1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.76 2005/01/27 15:00:39 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
7 A note about Exim versions 4.44 and 4.50
8 ----------------------------------------
10 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
11 changes of various kinds. As a consequence, a big documentation update was
12 needed. This delayed the release for rather longer than seemed good, especially
13 in the light of a couple of (minor) security issues. Therefore, the changes
14 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
15 release. So 4.44 and 4.50 are in effect two different branches that both start
18 I have left the 4.50 change log unchanged; it contains all the changes since
19 4.43. The change log for 4.44 is below; many of its items are identical to
20 those for 4.50. This seems to be the most sensible way to preserve the
21 historical information.
27 1. Minor wording change to the doc/README.SIEVE file.
29 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
30 computation of the current number of files was incorrect.
32 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
33 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
36 4. Give more explanation in the error message when the command for a transport
37 filter fails to execute.
39 5. There are several places where Exim runs a non-Exim command in a
40 subprocess. The SIGUSR1 signal should be disabled for these processes. This
41 was being done only for the command run by the queryprogram router. It is
42 now done for all such subprocesses. The other cases are: ${run, transport
43 filters, and the commands run by the lmtp and pipe transports.
45 6. Added CONFIGURE_GROUP build-time option.
47 7. Some older OS have a limit of 256 on the maximum number of file
48 descriptors. Exim was using setrlimit() to set 1000 as a large value
49 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
50 systems. I've change it so that if it can't get 1000, it tries for 256.
52 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
53 was an oversight, and furthermore, ever since the addition of extra
54 controls (e.g. 4.43/32), the checks on when to allow different forms of
55 "control" were broken. There should now be diagnostics for all cases when a
56 control that does not make sense is encountered.
58 9. Added the /retain_sender option to "control=submission".
60 10. $recipients is now available in the predata ACL (oversight).
62 11. Tidy the search cache before the fork to do a delivery from a message
63 received from the command line. Otherwise the child will trigger a lookup
64 failure and thereby defer the delivery if it tries to use (for example) a
65 cached ldap connection that the parent has called unbind on.
67 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
68 of $address_data from the recipient verification was clobbered by the
71 13. The value of address_data from a sender verification is now available in
72 $sender_address_data in subsequent conditions in the ACL statement.
74 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
76 15. Added a new option "connect=<time>" to callout options, to set a different
79 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
80 was its contents. (It was OK if the option was not defined at all.)
82 17. A "Completed" log line is now written for messages that are removed from
83 the spool by the -Mrm option.
85 18. New variables $sender_verify_failure and $recipient_verify_failure contain
86 information about exactly what failed.
88 19. Added -dd to debug only the daemon process.
90 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
91 handles timeouts, both on the server side and network timeouts. Renamed the
92 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
94 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
95 It is now set to "smtps".
97 22. $host_address is now set to the target address during the checking of
100 23. When checking ignore_target_hosts for an ipliteral router, no host name was
101 being passed; this would have caused $sender_host_name to have been used if
102 matching the list had actually called for a host name (not very likely,
103 since this list is usually IP addresses). A host name is now passed as
106 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
107 code that specifies a non-restarting handler (typically sigaction() in
108 modern systems) in an attempt to fix a rare and obscure crash bug.
110 25. Narrowed the window for a race in the daemon that could cause it to ignore
111 SIGCHLD signals. This is not a major problem, because they are used only to
112 wake it up if nothing else does.
114 26. A malformed maildirsize file could cause Exim to calculate negative values
115 for the mailbox size or file count. Odd effects could occur as a result.
116 The maildirsize information is now recalculated if the size or filecount
119 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
120 support for a long time. Removed HAVE_SYS_VFS_H.
122 28. Installed the latest version of exipick from John Jetmore.
124 29. In an address list, if the pattern was not a regular expression, an empty
125 subject address (from a bounce message) matched only if the pattern was an
126 empty string. Non-empty patterns were not even tested. This was the wrong
127 because it is perfectly reasonable to use an empty address as part of a
128 database query. An empty address is now tested by patterns that are
129 lookups. However, all the other forms of pattern expect the subject to
130 contain a local part and a domain, and therefore, for them, an empty
131 address still always fails if the pattern is not itself empty.
133 30. Exim went into a mad DNS loop when attempting to do a callout where the
134 host was specified on an smtp transport, and looking it up yielded more
137 31. Re-factored the code for checking spool and log partition space into a
138 function that finds that data and another that does the check. The former
139 is then used to implement four new variables: $spool_space, $log_space,
140 $spool_inodes, and $log_inodes.
142 32. The RFC2047 encoding function was originally intended for short strings
143 such as real names; it was not keeping to the 75-character limit for
144 encoded words that the RFC imposes. It now respects the limit, and
145 generates multiple encoded words if necessary. To be on the safe side, I
146 have increased the buffer size for the ${rfc2047: expansion operator from
149 33. It is now permitted to omit both strings after an "if" condition; if the
150 condition is true, the result is "true". As before, when the second string
151 is omitted, a false condition yields an empty string. This makes it less
152 cumbersome to write custom ACL and router conditions.
154 34. Failure to deliver a bounce message always caused it to be frozen, even if
155 there was an errors_to setting on the router. The errors_to setting is now
158 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
159 canonical form (fully expanded) before being placed in
160 $sender_host_address.
162 36. The table in the code that translates DNS record types into text (T_A to
163 "A" for instance) was missing entries for NS and CNAME. It is just possible
164 that this could have caused confusion if both these types were looked up
165 for the same domain, because the text type is used as part of Exim's
166 per-process caching. But the chance of anyone hitting this buglet seems
169 37. The dnsdb lookup has been extended in a number of ways.
171 (1) There is a new type, "zns", which walks up the domain tree until it
172 finds some nameserver records. It should be used with care.
174 (2) There is a new type, "mxh", which is like "mx" except that it returns
175 just the host names, not the priorities.
177 (3) It is now possible to give a list of domains (or IP addresses) to be
178 looked up. The behaviour when one of the lookups defers can be
179 controlled by a keyword.
181 (4) It is now possible to specify the separator character for use when
182 multiple records are returned.
184 38. The dnslists ACL condition has been extended: it is now possible to supply
185 a list of IP addresses and/or domains to be looked up in a particular DNS
188 39. Added log_selector=+queue_time_overall.
190 40. When running the queue in the test harness, wait just a tad after forking a
191 delivery process, to get repeatability of debugging output.
193 41. Include certificate and key file names in error message when GnuTLS fails
194 to set them up, because the GnuTLS error message doesn't include the name
195 of the failing file when there is a problem reading it.
197 42. Allow both -bf and -bF in the same test run.
199 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
201 44. The "Exiscan patch" is now merged into the mainline Exim source.
203 45. Sometimes the final signoff response after QUIT could fail to get
204 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
205 < 0 before doing a fflush(). This bug looks as though it goes back to the
206 introduction of TLS in release 3.20, but "sometimes" must have been rare
207 because the tests only now provoked it.
209 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
210 (this can affect the format of dates).
212 47. exim_tidydb, when checking for the continued existence of a message for
213 which it has found a message-specific retry record, was not finding
214 messages that were in split spool directories. Consequently, it was
215 deleting retry records that should have stayed in existence.
217 48. Steve fixed some bugs in eximstats.
219 49. The SPA authentication driver was not abandoning authentication and moving
220 on to the next authenticator when an expansion was forced to fail,
221 contradicting the general specification for all authenticators. Instead it
222 was generating a temporary error. It now behaves as specified.
224 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
225 (the order specifies the preference for clients). The order is now AES256,
226 AES128, 3DES, ARCFOUR128.
228 51. Small patch to Sieve code - explicitly set From: when generating an
231 52. Exim crashed if a remote delivery caused a very long error message to be
232 recorded - for instance if somebody sent an entire SpamAssassin report back
233 as a large number of 550 error lines. This bug was coincidentally fixed by
234 increasing the size of one of Exim's internal buffers (big_buffer) that
235 happened as part of the Exiscan merge. However, to be on the safe side, I
236 have made the code more robust (and fixed the comments that describe what
239 53. Now that there can be additional text after "Completed" in log lines (if
240 the queue_time_overall log selector is set), a one-byte patch to exigrep
241 was needed to allow it to recognize "Completed" as not the last thing in
244 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
245 patch that reportedly fixes this has been added. I am not expert enough to
246 create a test for it. This is what the patch creator wrote:
248 "I found a little strange behaviour of ldap code when working with
249 Windows 2003 AD Domain, where users was placed in more than one
250 Organization Units. When I tried to give exim partial DN, the exit code
251 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
252 But simultaneously result of request was absolutely normal ldap result,
253 so I produce this patch..."
255 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
256 REFERENCE, so I have modified the code to exclude the patch when that macro
259 55. Some experimental protocols are using DNS PTR records for new purposes. The
260 keys for these records are domain names, not reversed IP addresses. The
261 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
262 leaves it alone. Component reversal etc. now happens only for IP addresses.
265 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
267 57. Double the size of the debug message buffer (to 2048) so that more of very
268 long debug lines gets shown.
270 58. The exicyclog utility now does better if the number of log files to keep
271 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
273 59. Two changes related to the smtp_active_hostname option:
275 (1) $smtp_active_hostname is now available as a variable.
276 (2) The default for smtp_banner uses $smtp_active_hostname instead
277 of $primary_hostname.
279 60. The host_aton() function is supposed to be passed a string that is known
280 to be a valid IP address. However, in the case of IPv6 addresses, it was
281 not checking this. This is a hostage to fortune. Exim now panics and dies
282 if the condition is not met. A case was found where this could be provoked
283 from a dnsdb PTR lookup with an IPv6 address that had more than 8
284 components; fortuitously, this particular loophole had already been fixed
285 by change 4.50/55 above.
287 If there are any other similar loopholes, the new check in host_aton()
288 itself should stop them being exploited. The report I received stated that
289 data on the command line could provoke the exploit when Exim was running as
290 exim, but did not say which command line option was involved. All I could
291 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
295 61. There was a buffer overflow vulnerability in the SPA authentication code
296 (which came originally from the Samba project). I have added a test to the
297 spa_base64_to_bits() function which I hope fixes it.
300 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
301 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
303 63. The daemon start-up calls getloadavg() while still root for those OS that
304 need the first call to be done as root, but it missed one case: when
305 deliver_queue_load_max is set with deliver_drop_privilege. This is
306 necessary for the benefit of the queue runner, because there is no re-exec
307 when deliver_drop_privilege is set.
309 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
312 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
313 was in use, was not putting the data itself into the right store pool;
314 consequently, it could be overwritten for a subsequent message in the same
315 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
318 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
320 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
321 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
322 the function were treating the return as a boolean value, which happened to
323 work because 0=false and not-0=true, but is not correct code.
325 68. The host_aton() function was not handling scoped IPv6 addresses (those
326 with, for example, "%eth0" on the end) correctly.
328 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
329 negated items (that is, ~something) in unsigned ints. Some compilers
330 apparently mutter when there is no cast.
332 70. If an address verification called from an ACL failed, and did not produce a
333 user-specific message (i.e. there was only a "system" message), nothing was
334 put in $acl_verify_message. In this situation, it now puts the system
337 71. Change 4.23/11 added synchronization checking at the start of an SMTP
338 session; change 4.31/43 added the unwanted input to the log line - except
339 that it did not do this in the start of session case. It now does.
341 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
342 This is wrong and can cause the other end to generate a synchronization
343 error if it is another Exim or anything else that does the synchronization
344 check. A QUIT command is no longer sent after a timeout.
346 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
349 74. The defer_ok option of callout verification was not working if it was used
350 when verifying addresses in header lines, that is, for this case:
352 verify = header_sender/callout=defer_ok
354 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
355 those file descriptors could be used for SMTP connections. If anything
356 wrote to stderr (the example that came up was "warn" in embedded Perl), it
357 could be sent to the SMTP client, causing chaos. The daemon now opens
358 stdin, stdout, and stderr to /dev/null when it puts itself into the
361 76. Arrange for output from Perl's "warn" command to be written to Exim's main
362 log by default. The user can override this with suitable Perl magic.
365 ----------------------------------------------------
366 See the note above about the 4.44 and 4.50 releases.
367 ----------------------------------------------------
373 1. Change 4.43/35 introduced a bug that caused file counts to be
374 incorrectly computed when quota_filecount was set in an appendfile
377 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
378 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
381 3. Give more explanation in the error message when the command for a transport
382 filter fails to execute.
384 4. There are several places where Exim runs a non-Exim command in a
385 subprocess. The SIGUSR1 signal should be disabled for these processes. This
386 was being done only for the command run by the queryprogram router. It is
387 now done for all such subprocesses. The other cases are: ${run, transport
388 filters, and the commands run by the lmtp and pipe transports.
390 5. Some older OS have a limit of 256 on the maximum number of file
391 descriptors. Exim was using setrlimit() to set 1000 as a large value
392 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
393 systems. I've change it so that if it can't get 1000, it tries for 256.
395 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
396 was an oversight, and furthermore, ever since the addition of extra
397 controls (e.g. 4.43/32), the checks on when to allow different forms of
398 "control" were broken. There should now be diagnostics for all cases when a
399 control that does not make sense is encountered.
401 7. $recipients is now available in the predata ACL (oversight).
403 8. Tidy the search cache before the fork to do a delivery from a message
404 received from the command line. Otherwise the child will trigger a lookup
405 failure and thereby defer the delivery if it tries to use (for example) a
406 cached ldap connection that the parent has called unbind on.
408 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
409 of $address_data from the recipient verification was clobbered by the
412 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
413 was its contents. (It was OK if the option was not defined at all.)
415 11. A "Completed" log line is now written for messages that are removed from
416 the spool by the -Mrm option.
418 12. $host_address is now set to the target address during the checking of
421 13. When checking ignore_target_hosts for an ipliteral router, no host name was
422 being passed; this would have caused $sender_host_name to have been used if
423 matching the list had actually called for a host name (not very likely,
424 since this list is usually IP addresses). A host name is now passed as
427 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
428 code that specifies a non-restarting handler (typically sigaction() in
429 modern systems) in an attempt to fix a rare and obscure crash bug.
431 15. Narrowed the window for a race in the daemon that could cause it to ignore
432 SIGCHLD signals. This is not a major problem, because they are used only to
433 wake it up if nothing else does.
435 16. A malformed maildirsize file could cause Exim to calculate negative values
436 for the mailbox size or file count. Odd effects could occur as a result.
437 The maildirsize information is now recalculated if the size or filecount
440 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
441 support for a long time. Removed HAVE_SYS_VFS_H.
443 18. Updated exipick to current release from John Jetmore.
445 19. Allow an empty sender to be matched against a lookup in an address list.
446 Previously the only cases considered were a regular expression, or an
449 20. Exim went into a mad DNS lookup loop when doing a callout where the
450 host was specified on the transport, if the DNS lookup yielded more than
453 21. The RFC2047 encoding function was originally intended for short strings
454 such as real names; it was not keeping to the 75-character limit for
455 encoded words that the RFC imposes. It now respects the limit, and
456 generates multiple encoded words if necessary. To be on the safe side, I
457 have increased the buffer size for the ${rfc2047: expansion operator from
460 22. Failure to deliver a bounce message always caused it to be frozen, even if
461 there was an errors_to setting on the router. The errors_to setting is now
464 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
465 canonical form (fully expanded) before being placed in
466 $sender_host_address.
468 24. Updated eximstats to version 1.33
470 25. Include certificate and key file names in error message when GnuTLS fails
471 to set them up, because the GnuTLS error message doesn't include the name
472 of the failing file when there is a problem reading it.
474 26. Expand error message when OpenSSL has problems setting up cert/key files.
477 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
478 (this can affect the format of dates).
480 28. exim_tidydb, when checking for the continued existence of a message for
481 which it has found a message-specific retry record, was not finding
482 messages that were in split spool directories. Consequently, it was
483 deleting retry records that should have stayed in existence.
485 29. eximstats updated to version 1.35
486 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
487 1.35 - bugfix such that pie charts by volume are generated correctly
489 30. The SPA authentication driver was not abandoning authentication and moving
490 on to the next authenticator when an expansion was forced to fail,
491 contradicting the general specification for all authenticators. Instead it
492 was generating a temporary error. It now behaves as specified.
494 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
495 (the order specifies the preference for clients). The order is now AES256,
496 AES128, 3DES, ARCFOUR128.
498 31. Small patch to Sieve code - explicitly set From: when generating an
501 32. Exim crashed if a remote delivery caused a very long error message to be
502 recorded - for instance if somebody sent an entire SpamAssassin report back
503 as a large number of 550 error lines. This bug was coincidentally fixed by
504 increasing the size of one of Exim's internal buffers (big_buffer) that
505 happened as part of the Exiscan merge. However, to be on the safe side, I
506 have made the code more robust (and fixed the comments that describe what
509 33. Some experimental protocols are using DNS PTR records for new purposes. The
510 keys for these records are domain names, not reversed IP addresses. The
511 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
512 leaves it alone. Component reversal etc. now happens only for IP addresses.
515 34. The host_aton() function is supposed to be passed a string that is known
516 to be a valid IP address. However, in the case of IPv6 addresses, it was
517 not checking this. This is a hostage to fortune. Exim now panics and dies
518 if the condition is not met. A case was found where this could be provoked
519 from a dnsdb PTR lookup with an IPv6 address that had more than 8
520 components; fortuitously, this particular loophole had already been fixed
521 by change 4.50/55 or 4.44/33 above.
523 If there are any other similar loopholes, the new check in host_aton()
524 itself should stop them being exploited. The report I received stated that
525 data on the command line could provoke the exploit when Exim was running as
526 exim, but did not say which command line option was involved. All I could
527 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
531 35. There was a buffer overflow vulnerability in the SPA authentication code
532 (which came originally from the Samba project). I have added a test to the
533 spa_base64_to_bits() function which I hope fixes it.
536 36. The daemon start-up calls getloadavg() while still root for those OS that
537 need the first call to be done as root, but it missed one case: when
538 deliver_queue_load_max is set with deliver_drop_privilege. This is
539 necessary for the benefit of the queue runner, because there is no re-exec
540 when deliver_drop_privilege is set.
542 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
543 was in use, was not putting the data itself into the right store pool;
544 consequently, it could be overwritten for a subsequent message in the same
545 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
548 38. Sometimes the final signoff response after QUIT could fail to get
549 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
550 < 0 before doing a fflush(). This bug looks as though it goes back to the
551 introduction of TLS in release 3.20, but "sometimes" must have been rare
552 because the tests only now provoked it.
558 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
559 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
560 Now it returns an integer. A number of calls were still expecting a T/F
561 return. Fortuitously, in all cases, the tests worked in OK situations,
562 which is the norm. However, things would have gone wrong on any write
563 failures on the smtp file descriptor. This function is used when sending
564 messages over SMTP and also when doing verify callouts.
566 2. When Exim is called to do synchronous delivery of a locally submitted
567 message (the -odf or -odi options), it no longer closes stderr before doing
570 3. Implemented the mua_wrapper option.
572 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
574 5. Implemented the functions header_remove(), header_testname(),
575 header_add_at_position(), and receive_remove_recipient(), and exported them
578 6. If an ACL "warn" statement specified the addition of headers, Exim already
579 inserted X-ACL-Warn: at the start if there was no header name. However, it
580 was not making this test for the second and subsequent header lines if
581 there were newlines in the string. This meant that an invalid header could
582 be inserted if Exim was badly configured.
584 7. Allow an ACL "warn" statement to add header lines at the start or after all
585 the Received: headers, as well as at the end.
587 8. Added the rcpt_4xx retry error code.
589 9. Added postmaster_mailfrom=xxx to callout verification option.
591 10. Added mailfrom=xxxx to the callout verification option, for verify=
594 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
595 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
597 12. Inserted some casts to stop certain compilers warning when using pointer
598 differences as field lengths or precisions in printf-type calls (mostly
599 affecting debugging statements).
601 13. Added optional readline() support for -be (dynamically loaded).
603 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
604 same clock tick as a message's arrival, so that its received time was the
605 same as the "first fail" time on the retry record, and that message
606 remained on the queue past the ultimate address timeout, every queue runner
607 would try a delivery (because it was past the ultimate address timeout) but
608 after another failure, the ultimate address timeout, which should have then
609 bounced the address, did not kick in. This was a "< instead of <=" error;
610 in most cases the first failure would have been in the next clock tick
611 after the received time, and all would be well.
613 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
614 being recognized when the domain list was tested by the match_domain
615 condition in an expansion string.
617 16. Added the ${str2b64: operator.
619 17. Exim was always calling setrlimit() to set a large limit for the number of
620 processes, without checking whether the existing limit was already
621 adequate. (It did check for the limit on file descriptors.) Furthermore,
622 errors from getrlimit() and setrlimit() were being ignored. Now they are
623 logged to the main and panic logs, but Exim does carry on, to try to do its
624 job under whatever limits there are.
626 18. Imported PCRE 5.0.
628 19. Trivial typo in log message " temporarily refused connection" (the leading
631 20. If the log selector return_path_on_delivery was set and an address was
632 redirected to /dev/null, the delivery process crashed because it assumed
633 that a return path would always be set for a "successful" delivery. In this
634 case, the whole delivery is bypassed as an optimization, and therefore no
637 21. Internal re-arrangement: the function for sending a challenge and reading
638 a response while authentication was assuming a zero-terminated challenge
639 string. It's now changed to take a pointer and a length, to allow for
640 binary data in such strings.
642 22. Added the cyrus_sasl authenticator (code supplied by MBM).
644 23. Exim was not respecting finduser_retries when seeking the login of the
645 uid under which it was called; it was always trying 10 times. (The default
646 setting of finduser_retries is zero.) Also, it was sleeping after the final
647 failure, which is pointless.
649 24. Implemented tls_on_connect_ports.
651 25. Implemented acl_smtp_predata.
653 26. If the domain in control=submission is set empty, Exim assumes that the
654 authenticated id is a complete email address when it generates From: or
655 Sender: header lines.
657 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
658 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
659 chown and chgrp in /bin and hostname in /usr/bin.
661 28. Exim was keeping the "process log" file open after each use, just as it
662 does for the main log. This opens the possibility of it remaining open for
663 long periods when the USR1 signal hits a daemon. Occasional processlog
664 errors were reported, that could have been caused by this. Anyway, it seems
665 much more sensible not to leave this file open at all, so that is what now
668 29. The long-running daemon process does not normally write to the log once it
669 has entered its main loop, and it closes the log before doing so. This is
670 so that log files can straightforwardly be renamed and moved. However,
671 there are a couple of unusual error situations where the daemon does write
672 log entries, and I had neglected to close the log afterwards.
674 30. The text of an SMTP error response that was received during a remote
675 delivery was being truncated at 512 bytes. This is too short for some of
676 the long messages that one sometimes sees. I've increased the limit to
679 31. It is now possible to make retry rules that apply only when a message has a
680 specific sender, in particular, an empty sender.
682 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
683 it possible to be selective about when SMTP synchronization is enforced.
685 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
687 32. Implemented hosts_connection_nolog.
689 33. Added an ACL for QUIT.
691 34. Setting "delay_warning=" to disable warnings was not working; it gave a
694 35. Added mailbox_size and mailbox_filecount to appendfile.
696 36. Added control = no_multiline_responses to ACLs.
698 37. There was a bug in the logic of the code that waits for the clock to tick
699 in the case where the clock went backwards by a substantial amount such
700 that the microsecond fraction of "now" was more than the microsecond
701 fraction of "then" (but the whole seconds number was less).
703 38. Added support for the libradius Radius client library this is found on
704 FreeBSD (previously only the radiusclient library was supported).
710 1. When certain lookups returned multiple values in the form name=value, the
711 quoting of the values was not always being done properly. Specifically:
712 (a) If the value started with a double quote, but contained no whitespace,
714 (b) If the value contained whitespace other than a space character (i.e.
715 tabs or newlines or carriage returns) it was not quoted.
716 This fix has been applied to the mysql and pgsql lookups by writing a
717 separate quoting function and calling it from the lookup code. The fix
718 should probably also be applied to nisplus, ibase and oracle lookups, but
719 since I cannot test any of those, I have not disturbed their existing code.
721 2. A hit in the callout cache for a specific address caused a log line with no
722 reason for rejecting RCPT. Now it says "Previous (cached) callout
723 verification failure".
725 3. There was an off-by-one bug in the queryprogram router. An over-long
726 return line was truncated at 256 instead of 255 characters, thereby
727 overflowing its buffer with the terminating zero. As well as fixing this, I
728 have increased the buffer size to 1024 (and made a note to document this).
730 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
731 when Exim is waiting for an SMTP response from a remote server, Exim
732 restarts its select() call on the socket, thereby resetting its timeout.
733 This is not a problem when such interrupts are rare. Somebody set up a cron
734 job to run exiwhat every 2 minutes, which is less than the normal select()
735 timeout (5 or 10 minutes). This meant that the select() timeout never
736 kicked in because it was always reset. I have fixed this by comparing the
737 time when an interrupt arrives with the time at the start of the first call
738 to select(). If more time than the timeout has elapsed, the interrupt is
739 treated as a timeout.
741 5. Some internal re-factoring in preparation for the addition of Sieve
742 extensions (by MH). In particular, the "personal" test is moved to a
743 separate function, and given an option for scanning Cc: and Bcc: (which is
744 not set for Exim filters).
746 6. When Exim created an email address using the login of the caller as the
747 local part (e.g. when creating a From: or Sender: header line), it was not
748 quoting the local part when it contained special characters such as @.
750 7. Installed new OpenBSD configuration files.
752 8. Reworded some messages for syntax errors in "and" and "or" conditions to
753 try to make them clearer.
755 9. Callout options, other than the timeout value, were being ignored when
756 verifying sender addresses in header lines. For example, when using
758 verify = header_sender/callout=no_cache
760 the cache was (incorrectly) being used.
762 10. Added a missing instance of ${EXE} to the exim_install script; this affects
763 only the Cygwin environment.
765 11. When return_path_on_delivery was set as a log selector, if different remote
766 addresses in the same message used different return paths and parallel
767 remote delivery occurred, the wrong values would sometimes be logged.
768 (Whenever a remote delivery process finished, the return path value from
769 the most recently started remote delivery process was logged.)
771 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
772 lines when AUTH and/or TLS are in use. This is the "received protocol"
773 field. Exim used to use "asmtp" for authenticated SMTP, without any
774 indication (in the protocol name) for TLS use. Now it follows the RFC and
775 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
776 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
777 names appear in log lines as well as in Received: header lines.
779 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
780 extensions, and comparison tests, and to fix some bugs.
782 14. Changes to the "personal" filter test:
784 (1) The test was buggy in that it was just doing the equivalent of
785 "contains" tests on header lines. For example, if a user's address was
786 anne@some.where, the "personal" test would incorrectly be true for
788 To: susanne@some.where
790 This test is now done by extracting each address from the header in turn,
791 and checking the entire address. Other tests that are part of "personal"
792 are now done using regular expressions (for example, to check local parts
793 of addresses in From: header lines).
795 (2) The list of non-personal local parts in From: addresses has been
796 extended to include "listserv", "majordomo", "*-request", and "owner-*",
797 taken from the Sieve specification recommendations.
799 (3) If the message contains any header line starting with "List-" it is
800 treated as non-personal.
802 (4) The test for "circular" in the Subject: header line has been removed
803 because it now seems ill-conceived.
805 15. Minor typos in src/EDITME comments corrected.
807 16. Installed latest exipick from John Jetmore.
809 17. If headers_add on a router specified a text string that was too long for
810 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
811 of string_sprintf() is now avoided.
813 18. $message_body_size was not set (it was always zero) when running the DATA
814 ACL and the local_scan() function.
816 19. For the "mail" command in an Exim filter, no default was being set for
817 the once_repeat time, causing a random time value to be used if "once" was
818 specified. (If the value happened to be <= 0, no repeat happened.) The
819 default is now 0s, meaning "never repeat". The "vacation" command was OK
820 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
821 (I found it when inspecting the code).
823 20. There is now an overall timeout for performing a callout verification. It
824 defaults to 4 times the callout timeout, which applies to individual SMTP
825 commands during the callout. The overall timeout applies when there is more
826 than one host that can be tried. The timeout is checked before trying the
827 next host. This prevents very long delays if there are a large number of
828 hosts and all are timing out (e.g. when the network connections are timing
829 out). The value of the overall timeout can be changed by specifying an
830 additional sub-option for "callout", called "maxwait". For example:
832 verify = sender/callout=5s,maxwait=20s
834 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
835 to the end before writing, but this should make it even safer).
837 22. Exim was forgetting that it had advertised PIPELINING for the second and
838 subsequent messages on an SMTP connection. It was also not resetting its
839 memory on STARTTLS and an internal HELO.
841 23. When Exim logs an SMTP synchronization error within a session, it now
842 records whether PIPELINING has been advertised or not.
844 24. Added 3 instances of "(long int)" casts to time_t variables that were being
845 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
846 rather than long int.
848 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
850 26. Added the never_mail option to autoreply.
856 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
857 crash if the getsockname() call failed; this can happen if a connection is
858 closed very soon after it is established. The problem was simply in the
859 order in which certain operations were done, causing Exim to try to write
860 to the SMTP stream before it had set up the file descriptor. The bug has
861 been fixed by making things happen in the correct order.
867 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
868 before the connection was closed, thus losing the rejection response.
870 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
871 some early Solaris releases, but causes trouble in current releases where
872 socklen_t is defined.
874 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
877 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
880 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
881 that releasing the top of it at the end releases what was used for sub-
882 expansions (unless the block got too big). However, discard this block if
883 the first thing is a variable or header, so that we can use its block when
884 it is dynamic (useful for very large $message_headers, for example).
886 6. Lookups now cache *every* query, not just the most recent. A new, separate
887 store pool is used for this. It can be recovered when all lookup caches are
888 flushed. Lookups now release memory at the end of their result strings.
889 This has involved some general refactoring of the lookup sources.
891 7. Some code has been added to the store_xxx() functions to reduce the amount
892 of flapping under certain conditions.
894 8. log_incoming_interface used to affect only the <= reception log lines. Now
895 it causes the local interface and port to be added to several more SMTP log
896 lines, for example "SMTP connection from", and rejection lines.
898 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
900 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
902 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
903 could be overwritten at the end of the current message (or the start of a
904 new message if it was set in a HELO ACL). The value is now preserved for
905 the duration of the SMTP connection.
907 12. If a transport had a headers_rewrite setting, and a matching header line
908 contained an unqualified address, that address was qualified, even if it
909 did not match any rewriting rules. The underlying bug was that the values
910 of the flags that permit the existence of unqualified sender and recipient
911 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
912 non-local messages, and by -bnq for local messages) were not being
913 preserved with the message after it was received.
915 13. When Exim was logging an SMTP synchronization error, it could sometimes log
916 "next input=" as part of the text comprising the host identity instead of
917 the correct text. The code was using the same buffer for two different
918 strings. However, depending on which order the printing function evaluated
919 its arguments, the bug did not always show up. Under Linux, for example, my
920 test suite worked just fine.
922 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
923 doesn't work with some older versions of Perl. It has been changed to "my",
924 which in any case is probably the better facility to use.
926 15. A really picky compiler found some instances of statements for creating
927 error messages that either had too many or two few arguments for the format
930 16. The size of the buffer for calls to the DNS resolver has been increased
931 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
932 for addresses that have a lot of PTR records. This alleviates a problem; it
933 does not fully solve it.
935 17. A dnsdb lookup for PTR records that receives more data than will fit in the
936 buffer now truncates the list and logs the incident, which is the same
937 action as happens when Exim is looking up a host name and its aliases.
938 Previously in this situation something unpredictable would happen;
939 sometimes it was "internal error: store_reset failed".
941 18. If a server dropped the connection unexpectedly when an Exim client was
942 using GnuTLS and trying to read a response, the client delivery process
943 crashed while trying to generate an error log message.
945 19. If a "warn" verb in an ACL added multiple headers to a message in a single
948 warn message = H1: something\nH2: something
950 the text was added as a single header line from Exim's point of view
951 though it ended up OK in the delivered message. However, searching for the
952 second and subsequent header lines using $h_h2: did not work. This has been
953 fixed. Similarly, if a system filter added multiple headers in this way,
954 the routers could not see them.
956 20. Expanded the error message when iplsearch is called with an invalid key to
957 suggest using net-iplsearch in a host list.
959 21. When running tests using -bh, any delays imposed by "delay" modifiers in
960 ACLs are no longer actually imposed (and a message to that effect is
963 22. If a "gecos" field in a passwd entry contained escaped characters, in
964 particular, if it contained a \" sequence, Exim got it wrong when building
965 a From: or a Sender: header from that name. A second bug also caused
966 incorrect handling when an unquoted " was present following a character
969 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
970 was not being matched caselessly.
972 24. Arranged for all hyphens in the exim.8 source to be escaped with
975 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
976 properly. Recipient callout cache records were still being keyed to include
977 the sender, even when use_sender was set false. This led to far more
978 callouts that were necessary. The sender is no longer included in the key
979 when use_sender is false.
981 26. Added "control = submission" modifier to ACLs.
983 27. Added the ${base62d: operator to decode base 62 numbers.
985 28. dnsdb lookups can now access SRV records.
987 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
988 the configuration file.
990 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
991 (-v) mode. This makes the output for a verbose queue run more intelligible.
993 31. Added a use_postmaster feature to recipient callouts.
995 32. Added the $body_zerocount variable, containing the number of binary zero
996 bytes in the message body.
998 33. The time of last modification of the "new" subdirectory is now used as the
999 "mailbox time last read" when there is a quota error for a maildir
1002 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1004 35. Added +ignore_unknown as a special item in host lists.
1006 36. Code for decoding IPv6 addresses in host lists is now included, even if
1007 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1008 address was recognized as an IP address, but was then not correctly decoded
1009 into binary, causing unexpected and incorrect effects when compared with
1016 1. Very minor rewording of debugging text in manualroute to say "list of
1017 hosts" instead of "hostlist".
1019 2. If verify=header_syntax was set, and a header line with an unqualified
1020 address (no domain) and a large number of spaces between the end of the
1021 name and the colon was received, the reception process suffered a buffer
1022 overflow, and (when I tested it) crashed. This was caused by some obsolete
1023 code that should have been removed. The fix is to remove it!
1025 3. When running in the test harness, delay a bit after writing a bounce
1026 message to get a bit more predictability in the log output.
1028 4. Added a call to search_tidyup() just before forking a reception process. In
1029 theory, someone could use a lookup in the expansion of smtp_accept_max_
1030 per_host which, without the tidyup, could leave open a database connection.
1032 5. Added the variables $recipient_data and $sender_data which get set from a
1033 lookup success in an ACL "recipients" or "senders" condition, or a router
1034 "senders" option, similar to $domain_data and $local_part_data.
1036 6. Moved the writing of debug_print from before to after the "senders" test
1039 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1040 problems for message scanning, either using a data ACL, or using
1041 local_scan() because the Received: header was not generated till after they
1042 were called (in order to set the time as the time of reception completion).
1043 I have revised the way this works. The header is now generated after the
1044 body is received, but before the ACL or local_scan() are called. After they
1045 are run, the timestamp in the header is updated.
1051 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1052 before starting a queue runner without re-exec. This happened only when
1053 deliver_drop_privilege was set or when the Exim user was set to root. The
1054 effect of the bug was that timeouts during subsequent deliveries caused
1055 crashes instead of being properly handled. The handler is now left at its
1056 default (and expected) setting.
1058 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1059 message, again when deliver_drop_privilege is set or Exim is run as root.
1060 The bug described in (1) was not present in this case, but the tidying up
1061 of the other signals was missing. I have made the two cases consistent.
1063 3. The ignore_target_hosts setting on a manualroute router was being ignored
1064 for hosts that were looked up using the /MX notation.
1066 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1069 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1070 operated on the sender address. After changing the $sender_address to <>
1071 for the sender address verify, Exim was re-instated it as the original
1072 (before rewriting) address, but remembering that it had rewritten it, so it
1073 wasn't rewriting it again. This bug also had the effect of breaking the
1074 sender address verification caching when the sender address was rewritten.
1076 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1077 This has been changed so that if the ip literal address matches
1078 ignore_target_hosts, the router declines.
1080 7. Added expansion conditions match_domain, match_address, and match_local_
1081 part (NOT match_host).
1083 8. The placeholder for the Received: header didn't have a length field set.
1085 9. Added code to Exim itself and to exim_lock to test for a specific race
1086 condition that could lead to file corruption when using MBX delivery. The
1087 issue is with the lockfile that is created in /tmp. If this file is removed
1088 after a process has opened it but before that process has acquired a lock,
1089 there is the potential for a second process to recreate the file and also
1090 acquire a lock. This could lead to two Exim processes writing to the file
1091 at the same time. The added code performs the same test as UW imapd; it
1092 checks after acquiring the lock that its file descriptor still refers to
1093 the same named file.
1095 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1096 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1099 11. Added the smtp_active_hostname option. If used, this will typically be made
1100 to depend on the incoming interface address. Because $interface_address is
1101 not set up until the daemon has forked a reception process, error responses
1102 that can happen earlier (such as "too many connections") no longer contain
1105 12. If an expansion in a condition on a "warn" statement fails because a lookup
1106 defers, the "warn" statement is abandoned, and the next ACL statement is
1107 processed. Previously this caused the whole ACL to be aborted.
1109 13. Added the iplsearch lookup type.
1111 14. Added ident_timeout as a log selector.
1113 15. Added tls_certificate_verified as a log selector.
1115 16. Added a global option tls_require_ciphers (compare the smtp transport
1116 option of the same name). This controls incoming TLS connections.
1118 17. I finally figured out how to make tls_require_ciphers do a similar thing
1119 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1120 before starting the TLS session.
1122 18. Tabs are now shown as \t in -bP output.
1124 19. If the log selector return_path_on_delivery was set, Exim crashed when
1125 bouncing a message because it had too many Received: header lines.
1127 20. If two routers both had headers_remove settings, and the first one included
1128 a superfluous trailing colon, the final name in the first list and the
1129 first name in the second list were incorrectly joined into one item (with a
1130 colon in the middle).
1136 1. Added -C and -D options to the exinext utility, mainly to make it easier
1137 to include in the automated testing, but these could be helpful when
1138 multiple configurations are in use.
1140 2. The exinext utility was not formatting the output nicely when there was
1141 an alternate port involved in the retry record key, nor when there was a
1142 message id as well (for retries that were specific to a specific message
1143 and a specific host). It was also confused by IPv6 addresses, because of
1144 the additional colons they contain. I have fixed the IPv4 problem, and
1145 patched it up to do a reasonable job for IPv6.
1147 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1148 delivery, the log line now contains "pipelined" if PIPELINING was used.
1150 4. An SMTP transport process used to panic and die if the bind() call to set
1151 an explicit outgoing interface failed. This has been changed; it is now
1152 treated in the same way as a connect() failure.
1154 5. A reference to $sender_host_name in the part of a conditional expansion
1155 that was being skipped was still causing a DNS lookup. This no longer
1158 6. The def: expansion condition was not recognizing references to header lines
1159 that used bh_ and bheader_.
1161 7. Added the _cache feature to named lists.
1163 8. The code for checking quota_filecount in the appendfile transport was
1164 allowing one more file than it should have been.
1166 9. For compatibility with Sendmail, the command line option
1174 and sets the incoming protocol and host name (for trusted callers). The
1175 host name and its colon can be omitted when only the protocol is to be set.
1176 Note the Exim already has two private options, -pd and -ps, that refer to
1177 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1178 "s", but I don't think that's a major issue.
1180 10. A number of refactoring changes to the code, none of which should affect
1183 (a) The number of logging options was getting close to filling up the
1184 32-bit word that was used as a bit map. I have split them into two classes:
1185 those that are passed in the argument to log_write(), and those that are
1186 only ever tested independently outside of that function. These are now in
1187 separate 32-bit words, so there is plenty of room for expansion again.
1188 There is no change in the user interface or the logging behaviour.
1190 (b) When building, for example, log lines, the code previously used a
1191 macro that called string_cat() twice, in order to add two strings. This is
1192 not really sufficiently general. Furthermore, there was one instance where
1193 it was actually wrong because one of the argument was used twice, and in
1194 one call a function was used. (As it happened, calling the function twice
1195 did not affect the overall behaviour.) The macro has been replaced by a
1196 function that can join an arbitrary number of extra strings onto a growing
1199 (c) The code for expansion conditions now uses a table and a binary chop
1200 instead of a serial search (which was left over from when there were very
1201 few conditions). Also, it now recognizes conditions like "pam" even when
1202 the relevant support is not compiled in: a suitably worded error message is
1203 given if an attempt is made to use such a condition.
1205 11. Added ${time_interval:xxxxx}.
1207 12. A bug was causing one of the ddress fields not to be passed back correctly
1208 from remote delivery subprocesses. The field in question was not being
1209 subsequently used, so this caused to problems in practice.
1211 13. Added new log selectors queue_time and deliver_time.
1213 14. Might have fixed a bug in maildirsizefile handling that threw up
1214 "unexpected character" debug warnings, and recalculated the data
1215 unnecessarily. In any case, I expanded the warning message to give more
1218 15. Added the message "Restricted characters in address" to the statements in
1219 the default ACL that block characters like @ and % in local parts.
1221 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1222 Three changes have been made:
1224 (a) There was a serious bug; a negative response to MAIL caused the whole
1225 recipient domain to be cached as invalid, thereby blocking all messages
1226 to all local parts at the same domain, from all senders. This bug has
1227 been fixed. The domain is no longer cached after a negative response to
1228 MAIL if the sender used is not empty.
1230 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1233 (c) A new callout option, "use_sender" has been added for people who want
1234 the modified behaviour.
1240 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1243 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1244 indeed breaks things for older releases.
1246 3. Added additional logging to the case where there is a problem reading data
1247 from a filter that is running in a subprocess using a pipe, in order to
1248 try to track down a specific problem.
1250 4. Testing facility fudge: when running in the test harness and attempting
1251 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1252 getting "No route to host". Convert this to a timeout.
1254 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1257 6. Some OS don't have socklen_t but use size_t instead. This affects the
1258 fifth argument of getsockopt() amongst other things. This is now
1259 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1260 can be set for individual OS. I have set it for SunOS5, OSF1, and
1261 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1262 some earlier ones do not.
1264 7. Change 4.30/15 was not doing the test caselessly.
1266 8. The standard form for an IPv6 address literal was being rejected by address
1267 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1268 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1269 this, as well as the form without the "IPv6" on the front (but only when
1270 address literals are enabled, of course).
1272 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1274 10. Exim crashed if a message with an empty sender address specified by -f
1275 encountered a router with an errors_to setting. This could be provoked only
1276 by a command such as
1280 where an empty string was supplied; "<>" did not hit this bug.
1282 11. Installed PCRE release 4.5.
1284 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1285 remained set. It is now erased.
1287 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1288 times from message ids (which are base 36 rather than the normal 62).
1290 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1291 were being counted as actual protocol errors, and logged if the log
1292 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1293 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1294 and DATA following a set of rejected RCPTs do not count as protocol errors.
1295 In other words, Exim assumes they were pipelined, though this may not
1296 actually be the case. Of course, in all cases the client gets an
1297 appropriate error code.
1299 15. If a lookup fails in an ACL condition, a message about the failure may
1300 be available; it is used if testing the ACL cannot continue, because most
1301 such messages specify what the cause of the deferral is. However, some
1302 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1303 that caused an old message to be retained and used if a later statement
1304 caused a defer, replacing the real cause of the deferral.
1306 16. If an IP address had so many PTR records that the DNS lookup buffer
1307 was not large enough to hold them, Exim could crash while trying to process
1308 the truncated data. It now detects and logs this case.
1310 17. Further to 4.21/58, another change has been made: if (and only if) the
1311 first line of a message (the first header line) ends with CRLF, a bare LF
1312 in a subsequent header line has a space inserted after it, so as not to
1313 terminate the header.
1315 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1316 unnecessarily, used atoi() instead of strtol(), and didn't check the
1317 termination when getting file sizes from file names by regex.
1319 19. Completely re-implemented the support for maildirsize files, in the light
1320 of a number of problems with the previous contributed implementation
1321 (4.30/29). In particular:
1323 . If the quota is zero, the maildirsize file is maintained, but no quota is
1326 . If the maildir directory does not exist, it is created before any attempt
1327 to write a maildirsize file.
1329 . The quota value in the file is just a cache; if the quota is changed in
1330 the transport, the new value overrides.
1332 . A regular expression is available for excluding directories from the
1335 20. The autoreply transport checks the characters in options that define the
1336 message's headers; it allows continued headers, but it was checking with
1337 isspace() after an embedded newline instead of explicitly looking for a
1340 21. If all the "regular" hosts to which an address was routed had passed their
1341 expiry times, and had not reached their retry times, the address was
1342 bounced, even if fallback hosts were defined. Now Exim should go on to try
1345 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1346 equivalent code in the SMTP transport. Some hosts send humungous responses
1347 to HELO/EHLO, more than 1024 it seems.
1349 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1350 gives compiler warnings in some environments. I've now done it "properly",
1353 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1354 (because of problems with the built-in one) was declared to return uschar *
1355 instead of char *, causing compiler failure.
1357 25. Fixed a file descriptor leak when processing alias/forward files.
1359 26. Fixed a minor format string issue in dbfn.c.
1361 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1363 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1364 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1367 29. When a sender address is verified, it is cached, to save repeating the test
1368 when there is more than one recipient in a message. However, when the
1369 verification involves a callout, it is possible for different callout
1370 options to be set for different recipients. It is too complicated to keep
1371 track of this in the cache, so now Exim always runs a verification when a
1372 callout is required, relying on the callout cache for the optimization.
1373 The overhead is duplication of the address routing, but this should not be
1376 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1377 to be verified with callout=postmaster, and the main callout worked but the
1378 postmaster check failed, the verification correctly failed. However, if a
1379 subsequent RCPT command asked for sender verification *without* the
1380 postmaster check, incorrect caching caused this verification also to fail,
1383 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1384 it was not caching the DNS options (qualify_single, search_parents) that
1385 were used when the lookup failed. A subsequent lookup with different
1386 options therefore always gave the same answer, though there were cases
1387 where it should not have. (Example: a "domains = !$mx_any" option on a
1388 dnslookup router: the "domains" option is always processed without any
1389 widening, but the router might have qualify_single set.) Now Exim uses the
1390 cached value only when the same options are set.
1392 32. Added John Jetmore's "exipick" utility to the distribution.
1394 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1395 than a timeout (e.g. a certificate is required, and is not provided), an
1396 Exim server now closes the connection immediately. Previously it waited for
1397 the client to close - but if the client is SSL, it seems that they each
1398 wait for each other, leading to a delay before one of them times out.
1400 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1401 maintained 0.8.x compatibility because I don't think many are using it, and
1402 it is clearly obsolete.
1404 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1407 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1408 client certificate was expired. A simple patch fixes this, though I don't
1409 understand the full logic of why the verify callback is called multiple
1412 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1413 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1414 which causes problems with some clients (such as the Certicom SSL Plus
1415 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1416 disables the coutermeasure allowing Eudora to connect."
1418 38. Exim was not checking that a write() to a log file succeeded. This could
1419 lead to Bad Things if a log got too big, in particular if it hit a file
1420 size limit. Exim now panics and dies if it cannot write to a log file, just
1421 as it does if it cannot open a log file.
1423 39. Modified OS/Makefile-Linux so that it now contains
1425 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1427 The two -D definitions ensure that Exim is compiled with large file
1428 support, which makes it possible to handle log files that are bigger than
1431 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1432 instance) a domain was checked against a named list that involved a lookup,
1433 causing $domain_data to be set, then another domain was checked against the
1434 same list, then the first domain was re-checked, the value of $domain_data
1435 after the final check could be wrong. In particular, if the second check
1436 failed, it could be set empty. This bug probably also applied to
1439 41. The strip_trailing_dot option was not being applied to the address given
1440 with the -f command-line option.
1442 42. The code for reading a message's header from the spool was incrementing
1443 $received_count, but never initializing it. This meant that the value was
1444 incorrect (doubled) while delivering a message in the same process in which
1445 it was received. In the most common configuration of Exim, this never
1446 happens - a fresh exec is done - but it can happen when
1447 deliver_drop_privilege is set.
1449 43. When Exim logs an SMTP synchronization error - client data sent too soon -
1450 it now includes up to 150 characters of the unexpected data in the log
1453 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1454 and building data strings. The size of both of these buffers was 10 000
1455 bytes - far larger than anybody would *ever* want, thought I. Needless to
1456 say, somebody hit the limit. I have increased the maximum line length to
1457 20 000 and the maximum data length of concatenated lines to 100 000. I have
1458 also fixed two bugs, because there was no checking on these buffers. Tsk,
1459 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1460 buffer is too small.
1462 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1463 lsearch lookups. Now it does.
1465 46. When parsing a route_list item in a manualroute router, a fixed-length
1466 buffer was used for the list of hosts. I made this 1024 bytes long,
1467 thinking that nobody would ever have a list of hosts that long. Wrong.
1468 Somebody had a whole pile of complicated expansion conditions, and the
1469 string was silently truncated, leading to an expansion error. It turns out
1470 that it is easier to change to an unlimited length (owing to other changes
1471 that have happened since this code was originally written) than to build
1472 structure for giving a limitation error. The length of the item that
1473 expands into the list of hosts is now unlimited.
1475 47. The lsearch lookup could not handle data where the length of text line was
1476 more than 4095 characters. Such lines were truncated, leading to shortened
1477 data being returned. It should now handle lines of any length.
1479 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1480 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1483 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
1484 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1485 no longer happens in comment lines. A list of replacements is now placed
1486 at the head of all of the source files, except those whose only change is
1487 to replace PERL_COMMAND in the very first #! line.
1489 50. Replaced the slow insertion sort in queue.c, for sorting the list of
1490 messages on the queue, with a bottom-up merge sort, using code contributed
1491 by Michael Haardt. This should make operations like -bp somewhat faster on
1492 large queues. It won't affect queue runners, except when queue_run_in_order
1495 51. Installed eximstats 1.31 in the distribution.
1497 52. Added support for SRV lookups to the dnslookup router.
1499 53. If an ACL referred to $message_body or $message_body_end, the value was not
1500 reset for any messages that followed in the same SMTP session.
1502 54. The store-handling optimization for building very long strings was not
1503 differentiating between the different store pools. I don't think this
1504 actually made any difference in practice, but I've tidied it.
1506 55. While running the routers to verify a sender address, $sender_address
1507 was still set to the sender address. This is wrong, because when routing to
1508 send a bounce to the sender, it would be empty. Therefore, I have changed
1509 it so that, while verifying a sender address, $sender_address is set to <>.
1510 (There is no change to what happens when verifying a recipient address.)
1512 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1513 target A or AAAA records (if not already returned) without resetting the
1514 qualify_single or search_parents options of the DNS resolver. These are
1515 inappropriate in this case because the targets of MX and SRV records must
1516 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1517 target that, when qualified, matched something in the local domain. These
1518 two options are now turned off when doing these lookups.
1520 57. It seems that at least some releases of Reiserfs (which does not have the
1521 concept of a fixed number of inodes) returns zero and not -1 for the
1522 number of available inodes. This interacted badly with check_spool_inodes,
1523 which assumed that -1 was the "no such thing" setting. What I have done is
1524 to check that the total number of inodes is greater than zero before doing
1525 the test of how many are available.
1527 58. When a "warn" ACL statement has a log_message modifier, the message is
1528 remembered, and not repeated. This is to avoid a lot of repetition when a
1529 message has many recipients that cause the same warning to be written.
1530 Howewer, Exim was preserving the list of already written lines for an
1531 entire SMTP session, which doesn't seem right. The memory is now reset if a
1532 new message is started.
1534 59. The "rewrite" debugging flag was not showing the result of rewriting in the
1535 debugging output unless log_rewrite was also set.
1537 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1538 of (int)(handle) when we know that handle contains (void *)(-1).
1540 61. The Exim daemon panic-logs an error return when it closes the incoming
1541 connection. However "connection reset by peer" seems to be common, and
1542 isn't really an error worthy of noting specially, so that particular error
1545 62. When Exim is trying to find all the local interfaces, it used to panic and
1546 die if the ioctl to get the interface flags failed. However, it seems that
1547 on at least one OS (Solaris 9) it is possible to have an interface that is
1548 included in the list of interfaces, but for which you get a failure error
1549 for this call. This happens when the interface is not "plumbed" into a
1550 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1551 failure of the "get flags" call assumes that the interface is down.
1553 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
1554 makes life easier for people who are doing arithmetic on fields extracted
1555 from dates, where you often get leading zeros that should not be
1556 interpreted as octal.
1558 64. Added qualify_domain to the redirect router, to override the global
1561 65. If a pathologically long header line contained very many addresses (the
1562 report of this problem mentioned 10 000) and each of them was rewritten,
1563 Exim could use up a very large amount of memory. (It kept on making new
1564 copies of the header line as it rewrote, and never released the old ones.)
1565 At the expense of a bit more processing, the header rewriting function has
1566 been changed so that it no longer eats memory in this way.
1568 66. The generation of the Received: header has been moved from the time that a
1569 message starts to be received, to the time that it finishes. The timestamp
1570 in the Received: header should now be very close to that of the <= log
1571 line. There are two side-effects of this change:
1573 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1574 logged header lines no longer include the local Received: line, because
1575 it has not yet been created. The same applies to a copy of the message
1576 that is returned to a non-SMTP sender when a message is rejected.
1578 (b) When a filter file is tested using -bf, no additional Received: header
1579 is added to the test message. After some thought, I decided that this
1582 This change does not affect the value of $received_for. It is still set
1583 after address rewriting, but before local_scan() is called.
1585 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1587 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1588 gave an unhelpful panic error message, and a defer error. I have managed to
1589 change this behaviour so that it now rejects any supplied certificate,
1590 which seems right, as the list of acceptable certificates is empty.
1592 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1593 gave an unhelpful defer error. I have not managed to make this reject any
1594 supplied certificates, but the error message it gives is "no certificate
1595 supplied", which is not helpful.
1597 70. exigrep's output now also includes lines that are not associated with any
1598 message, but which match the given pattern. Implemented by a patch from
1599 Martin Sluka, which also tidied up the Perl a bit.
1601 71. Recipient callout verification, like sender verification, was using <> in
1602 the MAIL FROM command. This isn't really the right thing, since the actual
1603 sender may affect whether the remote host accepts the recipient or not. I
1604 have changed it to use the actual sender in the callout; this means that
1605 the cache record is now keyed on a recipient/sender pair, not just the
1606 recipient address. There doesn't seem to be a real danger of callout loops,
1607 since a callout by the remote host to check the sender would use <>.
1608 [SEE ABOVE: changed after hitting problems.]
1610 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1611 temporary errors. However, in the case of such a code being given after
1612 the end of a data transmission (i.e. after ".") Exim was failing to write
1613 a retry record for the message. (Yes, there was some broken host that was
1614 actually sending 8xx at this point.)
1616 73. An unknown lookup type in a host list could cause Exim to panic-die when
1617 the list was checked. (An example that provoked this was putting <; in the
1618 middle of a list instead of at the start.) If this happened during a DATA
1619 ACL check, a -D file could be left lying around. This kind of configuration
1620 error no longer causes Exim to die; instead it causes a defer errror. The
1621 incident is still logged to the main and panic logs.
1623 74. Buglet left over from Exim 3 conversion. The message "too many messages
1624 in one connection" was written to the rejectlog but not the mainlog, except
1625 when address rewriting (yes!) was being logged.
1627 75. Added write_rejectlog option.
1629 76. When a system filter was run not as root (that is, when system_filter_user
1630 was set), the values of the $n variables were not being returned to the
1631 main process; thus, they were not subsequently available in the $sn
1634 77. Added +return_path_on_delivery log selector.
1636 78. A connection timeout was being treated differently from recipients deferred
1637 when testing hosts_max_try with a message that was older than the host's
1638 retry timeout. (The host should not be counted, thus allowing all hosts to
1639 be tried at least once before bouncing.) This may have been the cause of an
1640 occasionally reported bug whereby a message would remain on the queue
1641 longer than the retry timeout, but would be bounced if a delivery was
1642 forced. I say "may" because I never totally pinned down the problem;
1643 setting up timeout/retry tests is difficult. See also the next item.
1645 79. The ultimate address timeout was not being applied to errors that involved
1646 a combination of host plus message (for example, a timeout on a MAIL
1647 command). When an address resolved to a number of possible hosts, and they
1648 were not all tried for each delivery (e.g. because of hosts_max_try), a
1649 message could remain on the queue longer than the retry timeout.
1651 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1654 81. Fixed an obscure SMTP outgoing bug which required at least the following
1655 conditions: (a) there was another message waiting for the same server;
1656 (b) the server returned 5xx to all RCPT commands in the first message so
1657 that the message was not completed; (c) the server dropped the connection
1658 or gave a negative response to the RSET that Exim sends to abort the
1659 transaction. The observed case was a dropped connection after DATA that had
1660 been sent in pipelining mode. That is, the server had advertised PIPELINING
1661 but was not implementing it correctly. The effect of the bug was incorrect
1662 behaviour, such as trying another host, and this could lead to a crash.
1668 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1669 and daemon.c were passed as pointers to ints; they should have been
1670 pointers to socklen_t variables (which are typically unsigned ints).
1672 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1675 3. Fixed a really odd bug that affected only the testing scheme; patching a
1676 certain fixed string in the binary changed the value of another string that
1677 happened to be identical to the end of the original first string.
1679 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1680 name", it returns that address as the IP address. On some operating
1681 systems (e.g. Solaris), it also passes back the IP address string as the
1682 "host name". However, on others (e.g. Linux), it passes back an empty
1683 string. Exim wasn't checking for this, and was changing the host name to an
1684 empty string, assuming it had been canonicized.
1686 5. Although rare, it is permitted to have more than one PTR record for a given
1687 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1688 all the names associated with an address, because they do in Solaris.
1689 However, it seems that they do not in Linux for data that comes from the
1690 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1691 I found this out when I moved to a new Linux workstation and tried to run
1692 the Exim test suite.
1694 To get round this problem I have changed the code so that it now does its
1695 own call to the DNS to look up PTR records when searching for a host name.
1696 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1697 addresses that are only in /etc/hosts are still found.
1699 This behaviour is, however, controlled by an option called host_lookup_
1700 order, which defaults to "bydns:byaddr". If people want to use the other
1701 order, or indeed, just use one or the other means of lookup, they can
1702 specify it in this variable.
1704 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1705 some operating systems, this comes back from gethostbyaddr() as an empty
1706 string, and this is what Exim used to test for. However, it seems that in
1707 other systems, "." is yielded. Exim now tests for this case too.
1709 7. The values of check_spool_space and check_log_space are now held internally
1710 as a number of kilobytes instead of an absolute number of bytes. If a
1711 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1712 kilobyte. This means that much larger values can be stored.
1714 8. Exim monitor: an attempt to get the action menu when not actually pointing
1715 at a message produces an empty menu entitled "No message selected". This
1716 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1717 no entries in it ("Shell widget menu has zero width and/or height"). So I
1718 have added a single, blank menu entry in this case.
1720 9. Added ${quote_local_part.
1722 10. MIME decoding is now applied to the contents of Subject: header lines when
1725 11. Now that a reference to $sender_host_address automatically causes a reverse
1726 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1727 host lookup before query-style lookups in lists that might use this
1728 variable. This has therefore been abolished, and the "net-" prefix is no
1729 longer necessary for query-style lookups.
1731 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1732 have been a typo for LFLAGS, so it has been changed.
1734 13. The install script calls Exim with "-C /dev/null" in order to find the
1735 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1736 to be output. Howeve, since Exim outputs its version number before the
1737 error, it didn't break the script. It just looked ugly. I fixed this by
1738 always allowing "-C /dev/null" if the caller is root.
1740 14. Ignore overlarge ACL variable number when reading spool file - insurance
1741 against a later release with more variables having written the file.
1743 15. The standard form for an IPv6 address literal was being rejected by EHLO.
1744 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1745 this, as well as the form without the "IPv6" on the front.
1747 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1748 OS/Makefile-Darwin file.
1750 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1751 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1753 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
1754 variable. However, some OS use ints for inodes. I've added cast to long int
1755 to get rid of the compiler warning.
1757 19. I had forgotten to lock out "/../" in configuration file names when
1758 ALT_CONFIG_PREFIX was set.
1760 20. Routers used for verification do not need to specify transports. However,
1761 if such a router generated a host list, and callout was configured, Exim
1762 crashed, because it could not find a port number from the (non-existent)
1763 transport. It now assumes port 25 in this circumstance.
1765 21. Added the -t option to exigrep.
1767 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1768 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1769 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1770 them is set, LOOKUP_LSEARCH is forced.
1772 23. "exim -bV" now outputs a list of lookups that are included in the binary.
1774 24. Added sender and host information to the "rejected by local_scan()" log
1775 line; previously there was no indication of these.
1777 25. Added .include_if_exists.
1779 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
1780 Linux. It turns out that not all file systems support this. Apparently some
1781 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1782 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1783 not supported on the file descriptor, is now ignored when Exim is trying to
1784 sync a directory. This applies only to Linux.
1786 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1788 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1789 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1792 29. Added support for maildirsize files from supplied patch (modified a bit).
1794 30. The use of :fail: followed by an empty string could lead Exim to respond to
1795 sender verification failures with (e.g.):
1797 550 Verification failed for <xxx>
1798 550 Sender verify failed
1800 where the first response line was missing the '-' that indicates it is not
1801 the final line of the response.
1803 31. The loop for finding the name of the user that called Exim had a hardwired
1804 limit of 10; it now uses the value of finduser_retries, which is used for
1805 all other user lookups.
1807 32. Added $received_count variable, available in data and not_smtp ACLs, and at
1810 33. Exim was neglecting to zero errno before one call of strtol() when
1811 expanding a string and expecting an integer value. On some systems this
1812 resulted in spurious "integer overflow" errors. Also, it was casting the
1813 result into an int without checking.
1815 34. Testing for a connection timeout using "timeout_connect" in the retry rules
1816 did not work. The code looks as if it has *never* worked, though it appears
1817 to have been documented since at least releast 1.62. I have made it work.
1819 35. The "timeout_DNS" error in retry rules, also documented since at least
1820 1.62, also never worked. As it isn't clear exactly what this means, and
1821 clearly it isn't a major issue, I have abolished the feature by treating it
1822 as "timeout", and writing a warning to the main and panic logs.
1824 36. The display of retry rules for -brt wasn't always showing the error code
1827 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
1828 timeout_connect_A, timeout_connect_MX.
1830 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
1831 to the empty sender.
1833 39. The daemon was not analysing the content of -oX till after it had closed
1834 stderr and disconnected from the controlling terminal. This meant that any
1835 syntax errors were only noted on the panic log, and the return code from
1836 the command was 0. By re-arranging the code a little, I've made the
1837 decoding happen first, so such errors now appear on stderr, and the return
1838 code is 1. However, the actual setting up of the sockets still happens in
1839 the disconnected process, so errors there are still only recorded on the
1842 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
1843 connections (as happens on some IP stacks) was logged at start up time as
1844 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
1845 it from "IPv6 and IPv4", which means that two separate sockets are being
1848 41. The debug output for gethostbyname2() or getipnodebyname() failures now
1849 says whether AF_INET or AF_INET6 was passed as an argument.
1851 42. Exiwhat output was messed up when time zones were included in log
1854 43. Exiwhat now gives more information about the daemon's listening ports,
1855 and whether -tls-on-connect was used.
1857 44. The "port" option of the smtp transport is now expanded.
1859 45. A "message" modifier in a "warn" statement in a non-message ACL was being
1860 silently ignored. Now an error message is written to the main and panic
1863 46. There's a new ACL modifier called "logwrite" which writes to a log file
1864 as soon as it is encountered.
1866 47. Added $local_user_uid and $local_user_gid at routing time.
1868 48. Exim crashed when trying to verify a sender address that was being
1871 49. Exim was recognizing only a space character after ".include". It now also
1872 recognizes a tab character.
1874 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
1875 extracting the relevant information from the specification. The man page no
1876 longer contains scrambled data for the -d option, and I've added a section
1877 at the front about calling Exim under different names.
1879 51. Added "extra_headers" argument to the "mail" command in filter files.
1881 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
1884 53. Installed eximstats 1.29.
1886 54. Added transport_filter_timeout as a generic transport option.
1888 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
1889 Cc: header lines. This was required by RFC 822, but it not required by RFC
1892 56. Exim used to add From:, Date:, and Message-Id: header lines to any
1893 incoming messages that did not have them. Now it does so only if the
1894 message originates locally, that is, if there is no associated remote host
1895 address. When Resent- header lines are present, this applies to the Resent-
1896 lines rather than the non-Resent- lines.
1898 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
1899 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
1901 58. Messages for configuration errors now include the name of the main
1902 configuration file - useful now that there may be more than one file in a
1903 list (.included file names were always shown).
1905 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
1906 for those rare installations that do not start the daemon as root or run it
1907 setuid root. I've cut out the call to initgroups() if the daemon is not
1910 60. The Exim user and group can now be bound into the binary as text strings
1911 that are looked up at the start of Exim's processing.
1913 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
1915 62. Added $mailstore_basename variable.
1917 63. Installed patch to sieve.c from Michael Haardt.
1919 64. When Exim failed to open the panic log after failing to open the main log,
1920 the original message it was trying to log was written to stderr and debug
1921 output, but if they were not available (the usual case in production), it
1922 was lost. Now it is written to syslog before the two lines that record the
1923 failures to open the logs.
1925 65. Users' Exim filters run in subprocesses under the user's uid. It is
1926 possible for a "deliver" command or an alias in a "personal" command to
1927 provoke an address rewrite. If logging of address rewriting is configured,
1928 this fails because the process is not running as root or exim. There may be
1929 a better way of dealing with this, but for the moment (because 4.30 needs
1930 to be released), I have disabled address rewrite logging when running a
1931 filter in a non-root, non-exim process.
1937 1. The buildconfig auxiliary program wasn't quoting the value set for
1938 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
1939 not defined. This bug was masked in 4.22 by the effect that was fixed in
1942 2. Some messages that were rejected after a message id was allocated were
1943 shown as "incomplete" by exigrep. It no longer does this for messages that
1944 are rejected by local_scan() or the DATA or non-SMTP ACLs.
1946 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
1947 have allow_domain_literals set, the ID did not get logged in the <= line.
1948 Domain literals are now always recognized in Message-ID: header lines.
1950 4. The first argument for a ${extract expansion item is the key name or field
1951 number. Leading and trailing spaces in this item were not being ignored,
1952 causing some misleading effects.
1954 5. When deliver_drop_privilege was set, single queue runner processes started
1955 manually (i.e. by the command "exim -q") or by the daemon (which uses the
1956 same command in the process it spins off) were not dropping privilege.
1958 6. When the daemon running as "exim" started a queue runner, it always
1959 re-executed Exim in the spun-off process. This is a waste of effort when
1960 deliver_drop_privilege is set. The new process now just calls the
1961 queue-runner function directly.
1967 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
1970 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
1971 ignored. Though the use of -oP was forcing the writing of a pid file, it
1972 was always written to the default place.
1974 3. If the message "no IP address found for host xxxx" is generated during
1975 incoming verification, it is now followed by identification of the incoming
1976 connection (so you can more easily find what provoked it).
1978 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
1980 5. Added some features to "harden" Exim a bit more against certain attacks:
1982 (a) There is now a build-time option called FIXED_NEVER_USERS that can
1983 be put in Local/Makefile. This is like the never_users runtime option,
1984 but it cannot be overridden. The default setting is "root".
1986 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
1987 prefix string with which any file named in a -C command line option
1990 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
1991 is retained for -C and -D only if the caller of Exim is root. Without
1992 it, the exim user may also use -C and -D and retain privilege.
1994 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
1995 command line option is disabled.
1997 6. Macro names set by the -D option must start with an upper case letter, just
1998 like macro names defined in the configuration file.
2000 7. Added "dereference=" facility to LDAP.
2002 8. Two instances of the typo "uknown" in the source files are fixed.
2004 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2005 the Configure-Makefile script screwed up while processing it.
2007 10. Incorporated PCRE 4.4.
2009 11. The SMTP synchronization check was not operating right at the start of an
2010 SMTP session. For example, it could not catch a HELO sent before the client
2011 waited for the greeting. There is now a check for outstanding input at the
2012 point when the greeting is written. Because of the duplex, asynchronous
2013 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2014 way, but not yet received, when the check is performed.
2016 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2017 on TCP/IP sockets, because this apparently causes some broken clients to
2020 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2021 unchanged) from the Cygwin maintainer.
2023 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2024 was supported instead of testing for mbx. Effectively a typo.
2026 15. The spa authenticator server code was not checking that the input it
2027 received was valid base64.
2029 16. The debug output line for the "set" modifier in ACLs was not showing the
2030 name of the variable that was being set.
2032 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2034 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2035 lookup on its own. Something else was needed to trigger the lookup. For
2036 example, a match in host_lookup or the need for a host name in a host list.
2037 Now, if $sender_host_name is referenced and the host name has not yet been
2038 looked up, a lookup is performed. If the lookup fails, the variable remains
2039 empty, and $host_lookup_failed is set to "1".
2041 19. Added "eqi" as a case-independent comparison operator.
2043 20. The saslauthd authentication condition could segfault if neither service
2044 nor realm was specified.
2046 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2047 error message that was logged was misleading, and incoming SMTP
2048 connections were dropped. The message is now more accurate, and temporary
2049 errors are given to SMTP connections.
2051 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2052 (except RSET) with a 421 error, until QUIT is received. However, it was
2053 failing to send a response to QUIT.
2055 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2056 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2057 ACL is now run after the helo_try_verify_hosts code.
2059 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2060 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2061 case-independent, so other case variants are also recognized). Apparently
2062 some systems use these upper case variants.
2064 25. If more than two messages were waiting for the same host, and a transport
2065 filter was specified for the transport, Exim sent two messages over the
2066 same TCP/IP connection, and then failed with "socket operation on non-
2067 socket" when it tried to send the third.
2069 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2071 27. The extern definition of crypt16() in expand.c was not being excluded when
2072 the OS had its own crypt16() function.
2074 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2075 as a preferred synonym for return_size_limit, both as an option and as an
2078 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2080 30. Changed the default configuration ACL to relax the local part checking rule
2081 for addresses that are not in any local domains. For these addresses,
2082 slashes and pipe symbols are allowed within local parts, but the sequence
2083 /../ is explicitly forbidden.
2085 31. SPA server authentication was not clearing the challenge buffer before
2088 32. log_message in a "warn" ACL statement was writing to the reject log as
2089 well as to the main log, which contradicts the documentation and doesn't
2090 seem right (because no rejection is happening). So I have stopped it.
2092 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2093 However, I am unable to do any testing of this.
2095 34. Fixed an infelicity in the appendfile transport. When checking directories
2096 for a mailbox, to see if any needed to be created, it was accidentally
2097 using path names with one or more superfluous leading slashes; tracing
2098 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2100 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2101 contents are added to the log line that is written for every discarded
2102 recipient. (Previously a log_message setting was ignored.)
2104 36. The ${quote: operator now quotes the string if it is empty.
2106 37. The install script runs exim in order to find its version number. If for
2107 some reason other than non-existence or emptiness, which it checks, it
2108 could not run './exim', it was installing it with an empty version number,
2109 i.e. as "exim-". This error state is now caught, and the installation is
2112 38. An argument was missing from the function that creates an error message
2113 when Exim fails to connect to the socket for saslauthd authentication.
2114 This could cause Exim to crash, or give a corrupted message.
2116 39. Added isip, isip4, and isip6 to ${if conditions.
2118 40. The ACL variables $acl_xx are now saved with the message, and can be
2119 accessed later in routers, transports, and filters.
2121 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2122 strings in the file are not string-expanded.
2124 42. If a MAIL command specified a SIZE value that was too large to fit into an
2125 int variable, the check against message_size_limit failed. Such values are
2126 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2127 day this will have to be increased, but I don't think I want to be around
2128 when emails are that large.
2135 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2136 iconv() is not standard in FreeBSD.
2138 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2139 IPv6 enabled. The observed symptom was a segmentation fault on return from
2140 the function os_common_find_running_interfaces() in src/os.c.
2142 3. In the check_special_case() function in daemon.c I had used "errno" as an
2143 argument name, which causes warnings on some systems. This was basically a
2144 typo, since it was named "eno" in the comments!
2146 4. The code that waits for the clock to tick (at a resolution of some fraction
2147 of a second) so as to ensure message-id uniqueness was always waiting for
2148 at least one whole tick, when it could have waited for less. [This is
2149 almost certainly not relevant at current processor speeds, where it is
2150 unlikely to ever wait at all. But we try to future-proof.]
2152 5. The function that sleeps for a time interval that includes fractions of a
2153 second contained a race. It did not block SIGALRM between setting the
2154 timer, and suspending (a couple of lines later). If the interval was short
2155 and the sigsuspend() was delayed until after it had expired, the suspension
2156 never ended. On busy systems this could lead to processes getting stuck for
2159 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2160 process, before it forks any delivery processes. The open lookup caching
2161 mechanism meant that the open file or database connection was passed into
2162 the delivery process. The problem was that delivery processes always tidy
2163 up cached lookup data. This could cause a problem for the next delivery
2164 process started by the queue runner, because the external queue runner
2165 process does not know about the closure. So the next delivery process
2166 still has data in the lookup cache. In the case of a file lookup, there was
2167 no problem because closing a file descriptor in a subprocess doesn't affect
2168 the parent. However, if the lookup was caching a connection to a database,
2169 the connection was closed, and the second delivery process was likely to
2170 see errors such as "PGSQL: query failed: server closed the connection
2171 unexpectedly". The problem has been fixed by closing all cached lookups
2172 in a queue runner before running a delivery process.
2174 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2175 seem to have the "const" qualifier which it has on other OS. I've
2178 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2179 *of the same type* (client or server) with the same public name that an
2180 error should be diagnosed.
2182 9. When Exim looked up a host name for an IP address, but failed to find the
2183 original IP address when looking up the host name (a safety check), it
2184 output the message "<ip address> does not match any IP for NULL", which was
2185 confusing, to say the least. The bug was that the host name should have
2186 appeared instead of "NULL".
2188 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2189 user that is built into the binary, and the -C or -D options is used, root
2190 privilege is dropped before the configuration file is read. In addition,
2191 logging is switched to stderr instead of the normal log files. If the
2192 configuration then re-defines the Exim user, the unprivileged environment
2193 is probably not what is expected, so Exim logs a panic warning message (but
2196 However, if deliver_drop_privilege is set, the unprivileged state may well
2197 be exactly what is intended, so the warning has been cut out in that case,
2198 and Exim is allowed to try to write to its normal log files.
2204 1. smtp_return_error_details was not giving details for temporary sender
2205 or receiver verification errors.
2207 2. Diagnose a configuration error if two authenticators have the same public
2210 3. Exim used not to create the message log file for a message until the first
2211 delivery attempt. This could be confusing when incoming messages were held
2212 for policy or load reasons. The message log file is now created at the time
2213 the message is received, and an initial "Received" line is written to it.
2215 4. The automatically generated man page for command line options had a minor
2216 bug that caused no ill effects; however, a more serious problem was that
2217 the procedure for building the man page automatically didn't always
2218 operate. Consequently, release 4.20 contains an out-of-date version. This
2219 shouldn't happen again.
2221 5. When building Exim with embedded Perl support, the script that builds the
2222 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2223 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2225 6. The freeze_tell option was not being used for messages that were frozen on
2226 arrival, either by an ACL or by local_scan().
2228 7. Added the smtp_incomplete_transaction log selector.
2230 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2231 was accepting AUTH without a new EHLO.
2233 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2234 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2237 10. Logging of TCP/IP connections (when configured) now happens in the main
2238 daemon process instead of the child process, so that the TCP/IP connection
2239 count is more accurate (but it can never be perfect).
2241 11. The use of "drop" in a nested ACL was not being handled correctly in the
2242 outer ACL. Now, if condition failure induced by the nested "drop" causes
2243 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2244 or "require"), the connection is dropped.
2246 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2247 that yield "discard" can now be used with an "accept" or a "discard" verb,
2248 but an error is generated for any others (because I can't see a useful way
2249 to define what should happen).
2251 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2252 are now processed in the same way as lines in the Exim configuration file.
2253 In particular, continuation lines are supported.
2255 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2257 15. Added -ti meaning -t -i.
2259 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2260 domains, and warn by logging if others are found.
2262 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2263 returned from GIFCONF when Exim is trying to find the list of interfaces on
2264 a host. The code in os.c has been modified to copy each ifreq to an aligned
2265 structure in all cases.
2267 Also, in some cases, the returned ifreq's were being copied to a 'struct
2268 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2269 means the last couple of bytes of an IPv6 address could be chopped if the
2270 ifreq contained only a normal sockaddr (14 bytes storage).
2272 18. Named domain lists were not supported in the hosts_treat_as_local option.
2273 An entry such as +xxxx was not recognized, and was treated as a literal
2276 19. Ensure that header lines added by a DATA ACL are included in the reject log
2277 if the ACL subsequently rejects the message.
2279 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2280 MD5 (which is deprecated).
2282 21. When testing a filter file using -bf, Exim was writing a message when it
2283 took the sender from a "From " line in the message, but it was not doing so
2284 when it took $return_path from a Return-Path: header line. It now does.
2286 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2287 with a valid header line field name (a series of printing characters
2288 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2290 23. Changed "disc" in the source to "disk" to conform to the documentation and
2291 the book and for uniformity.
2293 24. Ignore Sendmail's -Ooption=value command line item.
2295 25. When execve() failed while trying to run a command in a pipe transport,
2296 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2297 could be confused with a return value of 69 from the command itself. This
2298 has been changed to 127, the value the shell returns if it is asked to run
2299 a non-existent command. The wording for the related log line suggests a
2300 non-existent command as the problem.
2302 26. If received_header_text expands to an empty string, do not add a Received:
2303 header line to the message. (Well, it adds a token one on the spool, but
2304 marks it "old" so that it doesn't get used or transmitted.)
2306 27. Installed eximstats 1.28 (addition of -nt option).
2308 28. There was no check for failure on the call to getsockname() in the daemon
2309 code. This can fail if there is a shortage of resources on the system, with
2310 ENOMEM, for example. A temporary error is now given on failure.
2312 29. Contrary to the C standard, it seems that in some environments, the
2313 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2314 program. Exim now does this explicitly; it affects the formatting of
2315 timestamps using strftime().
2317 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2318 complaints. I've now initialized all the variables, to avoid this.
2320 32. Header lines added by a system filter were not being "seen" during
2321 transport-time rewrites.
2323 33. The info_callback() function passed to OpenSSL is set up with type void
2324 (*)(SSL *, int, int), as described somewhere. However, when calling the
2325 function (actually a macro) that sets it up, the type void(*)() is
2326 expected. I've put in a cast to prevent warnings from picky compilers.
2328 34. If a DNS black list lookup found a CNAME record, but there were no A
2329 records associated with the domain it pointed at, Exim crashed.
2331 35. If a DNS black list lookup returned more than one A record, Exim ignored
2332 all but the first. It now scans all returned addresses if a particular IP
2333 value is being sought. In this situation, the contents of the
2334 $dnslist_value variable are a list of all the addresses, separated by a
2337 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2338 to accept a host name and all its aliases if the forward lookup for any of
2339 them yielded the IP address of the incoming connection. Now it accepts only
2340 those names whose forward lookup yields the correct IP address. Any other
2341 names are discarded. This closes a loophole whereby a rogue DNS
2342 administrator could create reverse DNS records to break through a
2343 wildcarded host restriction in an ACL.
2345 37. If a user filter or a system filter that ran in a subprocess used any of
2346 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2347 the wrong values were passed to the pipe command ($thisaddress had the
2348 value of $0, $0 had the value of $1, etc). This bug was introduced by
2349 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2351 38. Improved the line breaking for long SMTP error messages from ACLs.
2352 Previously, if there was no break point between 40 and 75 characters, Exim
2353 left the rest of the message alone. Two changes have been made: (a) I've
2354 reduced the minimum length to 35 characters; (b) if it can't find a break
2355 point between 35 and 75 characters, it looks ahead and uses the first one
2356 that it finds. This may give the occasional overlong line, but at least the
2357 remaining text gets split now.
2359 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2360 file descriptors might be low, and that setting 1000 would always raise it.
2361 It turns out that in some environments, the limit is already over 1000 and
2362 that lowering it causes trouble. So now Exim takes care not to decrease it.
2364 40. When delivering a message, the value of $return_path is set to $sender_
2365 address at the start of routing (routers may change the value). By an
2366 oversight, this default was not being set up when an address was tested by
2367 -bt or -bv, which affected the outcome if any router or filter referred to
2370 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2371 log only when "message" or "log_message" are set. However, if one of the
2372 conditions was an address verification, or a call to a nested ACL, the
2373 messages generated by the underlying test were being passed through. This
2374 no longer happens. The underlying message is available in $acl_verify_
2375 message for both "message" and "log_message" expansions, so it can be
2376 passed through if needed.
2378 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2379 new expansion $bh_ to give the encoded byte string without charset
2380 translation. Translation happens only if iconv() is available; HAVE_ICONV
2381 indicates this at build time. HEADERS_CHARSET gives the charset to
2382 translate to; headers_charset can change it in the configuration, and
2383 "headers charset" can change it in an individual filter file.
2385 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2386 that creates RFC 2047 encoded "words" labels them as that charset instead
2387 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2388 expansion operator; (ii) when Exim creates a From: line for a local
2389 message; (iii) when a header line is rewritten to include a "phrase" part.
2391 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2392 buggy, causing it to skip the first lines of messages whose message ID
2393 ended in 'D'. This would not have bitten before Exim release 4.14, because
2394 message IDs were unlikely to end in 'D' before then. The effect was to have
2395 incorrect size information for certain domains.
2397 45. #include "config.h" was missing at the start of the crypt16.c module. This
2398 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2401 46. If there was a timeout during a "random" callout check, Exim treated it as
2402 a failure of the random address, and carried on sending RSET and the real
2403 address. If the delay was just some slowness somewhere, the response to the
2404 original RCPT would be taken as a response to RSET and so on, causing
2405 mayhem of various kinds.
2407 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2408 when I implemented it. It didn't allow for the fact that some option values
2409 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2410 the right test for positive values.
2412 48. Domain names in DNS records are case-independent. Exim always looks them up
2413 in lower case. Some resolvers return domain names in exactly the case they
2414 appear in the zone file, that is, they may contain uppercase letters. Not
2415 all resolvers do this - some return always lower case. Exim was treating a
2416 change of case by a resolver as a change of domain, similar to a widening
2417 of a domain abbreviation. This triggered its re-routing code and so it was
2418 trying to route what was effectively the same domain again. This normally
2419 caused routing to fail (because the router wouldn't handle the domain
2420 twice). Now Exim checks for this case specially, and just changes the
2421 casing of the domain that it ultimately uses when it transmits the message
2424 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2427 50. If a filter generated a file delivery with a non-absolute name (possible if
2428 no home directory exists for the router), the forbid_file option was not
2431 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2432 the existing equality matching.
2434 52. Exim was using ints instead of ino_t variables in some places where it was
2435 dealing with inode numbers.
2437 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2438 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2439 called TMPDIR, and if it finds it is different, it changes its value.
2441 54. The smtp_printf() function is now made available to local_scan() so
2442 additional output lines can be written before returning. There is also an
2443 smtp_fflush() function to enable the detection of a dropped connection.
2444 The variables smtp_input and smtp_batched_input are exported to
2447 55. Changed the default runtime configuration: the message "Unknown user"
2448 has been removed from the ACL, and instead placed on the localuser router,
2449 using the cannot_route_message feature. This means that any verification
2450 failures that generate their own messages won't get overridden. Similarly,
2451 the "Unrouteable address" message that was in the ACL for unverifiable
2452 relay addresses has also been removed.
2454 56. Added hosts_avoid_esmtp to the smtp transport.
2456 57. The exicyclog script was not checking for the esoteric option
2457 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2458 will work only if exicyclog is run under the appropriate euid.
2460 58. Following a discussion on the list, the rules by which Exim recognises line
2461 endings on incoming messages have been changed. The -dropcr and drop_cr
2462 options are now no-ops, retained only for backwards compatibility. The
2463 following line terminators are recognized: LF CRLF CR. However, special
2464 processing applies to CR:
2466 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2467 nor a local message in the state where . is a terminator.
2469 (ii) If a bare CR is encountered in a header line, an extra space is added
2470 after the line terminator so as not to end the header. The reasoning
2471 behind this is that bare CRs in header lines are most likely either
2472 to be mistakes, or people trying to play silly games.
2474 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
2475 was being incorrectly given as 18 bytes larger than it should have been.
2476 This is a VOB (very old bug).
2478 60. This may never have affected anything current, but just in case it has:
2479 When the local host is found other than at the start of a list of hosts,
2480 the local host, those with the same MX, and any that follow, are discarded.
2481 When the list in question was part of a longer list of hosts, the following
2482 hosts (not currently being processed) were also being discarded. This no
2483 longer happens. I'm not sure if this situation could ever has previously
2486 61. Added the "/MX" feature to lists of hosts in the manualroute and query
2489 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2490 header. This is something that is recommended in a new Internet Draft, and
2491 is something that is documented as being done by Sendmail. There are two
2492 possible values. For messages generated by the autoreply transport, Exim
2495 Auto-Submitted: auto-replied
2497 whereas for all other generated messages (e.g. bounces) it adds
2499 Auto-Submitted: auto-generated
2501 63. The "personal" condition in filters now includes a test for the
2502 Auto-Submitted: header. If it contains the string "auto-" the message it
2503 not considered personal.
2505 64. Added rcpt_include_affixes as a generic transport option.
2507 65. Added queue_only_override (default true).
2509 66. Added the syslog_duplication option.
2511 67. If what should have been the first header line of a message consisted of
2512 a space followed by a colon, Exim was mis-interpreting it as a header line.
2513 It isn't of course - it is syntactically invalid and should therefore be
2514 treated as the start of the message body. The misbehaviour could have
2515 caused a number of strange effects, including loss of data in subsequent
2516 header lines, and spool format errors.
2518 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2519 client host had authenticated. This control can now be exercised by an ACL
2520 for more flexibility.
2522 69. By default, callouts do not happen when testing with -bh. There is now a
2523 variant, -bhc, which does actually run the callout code, including
2524 consulting and updating the callout cache.
2526 70. Added support for saslauthd authentication, courtesy of Alexander
2529 71. If statvfs() failed on the spool or log directories while checking their
2530 size for availability, Exim confusingly gave the error "space shortage".
2531 Furthermore, in debugging mode it crashed with a floating point exception.
2532 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2533 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2534 serious problem, Exim now writes to the main and panic logs when this
2535 happens, with details of the failure. It then refuses to accept the
2536 incoming message, giving the message "spool directory problem" or "log
2537 directory problem" with a 421 code for SMTP messages.
2539 72. When Exim is about to re-exec itself, it ensures that the file descriptors
2540 0, 1, and 2 exist, because some OS complain for execs without them (see
2541 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2542 descriptors. However, the code omitted to check that the open succeeded,
2543 causing mysterious errors if for some reason the permissions on /dev/null
2544 got screwed. Now Exim writes a message to the main and panic logs, and
2545 bombs out if it can't open /dev/null.
2547 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2548 interact so that it is all more flexible. It is supposed to remain
2549 backwards compatible. Also added extra_local_interfaces.
2551 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2552 to bomb out with an assertion failure - to the client this appears as a
2553 connection drop. This problem occurs in the part of the code that was taken
2554 from the Samba project. Fortunately, the assertion is in a very simple
2555 function, so I have fixed this by reproducing the function inline in the
2556 one place where it is called, and arranging for authentication to fail
2557 instead of killing the process with assert().
2559 75. The SPA client code was not working when the server requested OEM rather
2560 than Unicode encoding.
2562 76. Added code to make require_files with a specific uid setting more usable in
2563 the case where statting the file as root fails - usually a non-root-mounted
2564 NFS file system. When this happens and the failure is EACCES, Exim now
2565 forks a subprocess and does the per-uid checking as the relevant uid.
2567 77. Added process_log_path.
2569 78. If log_file_path was not explicitly set, a setting of check_log_space or
2570 check_log_inodes was ignored.
2572 79. If a space check for the spool or log partitions fails, the incident is now
2573 logged. Of course, in the latter case the data may get lost...
2575 80. Added the %p formatting code to string_format() so that it can be used to
2576 print addresses in debug_print(). Adjusted all the address printing in the
2577 debugging in store.c to use %p rather than %d.
2579 81. There was a concern that a line of code in smtp_in.c could overflow a
2580 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2581 initially expressed, the concern was not well-founded, because trailing
2582 spaces are removed early. However, if the trailing spaces were followed by
2583 a NULL, they did not get removed, so the overflow was possible. Two fixes
2586 (a) I re-wrote the offending code in a cleaner fashion.
2587 (b) If an incoming SMTP command contains a NULL character, it is rejected
2590 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
2591 runs initgroups(), so that if the Exim user is in any additional groups,
2592 they will be used during message reception.
2598 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob