, to allow for
691 binary data in such strings.
693 22. Added the cyrus_sasl authenticator (code supplied by MBM).
695 23. Exim was not respecting finduser_retries when seeking the login of the
696 uid under which it was called; it was always trying 10 times. (The default
697 setting of finduser_retries is zero.) Also, it was sleeping after the final
698 failure, which is pointless.
700 24. Implemented tls_on_connect_ports.
702 25. Implemented acl_smtp_predata.
704 26. If the domain in control=submission is set empty, Exim assumes that the
705 authenticated id is a complete email address when it generates From: or
706 Sender: header lines.
708 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
709 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
710 chown and chgrp in /bin and hostname in /usr/bin.
712 28. Exim was keeping the "process log" file open after each use, just as it
713 does for the main log. This opens the possibility of it remaining open for
714 long periods when the USR1 signal hits a daemon. Occasional processlog
715 errors were reported, that could have been caused by this. Anyway, it seems
716 much more sensible not to leave this file open at all, so that is what now
719 29. The long-running daemon process does not normally write to the log once it
720 has entered its main loop, and it closes the log before doing so. This is
721 so that log files can straightforwardly be renamed and moved. However,
722 there are a couple of unusual error situations where the daemon does write
723 log entries, and I had neglected to close the log afterwards.
725 30. The text of an SMTP error response that was received during a remote
726 delivery was being truncated at 512 bytes. This is too short for some of
727 the long messages that one sometimes sees. I've increased the limit to
730 31. It is now possible to make retry rules that apply only when a message has a
731 specific sender, in particular, an empty sender.
733 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
734 it possible to be selective about when SMTP synchronization is enforced.
736 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
738 32. Implemented hosts_connection_nolog.
740 33. Added an ACL for QUIT.
742 34. Setting "delay_warning=" to disable warnings was not working; it gave a
745 35. Added mailbox_size and mailbox_filecount to appendfile.
747 36. Added control = no_multiline_responses to ACLs.
749 37. There was a bug in the logic of the code that waits for the clock to tick
750 in the case where the clock went backwards by a substantial amount such
751 that the microsecond fraction of "now" was more than the microsecond
752 fraction of "then" (but the whole seconds number was less).
754 38. Added support for the libradius Radius client library this is found on
755 FreeBSD (previously only the radiusclient library was supported).
761 1. When certain lookups returned multiple values in the form name=value, the
762 quoting of the values was not always being done properly. Specifically:
763 (a) If the value started with a double quote, but contained no whitespace,
765 (b) If the value contained whitespace other than a space character (i.e.
766 tabs or newlines or carriage returns) it was not quoted.
767 This fix has been applied to the mysql and pgsql lookups by writing a
768 separate quoting function and calling it from the lookup code. The fix
769 should probably also be applied to nisplus, ibase and oracle lookups, but
770 since I cannot test any of those, I have not disturbed their existing code.
772 2. A hit in the callout cache for a specific address caused a log line with no
773 reason for rejecting RCPT. Now it says "Previous (cached) callout
774 verification failure".
776 3. There was an off-by-one bug in the queryprogram router. An over-long
777 return line was truncated at 256 instead of 255 characters, thereby
778 overflowing its buffer with the terminating zero. As well as fixing this, I
779 have increased the buffer size to 1024 (and made a note to document this).
781 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
782 when Exim is waiting for an SMTP response from a remote server, Exim
783 restarts its select() call on the socket, thereby resetting its timeout.
784 This is not a problem when such interrupts are rare. Somebody set up a cron
785 job to run exiwhat every 2 minutes, which is less than the normal select()
786 timeout (5 or 10 minutes). This meant that the select() timeout never
787 kicked in because it was always reset. I have fixed this by comparing the
788 time when an interrupt arrives with the time at the start of the first call
789 to select(). If more time than the timeout has elapsed, the interrupt is
790 treated as a timeout.
792 5. Some internal re-factoring in preparation for the addition of Sieve
793 extensions (by MH). In particular, the "personal" test is moved to a
794 separate function, and given an option for scanning Cc: and Bcc: (which is
795 not set for Exim filters).
797 6. When Exim created an email address using the login of the caller as the
798 local part (e.g. when creating a From: or Sender: header line), it was not
799 quoting the local part when it contained special characters such as @.
801 7. Installed new OpenBSD configuration files.
803 8. Reworded some messages for syntax errors in "and" and "or" conditions to
804 try to make them clearer.
806 9. Callout options, other than the timeout value, were being ignored when
807 verifying sender addresses in header lines. For example, when using
809 verify = header_sender/callout=no_cache
811 the cache was (incorrectly) being used.
813 10. Added a missing instance of ${EXE} to the exim_install script; this affects
814 only the Cygwin environment.
816 11. When return_path_on_delivery was set as a log selector, if different remote
817 addresses in the same message used different return paths and parallel
818 remote delivery occurred, the wrong values would sometimes be logged.
819 (Whenever a remote delivery process finished, the return path value from
820 the most recently started remote delivery process was logged.)
822 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
823 lines when AUTH and/or TLS are in use. This is the "received protocol"
824 field. Exim used to use "asmtp" for authenticated SMTP, without any
825 indication (in the protocol name) for TLS use. Now it follows the RFC and
826 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
827 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
828 names appear in log lines as well as in Received: header lines.
830 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
831 extensions, and comparison tests, and to fix some bugs.
833 14. Changes to the "personal" filter test:
835 (1) The test was buggy in that it was just doing the equivalent of
836 "contains" tests on header lines. For example, if a user's address was
837 anne@some.where, the "personal" test would incorrectly be true for
839 To: susanne@some.where
841 This test is now done by extracting each address from the header in turn,
842 and checking the entire address. Other tests that are part of "personal"
843 are now done using regular expressions (for example, to check local parts
844 of addresses in From: header lines).
846 (2) The list of non-personal local parts in From: addresses has been
847 extended to include "listserv", "majordomo", "*-request", and "owner-*",
848 taken from the Sieve specification recommendations.
850 (3) If the message contains any header line starting with "List-" it is
851 treated as non-personal.
853 (4) The test for "circular" in the Subject: header line has been removed
854 because it now seems ill-conceived.
856 15. Minor typos in src/EDITME comments corrected.
858 16. Installed latest exipick from John Jetmore.
860 17. If headers_add on a router specified a text string that was too long for
861 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
862 of string_sprintf() is now avoided.
864 18. $message_body_size was not set (it was always zero) when running the DATA
865 ACL and the local_scan() function.
867 19. For the "mail" command in an Exim filter, no default was being set for
868 the once_repeat time, causing a random time value to be used if "once" was
869 specified. (If the value happened to be <= 0, no repeat happened.) The
870 default is now 0s, meaning "never repeat". The "vacation" command was OK
871 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
872 (I found it when inspecting the code).
874 20. There is now an overall timeout for performing a callout verification. It
875 defaults to 4 times the callout timeout, which applies to individual SMTP
876 commands during the callout. The overall timeout applies when there is more
877 than one host that can be tried. The timeout is checked before trying the
878 next host. This prevents very long delays if there are a large number of
879 hosts and all are timing out (e.g. when the network connections are timing
880 out). The value of the overall timeout can be changed by specifying an
881 additional sub-option for "callout", called "maxwait". For example:
883 verify = sender/callout=5s,maxwait=20s
885 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
886 to the end before writing, but this should make it even safer).
888 22. Exim was forgetting that it had advertised PIPELINING for the second and
889 subsequent messages on an SMTP connection. It was also not resetting its
890 memory on STARTTLS and an internal HELO.
892 23. When Exim logs an SMTP synchronization error within a session, it now
893 records whether PIPELINING has been advertised or not.
895 24. Added 3 instances of "(long int)" casts to time_t variables that were being
896 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
897 rather than long int.
899 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
901 26. Added the never_mail option to autoreply.
907 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
908 crash if the getsockname() call failed; this can happen if a connection is
909 closed very soon after it is established. The problem was simply in the
910 order in which certain operations were done, causing Exim to try to write
911 to the SMTP stream before it had set up the file descriptor. The bug has
912 been fixed by making things happen in the correct order.
918 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
919 before the connection was closed, thus losing the rejection response.
921 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
922 some early Solaris releases, but causes trouble in current releases where
923 socklen_t is defined.
925 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
928 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
931 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
932 that releasing the top of it at the end releases what was used for sub-
933 expansions (unless the block got too big). However, discard this block if
934 the first thing is a variable or header, so that we can use its block when
935 it is dynamic (useful for very large $message_headers, for example).
937 6. Lookups now cache *every* query, not just the most recent. A new, separate
938 store pool is used for this. It can be recovered when all lookup caches are
939 flushed. Lookups now release memory at the end of their result strings.
940 This has involved some general refactoring of the lookup sources.
942 7. Some code has been added to the store_xxx() functions to reduce the amount
943 of flapping under certain conditions.
945 8. log_incoming_interface used to affect only the <= reception log lines. Now
946 it causes the local interface and port to be added to several more SMTP log
947 lines, for example "SMTP connection from", and rejection lines.
949 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
951 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
953 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
954 could be overwritten at the end of the current message (or the start of a
955 new message if it was set in a HELO ACL). The value is now preserved for
956 the duration of the SMTP connection.
958 12. If a transport had a headers_rewrite setting, and a matching header line
959 contained an unqualified address, that address was qualified, even if it
960 did not match any rewriting rules. The underlying bug was that the values
961 of the flags that permit the existence of unqualified sender and recipient
962 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
963 non-local messages, and by -bnq for local messages) were not being
964 preserved with the message after it was received.
966 13. When Exim was logging an SMTP synchronization error, it could sometimes log
967 "next input=" as part of the text comprising the host identity instead of
968 the correct text. The code was using the same buffer for two different
969 strings. However, depending on which order the printing function evaluated
970 its arguments, the bug did not always show up. Under Linux, for example, my
971 test suite worked just fine.
973 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
974 doesn't work with some older versions of Perl. It has been changed to "my",
975 which in any case is probably the better facility to use.
977 15. A really picky compiler found some instances of statements for creating
978 error messages that either had too many or two few arguments for the format
981 16. The size of the buffer for calls to the DNS resolver has been increased
982 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
983 for addresses that have a lot of PTR records. This alleviates a problem; it
984 does not fully solve it.
986 17. A dnsdb lookup for PTR records that receives more data than will fit in the
987 buffer now truncates the list and logs the incident, which is the same
988 action as happens when Exim is looking up a host name and its aliases.
989 Previously in this situation something unpredictable would happen;
990 sometimes it was "internal error: store_reset failed".
992 18. If a server dropped the connection unexpectedly when an Exim client was
993 using GnuTLS and trying to read a response, the client delivery process
994 crashed while trying to generate an error log message.
996 19. If a "warn" verb in an ACL added multiple headers to a message in a single
999 warn message = H1: something\nH2: something
1001 the text was added as a single header line from Exim's point of view
1002 though it ended up OK in the delivered message. However, searching for the
1003 second and subsequent header lines using $h_h2: did not work. This has been
1004 fixed. Similarly, if a system filter added multiple headers in this way,
1005 the routers could not see them.
1007 20. Expanded the error message when iplsearch is called with an invalid key to
1008 suggest using net-iplsearch in a host list.
1010 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1011 ACLs are no longer actually imposed (and a message to that effect is
1014 22. If a "gecos" field in a passwd entry contained escaped characters, in
1015 particular, if it contained a \" sequence, Exim got it wrong when building
1016 a From: or a Sender: header from that name. A second bug also caused
1017 incorrect handling when an unquoted " was present following a character
1018 that needed quoting.
1020 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1021 was not being matched caselessly.
1023 24. Arranged for all hyphens in the exim.8 source to be escaped with
1026 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1027 properly. Recipient callout cache records were still being keyed to include
1028 the sender, even when use_sender was set false. This led to far more
1029 callouts that were necessary. The sender is no longer included in the key
1030 when use_sender is false.
1032 26. Added "control = submission" modifier to ACLs.
1034 27. Added the ${base62d: operator to decode base 62 numbers.
1036 28. dnsdb lookups can now access SRV records.
1038 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1039 the configuration file.
1041 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1042 (-v) mode. This makes the output for a verbose queue run more intelligible.
1044 31. Added a use_postmaster feature to recipient callouts.
1046 32. Added the $body_zerocount variable, containing the number of binary zero
1047 bytes in the message body.
1049 33. The time of last modification of the "new" subdirectory is now used as the
1050 "mailbox time last read" when there is a quota error for a maildir
1053 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1055 35. Added +ignore_unknown as a special item in host lists.
1057 36. Code for decoding IPv6 addresses in host lists is now included, even if
1058 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1059 address was recognized as an IP address, but was then not correctly decoded
1060 into binary, causing unexpected and incorrect effects when compared with
1067 1. Very minor rewording of debugging text in manualroute to say "list of
1068 hosts" instead of "hostlist".
1070 2. If verify=header_syntax was set, and a header line with an unqualified
1071 address (no domain) and a large number of spaces between the end of the
1072 name and the colon was received, the reception process suffered a buffer
1073 overflow, and (when I tested it) crashed. This was caused by some obsolete
1074 code that should have been removed. The fix is to remove it!
1076 3. When running in the test harness, delay a bit after writing a bounce
1077 message to get a bit more predictability in the log output.
1079 4. Added a call to search_tidyup() just before forking a reception process. In
1080 theory, someone could use a lookup in the expansion of smtp_accept_max_
1081 per_host which, without the tidyup, could leave open a database connection.
1083 5. Added the variables $recipient_data and $sender_data which get set from a
1084 lookup success in an ACL "recipients" or "senders" condition, or a router
1085 "senders" option, similar to $domain_data and $local_part_data.
1087 6. Moved the writing of debug_print from before to after the "senders" test
1090 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1091 problems for message scanning, either using a data ACL, or using
1092 local_scan() because the Received: header was not generated till after they
1093 were called (in order to set the time as the time of reception completion).
1094 I have revised the way this works. The header is now generated after the
1095 body is received, but before the ACL or local_scan() are called. After they
1096 are run, the timestamp in the header is updated.
1102 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1103 before starting a queue runner without re-exec. This happened only when
1104 deliver_drop_privilege was set or when the Exim user was set to root. The
1105 effect of the bug was that timeouts during subsequent deliveries caused
1106 crashes instead of being properly handled. The handler is now left at its
1107 default (and expected) setting.
1109 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1110 message, again when deliver_drop_privilege is set or Exim is run as root.
1111 The bug described in (1) was not present in this case, but the tidying up
1112 of the other signals was missing. I have made the two cases consistent.
1114 3. The ignore_target_hosts setting on a manualroute router was being ignored
1115 for hosts that were looked up using the /MX notation.
1117 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1120 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1121 operated on the sender address. After changing the $sender_address to <>
1122 for the sender address verify, Exim was re-instated it as the original
1123 (before rewriting) address, but remembering that it had rewritten it, so it
1124 wasn't rewriting it again. This bug also had the effect of breaking the
1125 sender address verification caching when the sender address was rewritten.
1127 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1128 This has been changed so that if the ip literal address matches
1129 ignore_target_hosts, the router declines.
1131 7. Added expansion conditions match_domain, match_address, and match_local_
1132 part (NOT match_host).
1134 8. The placeholder for the Received: header didn't have a length field set.
1136 9. Added code to Exim itself and to exim_lock to test for a specific race
1137 condition that could lead to file corruption when using MBX delivery. The
1138 issue is with the lockfile that is created in /tmp. If this file is removed
1139 after a process has opened it but before that process has acquired a lock,
1140 there is the potential for a second process to recreate the file and also
1141 acquire a lock. This could lead to two Exim processes writing to the file
1142 at the same time. The added code performs the same test as UW imapd; it
1143 checks after acquiring the lock that its file descriptor still refers to
1144 the same named file.
1146 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1147 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1150 11. Added the smtp_active_hostname option. If used, this will typically be made
1151 to depend on the incoming interface address. Because $interface_address is
1152 not set up until the daemon has forked a reception process, error responses
1153 that can happen earlier (such as "too many connections") no longer contain
1156 12. If an expansion in a condition on a "warn" statement fails because a lookup
1157 defers, the "warn" statement is abandoned, and the next ACL statement is
1158 processed. Previously this caused the whole ACL to be aborted.
1160 13. Added the iplsearch lookup type.
1162 14. Added ident_timeout as a log selector.
1164 15. Added tls_certificate_verified as a log selector.
1166 16. Added a global option tls_require_ciphers (compare the smtp transport
1167 option of the same name). This controls incoming TLS connections.
1169 17. I finally figured out how to make tls_require_ciphers do a similar thing
1170 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1171 before starting the TLS session.
1173 18. Tabs are now shown as \t in -bP output.
1175 19. If the log selector return_path_on_delivery was set, Exim crashed when
1176 bouncing a message because it had too many Received: header lines.
1178 20. If two routers both had headers_remove settings, and the first one included
1179 a superfluous trailing colon, the final name in the first list and the
1180 first name in the second list were incorrectly joined into one item (with a
1181 colon in the middle).
1187 1. Added -C and -D options to the exinext utility, mainly to make it easier
1188 to include in the automated testing, but these could be helpful when
1189 multiple configurations are in use.
1191 2. The exinext utility was not formatting the output nicely when there was
1192 an alternate port involved in the retry record key, nor when there was a
1193 message id as well (for retries that were specific to a specific message
1194 and a specific host). It was also confused by IPv6 addresses, because of
1195 the additional colons they contain. I have fixed the IPv4 problem, and
1196 patched it up to do a reasonable job for IPv6.
1198 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1199 delivery, the log line now contains "pipelined" if PIPELINING was used.
1201 4. An SMTP transport process used to panic and die if the bind() call to set
1202 an explicit outgoing interface failed. This has been changed; it is now
1203 treated in the same way as a connect() failure.
1205 5. A reference to $sender_host_name in the part of a conditional expansion
1206 that was being skipped was still causing a DNS lookup. This no longer
1209 6. The def: expansion condition was not recognizing references to header lines
1210 that used bh_ and bheader_.
1212 7. Added the _cache feature to named lists.
1214 8. The code for checking quota_filecount in the appendfile transport was
1215 allowing one more file than it should have been.
1217 9. For compatibility with Sendmail, the command line option
1225 and sets the incoming protocol and host name (for trusted callers). The
1226 host name and its colon can be omitted when only the protocol is to be set.
1227 Note the Exim already has two private options, -pd and -ps, that refer to
1228 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1229 "s", but I don't think that's a major issue.
1231 10. A number of refactoring changes to the code, none of which should affect
1234 (a) The number of logging options was getting close to filling up the
1235 32-bit word that was used as a bit map. I have split them into two classes:
1236 those that are passed in the argument to log_write(), and those that are
1237 only ever tested independently outside of that function. These are now in
1238 separate 32-bit words, so there is plenty of room for expansion again.
1239 There is no change in the user interface or the logging behaviour.
1241 (b) When building, for example, log lines, the code previously used a
1242 macro that called string_cat() twice, in order to add two strings. This is
1243 not really sufficiently general. Furthermore, there was one instance where
1244 it was actually wrong because one of the argument was used twice, and in
1245 one call a function was used. (As it happened, calling the function twice
1246 did not affect the overall behaviour.) The macro has been replaced by a
1247 function that can join an arbitrary number of extra strings onto a growing
1250 (c) The code for expansion conditions now uses a table and a binary chop
1251 instead of a serial search (which was left over from when there were very
1252 few conditions). Also, it now recognizes conditions like "pam" even when
1253 the relevant support is not compiled in: a suitably worded error message is
1254 given if an attempt is made to use such a condition.
1256 11. Added ${time_interval:xxxxx}.
1258 12. A bug was causing one of the ddress fields not to be passed back correctly
1259 from remote delivery subprocesses. The field in question was not being
1260 subsequently used, so this caused to problems in practice.
1262 13. Added new log selectors queue_time and deliver_time.
1264 14. Might have fixed a bug in maildirsizefile handling that threw up
1265 "unexpected character" debug warnings, and recalculated the data
1266 unnecessarily. In any case, I expanded the warning message to give more
1269 15. Added the message "Restricted characters in address" to the statements in
1270 the default ACL that block characters like @ and % in local parts.
1272 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1273 Three changes have been made:
1275 (a) There was a serious bug; a negative response to MAIL caused the whole
1276 recipient domain to be cached as invalid, thereby blocking all messages
1277 to all local parts at the same domain, from all senders. This bug has
1278 been fixed. The domain is no longer cached after a negative response to
1279 MAIL if the sender used is not empty.
1281 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1284 (c) A new callout option, "use_sender" has been added for people who want
1285 the modified behaviour.
1291 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1294 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1295 indeed breaks things for older releases.
1297 3. Added additional logging to the case where there is a problem reading data
1298 from a filter that is running in a subprocess using a pipe, in order to
1299 try to track down a specific problem.
1301 4. Testing facility fudge: when running in the test harness and attempting
1302 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1303 getting "No route to host". Convert this to a timeout.
1305 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1308 6. Some OS don't have socklen_t but use size_t instead. This affects the
1309 fifth argument of getsockopt() amongst other things. This is now
1310 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1311 can be set for individual OS. I have set it for SunOS5, OSF1, and
1312 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1313 some earlier ones do not.
1315 7. Change 4.30/15 was not doing the test caselessly.
1317 8. The standard form for an IPv6 address literal was being rejected by address
1318 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1319 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1320 this, as well as the form without the "IPv6" on the front (but only when
1321 address literals are enabled, of course).
1323 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1325 10. Exim crashed if a message with an empty sender address specified by -f
1326 encountered a router with an errors_to setting. This could be provoked only
1327 by a command such as
1331 where an empty string was supplied; "<>" did not hit this bug.
1333 11. Installed PCRE release 4.5.
1335 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1336 remained set. It is now erased.
1338 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1339 times from message ids (which are base 36 rather than the normal 62).
1341 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1342 were being counted as actual protocol errors, and logged if the log
1343 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1344 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1345 and DATA following a set of rejected RCPTs do not count as protocol errors.
1346 In other words, Exim assumes they were pipelined, though this may not
1347 actually be the case. Of course, in all cases the client gets an
1348 appropriate error code.
1350 15. If a lookup fails in an ACL condition, a message about the failure may
1351 be available; it is used if testing the ACL cannot continue, because most
1352 such messages specify what the cause of the deferral is. However, some
1353 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1354 that caused an old message to be retained and used if a later statement
1355 caused a defer, replacing the real cause of the deferral.
1357 16. If an IP address had so many PTR records that the DNS lookup buffer
1358 was not large enough to hold them, Exim could crash while trying to process
1359 the truncated data. It now detects and logs this case.
1361 17. Further to 4.21/58, another change has been made: if (and only if) the
1362 first line of a message (the first header line) ends with CRLF, a bare LF
1363 in a subsequent header line has a space inserted after it, so as not to
1364 terminate the header.
1366 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1367 unnecessarily, used atoi() instead of strtol(), and didn't check the
1368 termination when getting file sizes from file names by regex.
1370 19. Completely re-implemented the support for maildirsize files, in the light
1371 of a number of problems with the previous contributed implementation
1372 (4.30/29). In particular:
1374 . If the quota is zero, the maildirsize file is maintained, but no quota is
1377 . If the maildir directory does not exist, it is created before any attempt
1378 to write a maildirsize file.
1380 . The quota value in the file is just a cache; if the quota is changed in
1381 the transport, the new value overrides.
1383 . A regular expression is available for excluding directories from the
1386 20. The autoreply transport checks the characters in options that define the
1387 message's headers; it allows continued headers, but it was checking with
1388 isspace() after an embedded newline instead of explicitly looking for a
1391 21. If all the "regular" hosts to which an address was routed had passed their
1392 expiry times, and had not reached their retry times, the address was
1393 bounced, even if fallback hosts were defined. Now Exim should go on to try
1396 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1397 equivalent code in the SMTP transport. Some hosts send humungous responses
1398 to HELO/EHLO, more than 1024 it seems.
1400 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1401 gives compiler warnings in some environments. I've now done it "properly",
1404 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1405 (because of problems with the built-in one) was declared to return uschar *
1406 instead of char *, causing compiler failure.
1408 25. Fixed a file descriptor leak when processing alias/forward files.
1410 26. Fixed a minor format string issue in dbfn.c.
1412 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1414 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1415 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1418 29. When a sender address is verified, it is cached, to save repeating the test
1419 when there is more than one recipient in a message. However, when the
1420 verification involves a callout, it is possible for different callout
1421 options to be set for different recipients. It is too complicated to keep
1422 track of this in the cache, so now Exim always runs a verification when a
1423 callout is required, relying on the callout cache for the optimization.
1424 The overhead is duplication of the address routing, but this should not be
1427 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1428 to be verified with callout=postmaster, and the main callout worked but the
1429 postmaster check failed, the verification correctly failed. However, if a
1430 subsequent RCPT command asked for sender verification *without* the
1431 postmaster check, incorrect caching caused this verification also to fail,
1434 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1435 it was not caching the DNS options (qualify_single, search_parents) that
1436 were used when the lookup failed. A subsequent lookup with different
1437 options therefore always gave the same answer, though there were cases
1438 where it should not have. (Example: a "domains = !$mx_any" option on a
1439 dnslookup router: the "domains" option is always processed without any
1440 widening, but the router might have qualify_single set.) Now Exim uses the
1441 cached value only when the same options are set.
1443 32. Added John Jetmore's "exipick" utility to the distribution.
1445 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1446 than a timeout (e.g. a certificate is required, and is not provided), an
1447 Exim server now closes the connection immediately. Previously it waited for
1448 the client to close - but if the client is SSL, it seems that they each
1449 wait for each other, leading to a delay before one of them times out.
1451 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1452 maintained 0.8.x compatibility because I don't think many are using it, and
1453 it is clearly obsolete.
1455 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1458 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1459 client certificate was expired. A simple patch fixes this, though I don't
1460 understand the full logic of why the verify callback is called multiple
1463 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1464 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1465 which causes problems with some clients (such as the Certicom SSL Plus
1466 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1467 disables the coutermeasure allowing Eudora to connect."
1469 38. Exim was not checking that a write() to a log file succeeded. This could
1470 lead to Bad Things if a log got too big, in particular if it hit a file
1471 size limit. Exim now panics and dies if it cannot write to a log file, just
1472 as it does if it cannot open a log file.
1474 39. Modified OS/Makefile-Linux so that it now contains
1476 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1478 The two -D definitions ensure that Exim is compiled with large file
1479 support, which makes it possible to handle log files that are bigger than
1482 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1483 instance) a domain was checked against a named list that involved a lookup,
1484 causing $domain_data to be set, then another domain was checked against the
1485 same list, then the first domain was re-checked, the value of $domain_data
1486 after the final check could be wrong. In particular, if the second check
1487 failed, it could be set empty. This bug probably also applied to
1490 41. The strip_trailing_dot option was not being applied to the address given
1491 with the -f command-line option.
1493 42. The code for reading a message's header from the spool was incrementing
1494 $received_count, but never initializing it. This meant that the value was
1495 incorrect (doubled) while delivering a message in the same process in which
1496 it was received. In the most common configuration of Exim, this never
1497 happens - a fresh exec is done - but it can happen when
1498 deliver_drop_privilege is set.
1500 43. When Exim logs an SMTP synchronization error - client data sent too soon -
1501 it now includes up to 150 characters of the unexpected data in the log
1504 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1505 and building data strings. The size of both of these buffers was 10 000
1506 bytes - far larger than anybody would *ever* want, thought I. Needless to
1507 say, somebody hit the limit. I have increased the maximum line length to
1508 20 000 and the maximum data length of concatenated lines to 100 000. I have
1509 also fixed two bugs, because there was no checking on these buffers. Tsk,
1510 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1511 buffer is too small.
1513 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1514 lsearch lookups. Now it does.
1516 46. When parsing a route_list item in a manualroute router, a fixed-length
1517 buffer was used for the list of hosts. I made this 1024 bytes long,
1518 thinking that nobody would ever have a list of hosts that long. Wrong.
1519 Somebody had a whole pile of complicated expansion conditions, and the
1520 string was silently truncated, leading to an expansion error. It turns out
1521 that it is easier to change to an unlimited length (owing to other changes
1522 that have happened since this code was originally written) than to build
1523 structure for giving a limitation error. The length of the item that
1524 expands into the list of hosts is now unlimited.
1526 47. The lsearch lookup could not handle data where the length of text line was
1527 more than 4095 characters. Such lines were truncated, leading to shortened
1528 data being returned. It should now handle lines of any length.
1530 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1531 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1534 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
1535 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1536 no longer happens in comment lines. A list of replacements is now placed
1537 at the head of all of the source files, except those whose only change is
1538 to replace PERL_COMMAND in the very first #! line.
1540 50. Replaced the slow insertion sort in queue.c, for sorting the list of
1541 messages on the queue, with a bottom-up merge sort, using code contributed
1542 by Michael Haardt. This should make operations like -bp somewhat faster on
1543 large queues. It won't affect queue runners, except when queue_run_in_order
1546 51. Installed eximstats 1.31 in the distribution.
1548 52. Added support for SRV lookups to the dnslookup router.
1550 53. If an ACL referred to $message_body or $message_body_end, the value was not
1551 reset for any messages that followed in the same SMTP session.
1553 54. The store-handling optimization for building very long strings was not
1554 differentiating between the different store pools. I don't think this
1555 actually made any difference in practice, but I've tidied it.
1557 55. While running the routers to verify a sender address, $sender_address
1558 was still set to the sender address. This is wrong, because when routing to
1559 send a bounce to the sender, it would be empty. Therefore, I have changed
1560 it so that, while verifying a sender address, $sender_address is set to <>.
1561 (There is no change to what happens when verifying a recipient address.)
1563 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1564 target A or AAAA records (if not already returned) without resetting the
1565 qualify_single or search_parents options of the DNS resolver. These are
1566 inappropriate in this case because the targets of MX and SRV records must
1567 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1568 target that, when qualified, matched something in the local domain. These
1569 two options are now turned off when doing these lookups.
1571 57. It seems that at least some releases of Reiserfs (which does not have the
1572 concept of a fixed number of inodes) returns zero and not -1 for the
1573 number of available inodes. This interacted badly with check_spool_inodes,
1574 which assumed that -1 was the "no such thing" setting. What I have done is
1575 to check that the total number of inodes is greater than zero before doing
1576 the test of how many are available.
1578 58. When a "warn" ACL statement has a log_message modifier, the message is
1579 remembered, and not repeated. This is to avoid a lot of repetition when a
1580 message has many recipients that cause the same warning to be written.
1581 Howewer, Exim was preserving the list of already written lines for an
1582 entire SMTP session, which doesn't seem right. The memory is now reset if a
1583 new message is started.
1585 59. The "rewrite" debugging flag was not showing the result of rewriting in the
1586 debugging output unless log_rewrite was also set.
1588 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1589 of (int)(handle) when we know that handle contains (void *)(-1).
1591 61. The Exim daemon panic-logs an error return when it closes the incoming
1592 connection. However "connection reset by peer" seems to be common, and
1593 isn't really an error worthy of noting specially, so that particular error
1596 62. When Exim is trying to find all the local interfaces, it used to panic and
1597 die if the ioctl to get the interface flags failed. However, it seems that
1598 on at least one OS (Solaris 9) it is possible to have an interface that is
1599 included in the list of interfaces, but for which you get a failure error
1600 for this call. This happens when the interface is not "plumbed" into a
1601 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1602 failure of the "get flags" call assumes that the interface is down.
1604 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
1605 makes life easier for people who are doing arithmetic on fields extracted
1606 from dates, where you often get leading zeros that should not be
1607 interpreted as octal.
1609 64. Added qualify_domain to the redirect router, to override the global
1612 65. If a pathologically long header line contained very many addresses (the
1613 report of this problem mentioned 10 000) and each of them was rewritten,
1614 Exim could use up a very large amount of memory. (It kept on making new
1615 copies of the header line as it rewrote, and never released the old ones.)
1616 At the expense of a bit more processing, the header rewriting function has
1617 been changed so that it no longer eats memory in this way.
1619 66. The generation of the Received: header has been moved from the time that a
1620 message starts to be received, to the time that it finishes. The timestamp
1621 in the Received: header should now be very close to that of the <= log
1622 line. There are two side-effects of this change:
1624 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1625 logged header lines no longer include the local Received: line, because
1626 it has not yet been created. The same applies to a copy of the message
1627 that is returned to a non-SMTP sender when a message is rejected.
1629 (b) When a filter file is tested using -bf, no additional Received: header
1630 is added to the test message. After some thought, I decided that this
1633 This change does not affect the value of $received_for. It is still set
1634 after address rewriting, but before local_scan() is called.
1636 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1638 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1639 gave an unhelpful panic error message, and a defer error. I have managed to
1640 change this behaviour so that it now rejects any supplied certificate,
1641 which seems right, as the list of acceptable certificates is empty.
1643 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1644 gave an unhelpful defer error. I have not managed to make this reject any
1645 supplied certificates, but the error message it gives is "no certificate
1646 supplied", which is not helpful.
1648 70. exigrep's output now also includes lines that are not associated with any
1649 message, but which match the given pattern. Implemented by a patch from
1650 Martin Sluka, which also tidied up the Perl a bit.
1652 71. Recipient callout verification, like sender verification, was using <> in
1653 the MAIL FROM command. This isn't really the right thing, since the actual
1654 sender may affect whether the remote host accepts the recipient or not. I
1655 have changed it to use the actual sender in the callout; this means that
1656 the cache record is now keyed on a recipient/sender pair, not just the
1657 recipient address. There doesn't seem to be a real danger of callout loops,
1658 since a callout by the remote host to check the sender would use <>.
1659 [SEE ABOVE: changed after hitting problems.]
1661 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1662 temporary errors. However, in the case of such a code being given after
1663 the end of a data transmission (i.e. after ".") Exim was failing to write
1664 a retry record for the message. (Yes, there was some broken host that was
1665 actually sending 8xx at this point.)
1667 73. An unknown lookup type in a host list could cause Exim to panic-die when
1668 the list was checked. (An example that provoked this was putting <; in the
1669 middle of a list instead of at the start.) If this happened during a DATA
1670 ACL check, a -D file could be left lying around. This kind of configuration
1671 error no longer causes Exim to die; instead it causes a defer errror. The
1672 incident is still logged to the main and panic logs.
1674 74. Buglet left over from Exim 3 conversion. The message "too many messages
1675 in one connection" was written to the rejectlog but not the mainlog, except
1676 when address rewriting (yes!) was being logged.
1678 75. Added write_rejectlog option.
1680 76. When a system filter was run not as root (that is, when system_filter_user
1681 was set), the values of the $n variables were not being returned to the
1682 main process; thus, they were not subsequently available in the $sn
1685 77. Added +return_path_on_delivery log selector.
1687 78. A connection timeout was being treated differently from recipients deferred
1688 when testing hosts_max_try with a message that was older than the host's
1689 retry timeout. (The host should not be counted, thus allowing all hosts to
1690 be tried at least once before bouncing.) This may have been the cause of an
1691 occasionally reported bug whereby a message would remain on the queue
1692 longer than the retry timeout, but would be bounced if a delivery was
1693 forced. I say "may" because I never totally pinned down the problem;
1694 setting up timeout/retry tests is difficult. See also the next item.
1696 79. The ultimate address timeout was not being applied to errors that involved
1697 a combination of host plus message (for example, a timeout on a MAIL
1698 command). When an address resolved to a number of possible hosts, and they
1699 were not all tried for each delivery (e.g. because of hosts_max_try), a
1700 message could remain on the queue longer than the retry timeout.
1702 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1705 81. Fixed an obscure SMTP outgoing bug which required at least the following
1706 conditions: (a) there was another message waiting for the same server;
1707 (b) the server returned 5xx to all RCPT commands in the first message so
1708 that the message was not completed; (c) the server dropped the connection
1709 or gave a negative response to the RSET that Exim sends to abort the
1710 transaction. The observed case was a dropped connection after DATA that had
1711 been sent in pipelining mode. That is, the server had advertised PIPELINING
1712 but was not implementing it correctly. The effect of the bug was incorrect
1713 behaviour, such as trying another host, and this could lead to a crash.
1719 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1720 and daemon.c were passed as pointers to ints; they should have been
1721 pointers to socklen_t variables (which are typically unsigned ints).
1723 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1726 3. Fixed a really odd bug that affected only the testing scheme; patching a
1727 certain fixed string in the binary changed the value of another string that
1728 happened to be identical to the end of the original first string.
1730 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1731 name", it returns that address as the IP address. On some operating
1732 systems (e.g. Solaris), it also passes back the IP address string as the
1733 "host name". However, on others (e.g. Linux), it passes back an empty
1734 string. Exim wasn't checking for this, and was changing the host name to an
1735 empty string, assuming it had been canonicized.
1737 5. Although rare, it is permitted to have more than one PTR record for a given
1738 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1739 all the names associated with an address, because they do in Solaris.
1740 However, it seems that they do not in Linux for data that comes from the
1741 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1742 I found this out when I moved to a new Linux workstation and tried to run
1743 the Exim test suite.
1745 To get round this problem I have changed the code so that it now does its
1746 own call to the DNS to look up PTR records when searching for a host name.
1747 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1748 addresses that are only in /etc/hosts are still found.
1750 This behaviour is, however, controlled by an option called host_lookup_
1751 order, which defaults to "bydns:byaddr". If people want to use the other
1752 order, or indeed, just use one or the other means of lookup, they can
1753 specify it in this variable.
1755 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1756 some operating systems, this comes back from gethostbyaddr() as an empty
1757 string, and this is what Exim used to test for. However, it seems that in
1758 other systems, "." is yielded. Exim now tests for this case too.
1760 7. The values of check_spool_space and check_log_space are now held internally
1761 as a number of kilobytes instead of an absolute number of bytes. If a
1762 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1763 kilobyte. This means that much larger values can be stored.
1765 8. Exim monitor: an attempt to get the action menu when not actually pointing
1766 at a message produces an empty menu entitled "No message selected". This
1767 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1768 no entries in it ("Shell widget menu has zero width and/or height"). So I
1769 have added a single, blank menu entry in this case.
1771 9. Added ${quote_local_part.
1773 10. MIME decoding is now applied to the contents of Subject: header lines when
1776 11. Now that a reference to $sender_host_address automatically causes a reverse
1777 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1778 host lookup before query-style lookups in lists that might use this
1779 variable. This has therefore been abolished, and the "net-" prefix is no
1780 longer necessary for query-style lookups.
1782 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1783 have been a typo for LFLAGS, so it has been changed.
1785 13. The install script calls Exim with "-C /dev/null" in order to find the
1786 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1787 to be output. Howeve, since Exim outputs its version number before the
1788 error, it didn't break the script. It just looked ugly. I fixed this by
1789 always allowing "-C /dev/null" if the caller is root.
1791 14. Ignore overlarge ACL variable number when reading spool file - insurance
1792 against a later release with more variables having written the file.
1794 15. The standard form for an IPv6 address literal was being rejected by EHLO.
1795 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1796 this, as well as the form without the "IPv6" on the front.
1798 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1799 OS/Makefile-Darwin file.
1801 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1802 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1804 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
1805 variable. However, some OS use ints for inodes. I've added cast to long int
1806 to get rid of the compiler warning.
1808 19. I had forgotten to lock out "/../" in configuration file names when
1809 ALT_CONFIG_PREFIX was set.
1811 20. Routers used for verification do not need to specify transports. However,
1812 if such a router generated a host list, and callout was configured, Exim
1813 crashed, because it could not find a port number from the (non-existent)
1814 transport. It now assumes port 25 in this circumstance.
1816 21. Added the -t option to exigrep.
1818 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1819 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1820 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1821 them is set, LOOKUP_LSEARCH is forced.
1823 23. "exim -bV" now outputs a list of lookups that are included in the binary.
1825 24. Added sender and host information to the "rejected by local_scan()" log
1826 line; previously there was no indication of these.
1828 25. Added .include_if_exists.
1830 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
1831 Linux. It turns out that not all file systems support this. Apparently some
1832 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1833 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1834 not supported on the file descriptor, is now ignored when Exim is trying to
1835 sync a directory. This applies only to Linux.
1837 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1839 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1840 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1843 29. Added support for maildirsize files from supplied patch (modified a bit).
1845 30. The use of :fail: followed by an empty string could lead Exim to respond to
1846 sender verification failures with (e.g.):
1848 550 Verification failed for <xxx>
1849 550 Sender verify failed
1851 where the first response line was missing the '-' that indicates it is not
1852 the final line of the response.
1854 31. The loop for finding the name of the user that called Exim had a hardwired
1855 limit of 10; it now uses the value of finduser_retries, which is used for
1856 all other user lookups.
1858 32. Added $received_count variable, available in data and not_smtp ACLs, and at
1861 33. Exim was neglecting to zero errno before one call of strtol() when
1862 expanding a string and expecting an integer value. On some systems this
1863 resulted in spurious "integer overflow" errors. Also, it was casting the
1864 result into an int without checking.
1866 34. Testing for a connection timeout using "timeout_connect" in the retry rules
1867 did not work. The code looks as if it has *never* worked, though it appears
1868 to have been documented since at least releast 1.62. I have made it work.
1870 35. The "timeout_DNS" error in retry rules, also documented since at least
1871 1.62, also never worked. As it isn't clear exactly what this means, and
1872 clearly it isn't a major issue, I have abolished the feature by treating it
1873 as "timeout", and writing a warning to the main and panic logs.
1875 36. The display of retry rules for -brt wasn't always showing the error code
1878 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
1879 timeout_connect_A, timeout_connect_MX.
1881 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
1882 to the empty sender.
1884 39. The daemon was not analysing the content of -oX till after it had closed
1885 stderr and disconnected from the controlling terminal. This meant that any
1886 syntax errors were only noted on the panic log, and the return code from
1887 the command was 0. By re-arranging the code a little, I've made the
1888 decoding happen first, so such errors now appear on stderr, and the return
1889 code is 1. However, the actual setting up of the sockets still happens in
1890 the disconnected process, so errors there are still only recorded on the
1893 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
1894 connections (as happens on some IP stacks) was logged at start up time as
1895 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
1896 it from "IPv6 and IPv4", which means that two separate sockets are being
1899 41. The debug output for gethostbyname2() or getipnodebyname() failures now
1900 says whether AF_INET or AF_INET6 was passed as an argument.
1902 42. Exiwhat output was messed up when time zones were included in log
1905 43. Exiwhat now gives more information about the daemon's listening ports,
1906 and whether -tls-on-connect was used.
1908 44. The "port" option of the smtp transport is now expanded.
1910 45. A "message" modifier in a "warn" statement in a non-message ACL was being
1911 silently ignored. Now an error message is written to the main and panic
1914 46. There's a new ACL modifier called "logwrite" which writes to a log file
1915 as soon as it is encountered.
1917 47. Added $local_user_uid and $local_user_gid at routing time.
1919 48. Exim crashed when trying to verify a sender address that was being
1922 49. Exim was recognizing only a space character after ".include". It now also
1923 recognizes a tab character.
1925 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
1926 extracting the relevant information from the specification. The man page no
1927 longer contains scrambled data for the -d option, and I've added a section
1928 at the front about calling Exim under different names.
1930 51. Added "extra_headers" argument to the "mail" command in filter files.
1932 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
1935 53. Installed eximstats 1.29.
1937 54. Added transport_filter_timeout as a generic transport option.
1939 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
1940 Cc: header lines. This was required by RFC 822, but it not required by RFC
1943 56. Exim used to add From:, Date:, and Message-Id: header lines to any
1944 incoming messages that did not have them. Now it does so only if the
1945 message originates locally, that is, if there is no associated remote host
1946 address. When Resent- header lines are present, this applies to the Resent-
1947 lines rather than the non-Resent- lines.
1949 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
1950 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
1952 58. Messages for configuration errors now include the name of the main
1953 configuration file - useful now that there may be more than one file in a
1954 list (.included file names were always shown).
1956 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
1957 for those rare installations that do not start the daemon as root or run it
1958 setuid root. I've cut out the call to initgroups() if the daemon is not
1961 60. The Exim user and group can now be bound into the binary as text strings
1962 that are looked up at the start of Exim's processing.
1964 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
1966 62. Added $mailstore_basename variable.
1968 63. Installed patch to sieve.c from Michael Haardt.
1970 64. When Exim failed to open the panic log after failing to open the main log,
1971 the original message it was trying to log was written to stderr and debug
1972 output, but if they were not available (the usual case in production), it
1973 was lost. Now it is written to syslog before the two lines that record the
1974 failures to open the logs.
1976 65. Users' Exim filters run in subprocesses under the user's uid. It is
1977 possible for a "deliver" command or an alias in a "personal" command to
1978 provoke an address rewrite. If logging of address rewriting is configured,
1979 this fails because the process is not running as root or exim. There may be
1980 a better way of dealing with this, but for the moment (because 4.30 needs
1981 to be released), I have disabled address rewrite logging when running a
1982 filter in a non-root, non-exim process.
1988 1. The buildconfig auxiliary program wasn't quoting the value set for
1989 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
1990 not defined. This bug was masked in 4.22 by the effect that was fixed in
1993 2. Some messages that were rejected after a message id was allocated were
1994 shown as "incomplete" by exigrep. It no longer does this for messages that
1995 are rejected by local_scan() or the DATA or non-SMTP ACLs.
1997 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
1998 have allow_domain_literals set, the ID did not get logged in the <= line.
1999 Domain literals are now always recognized in Message-ID: header lines.
2001 4. The first argument for a ${extract expansion item is the key name or field
2002 number. Leading and trailing spaces in this item were not being ignored,
2003 causing some misleading effects.
2005 5. When deliver_drop_privilege was set, single queue runner processes started
2006 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2007 same command in the process it spins off) were not dropping privilege.
2009 6. When the daemon running as "exim" started a queue runner, it always
2010 re-executed Exim in the spun-off process. This is a waste of effort when
2011 deliver_drop_privilege is set. The new process now just calls the
2012 queue-runner function directly.
2018 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2021 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2022 ignored. Though the use of -oP was forcing the writing of a pid file, it
2023 was always written to the default place.
2025 3. If the message "no IP address found for host xxxx" is generated during
2026 incoming verification, it is now followed by identification of the incoming
2027 connection (so you can more easily find what provoked it).
2029 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2031 5. Added some features to "harden" Exim a bit more against certain attacks:
2033 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2034 be put in Local/Makefile. This is like the never_users runtime option,
2035 but it cannot be overridden. The default setting is "root".
2037 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2038 prefix string with which any file named in a -C command line option
2041 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2042 is retained for -C and -D only if the caller of Exim is root. Without
2043 it, the exim user may also use -C and -D and retain privilege.
2045 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2046 command line option is disabled.
2048 6. Macro names set by the -D option must start with an upper case letter, just
2049 like macro names defined in the configuration file.
2051 7. Added "dereference=" facility to LDAP.
2053 8. Two instances of the typo "uknown" in the source files are fixed.
2055 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2056 the Configure-Makefile script screwed up while processing it.
2058 10. Incorporated PCRE 4.4.
2060 11. The SMTP synchronization check was not operating right at the start of an
2061 SMTP session. For example, it could not catch a HELO sent before the client
2062 waited for the greeting. There is now a check for outstanding input at the
2063 point when the greeting is written. Because of the duplex, asynchronous
2064 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2065 way, but not yet received, when the check is performed.
2067 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2068 on TCP/IP sockets, because this apparently causes some broken clients to
2071 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2072 unchanged) from the Cygwin maintainer.
2074 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2075 was supported instead of testing for mbx. Effectively a typo.
2077 15. The spa authenticator server code was not checking that the input it
2078 received was valid base64.
2080 16. The debug output line for the "set" modifier in ACLs was not showing the
2081 name of the variable that was being set.
2083 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2085 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2086 lookup on its own. Something else was needed to trigger the lookup. For
2087 example, a match in host_lookup or the need for a host name in a host list.
2088 Now, if $sender_host_name is referenced and the host name has not yet been
2089 looked up, a lookup is performed. If the lookup fails, the variable remains
2090 empty, and $host_lookup_failed is set to "1".
2092 19. Added "eqi" as a case-independent comparison operator.
2094 20. The saslauthd authentication condition could segfault if neither service
2095 nor realm was specified.
2097 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2098 error message that was logged was misleading, and incoming SMTP
2099 connections were dropped. The message is now more accurate, and temporary
2100 errors are given to SMTP connections.
2102 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2103 (except RSET) with a 421 error, until QUIT is received. However, it was
2104 failing to send a response to QUIT.
2106 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2107 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2108 ACL is now run after the helo_try_verify_hosts code.
2110 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2111 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2112 case-independent, so other case variants are also recognized). Apparently
2113 some systems use these upper case variants.
2115 25. If more than two messages were waiting for the same host, and a transport
2116 filter was specified for the transport, Exim sent two messages over the
2117 same TCP/IP connection, and then failed with "socket operation on non-
2118 socket" when it tried to send the third.
2120 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2122 27. The extern definition of crypt16() in expand.c was not being excluded when
2123 the OS had its own crypt16() function.
2125 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2126 as a preferred synonym for return_size_limit, both as an option and as an
2129 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2131 30. Changed the default configuration ACL to relax the local part checking rule
2132 for addresses that are not in any local domains. For these addresses,
2133 slashes and pipe symbols are allowed within local parts, but the sequence
2134 /../ is explicitly forbidden.
2136 31. SPA server authentication was not clearing the challenge buffer before
2139 32. log_message in a "warn" ACL statement was writing to the reject log as
2140 well as to the main log, which contradicts the documentation and doesn't
2141 seem right (because no rejection is happening). So I have stopped it.
2143 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2144 However, I am unable to do any testing of this.
2146 34. Fixed an infelicity in the appendfile transport. When checking directories
2147 for a mailbox, to see if any needed to be created, it was accidentally
2148 using path names with one or more superfluous leading slashes; tracing
2149 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2151 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2152 contents are added to the log line that is written for every discarded
2153 recipient. (Previously a log_message setting was ignored.)
2155 36. The ${quote: operator now quotes the string if it is empty.
2157 37. The install script runs exim in order to find its version number. If for
2158 some reason other than non-existence or emptiness, which it checks, it
2159 could not run './exim', it was installing it with an empty version number,
2160 i.e. as "exim-". This error state is now caught, and the installation is
2163 38. An argument was missing from the function that creates an error message
2164 when Exim fails to connect to the socket for saslauthd authentication.
2165 This could cause Exim to crash, or give a corrupted message.
2167 39. Added isip, isip4, and isip6 to ${if conditions.
2169 40. The ACL variables $acl_xx are now saved with the message, and can be
2170 accessed later in routers, transports, and filters.
2172 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2173 strings in the file are not string-expanded.
2175 42. If a MAIL command specified a SIZE value that was too large to fit into an
2176 int variable, the check against message_size_limit failed. Such values are
2177 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2178 day this will have to be increased, but I don't think I want to be around
2179 when emails are that large.
2186 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2187 iconv() is not standard in FreeBSD.
2189 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2190 IPv6 enabled. The observed symptom was a segmentation fault on return from
2191 the function os_common_find_running_interfaces() in src/os.c.
2193 3. In the check_special_case() function in daemon.c I had used "errno" as an
2194 argument name, which causes warnings on some systems. This was basically a
2195 typo, since it was named "eno" in the comments!
2197 4. The code that waits for the clock to tick (at a resolution of some fraction
2198 of a second) so as to ensure message-id uniqueness was always waiting for
2199 at least one whole tick, when it could have waited for less. [This is
2200 almost certainly not relevant at current processor speeds, where it is
2201 unlikely to ever wait at all. But we try to future-proof.]
2203 5. The function that sleeps for a time interval that includes fractions of a
2204 second contained a race. It did not block SIGALRM between setting the
2205 timer, and suspending (a couple of lines later). If the interval was short
2206 and the sigsuspend() was delayed until after it had expired, the suspension
2207 never ended. On busy systems this could lead to processes getting stuck for
2210 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2211 process, before it forks any delivery processes. The open lookup caching
2212 mechanism meant that the open file or database connection was passed into
2213 the delivery process. The problem was that delivery processes always tidy
2214 up cached lookup data. This could cause a problem for the next delivery
2215 process started by the queue runner, because the external queue runner
2216 process does not know about the closure. So the next delivery process
2217 still has data in the lookup cache. In the case of a file lookup, there was
2218 no problem because closing a file descriptor in a subprocess doesn't affect
2219 the parent. However, if the lookup was caching a connection to a database,
2220 the connection was closed, and the second delivery process was likely to
2221 see errors such as "PGSQL: query failed: server closed the connection
2222 unexpectedly". The problem has been fixed by closing all cached lookups
2223 in a queue runner before running a delivery process.
2225 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2226 seem to have the "const" qualifier which it has on other OS. I've
2229 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2230 *of the same type* (client or server) with the same public name that an
2231 error should be diagnosed.
2233 9. When Exim looked up a host name for an IP address, but failed to find the
2234 original IP address when looking up the host name (a safety check), it
2235 output the message "<ip address> does not match any IP for NULL", which was
2236 confusing, to say the least. The bug was that the host name should have
2237 appeared instead of "NULL".
2239 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2240 user that is built into the binary, and the -C or -D options is used, root
2241 privilege is dropped before the configuration file is read. In addition,
2242 logging is switched to stderr instead of the normal log files. If the
2243 configuration then re-defines the Exim user, the unprivileged environment
2244 is probably not what is expected, so Exim logs a panic warning message (but
2247 However, if deliver_drop_privilege is set, the unprivileged state may well
2248 be exactly what is intended, so the warning has been cut out in that case,
2249 and Exim is allowed to try to write to its normal log files.
2255 1. smtp_return_error_details was not giving details for temporary sender
2256 or receiver verification errors.
2258 2. Diagnose a configuration error if two authenticators have the same public
2261 3. Exim used not to create the message log file for a message until the first
2262 delivery attempt. This could be confusing when incoming messages were held
2263 for policy or load reasons. The message log file is now created at the time
2264 the message is received, and an initial "Received" line is written to it.
2266 4. The automatically generated man page for command line options had a minor
2267 bug that caused no ill effects; however, a more serious problem was that
2268 the procedure for building the man page automatically didn't always
2269 operate. Consequently, release 4.20 contains an out-of-date version. This
2270 shouldn't happen again.
2272 5. When building Exim with embedded Perl support, the script that builds the
2273 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2274 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2276 6. The freeze_tell option was not being used for messages that were frozen on
2277 arrival, either by an ACL or by local_scan().
2279 7. Added the smtp_incomplete_transaction log selector.
2281 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2282 was accepting AUTH without a new EHLO.
2284 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2285 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2288 10. Logging of TCP/IP connections (when configured) now happens in the main
2289 daemon process instead of the child process, so that the TCP/IP connection
2290 count is more accurate (but it can never be perfect).
2292 11. The use of "drop" in a nested ACL was not being handled correctly in the
2293 outer ACL. Now, if condition failure induced by the nested "drop" causes
2294 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2295 or "require"), the connection is dropped.
2297 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2298 that yield "discard" can now be used with an "accept" or a "discard" verb,
2299 but an error is generated for any others (because I can't see a useful way
2300 to define what should happen).
2302 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2303 are now processed in the same way as lines in the Exim configuration file.
2304 In particular, continuation lines are supported.
2306 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2308 15. Added -ti meaning -t -i.
2310 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2311 domains, and warn by logging if others are found.
2313 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2314 returned from GIFCONF when Exim is trying to find the list of interfaces on
2315 a host. The code in os.c has been modified to copy each ifreq to an aligned
2316 structure in all cases.
2318 Also, in some cases, the returned ifreq's were being copied to a 'struct
2319 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2320 means the last couple of bytes of an IPv6 address could be chopped if the
2321 ifreq contained only a normal sockaddr (14 bytes storage).
2323 18. Named domain lists were not supported in the hosts_treat_as_local option.
2324 An entry such as +xxxx was not recognized, and was treated as a literal
2327 19. Ensure that header lines added by a DATA ACL are included in the reject log
2328 if the ACL subsequently rejects the message.
2330 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2331 MD5 (which is deprecated).
2333 21. When testing a filter file using -bf, Exim was writing a message when it
2334 took the sender from a "From " line in the message, but it was not doing so
2335 when it took $return_path from a Return-Path: header line. It now does.
2337 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2338 with a valid header line field name (a series of printing characters
2339 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2341 23. Changed "disc" in the source to "disk" to conform to the documentation and
2342 the book and for uniformity.
2344 24. Ignore Sendmail's -Ooption=value command line item.
2346 25. When execve() failed while trying to run a command in a pipe transport,
2347 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2348 could be confused with a return value of 69 from the command itself. This
2349 has been changed to 127, the value the shell returns if it is asked to run
2350 a non-existent command. The wording for the related log line suggests a
2351 non-existent command as the problem.
2353 26. If received_header_text expands to an empty string, do not add a Received:
2354 header line to the message. (Well, it adds a token one on the spool, but
2355 marks it "old" so that it doesn't get used or transmitted.)
2357 27. Installed eximstats 1.28 (addition of -nt option).
2359 28. There was no check for failure on the call to getsockname() in the daemon
2360 code. This can fail if there is a shortage of resources on the system, with
2361 ENOMEM, for example. A temporary error is now given on failure.
2363 29. Contrary to the C standard, it seems that in some environments, the
2364 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2365 program. Exim now does this explicitly; it affects the formatting of
2366 timestamps using strftime().
2368 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2369 complaints. I've now initialized all the variables, to avoid this.
2371 32. Header lines added by a system filter were not being "seen" during
2372 transport-time rewrites.
2374 33. The info_callback() function passed to OpenSSL is set up with type void
2375 (*)(SSL *, int, int), as described somewhere. However, when calling the
2376 function (actually a macro) that sets it up, the type void(*)() is
2377 expected. I've put in a cast to prevent warnings from picky compilers.
2379 34. If a DNS black list lookup found a CNAME record, but there were no A
2380 records associated with the domain it pointed at, Exim crashed.
2382 35. If a DNS black list lookup returned more than one A record, Exim ignored
2383 all but the first. It now scans all returned addresses if a particular IP
2384 value is being sought. In this situation, the contents of the
2385 $dnslist_value variable are a list of all the addresses, separated by a
2388 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2389 to accept a host name and all its aliases if the forward lookup for any of
2390 them yielded the IP address of the incoming connection. Now it accepts only
2391 those names whose forward lookup yields the correct IP address. Any other
2392 names are discarded. This closes a loophole whereby a rogue DNS
2393 administrator could create reverse DNS records to break through a
2394 wildcarded host restriction in an ACL.
2396 37. If a user filter or a system filter that ran in a subprocess used any of
2397 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2398 the wrong values were passed to the pipe command ($thisaddress had the
2399 value of $0, $0 had the value of $1, etc). This bug was introduced by
2400 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2402 38. Improved the line breaking for long SMTP error messages from ACLs.
2403 Previously, if there was no break point between 40 and 75 characters, Exim
2404 left the rest of the message alone. Two changes have been made: (a) I've
2405 reduced the minimum length to 35 characters; (b) if it can't find a break
2406 point between 35 and 75 characters, it looks ahead and uses the first one
2407 that it finds. This may give the occasional overlong line, but at least the
2408 remaining text gets split now.
2410 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2411 file descriptors might be low, and that setting 1000 would always raise it.
2412 It turns out that in some environments, the limit is already over 1000 and
2413 that lowering it causes trouble. So now Exim takes care not to decrease it.
2415 40. When delivering a message, the value of $return_path is set to $sender_
2416 address at the start of routing (routers may change the value). By an
2417 oversight, this default was not being set up when an address was tested by
2418 -bt or -bv, which affected the outcome if any router or filter referred to
2421 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2422 log only when "message" or "log_message" are set. However, if one of the
2423 conditions was an address verification, or a call to a nested ACL, the
2424 messages generated by the underlying test were being passed through. This
2425 no longer happens. The underlying message is available in $acl_verify_
2426 message for both "message" and "log_message" expansions, so it can be
2427 passed through if needed.
2429 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2430 new expansion $bh_ to give the encoded byte string without charset
2431 translation. Translation happens only if iconv() is available; HAVE_ICONV
2432 indicates this at build time. HEADERS_CHARSET gives the charset to
2433 translate to; headers_charset can change it in the configuration, and
2434 "headers charset" can change it in an individual filter file.
2436 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2437 that creates RFC 2047 encoded "words" labels them as that charset instead
2438 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2439 expansion operator; (ii) when Exim creates a From: line for a local
2440 message; (iii) when a header line is rewritten to include a "phrase" part.
2442 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2443 buggy, causing it to skip the first lines of messages whose message ID
2444 ended in 'D'. This would not have bitten before Exim release 4.14, because
2445 message IDs were unlikely to end in 'D' before then. The effect was to have
2446 incorrect size information for certain domains.
2448 45. #include "config.h" was missing at the start of the crypt16.c module. This
2449 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2452 46. If there was a timeout during a "random" callout check, Exim treated it as
2453 a failure of the random address, and carried on sending RSET and the real
2454 address. If the delay was just some slowness somewhere, the response to the
2455 original RCPT would be taken as a response to RSET and so on, causing
2456 mayhem of various kinds.
2458 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2459 when I implemented it. It didn't allow for the fact that some option values
2460 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2461 the right test for positive values.
2463 48. Domain names in DNS records are case-independent. Exim always looks them up
2464 in lower case. Some resolvers return domain names in exactly the case they
2465 appear in the zone file, that is, they may contain uppercase letters. Not
2466 all resolvers do this - some return always lower case. Exim was treating a
2467 change of case by a resolver as a change of domain, similar to a widening
2468 of a domain abbreviation. This triggered its re-routing code and so it was
2469 trying to route what was effectively the same domain again. This normally
2470 caused routing to fail (because the router wouldn't handle the domain
2471 twice). Now Exim checks for this case specially, and just changes the
2472 casing of the domain that it ultimately uses when it transmits the message
2475 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2478 50. If a filter generated a file delivery with a non-absolute name (possible if
2479 no home directory exists for the router), the forbid_file option was not
2482 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2483 the existing equality matching.
2485 52. Exim was using ints instead of ino_t variables in some places where it was
2486 dealing with inode numbers.
2488 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2489 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2490 called TMPDIR, and if it finds it is different, it changes its value.
2492 54. The smtp_printf() function is now made available to local_scan() so
2493 additional output lines can be written before returning. There is also an
2494 smtp_fflush() function to enable the detection of a dropped connection.
2495 The variables smtp_input and smtp_batched_input are exported to
2498 55. Changed the default runtime configuration: the message "Unknown user"
2499 has been removed from the ACL, and instead placed on the localuser router,
2500 using the cannot_route_message feature. This means that any verification
2501 failures that generate their own messages won't get overridden. Similarly,
2502 the "Unrouteable address" message that was in the ACL for unverifiable
2503 relay addresses has also been removed.
2505 56. Added hosts_avoid_esmtp to the smtp transport.
2507 57. The exicyclog script was not checking for the esoteric option
2508 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2509 will work only if exicyclog is run under the appropriate euid.
2511 58. Following a discussion on the list, the rules by which Exim recognises line
2512 endings on incoming messages have been changed. The -dropcr and drop_cr
2513 options are now no-ops, retained only for backwards compatibility. The
2514 following line terminators are recognized: LF CRLF CR. However, special
2515 processing applies to CR:
2517 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2518 nor a local message in the state where . is a terminator.
2520 (ii) If a bare CR is encountered in a header line, an extra space is added
2521 after the line terminator so as not to end the header. The reasoning
2522 behind this is that bare CRs in header lines are most likely either
2523 to be mistakes, or people trying to play silly games.
2525 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
2526 was being incorrectly given as 18 bytes larger than it should have been.
2527 This is a VOB (very old bug).
2529 60. This may never have affected anything current, but just in case it has:
2530 When the local host is found other than at the start of a list of hosts,
2531 the local host, those with the same MX, and any that follow, are discarded.
2532 When the list in question was part of a longer list of hosts, the following
2533 hosts (not currently being processed) were also being discarded. This no
2534 longer happens. I'm not sure if this situation could ever has previously
2537 61. Added the "/MX" feature to lists of hosts in the manualroute and query
2540 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2541 header. This is something that is recommended in a new Internet Draft, and
2542 is something that is documented as being done by Sendmail. There are two
2543 possible values. For messages generated by the autoreply transport, Exim
2546 Auto-Submitted: auto-replied
2548 whereas for all other generated messages (e.g. bounces) it adds
2550 Auto-Submitted: auto-generated
2552 63. The "personal" condition in filters now includes a test for the
2553 Auto-Submitted: header. If it contains the string "auto-" the message it
2554 not considered personal.
2556 64. Added rcpt_include_affixes as a generic transport option.
2558 65. Added queue_only_override (default true).
2560 66. Added the syslog_duplication option.
2562 67. If what should have been the first header line of a message consisted of
2563 a space followed by a colon, Exim was mis-interpreting it as a header line.
2564 It isn't of course - it is syntactically invalid and should therefore be
2565 treated as the start of the message body. The misbehaviour could have
2566 caused a number of strange effects, including loss of data in subsequent
2567 header lines, and spool format errors.
2569 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2570 client host had authenticated. This control can now be exercised by an ACL
2571 for more flexibility.
2573 69. By default, callouts do not happen when testing with -bh. There is now a
2574 variant, -bhc, which does actually run the callout code, including
2575 consulting and updating the callout cache.
2577 70. Added support for saslauthd authentication, courtesy of Alexander
2580 71. If statvfs() failed on the spool or log directories while checking their
2581 size for availability, Exim confusingly gave the error "space shortage".
2582 Furthermore, in debugging mode it crashed with a floating point exception.
2583 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2584 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2585 serious problem, Exim now writes to the main and panic logs when this
2586 happens, with details of the failure. It then refuses to accept the
2587 incoming message, giving the message "spool directory problem" or "log
2588 directory problem" with a 421 code for SMTP messages.
2590 72. When Exim is about to re-exec itself, it ensures that the file descriptors
2591 0, 1, and 2 exist, because some OS complain for execs without them (see
2592 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2593 descriptors. However, the code omitted to check that the open succeeded,
2594 causing mysterious errors if for some reason the permissions on /dev/null
2595 got screwed. Now Exim writes a message to the main and panic logs, and
2596 bombs out if it can't open /dev/null.
2598 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2599 interact so that it is all more flexible. It is supposed to remain
2600 backwards compatible. Also added extra_local_interfaces.
2602 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2603 to bomb out with an assertion failure - to the client this appears as a
2604 connection drop. This problem occurs in the part of the code that was taken
2605 from the Samba project. Fortunately, the assertion is in a very simple
2606 function, so I have fixed this by reproducing the function inline in the
2607 one place where it is called, and arranging for authentication to fail
2608 instead of killing the process with assert().
2610 75. The SPA client code was not working when the server requested OEM rather
2611 than Unicode encoding.
2613 76. Added code to make require_files with a specific uid setting more usable in
2614 the case where statting the file as root fails - usually a non-root-mounted
2615 NFS file system. When this happens and the failure is EACCES, Exim now
2616 forks a subprocess and does the per-uid checking as the relevant uid.
2618 77. Added process_log_path.
2620 78. If log_file_path was not explicitly set, a setting of check_log_space or
2621 check_log_inodes was ignored.
2623 79. If a space check for the spool or log partitions fails, the incident is now
2624 logged. Of course, in the latter case the data may get lost...
2626 80. Added the %p formatting code to string_format() so that it can be used to
2627 print addresses in debug_print(). Adjusted all the address printing in the
2628 debugging in store.c to use %p rather than %d.
2630 81. There was a concern that a line of code in smtp_in.c could overflow a
2631 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2632 initially expressed, the concern was not well-founded, because trailing
2633 spaces are removed early. However, if the trailing spaces were followed by
2634 a NULL, they did not get removed, so the overflow was possible. Two fixes
2637 (a) I re-wrote the offending code in a cleaner fashion.
2638 (b) If an incoming SMTP command contains a NULL character, it is rejected
2641 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
2642 runs initgroups(), so that if the Exim user is in any additional groups,
2643 they will be used during message reception.
2649 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob