1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.131 2005/05/03 10:02:27 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
10 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
11 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
13 TK/02 Fix ACL "control" statement not being available in MIME ACL.
15 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
17 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
18 to test Sieve filters that use "vacation".
20 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
21 that changes the way the GnuTLS parameters are stored in the cache file.
22 The new format can be generated externally. For backward compatibility,
23 if the data in the cache doesn't make sense, Exim assumes it has read an
24 old-format file, and it generates new data and writes a new file. This
25 means that you can't go back to an older release without removing the
28 PH/03 A redirect router that has both "unseen" and "one_time" set does not
29 work if there are any delivery delays because "one_time" forces the
30 parent to be marked "delivered", so its unseen clone is never tried
31 again. For this reason, Exim now forbids the simultaneous setting of
34 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
35 redirected to themselves ("homonym" addresses). Read the long ChangeLog
36 entry if you want to know the details. The fix, however, neglected to
37 consider the case when local delivery batching is involved. The test for
38 "previously delivered" was not happening when checking to see if an
39 address could be batched with a previous (undelivered) one; under
40 certain circumstances this could lead to multiple deliveries to the same
43 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
44 in its include files, and this causes problems building Exim.
46 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
47 header_syntax) but Exim was just ignoring anything given after a slash.
48 In particular, this caused confusion with an attempt to use "verify =
49 reverse_host_lookup/defer_ok". An error is now given when options are
50 supplied for verify items that do not have them. (Maybe reverse_host_
51 lookup should have a defer_ok option, but that's a different point.)
53 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
54 defined by RFC 821) to 2048, because there were problems with some AUTH
55 commands, and RFC 1869 says the size should be increased for extended
56 SMTP commands that take arguments.
58 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
61 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
62 "unknown" error; now it says that the functionality isn't in the binary.
64 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
65 an address' error message when a string expansion fails (syntax or
66 whatever). Otherwise the password may appear in the log. Following change
67 PH/42 below, there is no longer a chance of it appearing in a bounce
70 PH/11 Installed exipick version 20050225.0 from John Jetmore.
72 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
73 of its addresses was ever tried. (Bugzilla bug #2.)
75 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
76 the result incorrectly in the debug output. (It correctly added a newline
77 to what was transported.)
79 TF/01 Added $received_time.
81 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
82 commented out examples of how to interface to a virus scanner and to
83 SpamAssassin. Also added commented examples of av_scanner and
84 spamd_address settings.
86 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
87 and controls are allowed in which ACLs. There were a couple of minor
88 errors. Some of the entries in the conditions table (which is a table of
89 where they are NOT allowed) were getting very unwieldy; rewrote them as a
90 negation of where the condition IS allowed.
92 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
94 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
95 header file does not have a version number, so I've had to invent a new
96 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
97 API. The code is untested by me (my Linux distribution still has 0.3.2 of
98 radiusclient), but it was contributed by a Radius user.
100 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
101 files or directories, for OpenSSL.
103 PH/19 When an Exim process that is running as root has to create an Exim log
104 file, it does so in a subprocess that runs as exim:exim so as to get the
105 ownership right at creation (otherwise, other Exim processes might see
106 the file with the wrong ownership). There was no test for failure of this
107 fork() call, which would lead to the process getting stuck as it waited
108 for a non-existent subprocess. Forks do occasionally fail when resources
109 run out. I reviewed all the other calls to fork(); they all seem to check
112 PH/20 When checking for unexpected SMTP input at connect time (before writing
113 the banner), Exim was not dealing correctly with a non-positive return
114 from the read() function. If the client had disconnected by this time,
115 the result was a log entry for a synchronization error with an empty
116 string after "input=" when read() returned zero. If read() returned -1
117 (an event I could not check), uninitialized data bytes were printed.
118 There were reports of junk text (parts of files, etc) appearing after
121 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
123 PH/22 Added support for macro redefinition, and (re)definition in between
124 driver and ACL definitions.
126 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
127 forgetting to use the resulting value; it was using the unexpanded value.
129 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
130 hadn't been configured. The fix is from Juergen Kreileder, who
131 understands it better than I do:
133 "Here's what I see happening with three configured cyrus_sasl
134 authenticators configured (plain, login, cram-md5):
136 On startup auth_cyrus_sasl_init() gets called for each of these.
137 This means three calls to sasl_listmech() without a specified mech_list.
138 => SASL tests which mechs of all available mechs actually work
139 => three warnings about OTP not working
140 => the returned list contains: plain, login, cram-md5, digest-md5, ...
142 With the patch, sasl_listmech() also gets called three times. But now
143 SASL's mech_list option is set to the server_mech specified in the the
144 authenticator. Or in other words, the answer from sasl_listmech()
145 gets limited to just the mech you're testing for (which is different
147 => the return list contains just 'plain' or 'login', 'cram-md5' or
148 nothing depending on the value of ob->server_mech.
150 I've just tested the patch: Authentication still works fine,
151 unavailable mechs specified in the exim configuration are still
152 caught, and the auth.log warnings about OTP are gone."
154 PH/25 When debugging is enabled, the contents of the command line are added
155 to the debugging output, even when log_selector=+arguments is not
158 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
159 answer is "GNU", and only if the return is "GNU/something" is the answer
162 PH/27 $acl_verify_message is now set immediately after the failure of a
163 verification in an ACL, and so is available in subsequent modifiers. In
164 particular, the message can be preserved by coding like this:
166 warn !verify = sender
167 set acl_m0 = $acl_verify_message
169 Previously, $acl_verify_message was set only while expanding "message"
170 and "log_message" when a very denied access.
172 PH/28 Modified OS/os.c-Linux with
174 -#ifndef OS_LOAD_AVERAGE
175 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
177 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
178 nomenclature these days.)
180 PH/29 Installed patch from the Sieve maintainer that adds the options
181 sieve_useraddress and sieve_subaddress to the redirect router.
183 PH/30 In these circumstances:
184 . Two addresses routed to the same list of hosts;
185 . First host does not offer TLS;
186 . First host accepts first address;
187 . First host gives temporary error to second address;
188 . Second host offers TLS and a TLS session is established;
189 . Second host accepts second address.
190 Exim incorrectly logged both deliveries with the TLS parameters (cipher
191 and peerdn, if requested) that were in fact used only for the second
194 PH/31 When doing a callout as part of verifying an address, Exim was not paying
195 attention to any local part prefix or suffix that was matched by the
196 router that accepted the address. It now behaves in the same way as it
197 does for delivery: the affixes are removed from the local part unless
198 rcpt_include_affixes is set on the transport.
200 PH/32 Add the sender address, as F=<...>, to the log line when logging a
201 timeout during the DATA phase of an incoming message.
203 PH/33 Sieve envelope tests were broken for match types other than :is. I have
204 applied a patch sanctioned by the Sieve maintainer.
206 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
207 the uid or gid is negative. A case of a negative gid caused this to be
208 noticed. The fix allows for either to be negative.
210 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
211 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
214 PH/36 The change PH/12 above was broken. Fixed it.
216 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
217 the grounds that routing the same address twice would always produce the
218 same answer. This might have been true once, but it is certainly no
219 longer true now. Routing a child address may depend on the previous
220 routing that produced that child. Some complicated redirection strategies
221 went wrong when messages had multiple recipients, and made Exim's
222 behaviour dependent on the order in which the addresses were given.
224 I have moved the duplicate checking until after the routing is complete.
225 Exim scans the addresses that are assigned to local and remote
226 transports, and removes any duplicates. This means that more work will be
227 done, as duplicates will always all be routed, but duplicates are
228 presumably rare, so I don't expect this is of any significance.
230 For deliveries to pipes, files, and autoreplies, the duplicate checking
231 still happens during the routing process, since they are not going to be
234 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
235 It corrects a timeout issue with spamd. This is Ian's comment: "The
236 background is that sometimes spamd either never reads data from a
237 connection it has accepted, or it never writes response data. The exiscan
238 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
239 blindly assumes that writes won't block so it may never time out."
241 PH/39 Allow G after quota size as well as K and M.
243 PH/40 The value set for $authenticated_id in an authenticator may not contain
244 binary zeroes or newlines because the value is written to log lines and
245 to spool files. There was no check on this. Now the value is run through
246 the string_printing() function so that such characters are converted to
247 printable escape sequences.
249 PH/41 $message_linecount is a new variable that contains the total number of
250 lines in the message. Compare $body_linecount, which is the count for the
253 PH/42 Exim no longer gives details of delivery errors for specific addresses in
254 bounce and delay warning messages, except in certain special cases, which
257 (a) An SMTP error message from a remote host;
258 (b) A message specified in a :fail: redirection;
259 (c) A message specified in a "fail" command in a system filter;
260 (d) A message specified in a FAIL return from the queryprogram router;
261 (e) A message specified by the cannot_route_message router option.
263 In these cases only, Exim does include the error details in bounce and
264 warning messages. There are also a few cases where bland messages such
265 as "unrouteable address" or "local delivery error" are given.
267 PH/43 $value is now also set for the "else" part of a ${run expansion.
269 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
270 being worked on, but at least Exim now implements the latest version to
274 A note about Exim versions 4.44 and 4.50
275 ----------------------------------------
277 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
278 changes of various kinds. As a consequence, a big documentation update was
279 needed. This delayed the release for rather longer than seemed good, especially
280 in the light of a couple of (minor) security issues. Therefore, the changes
281 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
282 release. So 4.44 and 4.50 are in effect two different branches that both start
285 I have left the 4.50 change log unchanged; it contains all the changes since
286 4.43. The change log for 4.44 is below; many of its items are identical to
287 those for 4.50. This seems to be the most sensible way to preserve the
288 historical information.
294 1. Minor wording change to the doc/README.SIEVE file.
296 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
297 computation of the current number of files was incorrect.
299 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
300 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
303 4. Give more explanation in the error message when the command for a transport
304 filter fails to execute.
306 5. There are several places where Exim runs a non-Exim command in a
307 subprocess. The SIGUSR1 signal should be disabled for these processes. This
308 was being done only for the command run by the queryprogram router. It is
309 now done for all such subprocesses. The other cases are: ${run, transport
310 filters, and the commands run by the lmtp and pipe transports.
312 6. Added CONFIGURE_GROUP build-time option.
314 7. Some older OS have a limit of 256 on the maximum number of file
315 descriptors. Exim was using setrlimit() to set 1000 as a large value
316 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
317 systems. I've change it so that if it can't get 1000, it tries for 256.
319 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
320 was an oversight, and furthermore, ever since the addition of extra
321 controls (e.g. 4.43/32), the checks on when to allow different forms of
322 "control" were broken. There should now be diagnostics for all cases when a
323 control that does not make sense is encountered.
325 9. Added the /retain_sender option to "control=submission".
327 10. $recipients is now available in the predata ACL (oversight).
329 11. Tidy the search cache before the fork to do a delivery from a message
330 received from the command line. Otherwise the child will trigger a lookup
331 failure and thereby defer the delivery if it tries to use (for example) a
332 cached ldap connection that the parent has called unbind on.
334 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
335 of $address_data from the recipient verification was clobbered by the
338 13. The value of address_data from a sender verification is now available in
339 $sender_address_data in subsequent conditions in the ACL statement.
341 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
343 15. Added a new option "connect=<time>" to callout options, to set a different
346 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
347 was its contents. (It was OK if the option was not defined at all.)
349 17. A "Completed" log line is now written for messages that are removed from
350 the spool by the -Mrm option.
352 18. New variables $sender_verify_failure and $recipient_verify_failure contain
353 information about exactly what failed.
355 19. Added -dd to debug only the daemon process.
357 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
358 handles timeouts, both on the server side and network timeouts. Renamed the
359 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
361 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
362 It is now set to "smtps".
364 22. $host_address is now set to the target address during the checking of
367 23. When checking ignore_target_hosts for an ipliteral router, no host name was
368 being passed; this would have caused $sender_host_name to have been used if
369 matching the list had actually called for a host name (not very likely,
370 since this list is usually IP addresses). A host name is now passed as
373 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
374 code that specifies a non-restarting handler (typically sigaction() in
375 modern systems) in an attempt to fix a rare and obscure crash bug.
377 25. Narrowed the window for a race in the daemon that could cause it to ignore
378 SIGCHLD signals. This is not a major problem, because they are used only to
379 wake it up if nothing else does.
381 26. A malformed maildirsize file could cause Exim to calculate negative values
382 for the mailbox size or file count. Odd effects could occur as a result.
383 The maildirsize information is now recalculated if the size or filecount
386 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
387 support for a long time. Removed HAVE_SYS_VFS_H.
389 28. Installed the latest version of exipick from John Jetmore.
391 29. In an address list, if the pattern was not a regular expression, an empty
392 subject address (from a bounce message) matched only if the pattern was an
393 empty string. Non-empty patterns were not even tested. This was the wrong
394 because it is perfectly reasonable to use an empty address as part of a
395 database query. An empty address is now tested by patterns that are
396 lookups. However, all the other forms of pattern expect the subject to
397 contain a local part and a domain, and therefore, for them, an empty
398 address still always fails if the pattern is not itself empty.
400 30. Exim went into a mad DNS loop when attempting to do a callout where the
401 host was specified on an smtp transport, and looking it up yielded more
404 31. Re-factored the code for checking spool and log partition space into a
405 function that finds that data and another that does the check. The former
406 is then used to implement four new variables: $spool_space, $log_space,
407 $spool_inodes, and $log_inodes.
409 32. The RFC2047 encoding function was originally intended for short strings
410 such as real names; it was not keeping to the 75-character limit for
411 encoded words that the RFC imposes. It now respects the limit, and
412 generates multiple encoded words if necessary. To be on the safe side, I
413 have increased the buffer size for the ${rfc2047: expansion operator from
416 33. It is now permitted to omit both strings after an "if" condition; if the
417 condition is true, the result is "true". As before, when the second string
418 is omitted, a false condition yields an empty string. This makes it less
419 cumbersome to write custom ACL and router conditions.
421 34. Failure to deliver a bounce message always caused it to be frozen, even if
422 there was an errors_to setting on the router. The errors_to setting is now
425 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
426 canonical form (fully expanded) before being placed in
427 $sender_host_address.
429 36. The table in the code that translates DNS record types into text (T_A to
430 "A" for instance) was missing entries for NS and CNAME. It is just possible
431 that this could have caused confusion if both these types were looked up
432 for the same domain, because the text type is used as part of Exim's
433 per-process caching. But the chance of anyone hitting this buglet seems
436 37. The dnsdb lookup has been extended in a number of ways.
438 (1) There is a new type, "zns", which walks up the domain tree until it
439 finds some nameserver records. It should be used with care.
441 (2) There is a new type, "mxh", which is like "mx" except that it returns
442 just the host names, not the priorities.
444 (3) It is now possible to give a list of domains (or IP addresses) to be
445 looked up. The behaviour when one of the lookups defers can be
446 controlled by a keyword.
448 (4) It is now possible to specify the separator character for use when
449 multiple records are returned.
451 38. The dnslists ACL condition has been extended: it is now possible to supply
452 a list of IP addresses and/or domains to be looked up in a particular DNS
455 39. Added log_selector=+queue_time_overall.
457 40. When running the queue in the test harness, wait just a tad after forking a
458 delivery process, to get repeatability of debugging output.
460 41. Include certificate and key file names in error message when GnuTLS fails
461 to set them up, because the GnuTLS error message doesn't include the name
462 of the failing file when there is a problem reading it.
464 42. Allow both -bf and -bF in the same test run.
466 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
468 44. The "Exiscan patch" is now merged into the mainline Exim source.
470 45. Sometimes the final signoff response after QUIT could fail to get
471 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
472 < 0 before doing a fflush(). This bug looks as though it goes back to the
473 introduction of TLS in release 3.20, but "sometimes" must have been rare
474 because the tests only now provoked it.
476 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
477 (this can affect the format of dates).
479 47. exim_tidydb, when checking for the continued existence of a message for
480 which it has found a message-specific retry record, was not finding
481 messages that were in split spool directories. Consequently, it was
482 deleting retry records that should have stayed in existence.
484 48. Steve fixed some bugs in eximstats.
486 49. The SPA authentication driver was not abandoning authentication and moving
487 on to the next authenticator when an expansion was forced to fail,
488 contradicting the general specification for all authenticators. Instead it
489 was generating a temporary error. It now behaves as specified.
491 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
492 (the order specifies the preference for clients). The order is now AES256,
493 AES128, 3DES, ARCFOUR128.
495 51. Small patch to Sieve code - explicitly set From: when generating an
498 52. Exim crashed if a remote delivery caused a very long error message to be
499 recorded - for instance if somebody sent an entire SpamAssassin report back
500 as a large number of 550 error lines. This bug was coincidentally fixed by
501 increasing the size of one of Exim's internal buffers (big_buffer) that
502 happened as part of the Exiscan merge. However, to be on the safe side, I
503 have made the code more robust (and fixed the comments that describe what
506 53. Now that there can be additional text after "Completed" in log lines (if
507 the queue_time_overall log selector is set), a one-byte patch to exigrep
508 was needed to allow it to recognize "Completed" as not the last thing in
511 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
512 patch that reportedly fixes this has been added. I am not expert enough to
513 create a test for it. This is what the patch creator wrote:
515 "I found a little strange behaviour of ldap code when working with
516 Windows 2003 AD Domain, where users was placed in more than one
517 Organization Units. When I tried to give exim partial DN, the exit code
518 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
519 But simultaneously result of request was absolutely normal ldap result,
520 so I produce this patch..."
522 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
523 REFERENCE, so I have modified the code to exclude the patch when that macro
526 55. Some experimental protocols are using DNS PTR records for new purposes. The
527 keys for these records are domain names, not reversed IP addresses. The
528 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
529 leaves it alone. Component reversal etc. now happens only for IP addresses.
532 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
534 57. Double the size of the debug message buffer (to 2048) so that more of very
535 long debug lines gets shown.
537 58. The exicyclog utility now does better if the number of log files to keep
538 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
540 59. Two changes related to the smtp_active_hostname option:
542 (1) $smtp_active_hostname is now available as a variable.
543 (2) The default for smtp_banner uses $smtp_active_hostname instead
544 of $primary_hostname.
546 60. The host_aton() function is supposed to be passed a string that is known
547 to be a valid IP address. However, in the case of IPv6 addresses, it was
548 not checking this. This is a hostage to fortune. Exim now panics and dies
549 if the condition is not met. A case was found where this could be provoked
550 from a dnsdb PTR lookup with an IPv6 address that had more than 8
551 components; fortuitously, this particular loophole had already been fixed
552 by change 4.50/55 above.
554 If there are any other similar loopholes, the new check in host_aton()
555 itself should stop them being exploited. The report I received stated that
556 data on the command line could provoke the exploit when Exim was running as
557 exim, but did not say which command line option was involved. All I could
558 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
562 61. There was a buffer overflow vulnerability in the SPA authentication code
563 (which came originally from the Samba project). I have added a test to the
564 spa_base64_to_bits() function which I hope fixes it.
567 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
568 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
570 63. The daemon start-up calls getloadavg() while still root for those OS that
571 need the first call to be done as root, but it missed one case: when
572 deliver_queue_load_max is set with deliver_drop_privilege. This is
573 necessary for the benefit of the queue runner, because there is no re-exec
574 when deliver_drop_privilege is set.
576 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
579 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
580 was in use, was not putting the data itself into the right store pool;
581 consequently, it could be overwritten for a subsequent message in the same
582 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
585 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
587 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
588 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
589 the function were treating the return as a boolean value, which happened to
590 work because 0=false and not-0=true, but is not correct code.
592 68. The host_aton() function was not handling scoped IPv6 addresses (those
593 with, for example, "%eth0" on the end) correctly.
595 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
596 negated items (that is, ~something) in unsigned ints. Some compilers
597 apparently mutter when there is no cast.
599 70. If an address verification called from an ACL failed, and did not produce a
600 user-specific message (i.e. there was only a "system" message), nothing was
601 put in $acl_verify_message. In this situation, it now puts the system
604 71. Change 4.23/11 added synchronization checking at the start of an SMTP
605 session; change 4.31/43 added the unwanted input to the log line - except
606 that it did not do this in the start of session case. It now does.
608 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
609 This is wrong and can cause the other end to generate a synchronization
610 error if it is another Exim or anything else that does the synchronization
611 check. A QUIT command is no longer sent after a timeout.
613 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
616 74. The defer_ok option of callout verification was not working if it was used
617 when verifying addresses in header lines, that is, for this case:
619 verify = header_sender/callout=defer_ok
621 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
622 those file descriptors could be used for SMTP connections. If anything
623 wrote to stderr (the example that came up was "warn" in embedded Perl), it
624 could be sent to the SMTP client, causing chaos. The daemon now opens
625 stdin, stdout, and stderr to /dev/null when it puts itself into the
628 76. Arrange for output from Perl's "warn" command to be written to Exim's main
629 log by default. The user can override this with suitable Perl magic.
631 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
632 the log message when discard triggers, was not working for the DATA ACL or
633 for the non-SMTP ACL.
635 78. Error message wording change in sieve.c.
637 79. If smtp_accept_max_per_host was set, the number of connections could be
638 restricted to fewer than expected, because the daemon was trying to set up
639 a new connection before checking whether the processes handling previous
640 connections had finished. The check for completed processes is now done
641 earlier. On busy systems, this bug wouldn't be noticed because something
642 else would have woken the daemon, and it would have reaped the completed
645 80. If a message was submitted locally by a user whose login name contained one
646 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
647 It caused a spool format error. I have fixed the spool reading code. A
648 related problem was that the "from" clause in the Received: line became
649 illegal because of the space(s). It is now covered by ${quote_local_part.
651 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
654 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
656 83. Patch from Sieve maintainer to fix unterminated string problem in
659 84. Some minor changes to the Linux configuration files to help with other
660 OS variants using glibc.
662 85. One more patch for Sieve to update vacation handling to latest spec.
665 ----------------------------------------------------
666 See the note above about the 4.44 and 4.50 releases.
667 ----------------------------------------------------
673 1. Change 4.43/35 introduced a bug that caused file counts to be
674 incorrectly computed when quota_filecount was set in an appendfile
677 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
678 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
681 3. Give more explanation in the error message when the command for a transport
682 filter fails to execute.
684 4. There are several places where Exim runs a non-Exim command in a
685 subprocess. The SIGUSR1 signal should be disabled for these processes. This
686 was being done only for the command run by the queryprogram router. It is
687 now done for all such subprocesses. The other cases are: ${run, transport
688 filters, and the commands run by the lmtp and pipe transports.
690 5. Some older OS have a limit of 256 on the maximum number of file
691 descriptors. Exim was using setrlimit() to set 1000 as a large value
692 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
693 systems. I've change it so that if it can't get 1000, it tries for 256.
695 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
696 was an oversight, and furthermore, ever since the addition of extra
697 controls (e.g. 4.43/32), the checks on when to allow different forms of
698 "control" were broken. There should now be diagnostics for all cases when a
699 control that does not make sense is encountered.
701 7. $recipients is now available in the predata ACL (oversight).
703 8. Tidy the search cache before the fork to do a delivery from a message
704 received from the command line. Otherwise the child will trigger a lookup
705 failure and thereby defer the delivery if it tries to use (for example) a
706 cached ldap connection that the parent has called unbind on.
708 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
709 of $address_data from the recipient verification was clobbered by the
712 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
713 was its contents. (It was OK if the option was not defined at all.)
715 11. A "Completed" log line is now written for messages that are removed from
716 the spool by the -Mrm option.
718 12. $host_address is now set to the target address during the checking of
721 13. When checking ignore_target_hosts for an ipliteral router, no host name was
722 being passed; this would have caused $sender_host_name to have been used if
723 matching the list had actually called for a host name (not very likely,
724 since this list is usually IP addresses). A host name is now passed as
727 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
728 code that specifies a non-restarting handler (typically sigaction() in
729 modern systems) in an attempt to fix a rare and obscure crash bug.
731 15. Narrowed the window for a race in the daemon that could cause it to ignore
732 SIGCHLD signals. This is not a major problem, because they are used only to
733 wake it up if nothing else does.
735 16. A malformed maildirsize file could cause Exim to calculate negative values
736 for the mailbox size or file count. Odd effects could occur as a result.
737 The maildirsize information is now recalculated if the size or filecount
740 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
741 support for a long time. Removed HAVE_SYS_VFS_H.
743 18. Updated exipick to current release from John Jetmore.
745 19. Allow an empty sender to be matched against a lookup in an address list.
746 Previously the only cases considered were a regular expression, or an
749 20. Exim went into a mad DNS lookup loop when doing a callout where the
750 host was specified on the transport, if the DNS lookup yielded more than
753 21. The RFC2047 encoding function was originally intended for short strings
754 such as real names; it was not keeping to the 75-character limit for
755 encoded words that the RFC imposes. It now respects the limit, and
756 generates multiple encoded words if necessary. To be on the safe side, I
757 have increased the buffer size for the ${rfc2047: expansion operator from
760 22. Failure to deliver a bounce message always caused it to be frozen, even if
761 there was an errors_to setting on the router. The errors_to setting is now
764 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
765 canonical form (fully expanded) before being placed in
766 $sender_host_address.
768 24. Updated eximstats to version 1.33
770 25. Include certificate and key file names in error message when GnuTLS fails
771 to set them up, because the GnuTLS error message doesn't include the name
772 of the failing file when there is a problem reading it.
774 26. Expand error message when OpenSSL has problems setting up cert/key files.
777 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
778 (this can affect the format of dates).
780 28. exim_tidydb, when checking for the continued existence of a message for
781 which it has found a message-specific retry record, was not finding
782 messages that were in split spool directories. Consequently, it was
783 deleting retry records that should have stayed in existence.
785 29. eximstats updated to version 1.35
786 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
787 1.35 - bugfix such that pie charts by volume are generated correctly
789 30. The SPA authentication driver was not abandoning authentication and moving
790 on to the next authenticator when an expansion was forced to fail,
791 contradicting the general specification for all authenticators. Instead it
792 was generating a temporary error. It now behaves as specified.
794 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
795 (the order specifies the preference for clients). The order is now AES256,
796 AES128, 3DES, ARCFOUR128.
798 31. Small patch to Sieve code - explicitly set From: when generating an
801 32. Exim crashed if a remote delivery caused a very long error message to be
802 recorded - for instance if somebody sent an entire SpamAssassin report back
803 as a large number of 550 error lines. This bug was coincidentally fixed by
804 increasing the size of one of Exim's internal buffers (big_buffer) that
805 happened as part of the Exiscan merge. However, to be on the safe side, I
806 have made the code more robust (and fixed the comments that describe what
809 33. Some experimental protocols are using DNS PTR records for new purposes. The
810 keys for these records are domain names, not reversed IP addresses. The
811 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
812 leaves it alone. Component reversal etc. now happens only for IP addresses.
815 34. The host_aton() function is supposed to be passed a string that is known
816 to be a valid IP address. However, in the case of IPv6 addresses, it was
817 not checking this. This is a hostage to fortune. Exim now panics and dies
818 if the condition is not met. A case was found where this could be provoked
819 from a dnsdb PTR lookup with an IPv6 address that had more than 8
820 components; fortuitously, this particular loophole had already been fixed
821 by change 4.50/55 or 4.44/33 above.
823 If there are any other similar loopholes, the new check in host_aton()
824 itself should stop them being exploited. The report I received stated that
825 data on the command line could provoke the exploit when Exim was running as
826 exim, but did not say which command line option was involved. All I could
827 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
831 35. There was a buffer overflow vulnerability in the SPA authentication code
832 (which came originally from the Samba project). I have added a test to the
833 spa_base64_to_bits() function which I hope fixes it.
836 36. The daemon start-up calls getloadavg() while still root for those OS that
837 need the first call to be done as root, but it missed one case: when
838 deliver_queue_load_max is set with deliver_drop_privilege. This is
839 necessary for the benefit of the queue runner, because there is no re-exec
840 when deliver_drop_privilege is set.
842 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
843 was in use, was not putting the data itself into the right store pool;
844 consequently, it could be overwritten for a subsequent message in the same
845 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
848 38. Sometimes the final signoff response after QUIT could fail to get
849 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
850 < 0 before doing a fflush(). This bug looks as though it goes back to the
851 introduction of TLS in release 3.20, but "sometimes" must have been rare
852 because the tests only now provoked it.
858 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
859 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
860 Now it returns an integer. A number of calls were still expecting a T/F
861 return. Fortuitously, in all cases, the tests worked in OK situations,
862 which is the norm. However, things would have gone wrong on any write
863 failures on the smtp file descriptor. This function is used when sending
864 messages over SMTP and also when doing verify callouts.
866 2. When Exim is called to do synchronous delivery of a locally submitted
867 message (the -odf or -odi options), it no longer closes stderr before doing
870 3. Implemented the mua_wrapper option.
872 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
874 5. Implemented the functions header_remove(), header_testname(),
875 header_add_at_position(), and receive_remove_recipient(), and exported them
878 6. If an ACL "warn" statement specified the addition of headers, Exim already
879 inserted X-ACL-Warn: at the start if there was no header name. However, it
880 was not making this test for the second and subsequent header lines if
881 there were newlines in the string. This meant that an invalid header could
882 be inserted if Exim was badly configured.
884 7. Allow an ACL "warn" statement to add header lines at the start or after all
885 the Received: headers, as well as at the end.
887 8. Added the rcpt_4xx retry error code.
889 9. Added postmaster_mailfrom=xxx to callout verification option.
891 10. Added mailfrom=xxxx to the callout verification option, for verify=
894 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
895 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
897 12. Inserted some casts to stop certain compilers warning when using pointer
898 differences as field lengths or precisions in printf-type calls (mostly
899 affecting debugging statements).
901 13. Added optional readline() support for -be (dynamically loaded).
903 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
904 same clock tick as a message's arrival, so that its received time was the
905 same as the "first fail" time on the retry record, and that message
906 remained on the queue past the ultimate address timeout, every queue runner
907 would try a delivery (because it was past the ultimate address timeout) but
908 after another failure, the ultimate address timeout, which should have then
909 bounced the address, did not kick in. This was a "< instead of <=" error;
910 in most cases the first failure would have been in the next clock tick
911 after the received time, and all would be well.
913 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
914 being recognized when the domain list was tested by the match_domain
915 condition in an expansion string.
917 16. Added the ${str2b64: operator.
919 17. Exim was always calling setrlimit() to set a large limit for the number of
920 processes, without checking whether the existing limit was already
921 adequate. (It did check for the limit on file descriptors.) Furthermore,
922 errors from getrlimit() and setrlimit() were being ignored. Now they are
923 logged to the main and panic logs, but Exim does carry on, to try to do its
924 job under whatever limits there are.
926 18. Imported PCRE 5.0.
928 19. Trivial typo in log message " temporarily refused connection" (the leading
931 20. If the log selector return_path_on_delivery was set and an address was
932 redirected to /dev/null, the delivery process crashed because it assumed
933 that a return path would always be set for a "successful" delivery. In this
934 case, the whole delivery is bypassed as an optimization, and therefore no
937 21. Internal re-arrangement: the function for sending a challenge and reading
938 a response while authentication was assuming a zero-terminated challenge
939 string. It's now changed to take a pointer and a length, to allow for
940 binary data in such strings.
942 22. Added the cyrus_sasl authenticator (code supplied by MBM).
944 23. Exim was not respecting finduser_retries when seeking the login of the
945 uid under which it was called; it was always trying 10 times. (The default
946 setting of finduser_retries is zero.) Also, it was sleeping after the final
947 failure, which is pointless.
949 24. Implemented tls_on_connect_ports.
951 25. Implemented acl_smtp_predata.
953 26. If the domain in control=submission is set empty, Exim assumes that the
954 authenticated id is a complete email address when it generates From: or
955 Sender: header lines.
957 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
958 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
959 chown and chgrp in /bin and hostname in /usr/bin.
961 28. Exim was keeping the "process log" file open after each use, just as it
962 does for the main log. This opens the possibility of it remaining open for
963 long periods when the USR1 signal hits a daemon. Occasional processlog
964 errors were reported, that could have been caused by this. Anyway, it seems
965 much more sensible not to leave this file open at all, so that is what now
968 29. The long-running daemon process does not normally write to the log once it
969 has entered its main loop, and it closes the log before doing so. This is
970 so that log files can straightforwardly be renamed and moved. However,
971 there are a couple of unusual error situations where the daemon does write
972 log entries, and I had neglected to close the log afterwards.
974 30. The text of an SMTP error response that was received during a remote
975 delivery was being truncated at 512 bytes. This is too short for some of
976 the long messages that one sometimes sees. I've increased the limit to
979 31. It is now possible to make retry rules that apply only when a message has a
980 specific sender, in particular, an empty sender.
982 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
983 it possible to be selective about when SMTP synchronization is enforced.
985 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
987 32. Implemented hosts_connection_nolog.
989 33. Added an ACL for QUIT.
991 34. Setting "delay_warning=" to disable warnings was not working; it gave a
994 35. Added mailbox_size and mailbox_filecount to appendfile.
996 36. Added control = no_multiline_responses to ACLs.
998 37. There was a bug in the logic of the code that waits for the clock to tick
999 in the case where the clock went backwards by a substantial amount such
1000 that the microsecond fraction of "now" was more than the microsecond
1001 fraction of "then" (but the whole seconds number was less).
1003 38. Added support for the libradius Radius client library this is found on
1004 FreeBSD (previously only the radiusclient library was supported).
1010 1. When certain lookups returned multiple values in the form name=value, the
1011 quoting of the values was not always being done properly. Specifically:
1012 (a) If the value started with a double quote, but contained no whitespace,
1014 (b) If the value contained whitespace other than a space character (i.e.
1015 tabs or newlines or carriage returns) it was not quoted.
1016 This fix has been applied to the mysql and pgsql lookups by writing a
1017 separate quoting function and calling it from the lookup code. The fix
1018 should probably also be applied to nisplus, ibase and oracle lookups, but
1019 since I cannot test any of those, I have not disturbed their existing code.
1021 2. A hit in the callout cache for a specific address caused a log line with no
1022 reason for rejecting RCPT. Now it says "Previous (cached) callout
1023 verification failure".
1025 3. There was an off-by-one bug in the queryprogram router. An over-long
1026 return line was truncated at 256 instead of 255 characters, thereby
1027 overflowing its buffer with the terminating zero. As well as fixing this, I
1028 have increased the buffer size to 1024 (and made a note to document this).
1030 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1031 when Exim is waiting for an SMTP response from a remote server, Exim
1032 restarts its select() call on the socket, thereby resetting its timeout.
1033 This is not a problem when such interrupts are rare. Somebody set up a cron
1034 job to run exiwhat every 2 minutes, which is less than the normal select()
1035 timeout (5 or 10 minutes). This meant that the select() timeout never
1036 kicked in because it was always reset. I have fixed this by comparing the
1037 time when an interrupt arrives with the time at the start of the first call
1038 to select(). If more time than the timeout has elapsed, the interrupt is
1039 treated as a timeout.
1041 5. Some internal re-factoring in preparation for the addition of Sieve
1042 extensions (by MH). In particular, the "personal" test is moved to a
1043 separate function, and given an option for scanning Cc: and Bcc: (which is
1044 not set for Exim filters).
1046 6. When Exim created an email address using the login of the caller as the
1047 local part (e.g. when creating a From: or Sender: header line), it was not
1048 quoting the local part when it contained special characters such as @.
1050 7. Installed new OpenBSD configuration files.
1052 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1053 try to make them clearer.
1055 9. Callout options, other than the timeout value, were being ignored when
1056 verifying sender addresses in header lines. For example, when using
1058 verify = header_sender/callout=no_cache
1060 the cache was (incorrectly) being used.
1062 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1063 only the Cygwin environment.
1065 11. When return_path_on_delivery was set as a log selector, if different remote
1066 addresses in the same message used different return paths and parallel
1067 remote delivery occurred, the wrong values would sometimes be logged.
1068 (Whenever a remote delivery process finished, the return path value from
1069 the most recently started remote delivery process was logged.)
1071 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1072 lines when AUTH and/or TLS are in use. This is the "received protocol"
1073 field. Exim used to use "asmtp" for authenticated SMTP, without any
1074 indication (in the protocol name) for TLS use. Now it follows the RFC and
1075 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1076 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1077 names appear in log lines as well as in Received: header lines.
1079 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1080 extensions, and comparison tests, and to fix some bugs.
1082 14. Changes to the "personal" filter test:
1084 (1) The test was buggy in that it was just doing the equivalent of
1085 "contains" tests on header lines. For example, if a user's address was
1086 anne@some.where, the "personal" test would incorrectly be true for
1088 To: susanne@some.where
1090 This test is now done by extracting each address from the header in turn,
1091 and checking the entire address. Other tests that are part of "personal"
1092 are now done using regular expressions (for example, to check local parts
1093 of addresses in From: header lines).
1095 (2) The list of non-personal local parts in From: addresses has been
1096 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1097 taken from the Sieve specification recommendations.
1099 (3) If the message contains any header line starting with "List-" it is
1100 treated as non-personal.
1102 (4) The test for "circular" in the Subject: header line has been removed
1103 because it now seems ill-conceived.
1105 15. Minor typos in src/EDITME comments corrected.
1107 16. Installed latest exipick from John Jetmore.
1109 17. If headers_add on a router specified a text string that was too long for
1110 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1111 of string_sprintf() is now avoided.
1113 18. $message_body_size was not set (it was always zero) when running the DATA
1114 ACL and the local_scan() function.
1116 19. For the "mail" command in an Exim filter, no default was being set for
1117 the once_repeat time, causing a random time value to be used if "once" was
1118 specified. (If the value happened to be <= 0, no repeat happened.) The
1119 default is now 0s, meaning "never repeat". The "vacation" command was OK
1120 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1121 (I found it when inspecting the code).
1123 20. There is now an overall timeout for performing a callout verification. It
1124 defaults to 4 times the callout timeout, which applies to individual SMTP
1125 commands during the callout. The overall timeout applies when there is more
1126 than one host that can be tried. The timeout is checked before trying the
1127 next host. This prevents very long delays if there are a large number of
1128 hosts and all are timing out (e.g. when the network connections are timing
1129 out). The value of the overall timeout can be changed by specifying an
1130 additional sub-option for "callout", called "maxwait". For example:
1132 verify = sender/callout=5s,maxwait=20s
1134 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1135 to the end before writing, but this should make it even safer).
1137 22. Exim was forgetting that it had advertised PIPELINING for the second and
1138 subsequent messages on an SMTP connection. It was also not resetting its
1139 memory on STARTTLS and an internal HELO.
1141 23. When Exim logs an SMTP synchronization error within a session, it now
1142 records whether PIPELINING has been advertised or not.
1144 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1145 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1146 rather than long int.
1148 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1150 26. Added the never_mail option to autoreply.
1156 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1157 crash if the getsockname() call failed; this can happen if a connection is
1158 closed very soon after it is established. The problem was simply in the
1159 order in which certain operations were done, causing Exim to try to write
1160 to the SMTP stream before it had set up the file descriptor. The bug has
1161 been fixed by making things happen in the correct order.
1167 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1168 before the connection was closed, thus losing the rejection response.
1170 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1171 some early Solaris releases, but causes trouble in current releases where
1172 socklen_t is defined.
1174 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1177 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1180 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1181 that releasing the top of it at the end releases what was used for sub-
1182 expansions (unless the block got too big). However, discard this block if
1183 the first thing is a variable or header, so that we can use its block when
1184 it is dynamic (useful for very large $message_headers, for example).
1186 6. Lookups now cache *every* query, not just the most recent. A new, separate
1187 store pool is used for this. It can be recovered when all lookup caches are
1188 flushed. Lookups now release memory at the end of their result strings.
1189 This has involved some general refactoring of the lookup sources.
1191 7. Some code has been added to the store_xxx() functions to reduce the amount
1192 of flapping under certain conditions.
1194 8. log_incoming_interface used to affect only the <= reception log lines. Now
1195 it causes the local interface and port to be added to several more SMTP log
1196 lines, for example "SMTP connection from", and rejection lines.
1198 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1200 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1202 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1203 could be overwritten at the end of the current message (or the start of a
1204 new message if it was set in a HELO ACL). The value is now preserved for
1205 the duration of the SMTP connection.
1207 12. If a transport had a headers_rewrite setting, and a matching header line
1208 contained an unqualified address, that address was qualified, even if it
1209 did not match any rewriting rules. The underlying bug was that the values
1210 of the flags that permit the existence of unqualified sender and recipient
1211 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1212 non-local messages, and by -bnq for local messages) were not being
1213 preserved with the message after it was received.
1215 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1216 "next input=" as part of the text comprising the host identity instead of
1217 the correct text. The code was using the same buffer for two different
1218 strings. However, depending on which order the printing function evaluated
1219 its arguments, the bug did not always show up. Under Linux, for example, my
1220 test suite worked just fine.
1222 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1223 doesn't work with some older versions of Perl. It has been changed to "my",
1224 which in any case is probably the better facility to use.
1226 15. A really picky compiler found some instances of statements for creating
1227 error messages that either had too many or two few arguments for the format
1230 16. The size of the buffer for calls to the DNS resolver has been increased
1231 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1232 for addresses that have a lot of PTR records. This alleviates a problem; it
1233 does not fully solve it.
1235 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1236 buffer now truncates the list and logs the incident, which is the same
1237 action as happens when Exim is looking up a host name and its aliases.
1238 Previously in this situation something unpredictable would happen;
1239 sometimes it was "internal error: store_reset failed".
1241 18. If a server dropped the connection unexpectedly when an Exim client was
1242 using GnuTLS and trying to read a response, the client delivery process
1243 crashed while trying to generate an error log message.
1245 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1246 string, for example:
1248 warn message = H1: something\nH2: something
1250 the text was added as a single header line from Exim's point of view
1251 though it ended up OK in the delivered message. However, searching for the
1252 second and subsequent header lines using $h_h2: did not work. This has been
1253 fixed. Similarly, if a system filter added multiple headers in this way,
1254 the routers could not see them.
1256 20. Expanded the error message when iplsearch is called with an invalid key to
1257 suggest using net-iplsearch in a host list.
1259 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1260 ACLs are no longer actually imposed (and a message to that effect is
1263 22. If a "gecos" field in a passwd entry contained escaped characters, in
1264 particular, if it contained a \" sequence, Exim got it wrong when building
1265 a From: or a Sender: header from that name. A second bug also caused
1266 incorrect handling when an unquoted " was present following a character
1267 that needed quoting.
1269 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1270 was not being matched caselessly.
1272 24. Arranged for all hyphens in the exim.8 source to be escaped with
1275 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1276 properly. Recipient callout cache records were still being keyed to include
1277 the sender, even when use_sender was set false. This led to far more
1278 callouts that were necessary. The sender is no longer included in the key
1279 when use_sender is false.
1281 26. Added "control = submission" modifier to ACLs.
1283 27. Added the ${base62d: operator to decode base 62 numbers.
1285 28. dnsdb lookups can now access SRV records.
1287 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1288 the configuration file.
1290 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1291 (-v) mode. This makes the output for a verbose queue run more intelligible.
1293 31. Added a use_postmaster feature to recipient callouts.
1295 32. Added the $body_zerocount variable, containing the number of binary zero
1296 bytes in the message body.
1298 33. The time of last modification of the "new" subdirectory is now used as the
1299 "mailbox time last read" when there is a quota error for a maildir
1302 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1304 35. Added +ignore_unknown as a special item in host lists.
1306 36. Code for decoding IPv6 addresses in host lists is now included, even if
1307 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1308 address was recognized as an IP address, but was then not correctly decoded
1309 into binary, causing unexpected and incorrect effects when compared with
1316 1. Very minor rewording of debugging text in manualroute to say "list of
1317 hosts" instead of "hostlist".
1319 2. If verify=header_syntax was set, and a header line with an unqualified
1320 address (no domain) and a large number of spaces between the end of the
1321 name and the colon was received, the reception process suffered a buffer
1322 overflow, and (when I tested it) crashed. This was caused by some obsolete
1323 code that should have been removed. The fix is to remove it!
1325 3. When running in the test harness, delay a bit after writing a bounce
1326 message to get a bit more predictability in the log output.
1328 4. Added a call to search_tidyup() just before forking a reception process. In
1329 theory, someone could use a lookup in the expansion of smtp_accept_max_
1330 per_host which, without the tidyup, could leave open a database connection.
1332 5. Added the variables $recipient_data and $sender_data which get set from a
1333 lookup success in an ACL "recipients" or "senders" condition, or a router
1334 "senders" option, similar to $domain_data and $local_part_data.
1336 6. Moved the writing of debug_print from before to after the "senders" test
1339 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1340 problems for message scanning, either using a data ACL, or using
1341 local_scan() because the Received: header was not generated till after they
1342 were called (in order to set the time as the time of reception completion).
1343 I have revised the way this works. The header is now generated after the
1344 body is received, but before the ACL or local_scan() are called. After they
1345 are run, the timestamp in the header is updated.
1351 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1352 before starting a queue runner without re-exec. This happened only when
1353 deliver_drop_privilege was set or when the Exim user was set to root. The
1354 effect of the bug was that timeouts during subsequent deliveries caused
1355 crashes instead of being properly handled. The handler is now left at its
1356 default (and expected) setting.
1358 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1359 message, again when deliver_drop_privilege is set or Exim is run as root.
1360 The bug described in (1) was not present in this case, but the tidying up
1361 of the other signals was missing. I have made the two cases consistent.
1363 3. The ignore_target_hosts setting on a manualroute router was being ignored
1364 for hosts that were looked up using the /MX notation.
1366 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1369 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1370 operated on the sender address. After changing the $sender_address to <>
1371 for the sender address verify, Exim was re-instated it as the original
1372 (before rewriting) address, but remembering that it had rewritten it, so it
1373 wasn't rewriting it again. This bug also had the effect of breaking the
1374 sender address verification caching when the sender address was rewritten.
1376 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1377 This has been changed so that if the ip literal address matches
1378 ignore_target_hosts, the router declines.
1380 7. Added expansion conditions match_domain, match_address, and match_local_
1381 part (NOT match_host).
1383 8. The placeholder for the Received: header didn't have a length field set.
1385 9. Added code to Exim itself and to exim_lock to test for a specific race
1386 condition that could lead to file corruption when using MBX delivery. The
1387 issue is with the lockfile that is created in /tmp. If this file is removed
1388 after a process has opened it but before that process has acquired a lock,
1389 there is the potential for a second process to recreate the file and also
1390 acquire a lock. This could lead to two Exim processes writing to the file
1391 at the same time. The added code performs the same test as UW imapd; it
1392 checks after acquiring the lock that its file descriptor still refers to
1393 the same named file.
1395 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1396 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1399 11. Added the smtp_active_hostname option. If used, this will typically be made
1400 to depend on the incoming interface address. Because $interface_address is
1401 not set up until the daemon has forked a reception process, error responses
1402 that can happen earlier (such as "too many connections") no longer contain
1405 12. If an expansion in a condition on a "warn" statement fails because a lookup
1406 defers, the "warn" statement is abandoned, and the next ACL statement is
1407 processed. Previously this caused the whole ACL to be aborted.
1409 13. Added the iplsearch lookup type.
1411 14. Added ident_timeout as a log selector.
1413 15. Added tls_certificate_verified as a log selector.
1415 16. Added a global option tls_require_ciphers (compare the smtp transport
1416 option of the same name). This controls incoming TLS connections.
1418 17. I finally figured out how to make tls_require_ciphers do a similar thing
1419 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1420 before starting the TLS session.
1422 18. Tabs are now shown as \t in -bP output.
1424 19. If the log selector return_path_on_delivery was set, Exim crashed when
1425 bouncing a message because it had too many Received: header lines.
1427 20. If two routers both had headers_remove settings, and the first one included
1428 a superfluous trailing colon, the final name in the first list and the
1429 first name in the second list were incorrectly joined into one item (with a
1430 colon in the middle).
1436 1. Added -C and -D options to the exinext utility, mainly to make it easier
1437 to include in the automated testing, but these could be helpful when
1438 multiple configurations are in use.
1440 2. The exinext utility was not formatting the output nicely when there was
1441 an alternate port involved in the retry record key, nor when there was a
1442 message id as well (for retries that were specific to a specific message
1443 and a specific host). It was also confused by IPv6 addresses, because of
1444 the additional colons they contain. I have fixed the IPv4 problem, and
1445 patched it up to do a reasonable job for IPv6.
1447 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1448 delivery, the log line now contains "pipelined" if PIPELINING was used.
1450 4. An SMTP transport process used to panic and die if the bind() call to set
1451 an explicit outgoing interface failed. This has been changed; it is now
1452 treated in the same way as a connect() failure.
1454 5. A reference to $sender_host_name in the part of a conditional expansion
1455 that was being skipped was still causing a DNS lookup. This no longer
1458 6. The def: expansion condition was not recognizing references to header lines
1459 that used bh_ and bheader_.
1461 7. Added the _cache feature to named lists.
1463 8. The code for checking quota_filecount in the appendfile transport was
1464 allowing one more file than it should have been.
1466 9. For compatibility with Sendmail, the command line option
1474 and sets the incoming protocol and host name (for trusted callers). The
1475 host name and its colon can be omitted when only the protocol is to be set.
1476 Note the Exim already has two private options, -pd and -ps, that refer to
1477 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1478 "s", but I don't think that's a major issue.
1480 10. A number of refactoring changes to the code, none of which should affect
1483 (a) The number of logging options was getting close to filling up the
1484 32-bit word that was used as a bit map. I have split them into two classes:
1485 those that are passed in the argument to log_write(), and those that are
1486 only ever tested independently outside of that function. These are now in
1487 separate 32-bit words, so there is plenty of room for expansion again.
1488 There is no change in the user interface or the logging behaviour.
1490 (b) When building, for example, log lines, the code previously used a
1491 macro that called string_cat() twice, in order to add two strings. This is
1492 not really sufficiently general. Furthermore, there was one instance where
1493 it was actually wrong because one of the argument was used twice, and in
1494 one call a function was used. (As it happened, calling the function twice
1495 did not affect the overall behaviour.) The macro has been replaced by a
1496 function that can join an arbitrary number of extra strings onto a growing
1499 (c) The code for expansion conditions now uses a table and a binary chop
1500 instead of a serial search (which was left over from when there were very
1501 few conditions). Also, it now recognizes conditions like "pam" even when
1502 the relevant support is not compiled in: a suitably worded error message is
1503 given if an attempt is made to use such a condition.
1505 11. Added ${time_interval:xxxxx}.
1507 12. A bug was causing one of the ddress fields not to be passed back correctly
1508 from remote delivery subprocesses. The field in question was not being
1509 subsequently used, so this caused to problems in practice.
1511 13. Added new log selectors queue_time and deliver_time.
1513 14. Might have fixed a bug in maildirsizefile handling that threw up
1514 "unexpected character" debug warnings, and recalculated the data
1515 unnecessarily. In any case, I expanded the warning message to give more
1518 15. Added the message "Restricted characters in address" to the statements in
1519 the default ACL that block characters like @ and % in local parts.
1521 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1522 Three changes have been made:
1524 (a) There was a serious bug; a negative response to MAIL caused the whole
1525 recipient domain to be cached as invalid, thereby blocking all messages
1526 to all local parts at the same domain, from all senders. This bug has
1527 been fixed. The domain is no longer cached after a negative response to
1528 MAIL if the sender used is not empty.
1530 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1533 (c) A new callout option, "use_sender" has been added for people who want
1534 the modified behaviour.
1540 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1543 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1544 indeed breaks things for older releases.
1546 3. Added additional logging to the case where there is a problem reading data
1547 from a filter that is running in a subprocess using a pipe, in order to
1548 try to track down a specific problem.
1550 4. Testing facility fudge: when running in the test harness and attempting
1551 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1552 getting "No route to host". Convert this to a timeout.
1554 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1557 6. Some OS don't have socklen_t but use size_t instead. This affects the
1558 fifth argument of getsockopt() amongst other things. This is now
1559 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1560 can be set for individual OS. I have set it for SunOS5, OSF1, and
1561 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1562 some earlier ones do not.
1564 7. Change 4.30/15 was not doing the test caselessly.
1566 8. The standard form for an IPv6 address literal was being rejected by address
1567 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1568 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1569 this, as well as the form without the "IPv6" on the front (but only when
1570 address literals are enabled, of course).
1572 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1574 10. Exim crashed if a message with an empty sender address specified by -f
1575 encountered a router with an errors_to setting. This could be provoked only
1576 by a command such as
1580 where an empty string was supplied; "<>" did not hit this bug.
1582 11. Installed PCRE release 4.5.
1584 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1585 remained set. It is now erased.
1587 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1588 times from message ids (which are base 36 rather than the normal 62).
1590 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1591 were being counted as actual protocol errors, and logged if the log
1592 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1593 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1594 and DATA following a set of rejected RCPTs do not count as protocol errors.
1595 In other words, Exim assumes they were pipelined, though this may not
1596 actually be the case. Of course, in all cases the client gets an
1597 appropriate error code.
1599 15. If a lookup fails in an ACL condition, a message about the failure may
1600 be available; it is used if testing the ACL cannot continue, because most
1601 such messages specify what the cause of the deferral is. However, some
1602 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1603 that caused an old message to be retained and used if a later statement
1604 caused a defer, replacing the real cause of the deferral.
1606 16. If an IP address had so many PTR records that the DNS lookup buffer
1607 was not large enough to hold them, Exim could crash while trying to process
1608 the truncated data. It now detects and logs this case.
1610 17. Further to 4.21/58, another change has been made: if (and only if) the
1611 first line of a message (the first header line) ends with CRLF, a bare LF
1612 in a subsequent header line has a space inserted after it, so as not to
1613 terminate the header.
1615 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1616 unnecessarily, used atoi() instead of strtol(), and didn't check the
1617 termination when getting file sizes from file names by regex.
1619 19. Completely re-implemented the support for maildirsize files, in the light
1620 of a number of problems with the previous contributed implementation
1621 (4.30/29). In particular:
1623 . If the quota is zero, the maildirsize file is maintained, but no quota is
1626 . If the maildir directory does not exist, it is created before any attempt
1627 to write a maildirsize file.
1629 . The quota value in the file is just a cache; if the quota is changed in
1630 the transport, the new value overrides.
1632 . A regular expression is available for excluding directories from the
1635 20. The autoreply transport checks the characters in options that define the
1636 message's headers; it allows continued headers, but it was checking with
1637 isspace() after an embedded newline instead of explicitly looking for a
1640 21. If all the "regular" hosts to which an address was routed had passed their
1641 expiry times, and had not reached their retry times, the address was
1642 bounced, even if fallback hosts were defined. Now Exim should go on to try
1645 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1646 equivalent code in the SMTP transport. Some hosts send humungous responses
1647 to HELO/EHLO, more than 1024 it seems.
1649 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1650 gives compiler warnings in some environments. I've now done it "properly",
1653 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1654 (because of problems with the built-in one) was declared to return uschar *
1655 instead of char *, causing compiler failure.
1657 25. Fixed a file descriptor leak when processing alias/forward files.
1659 26. Fixed a minor format string issue in dbfn.c.
1661 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1663 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1664 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1667 29. When a sender address is verified, it is cached, to save repeating the test
1668 when there is more than one recipient in a message. However, when the
1669 verification involves a callout, it is possible for different callout
1670 options to be set for different recipients. It is too complicated to keep
1671 track of this in the cache, so now Exim always runs a verification when a
1672 callout is required, relying on the callout cache for the optimization.
1673 The overhead is duplication of the address routing, but this should not be
1676 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1677 to be verified with callout=postmaster, and the main callout worked but the
1678 postmaster check failed, the verification correctly failed. However, if a
1679 subsequent RCPT command asked for sender verification *without* the
1680 postmaster check, incorrect caching caused this verification also to fail,
1683 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1684 it was not caching the DNS options (qualify_single, search_parents) that
1685 were used when the lookup failed. A subsequent lookup with different
1686 options therefore always gave the same answer, though there were cases
1687 where it should not have. (Example: a "domains = !$mx_any" option on a
1688 dnslookup router: the "domains" option is always processed without any
1689 widening, but the router might have qualify_single set.) Now Exim uses the
1690 cached value only when the same options are set.
1692 32. Added John Jetmore's "exipick" utility to the distribution.
1694 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1695 than a timeout (e.g. a certificate is required, and is not provided), an
1696 Exim server now closes the connection immediately. Previously it waited for
1697 the client to close - but if the client is SSL, it seems that they each
1698 wait for each other, leading to a delay before one of them times out.
1700 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1701 maintained 0.8.x compatibility because I don't think many are using it, and
1702 it is clearly obsolete.
1704 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1707 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1708 client certificate was expired. A simple patch fixes this, though I don't
1709 understand the full logic of why the verify callback is called multiple
1712 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1713 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1714 which causes problems with some clients (such as the Certicom SSL Plus
1715 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1716 disables the coutermeasure allowing Eudora to connect."
1718 38. Exim was not checking that a write() to a log file succeeded. This could
1719 lead to Bad Things if a log got too big, in particular if it hit a file
1720 size limit. Exim now panics and dies if it cannot write to a log file, just
1721 as it does if it cannot open a log file.
1723 39. Modified OS/Makefile-Linux so that it now contains
1725 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1727 The two -D definitions ensure that Exim is compiled with large file
1728 support, which makes it possible to handle log files that are bigger than
1731 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1732 instance) a domain was checked against a named list that involved a lookup,
1733 causing $domain_data to be set, then another domain was checked against the
1734 same list, then the first domain was re-checked, the value of $domain_data
1735 after the final check could be wrong. In particular, if the second check
1736 failed, it could be set empty. This bug probably also applied to
1739 41. The strip_trailing_dot option was not being applied to the address given
1740 with the -f command-line option.
1742 42. The code for reading a message's header from the spool was incrementing
1743 $received_count, but never initializing it. This meant that the value was
1744 incorrect (doubled) while delivering a message in the same process in which
1745 it was received. In the most common configuration of Exim, this never
1746 happens - a fresh exec is done - but it can happen when
1747 deliver_drop_privilege is set.
1749 43. When Exim logs an SMTP synchronization error - client data sent too soon -
1750 it now includes up to 150 characters of the unexpected data in the log
1753 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1754 and building data strings. The size of both of these buffers was 10 000
1755 bytes - far larger than anybody would *ever* want, thought I. Needless to
1756 say, somebody hit the limit. I have increased the maximum line length to
1757 20 000 and the maximum data length of concatenated lines to 100 000. I have
1758 also fixed two bugs, because there was no checking on these buffers. Tsk,
1759 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1760 buffer is too small.
1762 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1763 lsearch lookups. Now it does.
1765 46. When parsing a route_list item in a manualroute router, a fixed-length
1766 buffer was used for the list of hosts. I made this 1024 bytes long,
1767 thinking that nobody would ever have a list of hosts that long. Wrong.
1768 Somebody had a whole pile of complicated expansion conditions, and the
1769 string was silently truncated, leading to an expansion error. It turns out
1770 that it is easier to change to an unlimited length (owing to other changes
1771 that have happened since this code was originally written) than to build
1772 structure for giving a limitation error. The length of the item that
1773 expands into the list of hosts is now unlimited.
1775 47. The lsearch lookup could not handle data where the length of text line was
1776 more than 4095 characters. Such lines were truncated, leading to shortened
1777 data being returned. It should now handle lines of any length.
1779 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1780 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1783 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
1784 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1785 no longer happens in comment lines. A list of replacements is now placed
1786 at the head of all of the source files, except those whose only change is
1787 to replace PERL_COMMAND in the very first #! line.
1789 50. Replaced the slow insertion sort in queue.c, for sorting the list of
1790 messages on the queue, with a bottom-up merge sort, using code contributed
1791 by Michael Haardt. This should make operations like -bp somewhat faster on
1792 large queues. It won't affect queue runners, except when queue_run_in_order
1795 51. Installed eximstats 1.31 in the distribution.
1797 52. Added support for SRV lookups to the dnslookup router.
1799 53. If an ACL referred to $message_body or $message_body_end, the value was not
1800 reset for any messages that followed in the same SMTP session.
1802 54. The store-handling optimization for building very long strings was not
1803 differentiating between the different store pools. I don't think this
1804 actually made any difference in practice, but I've tidied it.
1806 55. While running the routers to verify a sender address, $sender_address
1807 was still set to the sender address. This is wrong, because when routing to
1808 send a bounce to the sender, it would be empty. Therefore, I have changed
1809 it so that, while verifying a sender address, $sender_address is set to <>.
1810 (There is no change to what happens when verifying a recipient address.)
1812 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1813 target A or AAAA records (if not already returned) without resetting the
1814 qualify_single or search_parents options of the DNS resolver. These are
1815 inappropriate in this case because the targets of MX and SRV records must
1816 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1817 target that, when qualified, matched something in the local domain. These
1818 two options are now turned off when doing these lookups.
1820 57. It seems that at least some releases of Reiserfs (which does not have the
1821 concept of a fixed number of inodes) returns zero and not -1 for the
1822 number of available inodes. This interacted badly with check_spool_inodes,
1823 which assumed that -1 was the "no such thing" setting. What I have done is
1824 to check that the total number of inodes is greater than zero before doing
1825 the test of how many are available.
1827 58. When a "warn" ACL statement has a log_message modifier, the message is
1828 remembered, and not repeated. This is to avoid a lot of repetition when a
1829 message has many recipients that cause the same warning to be written.
1830 Howewer, Exim was preserving the list of already written lines for an
1831 entire SMTP session, which doesn't seem right. The memory is now reset if a
1832 new message is started.
1834 59. The "rewrite" debugging flag was not showing the result of rewriting in the
1835 debugging output unless log_rewrite was also set.
1837 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1838 of (int)(handle) when we know that handle contains (void *)(-1).
1840 61. The Exim daemon panic-logs an error return when it closes the incoming
1841 connection. However "connection reset by peer" seems to be common, and
1842 isn't really an error worthy of noting specially, so that particular error
1845 62. When Exim is trying to find all the local interfaces, it used to panic and
1846 die if the ioctl to get the interface flags failed. However, it seems that
1847 on at least one OS (Solaris 9) it is possible to have an interface that is
1848 included in the list of interfaces, but for which you get a failure error
1849 for this call. This happens when the interface is not "plumbed" into a
1850 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1851 failure of the "get flags" call assumes that the interface is down.
1853 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
1854 makes life easier for people who are doing arithmetic on fields extracted
1855 from dates, where you often get leading zeros that should not be
1856 interpreted as octal.
1858 64. Added qualify_domain to the redirect router, to override the global
1861 65. If a pathologically long header line contained very many addresses (the
1862 report of this problem mentioned 10 000) and each of them was rewritten,
1863 Exim could use up a very large amount of memory. (It kept on making new
1864 copies of the header line as it rewrote, and never released the old ones.)
1865 At the expense of a bit more processing, the header rewriting function has
1866 been changed so that it no longer eats memory in this way.
1868 66. The generation of the Received: header has been moved from the time that a
1869 message starts to be received, to the time that it finishes. The timestamp
1870 in the Received: header should now be very close to that of the <= log
1871 line. There are two side-effects of this change:
1873 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1874 logged header lines no longer include the local Received: line, because
1875 it has not yet been created. The same applies to a copy of the message
1876 that is returned to a non-SMTP sender when a message is rejected.
1878 (b) When a filter file is tested using -bf, no additional Received: header
1879 is added to the test message. After some thought, I decided that this
1882 This change does not affect the value of $received_for. It is still set
1883 after address rewriting, but before local_scan() is called.
1885 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1887 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1888 gave an unhelpful panic error message, and a defer error. I have managed to
1889 change this behaviour so that it now rejects any supplied certificate,
1890 which seems right, as the list of acceptable certificates is empty.
1892 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1893 gave an unhelpful defer error. I have not managed to make this reject any
1894 supplied certificates, but the error message it gives is "no certificate
1895 supplied", which is not helpful.
1897 70. exigrep's output now also includes lines that are not associated with any
1898 message, but which match the given pattern. Implemented by a patch from
1899 Martin Sluka, which also tidied up the Perl a bit.
1901 71. Recipient callout verification, like sender verification, was using <> in
1902 the MAIL FROM command. This isn't really the right thing, since the actual
1903 sender may affect whether the remote host accepts the recipient or not. I
1904 have changed it to use the actual sender in the callout; this means that
1905 the cache record is now keyed on a recipient/sender pair, not just the
1906 recipient address. There doesn't seem to be a real danger of callout loops,
1907 since a callout by the remote host to check the sender would use <>.
1908 [SEE ABOVE: changed after hitting problems.]
1910 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1911 temporary errors. However, in the case of such a code being given after
1912 the end of a data transmission (i.e. after ".") Exim was failing to write
1913 a retry record for the message. (Yes, there was some broken host that was
1914 actually sending 8xx at this point.)
1916 73. An unknown lookup type in a host list could cause Exim to panic-die when
1917 the list was checked. (An example that provoked this was putting <; in the
1918 middle of a list instead of at the start.) If this happened during a DATA
1919 ACL check, a -D file could be left lying around. This kind of configuration
1920 error no longer causes Exim to die; instead it causes a defer errror. The
1921 incident is still logged to the main and panic logs.
1923 74. Buglet left over from Exim 3 conversion. The message "too many messages
1924 in one connection" was written to the rejectlog but not the mainlog, except
1925 when address rewriting (yes!) was being logged.
1927 75. Added write_rejectlog option.
1929 76. When a system filter was run not as root (that is, when system_filter_user
1930 was set), the values of the $n variables were not being returned to the
1931 main process; thus, they were not subsequently available in the $sn
1934 77. Added +return_path_on_delivery log selector.
1936 78. A connection timeout was being treated differently from recipients deferred
1937 when testing hosts_max_try with a message that was older than the host's
1938 retry timeout. (The host should not be counted, thus allowing all hosts to
1939 be tried at least once before bouncing.) This may have been the cause of an
1940 occasionally reported bug whereby a message would remain on the queue
1941 longer than the retry timeout, but would be bounced if a delivery was
1942 forced. I say "may" because I never totally pinned down the problem;
1943 setting up timeout/retry tests is difficult. See also the next item.
1945 79. The ultimate address timeout was not being applied to errors that involved
1946 a combination of host plus message (for example, a timeout on a MAIL
1947 command). When an address resolved to a number of possible hosts, and they
1948 were not all tried for each delivery (e.g. because of hosts_max_try), a
1949 message could remain on the queue longer than the retry timeout.
1951 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1954 81. Fixed an obscure SMTP outgoing bug which required at least the following
1955 conditions: (a) there was another message waiting for the same server;
1956 (b) the server returned 5xx to all RCPT commands in the first message so
1957 that the message was not completed; (c) the server dropped the connection
1958 or gave a negative response to the RSET that Exim sends to abort the
1959 transaction. The observed case was a dropped connection after DATA that had
1960 been sent in pipelining mode. That is, the server had advertised PIPELINING
1961 but was not implementing it correctly. The effect of the bug was incorrect
1962 behaviour, such as trying another host, and this could lead to a crash.
1968 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1969 and daemon.c were passed as pointers to ints; they should have been
1970 pointers to socklen_t variables (which are typically unsigned ints).
1972 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1975 3. Fixed a really odd bug that affected only the testing scheme; patching a
1976 certain fixed string in the binary changed the value of another string that
1977 happened to be identical to the end of the original first string.
1979 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1980 name", it returns that address as the IP address. On some operating
1981 systems (e.g. Solaris), it also passes back the IP address string as the
1982 "host name". However, on others (e.g. Linux), it passes back an empty
1983 string. Exim wasn't checking for this, and was changing the host name to an
1984 empty string, assuming it had been canonicized.
1986 5. Although rare, it is permitted to have more than one PTR record for a given
1987 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1988 all the names associated with an address, because they do in Solaris.
1989 However, it seems that they do not in Linux for data that comes from the
1990 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1991 I found this out when I moved to a new Linux workstation and tried to run
1992 the Exim test suite.
1994 To get round this problem I have changed the code so that it now does its
1995 own call to the DNS to look up PTR records when searching for a host name.
1996 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1997 addresses that are only in /etc/hosts are still found.
1999 This behaviour is, however, controlled by an option called host_lookup_
2000 order, which defaults to "bydns:byaddr". If people want to use the other
2001 order, or indeed, just use one or the other means of lookup, they can
2002 specify it in this variable.
2004 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2005 some operating systems, this comes back from gethostbyaddr() as an empty
2006 string, and this is what Exim used to test for. However, it seems that in
2007 other systems, "." is yielded. Exim now tests for this case too.
2009 7. The values of check_spool_space and check_log_space are now held internally
2010 as a number of kilobytes instead of an absolute number of bytes. If a
2011 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2012 kilobyte. This means that much larger values can be stored.
2014 8. Exim monitor: an attempt to get the action menu when not actually pointing
2015 at a message produces an empty menu entitled "No message selected". This
2016 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2017 no entries in it ("Shell widget menu has zero width and/or height"). So I
2018 have added a single, blank menu entry in this case.
2020 9. Added ${quote_local_part.
2022 10. MIME decoding is now applied to the contents of Subject: header lines when
2025 11. Now that a reference to $sender_host_address automatically causes a reverse
2026 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2027 host lookup before query-style lookups in lists that might use this
2028 variable. This has therefore been abolished, and the "net-" prefix is no
2029 longer necessary for query-style lookups.
2031 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2032 have been a typo for LFLAGS, so it has been changed.
2034 13. The install script calls Exim with "-C /dev/null" in order to find the
2035 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2036 to be output. Howeve, since Exim outputs its version number before the
2037 error, it didn't break the script. It just looked ugly. I fixed this by
2038 always allowing "-C /dev/null" if the caller is root.
2040 14. Ignore overlarge ACL variable number when reading spool file - insurance
2041 against a later release with more variables having written the file.
2043 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2044 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2045 this, as well as the form without the "IPv6" on the front.
2047 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2048 OS/Makefile-Darwin file.
2050 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2051 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2053 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2054 variable. However, some OS use ints for inodes. I've added cast to long int
2055 to get rid of the compiler warning.
2057 19. I had forgotten to lock out "/../" in configuration file names when
2058 ALT_CONFIG_PREFIX was set.
2060 20. Routers used for verification do not need to specify transports. However,
2061 if such a router generated a host list, and callout was configured, Exim
2062 crashed, because it could not find a port number from the (non-existent)
2063 transport. It now assumes port 25 in this circumstance.
2065 21. Added the -t option to exigrep.
2067 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2068 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2069 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2070 them is set, LOOKUP_LSEARCH is forced.
2072 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2074 24. Added sender and host information to the "rejected by local_scan()" log
2075 line; previously there was no indication of these.
2077 25. Added .include_if_exists.
2079 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2080 Linux. It turns out that not all file systems support this. Apparently some
2081 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2082 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2083 not supported on the file descriptor, is now ignored when Exim is trying to
2084 sync a directory. This applies only to Linux.
2086 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2088 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2089 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2092 29. Added support for maildirsize files from supplied patch (modified a bit).
2094 30. The use of :fail: followed by an empty string could lead Exim to respond to
2095 sender verification failures with (e.g.):
2097 550 Verification failed for <xxx>
2098 550 Sender verify failed
2100 where the first response line was missing the '-' that indicates it is not
2101 the final line of the response.
2103 31. The loop for finding the name of the user that called Exim had a hardwired
2104 limit of 10; it now uses the value of finduser_retries, which is used for
2105 all other user lookups.
2107 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2110 33. Exim was neglecting to zero errno before one call of strtol() when
2111 expanding a string and expecting an integer value. On some systems this
2112 resulted in spurious "integer overflow" errors. Also, it was casting the
2113 result into an int without checking.
2115 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2116 did not work. The code looks as if it has *never* worked, though it appears
2117 to have been documented since at least releast 1.62. I have made it work.
2119 35. The "timeout_DNS" error in retry rules, also documented since at least
2120 1.62, also never worked. As it isn't clear exactly what this means, and
2121 clearly it isn't a major issue, I have abolished the feature by treating it
2122 as "timeout", and writing a warning to the main and panic logs.
2124 36. The display of retry rules for -brt wasn't always showing the error code
2127 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2128 timeout_connect_A, timeout_connect_MX.
2130 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2131 to the empty sender.
2133 39. The daemon was not analysing the content of -oX till after it had closed
2134 stderr and disconnected from the controlling terminal. This meant that any
2135 syntax errors were only noted on the panic log, and the return code from
2136 the command was 0. By re-arranging the code a little, I've made the
2137 decoding happen first, so such errors now appear on stderr, and the return
2138 code is 1. However, the actual setting up of the sockets still happens in
2139 the disconnected process, so errors there are still only recorded on the
2142 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2143 connections (as happens on some IP stacks) was logged at start up time as
2144 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2145 it from "IPv6 and IPv4", which means that two separate sockets are being
2148 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2149 says whether AF_INET or AF_INET6 was passed as an argument.
2151 42. Exiwhat output was messed up when time zones were included in log
2154 43. Exiwhat now gives more information about the daemon's listening ports,
2155 and whether -tls-on-connect was used.
2157 44. The "port" option of the smtp transport is now expanded.
2159 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2160 silently ignored. Now an error message is written to the main and panic
2163 46. There's a new ACL modifier called "logwrite" which writes to a log file
2164 as soon as it is encountered.
2166 47. Added $local_user_uid and $local_user_gid at routing time.
2168 48. Exim crashed when trying to verify a sender address that was being
2171 49. Exim was recognizing only a space character after ".include". It now also
2172 recognizes a tab character.
2174 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2175 extracting the relevant information from the specification. The man page no
2176 longer contains scrambled data for the -d option, and I've added a section
2177 at the front about calling Exim under different names.
2179 51. Added "extra_headers" argument to the "mail" command in filter files.
2181 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2184 53. Installed eximstats 1.29.
2186 54. Added transport_filter_timeout as a generic transport option.
2188 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2189 Cc: header lines. This was required by RFC 822, but it not required by RFC
2192 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2193 incoming messages that did not have them. Now it does so only if the
2194 message originates locally, that is, if there is no associated remote host
2195 address. When Resent- header lines are present, this applies to the Resent-
2196 lines rather than the non-Resent- lines.
2198 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2199 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2201 58. Messages for configuration errors now include the name of the main
2202 configuration file - useful now that there may be more than one file in a
2203 list (.included file names were always shown).
2205 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2206 for those rare installations that do not start the daemon as root or run it
2207 setuid root. I've cut out the call to initgroups() if the daemon is not
2210 60. The Exim user and group can now be bound into the binary as text strings
2211 that are looked up at the start of Exim's processing.
2213 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2215 62. Added $mailstore_basename variable.
2217 63. Installed patch to sieve.c from Michael Haardt.
2219 64. When Exim failed to open the panic log after failing to open the main log,
2220 the original message it was trying to log was written to stderr and debug
2221 output, but if they were not available (the usual case in production), it
2222 was lost. Now it is written to syslog before the two lines that record the
2223 failures to open the logs.
2225 65. Users' Exim filters run in subprocesses under the user's uid. It is
2226 possible for a "deliver" command or an alias in a "personal" command to
2227 provoke an address rewrite. If logging of address rewriting is configured,
2228 this fails because the process is not running as root or exim. There may be
2229 a better way of dealing with this, but for the moment (because 4.30 needs
2230 to be released), I have disabled address rewrite logging when running a
2231 filter in a non-root, non-exim process.
2237 1. The buildconfig auxiliary program wasn't quoting the value set for
2238 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2239 not defined. This bug was masked in 4.22 by the effect that was fixed in
2242 2. Some messages that were rejected after a message id was allocated were
2243 shown as "incomplete" by exigrep. It no longer does this for messages that
2244 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2246 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2247 have allow_domain_literals set, the ID did not get logged in the <= line.
2248 Domain literals are now always recognized in Message-ID: header lines.
2250 4. The first argument for a ${extract expansion item is the key name or field
2251 number. Leading and trailing spaces in this item were not being ignored,
2252 causing some misleading effects.
2254 5. When deliver_drop_privilege was set, single queue runner processes started
2255 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2256 same command in the process it spins off) were not dropping privilege.
2258 6. When the daemon running as "exim" started a queue runner, it always
2259 re-executed Exim in the spun-off process. This is a waste of effort when
2260 deliver_drop_privilege is set. The new process now just calls the
2261 queue-runner function directly.
2267 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2270 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2271 ignored. Though the use of -oP was forcing the writing of a pid file, it
2272 was always written to the default place.
2274 3. If the message "no IP address found for host xxxx" is generated during
2275 incoming verification, it is now followed by identification of the incoming
2276 connection (so you can more easily find what provoked it).
2278 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2280 5. Added some features to "harden" Exim a bit more against certain attacks:
2282 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2283 be put in Local/Makefile. This is like the never_users runtime option,
2284 but it cannot be overridden. The default setting is "root".
2286 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2287 prefix string with which any file named in a -C command line option
2290 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2291 is retained for -C and -D only if the caller of Exim is root. Without
2292 it, the exim user may also use -C and -D and retain privilege.
2294 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2295 command line option is disabled.
2297 6. Macro names set by the -D option must start with an upper case letter, just
2298 like macro names defined in the configuration file.
2300 7. Added "dereference=" facility to LDAP.
2302 8. Two instances of the typo "uknown" in the source files are fixed.
2304 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2305 the Configure-Makefile script screwed up while processing it.
2307 10. Incorporated PCRE 4.4.
2309 11. The SMTP synchronization check was not operating right at the start of an
2310 SMTP session. For example, it could not catch a HELO sent before the client
2311 waited for the greeting. There is now a check for outstanding input at the
2312 point when the greeting is written. Because of the duplex, asynchronous
2313 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2314 way, but not yet received, when the check is performed.
2316 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2317 on TCP/IP sockets, because this apparently causes some broken clients to
2320 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2321 unchanged) from the Cygwin maintainer.
2323 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2324 was supported instead of testing for mbx. Effectively a typo.
2326 15. The spa authenticator server code was not checking that the input it
2327 received was valid base64.
2329 16. The debug output line for the "set" modifier in ACLs was not showing the
2330 name of the variable that was being set.
2332 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2334 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2335 lookup on its own. Something else was needed to trigger the lookup. For
2336 example, a match in host_lookup or the need for a host name in a host list.
2337 Now, if $sender_host_name is referenced and the host name has not yet been
2338 looked up, a lookup is performed. If the lookup fails, the variable remains
2339 empty, and $host_lookup_failed is set to "1".
2341 19. Added "eqi" as a case-independent comparison operator.
2343 20. The saslauthd authentication condition could segfault if neither service
2344 nor realm was specified.
2346 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2347 error message that was logged was misleading, and incoming SMTP
2348 connections were dropped. The message is now more accurate, and temporary
2349 errors are given to SMTP connections.
2351 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2352 (except RSET) with a 421 error, until QUIT is received. However, it was
2353 failing to send a response to QUIT.
2355 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2356 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2357 ACL is now run after the helo_try_verify_hosts code.
2359 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2360 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2361 case-independent, so other case variants are also recognized). Apparently
2362 some systems use these upper case variants.
2364 25. If more than two messages were waiting for the same host, and a transport
2365 filter was specified for the transport, Exim sent two messages over the
2366 same TCP/IP connection, and then failed with "socket operation on non-
2367 socket" when it tried to send the third.
2369 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2371 27. The extern definition of crypt16() in expand.c was not being excluded when
2372 the OS had its own crypt16() function.
2374 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2375 as a preferred synonym for return_size_limit, both as an option and as an
2378 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2380 30. Changed the default configuration ACL to relax the local part checking rule
2381 for addresses that are not in any local domains. For these addresses,
2382 slashes and pipe symbols are allowed within local parts, but the sequence
2383 /../ is explicitly forbidden.
2385 31. SPA server authentication was not clearing the challenge buffer before
2388 32. log_message in a "warn" ACL statement was writing to the reject log as
2389 well as to the main log, which contradicts the documentation and doesn't
2390 seem right (because no rejection is happening). So I have stopped it.
2392 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2393 However, I am unable to do any testing of this.
2395 34. Fixed an infelicity in the appendfile transport. When checking directories
2396 for a mailbox, to see if any needed to be created, it was accidentally
2397 using path names with one or more superfluous leading slashes; tracing
2398 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2400 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2401 contents are added to the log line that is written for every discarded
2402 recipient. (Previously a log_message setting was ignored.)
2404 36. The ${quote: operator now quotes the string if it is empty.
2406 37. The install script runs exim in order to find its version number. If for
2407 some reason other than non-existence or emptiness, which it checks, it
2408 could not run './exim', it was installing it with an empty version number,
2409 i.e. as "exim-". This error state is now caught, and the installation is
2412 38. An argument was missing from the function that creates an error message
2413 when Exim fails to connect to the socket for saslauthd authentication.
2414 This could cause Exim to crash, or give a corrupted message.
2416 39. Added isip, isip4, and isip6 to ${if conditions.
2418 40. The ACL variables $acl_xx are now saved with the message, and can be
2419 accessed later in routers, transports, and filters.
2421 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2422 strings in the file are not string-expanded.
2424 42. If a MAIL command specified a SIZE value that was too large to fit into an
2425 int variable, the check against message_size_limit failed. Such values are
2426 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2427 day this will have to be increased, but I don't think I want to be around
2428 when emails are that large.
2435 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2436 iconv() is not standard in FreeBSD.
2438 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2439 IPv6 enabled. The observed symptom was a segmentation fault on return from
2440 the function os_common_find_running_interfaces() in src/os.c.
2442 3. In the check_special_case() function in daemon.c I had used "errno" as an
2443 argument name, which causes warnings on some systems. This was basically a
2444 typo, since it was named "eno" in the comments!
2446 4. The code that waits for the clock to tick (at a resolution of some fraction
2447 of a second) so as to ensure message-id uniqueness was always waiting for
2448 at least one whole tick, when it could have waited for less. [This is
2449 almost certainly not relevant at current processor speeds, where it is
2450 unlikely to ever wait at all. But we try to future-proof.]
2452 5. The function that sleeps for a time interval that includes fractions of a
2453 second contained a race. It did not block SIGALRM between setting the
2454 timer, and suspending (a couple of lines later). If the interval was short
2455 and the sigsuspend() was delayed until after it had expired, the suspension
2456 never ended. On busy systems this could lead to processes getting stuck for
2459 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2460 process, before it forks any delivery processes. The open lookup caching
2461 mechanism meant that the open file or database connection was passed into
2462 the delivery process. The problem was that delivery processes always tidy
2463 up cached lookup data. This could cause a problem for the next delivery
2464 process started by the queue runner, because the external queue runner
2465 process does not know about the closure. So the next delivery process
2466 still has data in the lookup cache. In the case of a file lookup, there was
2467 no problem because closing a file descriptor in a subprocess doesn't affect
2468 the parent. However, if the lookup was caching a connection to a database,
2469 the connection was closed, and the second delivery process was likely to
2470 see errors such as "PGSQL: query failed: server closed the connection
2471 unexpectedly". The problem has been fixed by closing all cached lookups
2472 in a queue runner before running a delivery process.
2474 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2475 seem to have the "const" qualifier which it has on other OS. I've
2478 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2479 *of the same type* (client or server) with the same public name that an
2480 error should be diagnosed.
2482 9. When Exim looked up a host name for an IP address, but failed to find the
2483 original IP address when looking up the host name (a safety check), it
2484 output the message "<ip address> does not match any IP for NULL", which was
2485 confusing, to say the least. The bug was that the host name should have
2486 appeared instead of "NULL".
2488 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2489 user that is built into the binary, and the -C or -D options is used, root
2490 privilege is dropped before the configuration file is read. In addition,
2491 logging is switched to stderr instead of the normal log files. If the
2492 configuration then re-defines the Exim user, the unprivileged environment
2493 is probably not what is expected, so Exim logs a panic warning message (but
2496 However, if deliver_drop_privilege is set, the unprivileged state may well
2497 be exactly what is intended, so the warning has been cut out in that case,
2498 and Exim is allowed to try to write to its normal log files.
2504 1. smtp_return_error_details was not giving details for temporary sender
2505 or receiver verification errors.
2507 2. Diagnose a configuration error if two authenticators have the same public
2510 3. Exim used not to create the message log file for a message until the first
2511 delivery attempt. This could be confusing when incoming messages were held
2512 for policy or load reasons. The message log file is now created at the time
2513 the message is received, and an initial "Received" line is written to it.
2515 4. The automatically generated man page for command line options had a minor
2516 bug that caused no ill effects; however, a more serious problem was that
2517 the procedure for building the man page automatically didn't always
2518 operate. Consequently, release 4.20 contains an out-of-date version. This
2519 shouldn't happen again.
2521 5. When building Exim with embedded Perl support, the script that builds the
2522 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2523 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2525 6. The freeze_tell option was not being used for messages that were frozen on
2526 arrival, either by an ACL or by local_scan().
2528 7. Added the smtp_incomplete_transaction log selector.
2530 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2531 was accepting AUTH without a new EHLO.
2533 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2534 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2537 10. Logging of TCP/IP connections (when configured) now happens in the main
2538 daemon process instead of the child process, so that the TCP/IP connection
2539 count is more accurate (but it can never be perfect).
2541 11. The use of "drop" in a nested ACL was not being handled correctly in the
2542 outer ACL. Now, if condition failure induced by the nested "drop" causes
2543 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2544 or "require"), the connection is dropped.
2546 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2547 that yield "discard" can now be used with an "accept" or a "discard" verb,
2548 but an error is generated for any others (because I can't see a useful way
2549 to define what should happen).
2551 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2552 are now processed in the same way as lines in the Exim configuration file.
2553 In particular, continuation lines are supported.
2555 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2557 15. Added -ti meaning -t -i.
2559 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2560 domains, and warn by logging if others are found.
2562 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2563 returned from GIFCONF when Exim is trying to find the list of interfaces on
2564 a host. The code in os.c has been modified to copy each ifreq to an aligned
2565 structure in all cases.
2567 Also, in some cases, the returned ifreq's were being copied to a 'struct
2568 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2569 means the last couple of bytes of an IPv6 address could be chopped if the
2570 ifreq contained only a normal sockaddr (14 bytes storage).
2572 18. Named domain lists were not supported in the hosts_treat_as_local option.
2573 An entry such as +xxxx was not recognized, and was treated as a literal
2576 19. Ensure that header lines added by a DATA ACL are included in the reject log
2577 if the ACL subsequently rejects the message.
2579 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2580 MD5 (which is deprecated).
2582 21. When testing a filter file using -bf, Exim was writing a message when it
2583 took the sender from a "From " line in the message, but it was not doing so
2584 when it took $return_path from a Return-Path: header line. It now does.
2586 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2587 with a valid header line field name (a series of printing characters
2588 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2590 23. Changed "disc" in the source to "disk" to conform to the documentation and
2591 the book and for uniformity.
2593 24. Ignore Sendmail's -Ooption=value command line item.
2595 25. When execve() failed while trying to run a command in a pipe transport,
2596 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2597 could be confused with a return value of 69 from the command itself. This
2598 has been changed to 127, the value the shell returns if it is asked to run
2599 a non-existent command. The wording for the related log line suggests a
2600 non-existent command as the problem.
2602 26. If received_header_text expands to an empty string, do not add a Received:
2603 header line to the message. (Well, it adds a token one on the spool, but
2604 marks it "old" so that it doesn't get used or transmitted.)
2606 27. Installed eximstats 1.28 (addition of -nt option).
2608 28. There was no check for failure on the call to getsockname() in the daemon
2609 code. This can fail if there is a shortage of resources on the system, with
2610 ENOMEM, for example. A temporary error is now given on failure.
2612 29. Contrary to the C standard, it seems that in some environments, the
2613 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2614 program. Exim now does this explicitly; it affects the formatting of
2615 timestamps using strftime().
2617 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2618 complaints. I've now initialized all the variables, to avoid this.
2620 32. Header lines added by a system filter were not being "seen" during
2621 transport-time rewrites.
2623 33. The info_callback() function passed to OpenSSL is set up with type void
2624 (*)(SSL *, int, int), as described somewhere. However, when calling the
2625 function (actually a macro) that sets it up, the type void(*)() is
2626 expected. I've put in a cast to prevent warnings from picky compilers.
2628 34. If a DNS black list lookup found a CNAME record, but there were no A
2629 records associated with the domain it pointed at, Exim crashed.
2631 35. If a DNS black list lookup returned more than one A record, Exim ignored
2632 all but the first. It now scans all returned addresses if a particular IP
2633 value is being sought. In this situation, the contents of the
2634 $dnslist_value variable are a list of all the addresses, separated by a
2637 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2638 to accept a host name and all its aliases if the forward lookup for any of
2639 them yielded the IP address of the incoming connection. Now it accepts only
2640 those names whose forward lookup yields the correct IP address. Any other
2641 names are discarded. This closes a loophole whereby a rogue DNS
2642 administrator could create reverse DNS records to break through a
2643 wildcarded host restriction in an ACL.
2645 37. If a user filter or a system filter that ran in a subprocess used any of
2646 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2647 the wrong values were passed to the pipe command ($thisaddress had the
2648 value of $0, $0 had the value of $1, etc). This bug was introduced by
2649 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2651 38. Improved the line breaking for long SMTP error messages from ACLs.
2652 Previously, if there was no break point between 40 and 75 characters, Exim
2653 left the rest of the message alone. Two changes have been made: (a) I've
2654 reduced the minimum length to 35 characters; (b) if it can't find a break
2655 point between 35 and 75 characters, it looks ahead and uses the first one
2656 that it finds. This may give the occasional overlong line, but at least the
2657 remaining text gets split now.
2659 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2660 file descriptors might be low, and that setting 1000 would always raise it.
2661 It turns out that in some environments, the limit is already over 1000 and
2662 that lowering it causes trouble. So now Exim takes care not to decrease it.
2664 40. When delivering a message, the value of $return_path is set to $sender_
2665 address at the start of routing (routers may change the value). By an
2666 oversight, this default was not being set up when an address was tested by
2667 -bt or -bv, which affected the outcome if any router or filter referred to
2670 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2671 log only when "message" or "log_message" are set. However, if one of the
2672 conditions was an address verification, or a call to a nested ACL, the
2673 messages generated by the underlying test were being passed through. This
2674 no longer happens. The underlying message is available in $acl_verify_
2675 message for both "message" and "log_message" expansions, so it can be
2676 passed through if needed.
2678 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2679 new expansion $bh_ to give the encoded byte string without charset
2680 translation. Translation happens only if iconv() is available; HAVE_ICONV
2681 indicates this at build time. HEADERS_CHARSET gives the charset to
2682 translate to; headers_charset can change it in the configuration, and
2683 "headers charset" can change it in an individual filter file.
2685 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2686 that creates RFC 2047 encoded "words" labels them as that charset instead
2687 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2688 expansion operator; (ii) when Exim creates a From: line for a local
2689 message; (iii) when a header line is rewritten to include a "phrase" part.
2691 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2692 buggy, causing it to skip the first lines of messages whose message ID
2693 ended in 'D'. This would not have bitten before Exim release 4.14, because
2694 message IDs were unlikely to end in 'D' before then. The effect was to have
2695 incorrect size information for certain domains.
2697 45. #include "config.h" was missing at the start of the crypt16.c module. This
2698 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2701 46. If there was a timeout during a "random" callout check, Exim treated it as
2702 a failure of the random address, and carried on sending RSET and the real
2703 address. If the delay was just some slowness somewhere, the response to the
2704 original RCPT would be taken as a response to RSET and so on, causing
2705 mayhem of various kinds.
2707 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2708 when I implemented it. It didn't allow for the fact that some option values
2709 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2710 the right test for positive values.
2712 48. Domain names in DNS records are case-independent. Exim always looks them up
2713 in lower case. Some resolvers return domain names in exactly the case they
2714 appear in the zone file, that is, they may contain uppercase letters. Not
2715 all resolvers do this - some return always lower case. Exim was treating a
2716 change of case by a resolver as a change of domain, similar to a widening
2717 of a domain abbreviation. This triggered its re-routing code and so it was
2718 trying to route what was effectively the same domain again. This normally
2719 caused routing to fail (because the router wouldn't handle the domain
2720 twice). Now Exim checks for this case specially, and just changes the
2721 casing of the domain that it ultimately uses when it transmits the message
2724 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2727 50. If a filter generated a file delivery with a non-absolute name (possible if
2728 no home directory exists for the router), the forbid_file option was not
2731 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2732 the existing equality matching.
2734 52. Exim was using ints instead of ino_t variables in some places where it was
2735 dealing with inode numbers.
2737 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2738 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2739 called TMPDIR, and if it finds it is different, it changes its value.
2741 54. The smtp_printf() function is now made available to local_scan() so
2742 additional output lines can be written before returning. There is also an
2743 smtp_fflush() function to enable the detection of a dropped connection.
2744 The variables smtp_input and smtp_batched_input are exported to
2747 55. Changed the default runtime configuration: the message "Unknown user"
2748 has been removed from the ACL, and instead placed on the localuser router,
2749 using the cannot_route_message feature. This means that any verification
2750 failures that generate their own messages won't get overridden. Similarly,
2751 the "Unrouteable address" message that was in the ACL for unverifiable
2752 relay addresses has also been removed.
2754 56. Added hosts_avoid_esmtp to the smtp transport.
2756 57. The exicyclog script was not checking for the esoteric option
2757 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2758 will work only if exicyclog is run under the appropriate euid.
2760 58. Following a discussion on the list, the rules by which Exim recognises line
2761 endings on incoming messages have been changed. The -dropcr and drop_cr
2762 options are now no-ops, retained only for backwards compatibility. The
2763 following line terminators are recognized: LF CRLF CR. However, special
2764 processing applies to CR:
2766 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2767 nor a local message in the state where . is a terminator.
2769 (ii) If a bare CR is encountered in a header line, an extra space is added
2770 after the line terminator so as not to end the header. The reasoning
2771 behind this is that bare CRs in header lines are most likely either
2772 to be mistakes, or people trying to play silly games.
2774 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
2775 was being incorrectly given as 18 bytes larger than it should have been.
2776 This is a VOB (very old bug).
2778 60. This may never have affected anything current, but just in case it has:
2779 When the local host is found other than at the start of a list of hosts,
2780 the local host, those with the same MX, and any that follow, are discarded.
2781 When the list in question was part of a longer list of hosts, the following
2782 hosts (not currently being processed) were also being discarded. This no
2783 longer happens. I'm not sure if this situation could ever has previously
2786 61. Added the "/MX" feature to lists of hosts in the manualroute and query
2789 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2790 header. This is something that is recommended in a new Internet Draft, and
2791 is something that is documented as being done by Sendmail. There are two
2792 possible values. For messages generated by the autoreply transport, Exim
2795 Auto-Submitted: auto-replied
2797 whereas for all other generated messages (e.g. bounces) it adds
2799 Auto-Submitted: auto-generated
2801 63. The "personal" condition in filters now includes a test for the
2802 Auto-Submitted: header. If it contains the string "auto-" the message it
2803 not considered personal.
2805 64. Added rcpt_include_affixes as a generic transport option.
2807 65. Added queue_only_override (default true).
2809 66. Added the syslog_duplication option.
2811 67. If what should have been the first header line of a message consisted of
2812 a space followed by a colon, Exim was mis-interpreting it as a header line.
2813 It isn't of course - it is syntactically invalid and should therefore be
2814 treated as the start of the message body. The misbehaviour could have
2815 caused a number of strange effects, including loss of data in subsequent
2816 header lines, and spool format errors.
2818 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2819 client host had authenticated. This control can now be exercised by an ACL
2820 for more flexibility.
2822 69. By default, callouts do not happen when testing with -bh. There is now a
2823 variant, -bhc, which does actually run the callout code, including
2824 consulting and updating the callout cache.
2826 70. Added support for saslauthd authentication, courtesy of Alexander
2829 71. If statvfs() failed on the spool or log directories while checking their
2830 size for availability, Exim confusingly gave the error "space shortage".
2831 Furthermore, in debugging mode it crashed with a floating point exception.
2832 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2833 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2834 serious problem, Exim now writes to the main and panic logs when this
2835 happens, with details of the failure. It then refuses to accept the
2836 incoming message, giving the message "spool directory problem" or "log
2837 directory problem" with a 421 code for SMTP messages.
2839 72. When Exim is about to re-exec itself, it ensures that the file descriptors
2840 0, 1, and 2 exist, because some OS complain for execs without them (see
2841 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2842 descriptors. However, the code omitted to check that the open succeeded,
2843 causing mysterious errors if for some reason the permissions on /dev/null
2844 got screwed. Now Exim writes a message to the main and panic logs, and
2845 bombs out if it can't open /dev/null.
2847 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2848 interact so that it is all more flexible. It is supposed to remain
2849 backwards compatible. Also added extra_local_interfaces.
2851 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2852 to bomb out with an assertion failure - to the client this appears as a
2853 connection drop. This problem occurs in the part of the code that was taken
2854 from the Samba project. Fortunately, the assertion is in a very simple
2855 function, so I have fixed this by reproducing the function inline in the
2856 one place where it is called, and arranging for authentication to fail
2857 instead of killing the process with assert().
2859 75. The SPA client code was not working when the server requested OEM rather
2860 than Unicode encoding.
2862 76. Added code to make require_files with a specific uid setting more usable in
2863 the case where statting the file as root fails - usually a non-root-mounted
2864 NFS file system. When this happens and the failure is EACCES, Exim now
2865 forks a subprocess and does the per-uid checking as the relevant uid.
2867 77. Added process_log_path.
2869 78. If log_file_path was not explicitly set, a setting of check_log_space or
2870 check_log_inodes was ignored.
2872 79. If a space check for the spool or log partitions fails, the incident is now
2873 logged. Of course, in the latter case the data may get lost...
2875 80. Added the %p formatting code to string_format() so that it can be used to
2876 print addresses in debug_print(). Adjusted all the address printing in the
2877 debugging in store.c to use %p rather than %d.
2879 81. There was a concern that a line of code in smtp_in.c could overflow a
2880 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2881 initially expressed, the concern was not well-founded, because trailing
2882 spaces are removed early. However, if the trailing spaces were followed by
2883 a NULL, they did not get removed, so the overflow was possible. Two fixes
2886 (a) I re-wrote the offending code in a cleaner fashion.
2887 (b) If an incoming SMTP command contains a NULL character, it is rejected
2890 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
2891 runs initgroups(), so that if the Exim user is in any additional groups,
2892 they will be used during message reception.
2898 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / jgh / exim.git / blob