Increase incoming SMTP command buffer size from 512 to 2048 (see RFC
[exim.git] / doc / doc-txt / ChangeLog
CommitLineData
8d67ada3 1$Cambridge: exim/doc/doc-txt/ChangeLog,v 1.93 2005/03/22 10:11:42 ph10 Exp $
495ae4b0
PH
2
3Change log file for Exim from version 4.21
4-------------------------------------------
5
bbe902f0 6
7982096b
PH
7Exim version 4.51
8-----------------
9
212d47ec
TK
10TK/01. Added Yahoo DomainKeys support via libdomainkeys. See
11 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
12
ebe2da97
TK
13TK/02. Fix ACL "control" statment not being available in MIME ACL.
14
15TK/03. Fix ACL "regex" condition not being available in MIME ACL.
16
7982096b
PH
17PH/01. Installed a patch from the Sieve maintainer that allows -bf to be used
18 to test Sieve filters that use "vacation".
19
b5aea5e1
PH
20PH/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
26 file.
7982096b 27
7f45268c
PH
28PH/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
32 these two options.
33
7816e254
PH
34PH/04. Change 4.11/85 fixed an obscure bug concerned with addresses that are
35 redirected to themselves ("homonym" addresses). Read the long ChangeLog
36 entry if you want to know the details. The fix, however, neglected to
37 consider the case when local delivery batching is involved. The test for
38 "previously delivered" was not happening when checking to see if an
39 address could be batched with a previous (undelivered) one; under
40 certain circumstances this could lead to multiple deliveries to the same
41 address. A one-line patch to add the appropriate test fixes the bug.
42
36a3b041
PH
43PH/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.
45
596875b3
PH
46PH/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.)
52
8d67ada3
PH
53PH/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.
57
7982096b 58
bbe902f0
PH
59A note about Exim versions 4.44 and 4.50
60----------------------------------------
61
62Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
63changes of various kinds. As a consequence, a big documentation update was
64needed. This delayed the release for rather longer than seemed good, especially
65in the light of a couple of (minor) security issues. Therefore, the changes
66that fixed bugs were backported into 4.43, to create a 4.44 maintenance
67release. So 4.44 and 4.50 are in effect two different branches that both start
68from 4.43.
69
70I have left the 4.50 change log unchanged; it contains all the changes since
714.43. The change log for 4.44 is below; many of its items are identical to
72those for 4.50. This seems to be the most sensible way to preserve the
73historical information.
74
75
f7b63901 76Exim version 4.50
495ae4b0
PH
77-----------------
78
5fe762f6
PH
79 1. Minor wording change to the doc/README.SIEVE file.
80
139059f6 81 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
5fe762f6 82 computation of the current number of files was incorrect.
495ae4b0 83
7086e875
PH
84 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
85 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
86 place.
87
35af9f61
PH
88 4. Give more explanation in the error message when the command for a transport
89 filter fails to execute.
90
b668c215
PH
91 5. There are several places where Exim runs a non-Exim command in a
92 subprocess. The SIGUSR1 signal should be disabled for these processes. This
93 was being done only for the command run by the queryprogram router. It is
94 now done for all such subprocesses. The other cases are: ${run, transport
95 filters, and the commands run by the lmtp and pipe transports.
96
a494b1e1
PH
97 6. Added CONFIGURE_GROUP build-time option.
98
99 7. Some older OS have a limit of 256 on the maximum number of file
100 descriptors. Exim was using setrlimit() to set 1000 as a large value
101 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
102 systems. I've change it so that if it can't get 1000, it tries for 256.
35edf2ff 103
c5fcb476
PH
104 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
105 was an oversight, and furthermore, ever since the addition of extra
106 controls (e.g. 4.43/32), the checks on when to allow different forms of
107 "control" were broken. There should now be diagnostics for all cases when a
108 control that does not make sense is encountered.
109
69358f02
PH
110 9. Added the /retain_sender option to "control=submission".
111
5be20824
PH
11210. $recipients is now available in the predata ACL (oversight).
113
eb2c0248
PH
11411. Tidy the search cache before the fork to do a delivery from a message
115 received from the command line. Otherwise the child will trigger a lookup
116 failure and thereby defer the delivery if it tries to use (for example) a
117 cached ldap connection that the parent has called unbind on.
118
2a3eea10
PH
11912. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
120 of $address_data from the recipient verification was clobbered by the
121 sender verification.
122
12313. The value of address_data from a sender verification is now available in
124 $sender_address_data in subsequent conditions in the ACL statement.
125
23c7ff99
PH
12614. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
127
4deaf07d
PH
12815. Added a new option "connect=<time>" to callout options, to set a different
129 connection timeout.
130
926e1192
PH
13116. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
132 was its contents. (It was OK if the option was not defined at all.)
133
650edc6f
PH
13417. A "Completed" log line is now written for messages that are removed from
135 the spool by the -Mrm option.
136
2c7db3f5
PH
13718. New variables $sender_verify_failure and $recipient_verify_failure contain
138 information about exactly what failed.
139
3d235903
PH
14019. Added -dd to debug only the daemon process.
141
7c7ad977
PH
14220. Incorporated Michael Haardt's patch to ldap.c for improving the way it
143 handles timeouts, both on the server side and network timeouts. Renamed the
144 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
145
981756db
PH
14621. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
147 It is now set to "smtps".
148
d4eb88df
PH
14922. $host_address is now set to the target address during the checking of
150 ignore_target_hosts.
151
15223. When checking ignore_target_hosts for an ipliteral router, no host name was
153 being passed; this would have caused $sender_host_name to have been used if
154 matching the list had actually called for a host name (not very likely,
155 since this list is usually IP addresses). A host name is now passed as
156 "[x.x.x.x]".
157
7d468ab8
PH
15824. Changed the calls that set up the SIGCHLD handler in the daemon to use the
159 code that specifies a non-restarting handler (typically sigaction() in
160 modern systems) in an attempt to fix a rare and obscure crash bug.
161
16225. Narrowed the window for a race in the daemon that could cause it to ignore
163 SIGCHLD signals. This is not a major problem, because they are used only to
164 wake it up if nothing else does.
165
62c0818f
PH
16626. A malformed maildirsize file could cause Exim to calculate negative values
167 for the mailbox size or file count. Odd effects could occur as a result.
168 The maildirsize information is now recalculated if the size or filecount
169 end up negative.
170
26034054
PH
17127. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
172 support for a long time. Removed HAVE_SYS_VFS_H.
173
af66f652
PH
17428. Installed the latest version of exipick from John Jetmore.
175
90af77f4
PH
17629. In an address list, if the pattern was not a regular expression, an empty
177 subject address (from a bounce message) matched only if the pattern was an
178 empty string. Non-empty patterns were not even tested. This was the wrong
179 because it is perfectly reasonable to use an empty address as part of a
180 database query. An empty address is now tested by patterns that are
181 lookups. However, all the other forms of pattern expect the subject to
182 contain a local part and a domain, and therefore, for them, an empty
183 address still always fails if the pattern is not itself empty.
184
d8ef3577
PH
18530. Exim went into a mad DNS loop when attempting to do a callout where the
186 host was specified on an smtp transport, and looking it up yielded more
187 than one IP address.
188
5cb8cbc6
PH
18931. Re-factored the code for checking spool and log partition space into a
190 function that finds that data and another that does the check. The former
191 is then used to implement four new variables: $spool_space, $log_space,
192 $spool_inodes, and $log_inodes.
193
14702f5b
PH
19432. The RFC2047 encoding function was originally intended for short strings
195 such as real names; it was not keeping to the 75-character limit for
196 encoded words that the RFC imposes. It now respects the limit, and
197 generates multiple encoded words if necessary. To be on the safe side, I
198 have increased the buffer size for the ${rfc2047: expansion operator from
199 1024 to 2048 bytes.
200
063b1e99
PH
20133. It is now permitted to omit both strings after an "if" condition; if the
202 condition is true, the result is "true". As before, when the second string
203 is omitted, a false condition yields an empty string. This makes it less
204 cumbersome to write custom ACL and router conditions.
205
652e1b65
PH
20634. Failure to deliver a bounce message always caused it to be frozen, even if
207 there was an errors_to setting on the router. The errors_to setting is now
208 respected.
209
6f0c9a4f
PH
21035. If an IPv6 address is given for -bh or -bhc, it is now converted to the
211 canonical form (fully expanded) before being placed in
212 $sender_host_address.
213
33397d19
PH
21436. The table in the code that translates DNS record types into text (T_A to
215 "A" for instance) was missing entries for NS and CNAME. It is just possible
216 that this could have caused confusion if both these types were looked up
217 for the same domain, because the text type is used as part of Exim's
218 per-process caching. But the chance of anyone hitting this buglet seems
219 very small.
220
7bb56e1f
PH
22137. The dnsdb lookup has been extended in a number of ways.
222
223 (1) There is a new type, "zns", which walks up the domain tree until it
224 finds some nameserver records. It should be used with care.
225
ea3bc19b
PH
226 (2) There is a new type, "mxh", which is like "mx" except that it returns
227 just the host names, not the priorities.
228
229 (3) It is now possible to give a list of domains (or IP addresses) to be
ff4dbb19
PH
230 looked up. The behaviour when one of the lookups defers can be
231 controlled by a keyword.
7bb56e1f 232
ea3bc19b 233 (4) It is now possible to specify the separator character for use when
7bb56e1f 234 multiple records are returned.
33397d19 235
0bcb2a0e
PH
23638. The dnslists ACL condition has been extended: it is now possible to supply
237 a list of IP addresses and/or domains to be looked up in a particular DNS
238 domain.
239
2ac0e484
PH
24039. Added log_selector=+queue_time_overall.
241
4e1fde53
PH
24240. When running the queue in the test harness, wait just a tad after forking a
243 delivery process, to get repeatability of debugging output.
244
de365ded
PH
24541. Include certificate and key file names in error message when GnuTLS fails
246 to set them up, because the GnuTLS error message doesn't include the name
247 of the failing file when there is a problem reading it.
248
f05da2e8
PH
24942. Allow both -bf and -bF in the same test run.
250
d6453af2
PH
25143. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
252
f7b63901
PH
25344. The "Exiscan patch" is now merged into the mainline Exim source.
254
25545. Sometimes the final signoff response after QUIT could fail to get
256 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
257 < 0 before doing a fflush(). This bug looks as though it goes back to the
258 introduction of TLS in release 3.20, but "sometimes" must have been rare
259 because the tests only now provoked it.
260
a444213a
PH
26146. Reset the locale to "C" after calling embedded Perl, in case it was changed
262 (this can affect the format of dates).
263
0ec020ea
PH
26447. exim_tidydb, when checking for the continued existence of a message for
265 which it has found a message-specific retry record, was not finding
266 messages that were in split spool directories. Consequently, it was
267 deleting retry records that should have stayed in existence.
268
b1206957
PH
26948. Steve fixed some bugs in eximstats.
270
27149. The SPA authentication driver was not abandoning authentication and moving
272 on to the next authenticator when an expansion was forced to fail,
273 contradicting the general specification for all authenticators. Instead it
274 was generating a temporary error. It now behaves as specified.
275
26dd5a95
PH
27650. The default ordering of permitted cipher suites for GnuTLS was pessimal
277 (the order specifies the preference for clients). The order is now AES256,
278 AES128, 3DES, ARCFOUR128.
279
343b2385
PH
28051. Small patch to Sieve code - explicitly set From: when generating an
281 autoreply.
282
1c5466b9
PH
28352. Exim crashed if a remote delivery caused a very long error message to be
284 recorded - for instance if somebody sent an entire SpamAssassin report back
285 as a large number of 550 error lines. This bug was coincidentally fixed by
286 increasing the size of one of Exim's internal buffers (big_buffer) that
287 happened as part of the Exiscan merge. However, to be on the safe side, I
288 have made the code more robust (and fixed the comments that describe what
289 is going on).
290
55ee9ee3
PH
29153. Now that there can be additional text after "Completed" in log lines (if
292 the queue_time_overall log selector is set), a one-byte patch to exigrep
293 was needed to allow it to recognize "Completed" as not the last thing in
294 the line.
295
d38f8232
PH
29654. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
297 patch that reportedly fixes this has been added. I am not expert enough to
298 create a test for it. This is what the patch creator wrote:
299
300 "I found a little strange behaviour of ldap code when working with
301 Windows 2003 AD Domain, where users was placed in more than one
302 Organization Units. When I tried to give exim partial DN, the exit code
303 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
304 But simultaneously result of request was absolutely normal ldap result,
305 so I produce this patch..."
306
3295e65b
PH
307 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
308 REFERENCE, so I have modified the code to exclude the patch when that macro
309 is not defined.
310
7102e136
PH
31155. Some experimental protocols are using DNS PTR records for new purposes. The
312 keys for these records are domain names, not reversed IP addresses. The
b975ba52
PH
313 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
314 leaves it alone. Component reversal etc. now happens only for IP addresses.
ea3a6f44 315 CAN-2005-0021
7102e136 316
3ca0ba97
PH
31756. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
318
c2bcbe20
PH
31957. Double the size of the debug message buffer (to 2048) so that more of very
320 long debug lines gets shown.
321
18ce445d
PH
32258. The exicyclog utility now does better if the number of log files to keep
323 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
324
1f5b4c3d
PH
32559. Two changes related to the smtp_active_hostname option:
326
327 (1) $smtp_active_hostname is now available as a variable.
328 (2) The default for smtp_banner uses $smtp_active_hostname instead
329 of $primary_hostname.
330
b975ba52
PH
33160. The host_aton() function is supposed to be passed a string that is known
332 to be a valid IP address. However, in the case of IPv6 addresses, it was
333 not checking this. This is a hostage to fortune. Exim now panics and dies
334 if the condition is not met. A case was found where this could be provoked
85b87bc2
PH
335 from a dnsdb PTR lookup with an IPv6 address that had more than 8
336 components; fortuitously, this particular loophole had already been fixed
337 by change 4.50/55 above.
338
339 If there are any other similar loopholes, the new check in host_aton()
340 itself should stop them being exploited. The report I received stated that
341 data on the command line could provoke the exploit when Exim was running as
342 exim, but did not say which command line option was involved. All I could
343 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
344 running as the user.
ea3a6f44 345 CAN-2005-0021
85b87bc2
PH
346
34761. There was a buffer overflow vulnerability in the SPA authentication code
348 (which came originally from the Samba project). I have added a test to the
349 spa_base64_to_bits() function which I hope fixes it.
ea3a6f44 350 CAN-2005-0022
b975ba52 351
17ffcae7
PH
35262. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
353 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
354
d95f9fdb
PH
35563. The daemon start-up calls getloadavg() while still root for those OS that
356 need the first call to be done as root, but it missed one case: when
357 deliver_queue_load_max is set with deliver_drop_privilege. This is
358 necessary for the benefit of the queue runner, because there is no re-exec
359 when deliver_drop_privilege is set.
360
86b8287f
PH
36164. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
362 This has been fixed.
363
60dc5e56
PH
36465. Caching of lookup data for "hosts =" ACL conditions, when a named host list
365 was in use, was not putting the data itself into the right store pool;
366 consequently, it could be overwritten for a subsequent message in the same
367 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
368 the caching.)
369
533244af
PH
37066. Added hosts_max_try_hardlimit to the smtp transport, default 50.
371
a5a28604
PH
37267. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
373 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
374 the function were treating the return as a boolean value, which happened to
375 work because 0=false and not-0=true, but is not correct code.
376
7e634d24
PH
37768. The host_aton() function was not handling scoped IPv6 addresses (those
378 with, for example, "%eth0" on the end) correctly.
379
3e11c26b
PH
38069. Fixed some compiler warnings in acl.c for the bitmaps specified with
381 negated items (that is, ~something) in unsigned ints. Some compilers
382 apparently mutter when there is no cast.
383
6729cf78
PH
38470. If an address verification called from an ACL failed, and did not produce a
385 user-specific message (i.e. there was only a "system" message), nothing was
386 put in $acl_verify_message. In this situation, it now puts the system
387 message there.
388
00f00ca5
PH
38971. Change 4.23/11 added synchronization checking at the start of an SMTP
390 session; change 4.31/43 added the unwanted input to the log line - except
391 that it did not do this in the start of session case. It now does.
392
c9bdd01c
PH
39372. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
394 This is wrong and can cause the other end to generate a synchronization
395 error if it is another Exim or anything else that does the synchronization
396 check. A QUIT command is no longer sent after a timeout.
397
d43194df
PH
39873. $host_lookup_deferred has been added, to make it easier to detect DEFERs
399 during host lookups.
400
fe5b5d0b
PH
40174. The defer_ok option of callout verification was not working if it was used
402 when verifying addresses in header lines, that is, for this case:
403
404 verify = header_sender/callout=defer_ok
405
76a2d7ba
PH
40675. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
407 those file descriptors could be used for SMTP connections. If anything
408 wrote to stderr (the example that came up was "warn" in embedded Perl), it
409 could be sent to the SMTP client, causing chaos. The daemon now opens
410 stdin, stdout, and stderr to /dev/null when it puts itself into the
411 background.
412
41376. Arrange for output from Perl's "warn" command to be written to Exim's main
414 log by default. The user can override this with suitable Perl magic.
415
04f7d5b9
PH
41677. The use of log_message on a "discard" ACL verb, which is supposed to add to
417 the log message when discard triggers, was not working for the DATA ACL or
418 for the non-SMTP ACL.
419
bc60667e
PH
42078. Error message wording change in sieve.c.
421
bb6e88ff
PH
42279. If smtp_accept_max_per_host was set, the number of connections could be
423 restricted to fewer than expected, because the daemon was trying to set up
424 a new connection before checking whether the processes handling previous
425 connections had finished. The check for completed processes is now done
426 earlier. On busy systems, this bug wouldn't be noticed because something
427 else would have woken the daemon, and it would have reaped the completed
428 process earlier.
429
1e70f85b
PH
43080. If a message was submitted locally by a user whose login name contained one
431 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
432 It caused a spool format error. I have fixed the spool reading code. A
433 related problem was that the "from" clause in the Received: line became
434 illegal because of the space(s). It is now covered by ${quote_local_part.
435
43681. Included the latest eximstats from Steve (adds average sizes to HTML Top
437 tables).
438
4e01f9d6
PH
43982. Updated OS/Makefile-AIX as per message from Mike Meredith.
440
1ee1cef2
PH
44183. Patch from Sieve maintainer to fix unterminated string problem in
442 "vacation" handling.
443
6e2b4ccc
PH
44484. Some minor changes to the Linux configuration files to help with other
445 OS variants using glibc.
446
8e669ac1
PH
44785. One more patch for Sieve to update vacation handling to latest spec.
448
495ae4b0 449
bbe902f0
PH
450----------------------------------------------------
451See the note above about the 4.44 and 4.50 releases.
452----------------------------------------------------
453
454
455Exim version 4.44
456-----------------
457
458 1. Change 4.43/35 introduced a bug that caused file counts to be
459 incorrectly computed when quota_filecount was set in an appendfile
460 transport
461
462 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
463 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
464 place.
465
466 3. Give more explanation in the error message when the command for a transport
467 filter fails to execute.
468
469 4. There are several places where Exim runs a non-Exim command in a
470 subprocess. The SIGUSR1 signal should be disabled for these processes. This
471 was being done only for the command run by the queryprogram router. It is
472 now done for all such subprocesses. The other cases are: ${run, transport
473 filters, and the commands run by the lmtp and pipe transports.
474
475 5. Some older OS have a limit of 256 on the maximum number of file
476 descriptors. Exim was using setrlimit() to set 1000 as a large value
477 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
478 systems. I've change it so that if it can't get 1000, it tries for 256.
479
480 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
481 was an oversight, and furthermore, ever since the addition of extra
482 controls (e.g. 4.43/32), the checks on when to allow different forms of
483 "control" were broken. There should now be diagnostics for all cases when a
484 control that does not make sense is encountered.
485
486 7. $recipients is now available in the predata ACL (oversight).
487
488 8. Tidy the search cache before the fork to do a delivery from a message
489 received from the command line. Otherwise the child will trigger a lookup
490 failure and thereby defer the delivery if it tries to use (for example) a
491 cached ldap connection that the parent has called unbind on.
492
493 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
494 of $address_data from the recipient verification was clobbered by the
495 sender verification.
496
49710. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
498 was its contents. (It was OK if the option was not defined at all.)
499
50011. A "Completed" log line is now written for messages that are removed from
501 the spool by the -Mrm option.
502
50312. $host_address is now set to the target address during the checking of
504 ignore_target_hosts.
505
50613. When checking ignore_target_hosts for an ipliteral router, no host name was
507 being passed; this would have caused $sender_host_name to have been used if
508 matching the list had actually called for a host name (not very likely,
509 since this list is usually IP addresses). A host name is now passed as
510 "[x.x.x.x]".
511
51214. Changed the calls that set up the SIGCHLD handler in the daemon to use the
513 code that specifies a non-restarting handler (typically sigaction() in
514 modern systems) in an attempt to fix a rare and obscure crash bug.
515
51615. Narrowed the window for a race in the daemon that could cause it to ignore
517 SIGCHLD signals. This is not a major problem, because they are used only to
518 wake it up if nothing else does.
519
52016. A malformed maildirsize file could cause Exim to calculate negative values
521 for the mailbox size or file count. Odd effects could occur as a result.
522 The maildirsize information is now recalculated if the size or filecount
523 end up negative.
524
52517. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
526 support for a long time. Removed HAVE_SYS_VFS_H.
527
ea3a6f44 52818. Updated exipick to current release from John Jetmore.
bbe902f0
PH
529
53019. Allow an empty sender to be matched against a lookup in an address list.
531 Previously the only cases considered were a regular expression, or an
532 empty pattern.
533
53420. Exim went into a mad DNS lookup loop when doing a callout where the
535 host was specified on the transport, if the DNS lookup yielded more than
536 one IP address.
537
ea3a6f44
NM
53821. The RFC2047 encoding function was originally intended for short strings
539 such as real names; it was not keeping to the 75-character limit for
540 encoded words that the RFC imposes. It now respects the limit, and
541 generates multiple encoded words if necessary. To be on the safe side, I
542 have increased the buffer size for the ${rfc2047: expansion operator from
543 1024 to 2048 bytes.
bbe902f0 544
ea3a6f44
NM
54522. Failure to deliver a bounce message always caused it to be frozen, even if
546 there was an errors_to setting on the router. The errors_to setting is now
547 respected.
bbe902f0
PH
548
54923. If an IPv6 address is given for -bh or -bhc, it is now converted to the
550 canonical form (fully expanded) before being placed in
551 $sender_host_address.
552
55324. Updated eximstats to version 1.33
554
ea3a6f44
NM
55525. Include certificate and key file names in error message when GnuTLS fails
556 to set them up, because the GnuTLS error message doesn't include the name
557 of the failing file when there is a problem reading it.
bbe902f0
PH
558
55926. Expand error message when OpenSSL has problems setting up cert/key files.
ea3a6f44 560 As per change 25.
bbe902f0 561
ea3a6f44
NM
56227. Reset the locale to "C" after calling embedded Perl, in case it was changed
563 (this can affect the format of dates).
bbe902f0 564
ea3a6f44
NM
56528. exim_tidydb, when checking for the continued existence of a message for
566 which it has found a message-specific retry record, was not finding
567 messages that were in split spool directories. Consequently, it was
568 deleting retry records that should have stayed in existence.
bbe902f0
PH
569
57029. eximstats updated to version 1.35
571 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
572 1.35 - bugfix such that pie charts by volume are generated correctly
573
ea3a6f44
NM
57430. The SPA authentication driver was not abandoning authentication and moving
575 on to the next authenticator when an expansion was forced to fail,
576 contradicting the general specification for all authenticators. Instead it
577 was generating a temporary error. It now behaves as specified.
bbe902f0 578
ea3a6f44
NM
57931. The default ordering of permitted cipher suites for GnuTLS was pessimal
580 (the order specifies the preference for clients). The order is now AES256,
581 AES128, 3DES, ARCFOUR128.
bbe902f0 582
ea3a6f44
NM
58331. Small patch to Sieve code - explicitly set From: when generating an
584 autoreply.
bbe902f0 585
ea3a6f44
NM
58632. Exim crashed if a remote delivery caused a very long error message to be
587 recorded - for instance if somebody sent an entire SpamAssassin report back
588 as a large number of 550 error lines. This bug was coincidentally fixed by
589 increasing the size of one of Exim's internal buffers (big_buffer) that
590 happened as part of the Exiscan merge. However, to be on the safe side, I
591 have made the code more robust (and fixed the comments that describe what
592 is going on).
bbe902f0 593
ea3a6f44
NM
59433. Some experimental protocols are using DNS PTR records for new purposes. The
595 keys for these records are domain names, not reversed IP addresses. The
596 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
597 leaves it alone. Component reversal etc. now happens only for IP addresses.
bbe902f0
PH
598 CAN-2005-0021
599
ea3a6f44
NM
60034. The host_aton() function is supposed to be passed a string that is known
601 to be a valid IP address. However, in the case of IPv6 addresses, it was
602 not checking this. This is a hostage to fortune. Exim now panics and dies
603 if the condition is not met. A case was found where this could be provoked
604 from a dnsdb PTR lookup with an IPv6 address that had more than 8
605 components; fortuitously, this particular loophole had already been fixed
606 by change 4.50/55 or 4.44/33 above.
607
608 If there are any other similar loopholes, the new check in host_aton()
609 itself should stop them being exploited. The report I received stated that
610 data on the command line could provoke the exploit when Exim was running as
611 exim, but did not say which command line option was involved. All I could
612 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
613 running as the user.
bbe902f0
PH
614 CAN-2005-0021
615
ea3a6f44
NM
61635. There was a buffer overflow vulnerability in the SPA authentication code
617 (which came originally from the Samba project). I have added a test to the
618 spa_base64_to_bits() function which I hope fixes it.
bbe902f0
PH
619 CAN-2005-0022
620
ea3a6f44
NM
62136. The daemon start-up calls getloadavg() while still root for those OS that
622 need the first call to be done as root, but it missed one case: when
623 deliver_queue_load_max is set with deliver_drop_privilege. This is
624 necessary for the benefit of the queue runner, because there is no re-exec
625 when deliver_drop_privilege is set.
bbe902f0 626
ea3a6f44
NM
62737. Caching of lookup data for "hosts =" ACL conditions, when a named host list
628 was in use, was not putting the data itself into the right store pool;
629 consequently, it could be overwritten for a subsequent message in the same
630 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
631 the caching.)
bbe902f0 632
ea3a6f44
NM
63338. Sometimes the final signoff response after QUIT could fail to get
634 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
635 < 0 before doing a fflush(). This bug looks as though it goes back to the
636 introduction of TLS in release 3.20, but "sometimes" must have been rare
637 because the tests only now provoked it.
bbe902f0
PH
638
639
495ae4b0
PH
640Exim version 4.43
641-----------------
642
643 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
644 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
645 Now it returns an integer. A number of calls were still expecting a T/F
646 return. Fortuitously, in all cases, the tests worked in OK situations,
647 which is the norm. However, things would have gone wrong on any write
648 failures on the smtp file descriptor. This function is used when sending
649 messages over SMTP and also when doing verify callouts.
650
651 2. When Exim is called to do synchronous delivery of a locally submitted
652 message (the -odf or -odi options), it no longer closes stderr before doing
653 the delivery.
654
655 3. Implemented the mua_wrapper option.
656
657 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
658
659 5. Implemented the functions header_remove(), header_testname(),
660 header_add_at_position(), and receive_remove_recipient(), and exported them
661 to local_scan().
662
663 6. If an ACL "warn" statement specified the addition of headers, Exim already
664 inserted X-ACL-Warn: at the start if there was no header name. However, it
665 was not making this test for the second and subsequent header lines if
666 there were newlines in the string. This meant that an invalid header could
667 be inserted if Exim was badly configured.
668
669 7. Allow an ACL "warn" statement to add header lines at the start or after all
670 the Received: headers, as well as at the end.
671
672 8. Added the rcpt_4xx retry error code.
673
674 9. Added postmaster_mailfrom=xxx to callout verification option.
675
67610. Added mailfrom=xxxx to the callout verification option, for verify=
677 header_sender only.
678
67911. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
680 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
681
68212. Inserted some casts to stop certain compilers warning when using pointer
683 differences as field lengths or precisions in printf-type calls (mostly
684 affecting debugging statements).
685
68613. Added optional readline() support for -be (dynamically loaded).
687
68814. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
689 same clock tick as a message's arrival, so that its received time was the
690 same as the "first fail" time on the retry record, and that message
691 remained on the queue past the ultimate address timeout, every queue runner
692 would try a delivery (because it was past the ultimate address timeout) but
693 after another failure, the ultimate address timeout, which should have then
694 bounced the address, did not kick in. This was a "< instead of <=" error;
695 in most cases the first failure would have been in the next clock tick
696 after the received time, and all would be well.
697
69815. The special items beginning with @ in domain lists (e.g. @mx_any) were not
699 being recognized when the domain list was tested by the match_domain
700 condition in an expansion string.
701
70216. Added the ${str2b64: operator.
703
70417. Exim was always calling setrlimit() to set a large limit for the number of
705 processes, without checking whether the existing limit was already
706 adequate. (It did check for the limit on file descriptors.) Furthermore,
707 errors from getrlimit() and setrlimit() were being ignored. Now they are
708 logged to the main and panic logs, but Exim does carry on, to try to do its
709 job under whatever limits there are.
710
71118. Imported PCRE 5.0.
712
71319. Trivial typo in log message " temporarily refused connection" (the leading
714 space).
715
71620. If the log selector return_path_on_delivery was set and an address was
717 redirected to /dev/null, the delivery process crashed because it assumed
718 that a return path would always be set for a "successful" delivery. In this
719 case, the whole delivery is bypassed as an optimization, and therefore no
720 return path is set.
721
72221. Internal re-arrangement: the function for sending a challenge and reading
723 a response while authentication was assuming a zero-terminated challenge
724 string. It's now changed to take a pointer and a length, to allow for
725 binary data in such strings.
726
72722. Added the cyrus_sasl authenticator (code supplied by MBM).
728
72923. Exim was not respecting finduser_retries when seeking the login of the
730 uid under which it was called; it was always trying 10 times. (The default
731 setting of finduser_retries is zero.) Also, it was sleeping after the final
732 failure, which is pointless.
733
73424. Implemented tls_on_connect_ports.
735
73625. Implemented acl_smtp_predata.
737
73826. If the domain in control=submission is set empty, Exim assumes that the
739 authenticated id is a complete email address when it generates From: or
740 Sender: header lines.
741
74227. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
743 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
744 chown and chgrp in /bin and hostname in /usr/bin.
745
74628. Exim was keeping the "process log" file open after each use, just as it
747 does for the main log. This opens the possibility of it remaining open for
748 long periods when the USR1 signal hits a daemon. Occasional processlog
749 errors were reported, that could have been caused by this. Anyway, it seems
750 much more sensible not to leave this file open at all, so that is what now
751 happens.
752
75329. The long-running daemon process does not normally write to the log once it
754 has entered its main loop, and it closes the log before doing so. This is
755 so that log files can straightforwardly be renamed and moved. However,
756 there are a couple of unusual error situations where the daemon does write
757 log entries, and I had neglected to close the log afterwards.
758
75930. The text of an SMTP error response that was received during a remote
760 delivery was being truncated at 512 bytes. This is too short for some of
761 the long messages that one sometimes sees. I've increased the limit to
762 1024.
763
76431. It is now possible to make retry rules that apply only when a message has a
765 specific sender, in particular, an empty sender.
766
76732. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
768 it possible to be selective about when SMTP synchronization is enforced.
769
77033. Added "control = caseful_local_part" and "control = "caselower_local_part".
771
77232. Implemented hosts_connection_nolog.
773
77433. Added an ACL for QUIT.
775
77634. Setting "delay_warning=" to disable warnings was not working; it gave a
777 syntax error.
778
77935. Added mailbox_size and mailbox_filecount to appendfile.
780
78136. Added control = no_multiline_responses to ACLs.
782
78337. There was a bug in the logic of the code that waits for the clock to tick
784 in the case where the clock went backwards by a substantial amount such
785 that the microsecond fraction of "now" was more than the microsecond
786 fraction of "then" (but the whole seconds number was less).
787
78838. Added support for the libradius Radius client library this is found on
789 FreeBSD (previously only the radiusclient library was supported).
790
791
792Exim version 4.42
793-----------------
794
795 1. When certain lookups returned multiple values in the form name=value, the
796 quoting of the values was not always being done properly. Specifically:
797 (a) If the value started with a double quote, but contained no whitespace,
798 it was not quoted.
799 (b) If the value contained whitespace other than a space character (i.e.
800 tabs or newlines or carriage returns) it was not quoted.
801 This fix has been applied to the mysql and pgsql lookups by writing a
802 separate quoting function and calling it from the lookup code. The fix
803 should probably also be applied to nisplus, ibase and oracle lookups, but
804 since I cannot test any of those, I have not disturbed their existing code.
805
806 2. A hit in the callout cache for a specific address caused a log line with no
807 reason for rejecting RCPT. Now it says "Previous (cached) callout
808 verification failure".
809
810 3. There was an off-by-one bug in the queryprogram router. An over-long
811 return line was truncated at 256 instead of 255 characters, thereby
812 overflowing its buffer with the terminating zero. As well as fixing this, I
813 have increased the buffer size to 1024 (and made a note to document this).
814
815 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
816 when Exim is waiting for an SMTP response from a remote server, Exim
817 restarts its select() call on the socket, thereby resetting its timeout.
818 This is not a problem when such interrupts are rare. Somebody set up a cron
819 job to run exiwhat every 2 minutes, which is less than the normal select()
820 timeout (5 or 10 minutes). This meant that the select() timeout never
821 kicked in because it was always reset. I have fixed this by comparing the
822 time when an interrupt arrives with the time at the start of the first call
823 to select(). If more time than the timeout has elapsed, the interrupt is
824 treated as a timeout.
825
826 5. Some internal re-factoring in preparation for the addition of Sieve
827 extensions (by MH). In particular, the "personal" test is moved to a
828 separate function, and given an option for scanning Cc: and Bcc: (which is
829 not set for Exim filters).
830
831 6. When Exim created an email address using the login of the caller as the
832 local part (e.g. when creating a From: or Sender: header line), it was not
833 quoting the local part when it contained special characters such as @.
834
835 7. Installed new OpenBSD configuration files.
836
837 8. Reworded some messages for syntax errors in "and" and "or" conditions to
838 try to make them clearer.
839
840 9. Callout options, other than the timeout value, were being ignored when
841 verifying sender addresses in header lines. For example, when using
842
843 verify = header_sender/callout=no_cache
844
845 the cache was (incorrectly) being used.
846
84710. Added a missing instance of ${EXE} to the exim_install script; this affects
848 only the Cygwin environment.
849
85011. When return_path_on_delivery was set as a log selector, if different remote
851 addresses in the same message used different return paths and parallel
852 remote delivery occurred, the wrong values would sometimes be logged.
853 (Whenever a remote delivery process finished, the return path value from
854 the most recently started remote delivery process was logged.)
855
85612. RFC 3848 specifies standard names for the "with" phrase in Received: header
857 lines when AUTH and/or TLS are in use. This is the "received protocol"
858 field. Exim used to use "asmtp" for authenticated SMTP, without any
859 indication (in the protocol name) for TLS use. Now it follows the RFC and
860 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
861 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
862 names appear in log lines as well as in Received: header lines.
863
86413. Installed MH's patches for Sieve to add the "copy" and "vacation"
865 extensions, and comparison tests, and to fix some bugs.
866
86714. Changes to the "personal" filter test:
868
869 (1) The test was buggy in that it was just doing the equivalent of
870 "contains" tests on header lines. For example, if a user's address was
871 anne@some.where, the "personal" test would incorrectly be true for
872
873 To: susanne@some.where
874
875 This test is now done by extracting each address from the header in turn,
876 and checking the entire address. Other tests that are part of "personal"
877 are now done using regular expressions (for example, to check local parts
878 of addresses in From: header lines).
879
880 (2) The list of non-personal local parts in From: addresses has been
881 extended to include "listserv", "majordomo", "*-request", and "owner-*",
882 taken from the Sieve specification recommendations.
883
884 (3) If the message contains any header line starting with "List-" it is
885 treated as non-personal.
886
887 (4) The test for "circular" in the Subject: header line has been removed
888 because it now seems ill-conceived.
889
89015. Minor typos in src/EDITME comments corrected.
891
89216. Installed latest exipick from John Jetmore.
893
89417. If headers_add on a router specified a text string that was too long for
895 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
896 of string_sprintf() is now avoided.
897
89818. $message_body_size was not set (it was always zero) when running the DATA
899 ACL and the local_scan() function.
900
90119. For the "mail" command in an Exim filter, no default was being set for
902 the once_repeat time, causing a random time value to be used if "once" was
903 specified. (If the value happened to be <= 0, no repeat happened.) The
904 default is now 0s, meaning "never repeat". The "vacation" command was OK
905 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
906 (I found it when inspecting the code).
907
90820. There is now an overall timeout for performing a callout verification. It
909 defaults to 4 times the callout timeout, which applies to individual SMTP
910 commands during the callout. The overall timeout applies when there is more
911 than one host that can be tried. The timeout is checked before trying the
912 next host. This prevents very long delays if there are a large number of
913 hosts and all are timing out (e.g. when the network connections are timing
914 out). The value of the overall timeout can be changed by specifying an
915 additional sub-option for "callout", called "maxwait". For example:
916
917 verify = sender/callout=5s,maxwait=20s
918
91921. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
920 to the end before writing, but this should make it even safer).
921
92222. Exim was forgetting that it had advertised PIPELINING for the second and
923 subsequent messages on an SMTP connection. It was also not resetting its
924 memory on STARTTLS and an internal HELO.
925
92623. When Exim logs an SMTP synchronization error within a session, it now
927 records whether PIPELINING has been advertised or not.
928
92924. Added 3 instances of "(long int)" casts to time_t variables that were being
930 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
931 rather than long int.
932
93325. Installed the latest Cygwin configuration files from the Cygwin maintainer.
934
93526. Added the never_mail option to autoreply.
936
937
938Exim version 4.41
939-----------------
940
941 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
942 crash if the getsockname() call failed; this can happen if a connection is
943 closed very soon after it is established. The problem was simply in the
944 order in which certain operations were done, causing Exim to try to write
945 to the SMTP stream before it had set up the file descriptor. The bug has
946 been fixed by making things happen in the correct order.
947
948
949Exim version 4.40
950-----------------
951
952 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
953 before the connection was closed, thus losing the rejection response.
954
955 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
956 some early Solaris releases, but causes trouble in current releases where
957 socklen_t is defined.
958
959 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
960 always exist.
961
962 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
963 configured.
964
965 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
966 that releasing the top of it at the end releases what was used for sub-
967 expansions (unless the block got too big). However, discard this block if
968 the first thing is a variable or header, so that we can use its block when
969 it is dynamic (useful for very large $message_headers, for example).
970
971 6. Lookups now cache *every* query, not just the most recent. A new, separate
972 store pool is used for this. It can be recovered when all lookup caches are
973 flushed. Lookups now release memory at the end of their result strings.
974 This has involved some general refactoring of the lookup sources.
975
976 7. Some code has been added to the store_xxx() functions to reduce the amount
977 of flapping under certain conditions.
978
979 8. log_incoming_interface used to affect only the <= reception log lines. Now
980 it causes the local interface and port to be added to several more SMTP log
981 lines, for example "SMTP connection from", and rejection lines.
982
983 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
984
98510. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
986
98711. If $host_data was set by virtue of a hosts lookup in an ACL, its value
988 could be overwritten at the end of the current message (or the start of a
989 new message if it was set in a HELO ACL). The value is now preserved for
990 the duration of the SMTP connection.
991
99212. If a transport had a headers_rewrite setting, and a matching header line
993 contained an unqualified address, that address was qualified, even if it
994 did not match any rewriting rules. The underlying bug was that the values
995 of the flags that permit the existence of unqualified sender and recipient
996 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
997 non-local messages, and by -bnq for local messages) were not being
998 preserved with the message after it was received.
999
100013. When Exim was logging an SMTP synchronization error, it could sometimes log
1001 "next input=" as part of the text comprising the host identity instead of
1002 the correct text. The code was using the same buffer for two different
1003 strings. However, depending on which order the printing function evaluated
1004 its arguments, the bug did not always show up. Under Linux, for example, my
1005 test suite worked just fine.
1006
100714. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1008 doesn't work with some older versions of Perl. It has been changed to "my",
1009 which in any case is probably the better facility to use.
1010
101115. A really picky compiler found some instances of statements for creating
1012 error messages that either had too many or two few arguments for the format
1013 string.
1014
101516. The size of the buffer for calls to the DNS resolver has been increased
1016 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1017 for addresses that have a lot of PTR records. This alleviates a problem; it
1018 does not fully solve it.
1019
102017. A dnsdb lookup for PTR records that receives more data than will fit in the
1021 buffer now truncates the list and logs the incident, which is the same
1022 action as happens when Exim is looking up a host name and its aliases.
1023 Previously in this situation something unpredictable would happen;
1024 sometimes it was "internal error: store_reset failed".
1025
102618. If a server dropped the connection unexpectedly when an Exim client was
1027 using GnuTLS and trying to read a response, the client delivery process
1028 crashed while trying to generate an error log message.
1029
103019. If a "warn" verb in an ACL added multiple headers to a message in a single
1031 string, for example:
1032
1033 warn message = H1: something\nH2: something
1034
1035 the text was added as a single header line from Exim's point of view
1036 though it ended up OK in the delivered message. However, searching for the
1037 second and subsequent header lines using $h_h2: did not work. This has been
1038 fixed. Similarly, if a system filter added multiple headers in this way,
1039 the routers could not see them.
1040
104120. Expanded the error message when iplsearch is called with an invalid key to
1042 suggest using net-iplsearch in a host list.
1043
104421. When running tests using -bh, any delays imposed by "delay" modifiers in
1045 ACLs are no longer actually imposed (and a message to that effect is
1046 output).
1047
104822. If a "gecos" field in a passwd entry contained escaped characters, in
1049 particular, if it contained a \" sequence, Exim got it wrong when building
1050 a From: or a Sender: header from that name. A second bug also caused
1051 incorrect handling when an unquoted " was present following a character
1052 that needed quoting.
1053
105423. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1055 was not being matched caselessly.
1056
105724. Arranged for all hyphens in the exim.8 source to be escaped with
1058 backslashes.
1059
106025. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1061 properly. Recipient callout cache records were still being keyed to include
1062 the sender, even when use_sender was set false. This led to far more
1063 callouts that were necessary. The sender is no longer included in the key
1064 when use_sender is false.
1065
106626. Added "control = submission" modifier to ACLs.
1067
106827. Added the ${base62d: operator to decode base 62 numbers.
1069
107028. dnsdb lookups can now access SRV records.
1071
107229. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1073 the configuration file.
1074
107530. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1076 (-v) mode. This makes the output for a verbose queue run more intelligible.
1077
107831. Added a use_postmaster feature to recipient callouts.
1079
108032. Added the $body_zerocount variable, containing the number of binary zero
1081 bytes in the message body.
1082
108333. The time of last modification of the "new" subdirectory is now used as the
1084 "mailbox time last read" when there is a quota error for a maildir
1085 delivery.
1086
108734. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1088
108935. Added +ignore_unknown as a special item in host lists.
1090
109136. Code for decoding IPv6 addresses in host lists is now included, even if
1092 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1093 address was recognized as an IP address, but was then not correctly decoded
1094 into binary, causing unexpected and incorrect effects when compared with
1095 another IP address.
1096
1097
1098Exim version 4.34
1099-----------------
1100
1101 1. Very minor rewording of debugging text in manualroute to say "list of
1102 hosts" instead of "hostlist".
1103
1104 2. If verify=header_syntax was set, and a header line with an unqualified
1105 address (no domain) and a large number of spaces between the end of the
1106 name and the colon was received, the reception process suffered a buffer
1107 overflow, and (when I tested it) crashed. This was caused by some obsolete
1108 code that should have been removed. The fix is to remove it!
1109
1110 3. When running in the test harness, delay a bit after writing a bounce
1111 message to get a bit more predictability in the log output.
1112
1113 4. Added a call to search_tidyup() just before forking a reception process. In
1114 theory, someone could use a lookup in the expansion of smtp_accept_max_
1115 per_host which, without the tidyup, could leave open a database connection.
1116
1117 5. Added the variables $recipient_data and $sender_data which get set from a
1118 lookup success in an ACL "recipients" or "senders" condition, or a router
1119 "senders" option, similar to $domain_data and $local_part_data.
1120
1121 6. Moved the writing of debug_print from before to after the "senders" test
1122 for routers.
1123
1124 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1125 problems for message scanning, either using a data ACL, or using
1126 local_scan() because the Received: header was not generated till after they
1127 were called (in order to set the time as the time of reception completion).
1128 I have revised the way this works. The header is now generated after the
1129 body is received, but before the ACL or local_scan() are called. After they
1130 are run, the timestamp in the header is updated.
1131
1132
1133Exim version 4.33
1134-----------------
1135
1136 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1137 before starting a queue runner without re-exec. This happened only when
1138 deliver_drop_privilege was set or when the Exim user was set to root. The
1139 effect of the bug was that timeouts during subsequent deliveries caused
1140 crashes instead of being properly handled. The handler is now left at its
1141 default (and expected) setting.
1142
1143 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1144 message, again when deliver_drop_privilege is set or Exim is run as root.
1145 The bug described in (1) was not present in this case, but the tidying up
1146 of the other signals was missing. I have made the two cases consistent.
1147
1148 3. The ignore_target_hosts setting on a manualroute router was being ignored
1149 for hosts that were looked up using the /MX notation.
1150
1151 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1152 in domain lists.
1153
1154 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1155 operated on the sender address. After changing the $sender_address to <>
1156 for the sender address verify, Exim was re-instated it as the original
1157 (before rewriting) address, but remembering that it had rewritten it, so it
1158 wasn't rewriting it again. This bug also had the effect of breaking the
1159 sender address verification caching when the sender address was rewritten.
1160
1161 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1162 This has been changed so that if the ip literal address matches
1163 ignore_target_hosts, the router declines.
1164
1165 7. Added expansion conditions match_domain, match_address, and match_local_
1166 part (NOT match_host).
1167
1168 8. The placeholder for the Received: header didn't have a length field set.
1169
1170 9. Added code to Exim itself and to exim_lock to test for a specific race
1171 condition that could lead to file corruption when using MBX delivery. The
1172 issue is with the lockfile that is created in /tmp. If this file is removed
1173 after a process has opened it but before that process has acquired a lock,
1174 there is the potential for a second process to recreate the file and also
1175 acquire a lock. This could lead to two Exim processes writing to the file
1176 at the same time. The added code performs the same test as UW imapd; it
1177 checks after acquiring the lock that its file descriptor still refers to
1178 the same named file.
1179
118010. The buffer for building added header lines was of fixed size, 8192 bytes.
1181 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1182 when Exim is built.
1183
118411. Added the smtp_active_hostname option. If used, this will typically be made
1185 to depend on the incoming interface address. Because $interface_address is
1186 not set up until the daemon has forked a reception process, error responses
1187 that can happen earlier (such as "too many connections") no longer contain
1188 a host name.
1189
119012. If an expansion in a condition on a "warn" statement fails because a lookup
1191 defers, the "warn" statement is abandoned, and the next ACL statement is
1192 processed. Previously this caused the whole ACL to be aborted.
1193
119413. Added the iplsearch lookup type.
1195
119614. Added ident_timeout as a log selector.
1197
119815. Added tls_certificate_verified as a log selector.
1199
120016. Added a global option tls_require_ciphers (compare the smtp transport
1201 option of the same name). This controls incoming TLS connections.
1202
120317. I finally figured out how to make tls_require_ciphers do a similar thing
1204 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1205 before starting the TLS session.
1206
120718. Tabs are now shown as \t in -bP output.
1208
120919. If the log selector return_path_on_delivery was set, Exim crashed when
1210 bouncing a message because it had too many Received: header lines.
1211
121220. If two routers both had headers_remove settings, and the first one included
1213 a superfluous trailing colon, the final name in the first list and the
1214 first name in the second list were incorrectly joined into one item (with a
1215 colon in the middle).
1216
1217
1218Exim version 4.32
1219-----------------
1220
1221 1. Added -C and -D options to the exinext utility, mainly to make it easier
1222 to include in the automated testing, but these could be helpful when
1223 multiple configurations are in use.
1224
1225 2. The exinext utility was not formatting the output nicely when there was
1226 an alternate port involved in the retry record key, nor when there was a
1227 message id as well (for retries that were specific to a specific message
1228 and a specific host). It was also confused by IPv6 addresses, because of
1229 the additional colons they contain. I have fixed the IPv4 problem, and
1230 patched it up to do a reasonable job for IPv6.
1231
1232 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1233 delivery, the log line now contains "pipelined" if PIPELINING was used.
1234
1235 4. An SMTP transport process used to panic and die if the bind() call to set
1236 an explicit outgoing interface failed. This has been changed; it is now
1237 treated in the same way as a connect() failure.
1238
1239 5. A reference to $sender_host_name in the part of a conditional expansion
1240 that was being skipped was still causing a DNS lookup. This no longer
1241 occurs.
1242
1243 6. The def: expansion condition was not recognizing references to header lines
1244 that used bh_ and bheader_.
1245
1246 7. Added the _cache feature to named lists.
1247
1248 8. The code for checking quota_filecount in the appendfile transport was
1249 allowing one more file than it should have been.
1250
1251 9. For compatibility with Sendmail, the command line option
1252
1253 -prval:sval
1254
1255 is equivalent to
1256
1257 -oMr rval -oMs sval
1258
1259 and sets the incoming protocol and host name (for trusted callers). The
1260 host name and its colon can be omitted when only the protocol is to be set.
1261 Note the Exim already has two private options, -pd and -ps, that refer to
1262 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1263 "s", but I don't think that's a major issue.
1264
126510. A number of refactoring changes to the code, none of which should affect
1266 Exim's behaviour:
1267
1268 (a) The number of logging options was getting close to filling up the
1269 32-bit word that was used as a bit map. I have split them into two classes:
1270 those that are passed in the argument to log_write(), and those that are
1271 only ever tested independently outside of that function. These are now in
1272 separate 32-bit words, so there is plenty of room for expansion again.
1273 There is no change in the user interface or the logging behaviour.
1274
1275 (b) When building, for example, log lines, the code previously used a
1276 macro that called string_cat() twice, in order to add two strings. This is
1277 not really sufficiently general. Furthermore, there was one instance where
1278 it was actually wrong because one of the argument was used twice, and in
1279 one call a function was used. (As it happened, calling the function twice
1280 did not affect the overall behaviour.) The macro has been replaced by a
1281 function that can join an arbitrary number of extra strings onto a growing
1282 string.
1283
1284 (c) The code for expansion conditions now uses a table and a binary chop
1285 instead of a serial search (which was left over from when there were very
1286 few conditions). Also, it now recognizes conditions like "pam" even when
1287 the relevant support is not compiled in: a suitably worded error message is
1288 given if an attempt is made to use such a condition.
1289
129011. Added ${time_interval:xxxxx}.
1291
129212. A bug was causing one of the ddress fields not to be passed back correctly
1293 from remote delivery subprocesses. The field in question was not being
1294 subsequently used, so this caused to problems in practice.
1295
129613. Added new log selectors queue_time and deliver_time.
1297
129814. Might have fixed a bug in maildirsizefile handling that threw up
1299 "unexpected character" debug warnings, and recalculated the data
1300 unnecessarily. In any case, I expanded the warning message to give more
1301 information.
1302
130315. Added the message "Restricted characters in address" to the statements in
1304 the default ACL that block characters like @ and % in local parts.
1305
130616. Change 71 for release 4.31 proved to be much less benign that I imagined.
1307 Three changes have been made:
1308
1309 (a) There was a serious bug; a negative response to MAIL caused the whole
1310 recipient domain to be cached as invalid, thereby blocking all messages
1311 to all local parts at the same domain, from all senders. This bug has
1312 been fixed. The domain is no longer cached after a negative response to
1313 MAIL if the sender used is not empty.
1314
1315 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1316 been restored.
1317
1318 (c) A new callout option, "use_sender" has been added for people who want
1319 the modified behaviour.
1320
1321
1322Exim version 4.31
1323-----------------
1324
1325 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1326 Larry Rosenman.
1327
1328 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1329 indeed breaks things for older releases.
1330
1331 3. Added additional logging to the case where there is a problem reading data
1332 from a filter that is running in a subprocess using a pipe, in order to
1333 try to track down a specific problem.
1334
1335 4. Testing facility fudge: when running in the test harness and attempting
1336 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1337 getting "No route to host". Convert this to a timeout.
1338
1339 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1340 warning.
1341
1342 6. Some OS don't have socklen_t but use size_t instead. This affects the
1343 fifth argument of getsockopt() amongst other things. This is now
1344 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1345 can be set for individual OS. I have set it for SunOS5, OSF1, and
1346 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1347 some earlier ones do not.
1348
1349 7. Change 4.30/15 was not doing the test caselessly.
1350
1351 8. The standard form for an IPv6 address literal was being rejected by address
1352 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1353 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1354 this, as well as the form without the "IPv6" on the front (but only when
1355 address literals are enabled, of course).
1356
1357 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1358
135910. Exim crashed if a message with an empty sender address specified by -f
1360 encountered a router with an errors_to setting. This could be provoked only
1361 by a command such as
1362
1363 exim -f "" ...
1364
1365 where an empty string was supplied; "<>" did not hit this bug.
1366
136711. Installed PCRE release 4.5.
1368
136912. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1370 remained set. It is now erased.
1371
137213. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1373 times from message ids (which are base 36 rather than the normal 62).
1374
137514. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1376 were being counted as actual protocol errors, and logged if the log
1377 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1378 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1379 and DATA following a set of rejected RCPTs do not count as protocol errors.
1380 In other words, Exim assumes they were pipelined, though this may not
1381 actually be the case. Of course, in all cases the client gets an
1382 appropriate error code.
1383
138415. If a lookup fails in an ACL condition, a message about the failure may
1385 be available; it is used if testing the ACL cannot continue, because most
1386 such messages specify what the cause of the deferral is. However, some
1387 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1388 that caused an old message to be retained and used if a later statement
1389 caused a defer, replacing the real cause of the deferral.
1390
139116. If an IP address had so many PTR records that the DNS lookup buffer
1392 was not large enough to hold them, Exim could crash while trying to process
1393 the truncated data. It now detects and logs this case.
1394
139517. Further to 4.21/58, another change has been made: if (and only if) the
1396 first line of a message (the first header line) ends with CRLF, a bare LF
1397 in a subsequent header line has a space inserted after it, so as not to
1398 terminate the header.
1399
140018. Refactoring: tidied an ugly bit of code in appendfile that copied data
1401 unnecessarily, used atoi() instead of strtol(), and didn't check the
1402 termination when getting file sizes from file names by regex.
1403
140419. Completely re-implemented the support for maildirsize files, in the light
1405 of a number of problems with the previous contributed implementation
1406 (4.30/29). In particular:
1407
1408 . If the quota is zero, the maildirsize file is maintained, but no quota is
1409 imposed.
1410
1411 . If the maildir directory does not exist, it is created before any attempt
1412 to write a maildirsize file.
1413
1414 . The quota value in the file is just a cache; if the quota is changed in
1415 the transport, the new value overrides.
1416
1417 . A regular expression is available for excluding directories from the
1418 count.
1419
142020. The autoreply transport checks the characters in options that define the
1421 message's headers; it allows continued headers, but it was checking with
1422 isspace() after an embedded newline instead of explicitly looking for a
1423 space or a tab.
1424
142521. If all the "regular" hosts to which an address was routed had passed their
1426 expiry times, and had not reached their retry times, the address was
1427 bounced, even if fallback hosts were defined. Now Exim should go on to try
1428 the fallback hosts.
1429
143022. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1431 equivalent code in the SMTP transport. Some hosts send humungous responses
1432 to HELO/EHLO, more than 1024 it seems.
1433
143423. Refactoring: code in filter.c used (void *) for "any old type" but this
1435 gives compiler warnings in some environments. I've now done it "properly",
1436 using a union.
1437
143824. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1439 (because of problems with the built-in one) was declared to return uschar *
1440 instead of char *, causing compiler failure.
1441
144225. Fixed a file descriptor leak when processing alias/forward files.
1443
144426. Fixed a minor format string issue in dbfn.c.
1445
144627. Typo in exim.c: ("dmbnz" for "dbmnz").
1447
144828. If a filter file refered to $h_xxx or $message_headers, and the headers
1449 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1450 become corrupted.
1451
145229. When a sender address is verified, it is cached, to save repeating the test
1453 when there is more than one recipient in a message. However, when the
1454 verification involves a callout, it is possible for different callout
1455 options to be set for different recipients. It is too complicated to keep
1456 track of this in the cache, so now Exim always runs a verification when a
1457 callout is required, relying on the callout cache for the optimization.
1458 The overhead is duplication of the address routing, but this should not be
1459 too great.
1460
146130. Fixed a bug in callout caching. If a RCPT command caused the sender address
1462 to be verified with callout=postmaster, and the main callout worked but the
1463 postmaster check failed, the verification correctly failed. However, if a
1464 subsequent RCPT command asked for sender verification *without* the
1465 postmaster check, incorrect caching caused this verification also to fail,
1466 incorrectly.
1467
146831. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1469 it was not caching the DNS options (qualify_single, search_parents) that
1470 were used when the lookup failed. A subsequent lookup with different
1471 options therefore always gave the same answer, though there were cases
1472 where it should not have. (Example: a "domains = !$mx_any" option on a
1473 dnslookup router: the "domains" option is always processed without any
1474 widening, but the router might have qualify_single set.) Now Exim uses the
1475 cached value only when the same options are set.
1476
147732. Added John Jetmore's "exipick" utility to the distribution.
1478
147933. GnuTLS: When an attempt to start a TLS session fails for any reason other
1480 than a timeout (e.g. a certificate is required, and is not provided), an
1481 Exim server now closes the connection immediately. Previously it waited for
1482 the client to close - but if the client is SSL, it seems that they each
1483 wait for each other, leading to a delay before one of them times out.
1484
148534: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1486 maintained 0.8.x compatibility because I don't think many are using it, and
1487 it is clearly obsolete.
1488
148935. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1490 transport.
1491
149236. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1493 client certificate was expired. A simple patch fixes this, though I don't
1494 understand the full logic of why the verify callback is called multiple
1495 times.
1496
149737. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1498 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1499 which causes problems with some clients (such as the Certicom SSL Plus
1500 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1501 disables the coutermeasure allowing Eudora to connect."
1502
150338. Exim was not checking that a write() to a log file succeeded. This could
1504 lead to Bad Things if a log got too big, in particular if it hit a file
1505 size limit. Exim now panics and dies if it cannot write to a log file, just
1506 as it does if it cannot open a log file.
1507
150839. Modified OS/Makefile-Linux so that it now contains
1509
1510 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1511
1512 The two -D definitions ensure that Exim is compiled with large file
1513 support, which makes it possible to handle log files that are bigger than
1514 2^31.
1515
151640. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1517 instance) a domain was checked against a named list that involved a lookup,
1518 causing $domain_data to be set, then another domain was checked against the
1519 same list, then the first domain was re-checked, the value of $domain_data
1520 after the final check could be wrong. In particular, if the second check
1521 failed, it could be set empty. This bug probably also applied to
1522 $localpart_data.
1523
152441. The strip_trailing_dot option was not being applied to the address given
1525 with the -f command-line option.
1526
152742. The code for reading a message's header from the spool was incrementing
1528 $received_count, but never initializing it. This meant that the value was
1529 incorrect (doubled) while delivering a message in the same process in which
1530 it was received. In the most common configuration of Exim, this never
1531 happens - a fresh exec is done - but it can happen when
1532 deliver_drop_privilege is set.
1533
153443. When Exim logs an SMTP synchronization error - client data sent too soon -
1535 it now includes up to 150 characters of the unexpected data in the log
1536 line.
1537
153844. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1539 and building data strings. The size of both of these buffers was 10 000
1540 bytes - far larger than anybody would *ever* want, thought I. Needless to
1541 say, somebody hit the limit. I have increased the maximum line length to
1542 20 000 and the maximum data length of concatenated lines to 100 000. I have
1543 also fixed two bugs, because there was no checking on these buffers. Tsk,
1544 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1545 buffer is too small.
1546
154745. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1548 lsearch lookups. Now it does.
1549
155046. When parsing a route_list item in a manualroute router, a fixed-length
1551 buffer was used for the list of hosts. I made this 1024 bytes long,
1552 thinking that nobody would ever have a list of hosts that long. Wrong.
1553 Somebody had a whole pile of complicated expansion conditions, and the
1554 string was silently truncated, leading to an expansion error. It turns out
1555 that it is easier to change to an unlimited length (owing to other changes
1556 that have happened since this code was originally written) than to build
1557 structure for giving a limitation error. The length of the item that
1558 expands into the list of hosts is now unlimited.
1559
156047. The lsearch lookup could not handle data where the length of text line was
1561 more than 4095 characters. Such lines were truncated, leading to shortened
1562 data being returned. It should now handle lines of any length.
1563
156448. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1565 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1566 ACL").
1567
156849. Cosmetic tidy to scripts like exicyclog that are generated by globally
1569 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1570 no longer happens in comment lines. A list of replacements is now placed
1571 at the head of all of the source files, except those whose only change is
1572 to replace PERL_COMMAND in the very first #! line.
1573
157450. Replaced the slow insertion sort in queue.c, for sorting the list of
1575 messages on the queue, with a bottom-up merge sort, using code contributed
1576 by Michael Haardt. This should make operations like -bp somewhat faster on
1577 large queues. It won't affect queue runners, except when queue_run_in_order
1578 is set.
1579
158051. Installed eximstats 1.31 in the distribution.
1581
158252. Added support for SRV lookups to the dnslookup router.
1583
158453. If an ACL referred to $message_body or $message_body_end, the value was not
1585 reset for any messages that followed in the same SMTP session.
1586
158754. The store-handling optimization for building very long strings was not
1588 differentiating between the different store pools. I don't think this
1589 actually made any difference in practice, but I've tidied it.
1590
159155. While running the routers to verify a sender address, $sender_address
1592 was still set to the sender address. This is wrong, because when routing to
1593 send a bounce to the sender, it would be empty. Therefore, I have changed
1594 it so that, while verifying a sender address, $sender_address is set to <>.
1595 (There is no change to what happens when verifying a recipient address.)
1596
159756. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1598 target A or AAAA records (if not already returned) without resetting the
1599 qualify_single or search_parents options of the DNS resolver. These are
1600 inappropriate in this case because the targets of MX and SRV records must
1601 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1602 target that, when qualified, matched something in the local domain. These
1603 two options are now turned off when doing these lookups.
1604
160557. It seems that at least some releases of Reiserfs (which does not have the
1606 concept of a fixed number of inodes) returns zero and not -1 for the
1607 number of available inodes. This interacted badly with check_spool_inodes,
1608 which assumed that -1 was the "no such thing" setting. What I have done is
1609 to check that the total number of inodes is greater than zero before doing
1610 the test of how many are available.
1611
161258. When a "warn" ACL statement has a log_message modifier, the message is
1613 remembered, and not repeated. This is to avoid a lot of repetition when a
1614 message has many recipients that cause the same warning to be written.
1615 Howewer, Exim was preserving the list of already written lines for an
1616 entire SMTP session, which doesn't seem right. The memory is now reset if a
1617 new message is started.
1618
161959. The "rewrite" debugging flag was not showing the result of rewriting in the
1620 debugging output unless log_rewrite was also set.
1621
162260. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1623 of (int)(handle) when we know that handle contains (void *)(-1).
1624
162561. The Exim daemon panic-logs an error return when it closes the incoming
1626 connection. However "connection reset by peer" seems to be common, and
1627 isn't really an error worthy of noting specially, so that particular error
1628 is no long logged.
1629
163062. When Exim is trying to find all the local interfaces, it used to panic and
1631 die if the ioctl to get the interface flags failed. However, it seems that
1632 on at least one OS (Solaris 9) it is possible to have an interface that is
1633 included in the list of interfaces, but for which you get a failure error
1634 for this call. This happens when the interface is not "plumbed" into a
1635 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1636 failure of the "get flags" call assumes that the interface is down.
1637
163863. Added a ${eval10: operator, which assumes all numbers are decimal. This
1639 makes life easier for people who are doing arithmetic on fields extracted
1640 from dates, where you often get leading zeros that should not be
1641 interpreted as octal.
1642
164364. Added qualify_domain to the redirect router, to override the global
1644 setting.
1645
164665. If a pathologically long header line contained very many addresses (the
1647 report of this problem mentioned 10 000) and each of them was rewritten,
1648 Exim could use up a very large amount of memory. (It kept on making new
1649 copies of the header line as it rewrote, and never released the old ones.)
1650 At the expense of a bit more processing, the header rewriting function has
1651 been changed so that it no longer eats memory in this way.
1652
165366. The generation of the Received: header has been moved from the time that a
1654 message starts to be received, to the time that it finishes. The timestamp
1655 in the Received: header should now be very close to that of the <= log
1656 line. There are two side-effects of this change:
1657
1658 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1659 logged header lines no longer include the local Received: line, because
1660 it has not yet been created. The same applies to a copy of the message
1661 that is returned to a non-SMTP sender when a message is rejected.
1662
1663 (b) When a filter file is tested using -bf, no additional Received: header
1664 is added to the test message. After some thought, I decided that this
1665 is a bug fix.
1666
1667 This change does not affect the value of $received_for. It is still set
1668 after address rewriting, but before local_scan() is called.
1669
167067. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1671
167268. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1673 gave an unhelpful panic error message, and a defer error. I have managed to
1674 change this behaviour so that it now rejects any supplied certificate,
1675 which seems right, as the list of acceptable certificates is empty.
1676
167769. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1678 gave an unhelpful defer error. I have not managed to make this reject any
1679 supplied certificates, but the error message it gives is "no certificate
1680 supplied", which is not helpful.
1681
168270. exigrep's output now also includes lines that are not associated with any
1683 message, but which match the given pattern. Implemented by a patch from
1684 Martin Sluka, which also tidied up the Perl a bit.
1685
168671. Recipient callout verification, like sender verification, was using <> in
1687 the MAIL FROM command. This isn't really the right thing, since the actual
1688 sender may affect whether the remote host accepts the recipient or not. I
1689 have changed it to use the actual sender in the callout; this means that
1690 the cache record is now keyed on a recipient/sender pair, not just the
1691 recipient address. There doesn't seem to be a real danger of callout loops,
1692 since a callout by the remote host to check the sender would use <>.
1693 [SEE ABOVE: changed after hitting problems.]
1694
169572. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1696 temporary errors. However, in the case of such a code being given after
1697 the end of a data transmission (i.e. after ".") Exim was failing to write
1698 a retry record for the message. (Yes, there was some broken host that was
1699 actually sending 8xx at this point.)
1700
170173. An unknown lookup type in a host list could cause Exim to panic-die when
1702 the list was checked. (An example that provoked this was putting <; in the
1703 middle of a list instead of at the start.) If this happened during a DATA
1704 ACL check, a -D file could be left lying around. This kind of configuration
1705 error no longer causes Exim to die; instead it causes a defer errror. The
1706 incident is still logged to the main and panic logs.
1707
170874. Buglet left over from Exim 3 conversion. The message "too many messages
1709 in one connection" was written to the rejectlog but not the mainlog, except
1710 when address rewriting (yes!) was being logged.
1711
171275. Added write_rejectlog option.
1713
171476. When a system filter was run not as root (that is, when system_filter_user
1715 was set), the values of the $n variables were not being returned to the
1716 main process; thus, they were not subsequently available in the $sn
1717 variables.
1718
171977. Added +return_path_on_delivery log selector.
1720
172178. A connection timeout was being treated differently from recipients deferred
1722 when testing hosts_max_try with a message that was older than the host's
1723 retry timeout. (The host should not be counted, thus allowing all hosts to
1724 be tried at least once before bouncing.) This may have been the cause of an
1725 occasionally reported bug whereby a message would remain on the queue
1726 longer than the retry timeout, but would be bounced if a delivery was
1727 forced. I say "may" because I never totally pinned down the problem;
1728 setting up timeout/retry tests is difficult. See also the next item.
1729
173079. The ultimate address timeout was not being applied to errors that involved
1731 a combination of host plus message (for example, a timeout on a MAIL
1732 command). When an address resolved to a number of possible hosts, and they
1733 were not all tried for each delivery (e.g. because of hosts_max_try), a
1734 message could remain on the queue longer than the retry timeout.
1735
173680. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1737 Haardt.
1738
173981. Fixed an obscure SMTP outgoing bug which required at least the following
1740 conditions: (a) there was another message waiting for the same server;
1741 (b) the server returned 5xx to all RCPT commands in the first message so
1742 that the message was not completed; (c) the server dropped the connection
1743 or gave a negative response to the RSET that Exim sends to abort the
1744 transaction. The observed case was a dropped connection after DATA that had
1745 been sent in pipelining mode. That is, the server had advertised PIPELINING
1746 but was not implementing it correctly. The effect of the bug was incorrect
1747 behaviour, such as trying another host, and this could lead to a crash.
1748
1749
1750Exim version 4.30
1751-----------------
1752
1753 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1754 and daemon.c were passed as pointers to ints; they should have been
1755 pointers to socklen_t variables (which are typically unsigned ints).
1756
1757 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1758 fixed.
1759
1760 3. Fixed a really odd bug that affected only the testing scheme; patching a
1761 certain fixed string in the binary changed the value of another string that
1762 happened to be identical to the end of the original first string.
1763
1764 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1765 name", it returns that address as the IP address. On some operating
1766 systems (e.g. Solaris), it also passes back the IP address string as the
1767 "host name". However, on others (e.g. Linux), it passes back an empty
1768 string. Exim wasn't checking for this, and was changing the host name to an
1769 empty string, assuming it had been canonicized.
1770
1771 5. Although rare, it is permitted to have more than one PTR record for a given
1772 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1773 all the names associated with an address, because they do in Solaris.
1774 However, it seems that they do not in Linux for data that comes from the
1775 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1776 I found this out when I moved to a new Linux workstation and tried to run
1777 the Exim test suite.
1778
1779 To get round this problem I have changed the code so that it now does its
1780 own call to the DNS to look up PTR records when searching for a host name.
1781 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1782 addresses that are only in /etc/hosts are still found.
1783
1784 This behaviour is, however, controlled by an option called host_lookup_
1785 order, which defaults to "bydns:byaddr". If people want to use the other
1786 order, or indeed, just use one or the other means of lookup, they can
1787 specify it in this variable.
1788
1789 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1790 some operating systems, this comes back from gethostbyaddr() as an empty
1791 string, and this is what Exim used to test for. However, it seems that in
1792 other systems, "." is yielded. Exim now tests for this case too.
1793
1794 7. The values of check_spool_space and check_log_space are now held internally
1795 as a number of kilobytes instead of an absolute number of bytes. If a
1796 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1797 kilobyte. This means that much larger values can be stored.
1798
1799 8. Exim monitor: an attempt to get the action menu when not actually pointing
1800 at a message produces an empty menu entitled "No message selected". This
1801 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1802 no entries in it ("Shell widget menu has zero width and/or height"). So I
1803 have added a single, blank menu entry in this case.
1804
1805 9. Added ${quote_local_part.
1806
180710. MIME decoding is now applied to the contents of Subject: header lines when
1808 they are logged.
1809
181011. Now that a reference to $sender_host_address automatically causes a reverse
1811 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1812 host lookup before query-style lookups in lists that might use this
1813 variable. This has therefore been abolished, and the "net-" prefix is no
1814 longer necessary for query-style lookups.
1815
181612. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1817 have been a typo for LFLAGS, so it has been changed.
1818
181913. The install script calls Exim with "-C /dev/null" in order to find the
1820 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1821 to be output. Howeve, since Exim outputs its version number before the
1822 error, it didn't break the script. It just looked ugly. I fixed this by
1823 always allowing "-C /dev/null" if the caller is root.
1824
182514. Ignore overlarge ACL variable number when reading spool file - insurance
1826 against a later release with more variables having written the file.
1827
182815. The standard form for an IPv6 address literal was being rejected by EHLO.
1829 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1830 this, as well as the form without the "IPv6" on the front.
1831
183216. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1833 OS/Makefile-Darwin file.
1834
183517. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1836 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1837
183818. After change 4.21/52, "%ld" was used to format the contents of the $inode
1839 variable. However, some OS use ints for inodes. I've added cast to long int
1840 to get rid of the compiler warning.
1841
184219. I had forgotten to lock out "/../" in configuration file names when
1843 ALT_CONFIG_PREFIX was set.
1844
184520. Routers used for verification do not need to specify transports. However,
1846 if such a router generated a host list, and callout was configured, Exim
1847 crashed, because it could not find a port number from the (non-existent)
1848 transport. It now assumes port 25 in this circumstance.
1849
185021. Added the -t option to exigrep.
1851
185222. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1853 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1854 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1855 them is set, LOOKUP_LSEARCH is forced.
1856
185723. "exim -bV" now outputs a list of lookups that are included in the binary.
1858
185924. Added sender and host information to the "rejected by local_scan()" log
1860 line; previously there was no indication of these.
1861
186225. Added .include_if_exists.
1863
186426. Change 3.952/11 added an explicit directory sync on top of a file sync for
1865 Linux. It turns out that not all file systems support this. Apparently some
1866 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1867 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1868 not supported on the file descriptor, is now ignored when Exim is trying to
1869 sync a directory. This applies only to Linux.
1870
187127. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1872
187328. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1874 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1875 this.
1876
187729. Added support for maildirsize files from supplied patch (modified a bit).
1878
187930. The use of :fail: followed by an empty string could lead Exim to respond to
1880 sender verification failures with (e.g.):
1881
1882 550 Verification failed for <xxx>
1883 550 Sender verify failed
1884
1885 where the first response line was missing the '-' that indicates it is not
1886 the final line of the response.
1887
188831. The loop for finding the name of the user that called Exim had a hardwired
1889 limit of 10; it now uses the value of finduser_retries, which is used for
1890 all other user lookups.
1891
189232. Added $received_count variable, available in data and not_smtp ACLs, and at
1893 delivery time.
1894
189533. Exim was neglecting to zero errno before one call of strtol() when
1896 expanding a string and expecting an integer value. On some systems this
1897 resulted in spurious "integer overflow" errors. Also, it was casting the
1898 result into an int without checking.
1899
190034. Testing for a connection timeout using "timeout_connect" in the retry rules
1901 did not work. The code looks as if it has *never* worked, though it appears
1902 to have been documented since at least releast 1.62. I have made it work.
1903
190435. The "timeout_DNS" error in retry rules, also documented since at least
1905 1.62, also never worked. As it isn't clear exactly what this means, and
1906 clearly it isn't a major issue, I have abolished the feature by treating it
1907 as "timeout", and writing a warning to the main and panic logs.
1908
190936. The display of retry rules for -brt wasn't always showing the error code
1910 correctly.
1911
191237. Added new error conditions to retry rules: timeout_A, timeout_MX,
1913 timeout_connect_A, timeout_connect_MX.
1914
191538. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
1916 to the empty sender.
1917
191839. The daemon was not analysing the content of -oX till after it had closed
1919 stderr and disconnected from the controlling terminal. This meant that any
1920 syntax errors were only noted on the panic log, and the return code from
1921 the command was 0. By re-arranging the code a little, I've made the
1922 decoding happen first, so such errors now appear on stderr, and the return
1923 code is 1. However, the actual setting up of the sockets still happens in
1924 the disconnected process, so errors there are still only recorded on the
1925 panic log.
1926
192740. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
1928 connections (as happens on some IP stacks) was logged at start up time as
1929 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
1930 it from "IPv6 and IPv4", which means that two separate sockets are being
1931 used.
1932
193341. The debug output for gethostbyname2() or getipnodebyname() failures now
1934 says whether AF_INET or AF_INET6 was passed as an argument.
1935
193642. Exiwhat output was messed up when time zones were included in log
1937 timestamps.
1938
193943. Exiwhat now gives more information about the daemon's listening ports,
1940 and whether -tls-on-connect was used.
1941
194244. The "port" option of the smtp transport is now expanded.
1943
194445. A "message" modifier in a "warn" statement in a non-message ACL was being
1945 silently ignored. Now an error message is written to the main and panic
1946 logs.
1947
194846. There's a new ACL modifier called "logwrite" which writes to a log file
1949 as soon as it is encountered.
1950
195147. Added $local_user_uid and $local_user_gid at routing time.
1952
195348. Exim crashed when trying to verify a sender address that was being
1954 rewritten to "<>".
1955
195649. Exim was recognizing only a space character after ".include". It now also
1957 recognizes a tab character.
1958
195950. Fixed several bugs in the Perl script that creates the exim.8 man page by
1960 extracting the relevant information from the specification. The man page no
1961 longer contains scrambled data for the -d option, and I've added a section
1962 at the front about calling Exim under different names.
1963
196451. Added "extra_headers" argument to the "mail" command in filter files.
1965
196652. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
1967 crash.
1968
196953. Installed eximstats 1.29.
1970
197154. Added transport_filter_timeout as a generic transport option.
1972
197355. Exim no longer adds an empty Bcc: header to messages that have no To: or
1974 Cc: header lines. This was required by RFC 822, but it not required by RFC
1975 2822.
1976
197756. Exim used to add From:, Date:, and Message-Id: header lines to any
1978 incoming messages that did not have them. Now it does so only if the
1979 message originates locally, that is, if there is no associated remote host
1980 address. When Resent- header lines are present, this applies to the Resent-
1981 lines rather than the non-Resent- lines.
1982
198357. Drop incoming SMTP connection after too many syntax or protocol errors. The
1984 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
1985
198658. Messages for configuration errors now include the name of the main
1987 configuration file - useful now that there may be more than one file in a
1988 list (.included file names were always shown).
1989
199059. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
1991 for those rare installations that do not start the daemon as root or run it
1992 setuid root. I've cut out the call to initgroups() if the daemon is not
1993 root at that time.
1994
199560. The Exim user and group can now be bound into the binary as text strings
1996 that are looked up at the start of Exim's processing.
1997
199861. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
1999
200062. Added $mailstore_basename variable.
2001
200263. Installed patch to sieve.c from Michael Haardt.
2003
200464. When Exim failed to open the panic log after failing to open the main log,
2005 the original message it was trying to log was written to stderr and debug
2006 output, but if they were not available (the usual case in production), it
2007 was lost. Now it is written to syslog before the two lines that record the
2008 failures to open the logs.
2009
201065. Users' Exim filters run in subprocesses under the user's uid. It is
2011 possible for a "deliver" command or an alias in a "personal" command to
2012 provoke an address rewrite. If logging of address rewriting is configured,
2013 this fails because the process is not running as root or exim. There may be
2014 a better way of dealing with this, but for the moment (because 4.30 needs
2015 to be released), I have disabled address rewrite logging when running a
2016 filter in a non-root, non-exim process.
2017
2018
2019Exim version 4.24
2020-----------------
2021
2022 1. The buildconfig auxiliary program wasn't quoting the value set for
2023 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2024 not defined. This bug was masked in 4.22 by the effect that was fixed in
2025 change 4.23/1.
2026
2027 2. Some messages that were rejected after a message id was allocated were
2028 shown as "incomplete" by exigrep. It no longer does this for messages that
2029 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2030
2031 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2032 have allow_domain_literals set, the ID did not get logged in the <= line.
2033 Domain literals are now always recognized in Message-ID: header lines.
2034
2035 4. The first argument for a ${extract expansion item is the key name or field
2036 number. Leading and trailing spaces in this item were not being ignored,
2037 causing some misleading effects.
2038
2039 5. When deliver_drop_privilege was set, single queue runner processes started
2040 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2041 same command in the process it spins off) were not dropping privilege.
2042
2043 6. When the daemon running as "exim" started a queue runner, it always
2044 re-executed Exim in the spun-off process. This is a waste of effort when
2045 deliver_drop_privilege is set. The new process now just calls the
2046 queue-runner function directly.
2047
2048
2049Exim version 4.23
2050-----------------
2051
2052 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2053 HEADERS_CHARSET.
2054
2055 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2056 ignored. Though the use of -oP was forcing the writing of a pid file, it
2057 was always written to the default place.
2058
2059 3. If the message "no IP address found for host xxxx" is generated during
2060 incoming verification, it is now followed by identification of the incoming
2061 connection (so you can more easily find what provoked it).
2062
2063 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2064
2065 5. Added some features to "harden" Exim a bit more against certain attacks:
2066
2067 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2068 be put in Local/Makefile. This is like the never_users runtime option,
2069 but it cannot be overridden. The default setting is "root".
2070
2071 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2072 prefix string with which any file named in a -C command line option
2073 must start.
2074
2075 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2076 is retained for -C and -D only if the caller of Exim is root. Without
2077 it, the exim user may also use -C and -D and retain privilege.
2078
2079 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2080 command line option is disabled.
2081
2082 6. Macro names set by the -D option must start with an upper case letter, just
2083 like macro names defined in the configuration file.
2084
2085 7. Added "dereference=" facility to LDAP.
2086
2087 8. Two instances of the typo "uknown" in the source files are fixed.
2088
2089 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2090 the Configure-Makefile script screwed up while processing it.
2091
209210. Incorporated PCRE 4.4.
2093
209411. The SMTP synchronization check was not operating right at the start of an
2095 SMTP session. For example, it could not catch a HELO sent before the client
2096 waited for the greeting. There is now a check for outstanding input at the
2097 point when the greeting is written. Because of the duplex, asynchronous
2098 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2099 way, but not yet received, when the check is performed.
2100
210112. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2102 on TCP/IP sockets, because this apparently causes some broken clients to
2103 timeout.
2104
210513. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2106 unchanged) from the Cygwin maintainer.
2107
210814. The code for -bV that shows what is in the binary showed "mbx" when maildir
2109 was supported instead of testing for mbx. Effectively a typo.
2110
211115. The spa authenticator server code was not checking that the input it
2112 received was valid base64.
2113
211416. The debug output line for the "set" modifier in ACLs was not showing the
2115 name of the variable that was being set.
2116
211717. Code tidy: the variable type "vtype_string" was never used. Removed it.
2118
211918. Previously, a reference to $sender_host_name did not cause a DNS reverse
2120 lookup on its own. Something else was needed to trigger the lookup. For
2121 example, a match in host_lookup or the need for a host name in a host list.
2122 Now, if $sender_host_name is referenced and the host name has not yet been
2123 looked up, a lookup is performed. If the lookup fails, the variable remains
2124 empty, and $host_lookup_failed is set to "1".
2125
212619. Added "eqi" as a case-independent comparison operator.
2127
212820. The saslauthd authentication condition could segfault if neither service
2129 nor realm was specified.
2130
213121. If an overflowing value such as "2048M" was set for message_size_limit, the
2132 error message that was logged was misleading, and incoming SMTP
2133 connections were dropped. The message is now more accurate, and temporary
2134 errors are given to SMTP connections.
2135
213622. In some error situations (such as 21 above) Exim rejects all SMTP commands
2137 (except RSET) with a 421 error, until QUIT is received. However, it was
2138 failing to send a response to QUIT.
2139
214023. The HELO ACL was being run before the code for helo_try_verify_hosts,
2141 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2142 ACL is now run after the helo_try_verify_hosts code.
2143
214424. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2145 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2146 case-independent, so other case variants are also recognized). Apparently
2147 some systems use these upper case variants.
2148
214925. If more than two messages were waiting for the same host, and a transport
2150 filter was specified for the transport, Exim sent two messages over the
2151 same TCP/IP connection, and then failed with "socket operation on non-
2152 socket" when it tried to send the third.
2153
215426. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2155
215627. The extern definition of crypt16() in expand.c was not being excluded when
2157 the OS had its own crypt16() function.
2158
215928. Added bounce_return_body as a new option, and bounce_return_size_limit
2160 as a preferred synonym for return_size_limit, both as an option and as an
2161 expansion variable.
2162
216329. Added LIBS=-liconv to OS/Makefile-OSF1.
2164
216530. Changed the default configuration ACL to relax the local part checking rule
2166 for addresses that are not in any local domains. For these addresses,
2167 slashes and pipe symbols are allowed within local parts, but the sequence
2168 /../ is explicitly forbidden.
2169
217031. SPA server authentication was not clearing the challenge buffer before
2171 using it.
2172
217332. log_message in a "warn" ACL statement was writing to the reject log as
2174 well as to the main log, which contradicts the documentation and doesn't
2175 seem right (because no rejection is happening). So I have stopped it.
2176
217733. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2178 However, I am unable to do any testing of this.
2179
218034. Fixed an infelicity in the appendfile transport. When checking directories
2181 for a mailbox, to see if any needed to be created, it was accidentally
2182 using path names with one or more superfluous leading slashes; tracing
2183 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2184
218535. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2186 contents are added to the log line that is written for every discarded
2187 recipient. (Previously a log_message setting was ignored.)
2188
218936. The ${quote: operator now quotes the string if it is empty.
2190
219137. The install script runs exim in order to find its version number. If for
2192 some reason other than non-existence or emptiness, which it checks, it
2193 could not run './exim', it was installing it with an empty version number,
2194 i.e. as "exim-". This error state is now caught, and the installation is
2195 aborted.
2196
219738. An argument was missing from the function that creates an error message
2198 when Exim fails to connect to the socket for saslauthd authentication.
2199 This could cause Exim to crash, or give a corrupted message.
2200
220139. Added isip, isip4, and isip6 to ${if conditions.
2202
220340. The ACL variables $acl_xx are now saved with the message, and can be
2204 accessed later in routers, transports, and filters.
2205
220641. The new lookup type nwildlsearch is like wildlsearch, except that the key
2207 strings in the file are not string-expanded.
2208
220942. If a MAIL command specified a SIZE value that was too large to fit into an
2210 int variable, the check against message_size_limit failed. Such values are
2211 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2212 day this will have to be increased, but I don't think I want to be around
2213 when emails are that large.
2214
2215
2216
2217Exim version 4.22
2218-----------------
2219
2220 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2221 iconv() is not standard in FreeBSD.
2222
2223 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2224 IPv6 enabled. The observed symptom was a segmentation fault on return from
2225 the function os_common_find_running_interfaces() in src/os.c.
2226
2227 3. In the check_special_case() function in daemon.c I had used "errno" as an
2228 argument name, which causes warnings on some systems. This was basically a
2229 typo, since it was named "eno" in the comments!
2230
2231 4. The code that waits for the clock to tick (at a resolution of some fraction
2232 of a second) so as to ensure message-id uniqueness was always waiting for
2233 at least one whole tick, when it could have waited for less. [This is
2234 almost certainly not relevant at current processor speeds, where it is
2235 unlikely to ever wait at all. But we try to future-proof.]
2236
2237 5. The function that sleeps for a time interval that includes fractions of a
2238 second contained a race. It did not block SIGALRM between setting the
2239 timer, and suspending (a couple of lines later). If the interval was short
2240 and the sigsuspend() was delayed until after it had expired, the suspension
2241 never ended. On busy systems this could lead to processes getting stuck for
2242 ever.
2243
2244 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2245 process, before it forks any delivery processes. The open lookup caching
2246 mechanism meant that the open file or database connection was passed into
2247 the delivery process. The problem was that delivery processes always tidy
2248 up cached lookup data. This could cause a problem for the next delivery
2249 process started by the queue runner, because the external queue runner
2250 process does not know about the closure. So the next delivery process
2251 still has data in the lookup cache. In the case of a file lookup, there was
2252 no problem because closing a file descriptor in a subprocess doesn't affect
2253 the parent. However, if the lookup was caching a connection to a database,
2254 the connection was closed, and the second delivery process was likely to
2255 see errors such as "PGSQL: query failed: server closed the connection
2256 unexpectedly". The problem has been fixed by closing all cached lookups
2257 in a queue runner before running a delivery process.
2258
2259 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2260 seem to have the "const" qualifier which it has on other OS. I've
2261 parameterised it.
2262
2263 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2264 *of the same type* (client or server) with the same public name that an
2265 error should be diagnosed.
2266
2267 9. When Exim looked up a host name for an IP address, but failed to find the
2268 original IP address when looking up the host name (a safety check), it
2269 output the message "<ip address> does not match any IP for NULL", which was
2270 confusing, to say the least. The bug was that the host name should have
2271 appeared instead of "NULL".
2272
227310. Since release 3.03, if Exim is called by a uid other than root or the Exim
2274 user that is built into the binary, and the -C or -D options is used, root
2275 privilege is dropped before the configuration file is read. In addition,
2276 logging is switched to stderr instead of the normal log files. If the
2277 configuration then re-defines the Exim user, the unprivileged environment
2278 is probably not what is expected, so Exim logs a panic warning message (but
2279 proceeds).
2280
2281 However, if deliver_drop_privilege is set, the unprivileged state may well
2282 be exactly what is intended, so the warning has been cut out in that case,
2283 and Exim is allowed to try to write to its normal log files.
2284
2285
2286Exim version 4.21
2287-----------------
2288
2289 1. smtp_return_error_details was not giving details for temporary sender
2290 or receiver verification errors.
2291
2292 2. Diagnose a configuration error if two authenticators have the same public
2293 name.
2294
2295 3. Exim used not to create the message log file for a message until the first
2296 delivery attempt. This could be confusing when incoming messages were held
2297 for policy or load reasons. The message log file is now created at the time
2298 the message is received, and an initial "Received" line is written to it.
2299
2300 4. The automatically generated man page for command line options had a minor
2301 bug that caused no ill effects; however, a more serious problem was that
2302 the procedure for building the man page automatically didn't always
2303 operate. Consequently, release 4.20 contains an out-of-date version. This
2304 shouldn't happen again.
2305
2306 5. When building Exim with embedded Perl support, the script that builds the
2307 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2308 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2309
2310 6. The freeze_tell option was not being used for messages that were frozen on
2311 arrival, either by an ACL or by local_scan().
2312
2313 7. Added the smtp_incomplete_transaction log selector.
2314
2315 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2316 was accepting AUTH without a new EHLO.
2317
2318 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2319 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2320 RFC.
2321
232210. Logging of TCP/IP connections (when configured) now happens in the main
2323 daemon process instead of the child process, so that the TCP/IP connection
2324 count is more accurate (but it can never be perfect).
2325
232611. The use of "drop" in a nested ACL was not being handled correctly in the
2327 outer ACL. Now, if condition failure induced by the nested "drop" causes
2328 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2329 or "require"), the connection is dropped.
2330
233112. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2332 that yield "discard" can now be used with an "accept" or a "discard" verb,
2333 but an error is generated for any others (because I can't see a useful way
2334 to define what should happen).
2335
233613. When an ACL is read dynamically from a file (or anywhere else), the lines
2337 are now processed in the same way as lines in the Exim configuration file.
2338 In particular, continuation lines are supported.
2339
234014. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2341
234215. Added -ti meaning -t -i.
2343
234416. Check for letters, digits, hyphens, and dots in the names of dnslist
2345 domains, and warn by logging if others are found.
2346
234717. At least on BSD, alignment is not guarenteed for the array of ifreq's
2348 returned from GIFCONF when Exim is trying to find the list of interfaces on
2349 a host. The code in os.c has been modified to copy each ifreq to an aligned
2350 structure in all cases.
2351
2352 Also, in some cases, the returned ifreq's were being copied to a 'struct
2353 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2354 means the last couple of bytes of an IPv6 address could be chopped if the
2355 ifreq contained only a normal sockaddr (14 bytes storage).
2356
235718. Named domain lists were not supported in the hosts_treat_as_local option.
2358 An entry such as +xxxx was not recognized, and was treated as a literal
2359 domain name.
2360
236119. Ensure that header lines added by a DATA ACL are included in the reject log
2362 if the ACL subsequently rejects the message.
2363
236420. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2365 MD5 (which is deprecated).
2366
236721. When testing a filter file using -bf, Exim was writing a message when it
2368 took the sender from a "From " line in the message, but it was not doing so
2369 when it took $return_path from a Return-Path: header line. It now does.
2370
237122. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2372 with a valid header line field name (a series of printing characters
2373 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2374
237523. Changed "disc" in the source to "disk" to conform to the documentation and
2376 the book and for uniformity.
2377
237824. Ignore Sendmail's -Ooption=value command line item.
2379
238025. When execve() failed while trying to run a command in a pipe transport,
2381 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2382 could be confused with a return value of 69 from the command itself. This
2383 has been changed to 127, the value the shell returns if it is asked to run
2384 a non-existent command. The wording for the related log line suggests a
2385 non-existent command as the problem.
2386
238726. If received_header_text expands to an empty string, do not add a Received:
2388 header line to the message. (Well, it adds a token one on the spool, but
2389 marks it "old" so that it doesn't get used or transmitted.)
2390
239127. Installed eximstats 1.28 (addition of -nt option).
2392
239328. There was no check for failure on the call to getsockname() in the daemon
2394 code. This can fail if there is a shortage of resources on the system, with
2395 ENOMEM, for example. A temporary error is now given on failure.
2396
239729. Contrary to the C standard, it seems that in some environments, the
2398 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2399 program. Exim now does this explicitly; it affects the formatting of
2400 timestamps using strftime().
2401
240230. If exiqsumm was given junk data, it threw up some uninitialized variable
2403 complaints. I've now initialized all the variables, to avoid this.
2404
240532. Header lines added by a system filter were not being "seen" during
2406 transport-time rewrites.
2407
240833. The info_callback() function passed to OpenSSL is set up with type void
2409 (*)(SSL *, int, int), as described somewhere. However, when calling the
2410 function (actually a macro) that sets it up, the type void(*)() is
2411 expected. I've put in a cast to prevent warnings from picky compilers.
2412
241334. If a DNS black list lookup found a CNAME record, but there were no A
2414 records associated with the domain it pointed at, Exim crashed.
2415
241635. If a DNS black list lookup returned more than one A record, Exim ignored
2417 all but the first. It now scans all returned addresses if a particular IP
2418 value is being sought. In this situation, the contents of the
2419 $dnslist_value variable are a list of all the addresses, separated by a
2420 comma and a space.
2421
242236. Tightened up the rules for host name lookups using reverse DNS. Exim used
2423 to accept a host name and all its aliases if the forward lookup for any of
2424 them yielded the IP address of the incoming connection. Now it accepts only
2425 those names whose forward lookup yields the correct IP address. Any other
2426 names are discarded. This closes a loophole whereby a rogue DNS
2427 administrator could create reverse DNS records to break through a
2428 wildcarded host restriction in an ACL.
2429
243037. If a user filter or a system filter that ran in a subprocess used any of
2431 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2432 the wrong values were passed to the pipe command ($thisaddress had the
2433 value of $0, $0 had the value of $1, etc). This bug was introduced by
2434 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2435
243638. Improved the line breaking for long SMTP error messages from ACLs.
2437 Previously, if there was no break point between 40 and 75 characters, Exim
2438 left the rest of the message alone. Two changes have been made: (a) I've
2439 reduced the minimum length to 35 characters; (b) if it can't find a break
2440 point between 35 and 75 characters, it looks ahead and uses the first one
2441 that it finds. This may give the occasional overlong line, but at least the
2442 remaining text gets split now.
2443
244439. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2445 file descriptors might be low, and that setting 1000 would always raise it.
2446 It turns out that in some environments, the limit is already over 1000 and
2447 that lowering it causes trouble. So now Exim takes care not to decrease it.
2448
244940. When delivering a message, the value of $return_path is set to $sender_
2450 address at the start of routing (routers may change the value). By an
2451 oversight, this default was not being set up when an address was tested by
2452 -bt or -bv, which affected the outcome if any router or filter referred to
2453 $return_path.
2454
245541. The idea of the "warn" ACL verb is that it adds a header or writes to the
2456 log only when "message" or "log_message" are set. However, if one of the
2457 conditions was an address verification, or a call to a nested ACL, the
2458 messages generated by the underlying test were being passed through. This
2459 no longer happens. The underlying message is available in $acl_verify_
2460 message for both "message" and "log_message" expansions, so it can be
2461 passed through if needed.
2462
246342. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2464 new expansion $bh_ to give the encoded byte string without charset
2465 translation. Translation happens only if iconv() is available; HAVE_ICONV
2466 indicates this at build time. HEADERS_CHARSET gives the charset to
2467 translate to; headers_charset can change it in the configuration, and
2468 "headers charset" can change it in an individual filter file.
2469
247043. Now that we have a default RFC 2047 charset (see above), the code in Exim
2471 that creates RFC 2047 encoded "words" labels them as that charset instead
2472 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2473 expansion operator; (ii) when Exim creates a From: line for a local
2474 message; (iii) when a header line is rewritten to include a "phrase" part.
2475
247644. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2477 buggy, causing it to skip the first lines of messages whose message ID
2478 ended in 'D'. This would not have bitten before Exim release 4.14, because
2479 message IDs were unlikely to end in 'D' before then. The effect was to have
2480 incorrect size information for certain domains.
2481
248245. #include "config.h" was missing at the start of the crypt16.c module. This
2483 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2484 noticed.
2485
248646. If there was a timeout during a "random" callout check, Exim treated it as
2487 a failure of the random address, and carried on sending RSET and the real
2488 address. If the delay was just some slowness somewhere, the response to the
2489 original RCPT would be taken as a response to RSET and so on, causing
2490 mayhem of various kinds.
2491
249247. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2493 when I implemented it. It didn't allow for the fact that some option values
2494 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2495 the right test for positive values.
2496
249748. Domain names in DNS records are case-independent. Exim always looks them up
2498 in lower case. Some resolvers return domain names in exactly the case they
2499 appear in the zone file, that is, they may contain uppercase letters. Not
2500 all resolvers do this - some return always lower case. Exim was treating a
2501 change of case by a resolver as a change of domain, similar to a widening
2502 of a domain abbreviation. This triggered its re-routing code and so it was
2503 trying to route what was effectively the same domain again. This normally
2504 caused routing to fail (because the router wouldn't handle the domain
2505 twice). Now Exim checks for this case specially, and just changes the
2506 casing of the domain that it ultimately uses when it transmits the message
2507 envelope.
2508
250949. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2510 module.
2511
251250. If a filter generated a file delivery with a non-absolute name (possible if
2513 no home directory exists for the router), the forbid_file option was not
2514 forbidding it.
2515
251651. Added '&' feature to dnslists, to provide bit mask matching in addition to
2517 the existing equality matching.
2518
251952. Exim was using ints instead of ino_t variables in some places where it was
2520 dealing with inode numbers.
2521
252253. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2523 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2524 called TMPDIR, and if it finds it is different, it changes its value.
2525
252654. The smtp_printf() function is now made available to local_scan() so
2527 additional output lines can be written before returning. There is also an
2528 smtp_fflush() function to enable the detection of a dropped connection.
2529 The variables smtp_input and smtp_batched_input are exported to
2530 local_scan().
2531
253255. Changed the default runtime configuration: the message "Unknown user"
2533 has been removed from the ACL, and instead placed on the localuser router,
2534 using the cannot_route_message feature. This means that any verification
2535 failures that generate their own messages won't get overridden. Similarly,
2536 the "Unrouteable address" message that was in the ACL for unverifiable
2537 relay addresses has also been removed.
2538
253956. Added hosts_avoid_esmtp to the smtp transport.
2540
254157. The exicyclog script was not checking for the esoteric option
2542 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2543 will work only if exicyclog is run under the appropriate euid.
2544
254558. Following a discussion on the list, the rules by which Exim recognises line
2546 endings on incoming messages have been changed. The -dropcr and drop_cr
2547 options are now no-ops, retained only for backwards compatibility. The
2548 following line terminators are recognized: LF CRLF CR. However, special
2549 processing applies to CR:
2550
2551 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2552 nor a local message in the state where . is a terminator.
2553
2554 (ii) If a bare CR is encountered in a header line, an extra space is added
2555 after the line terminator so as not to end the header. The reasoning
2556 behind this is that bare CRs in header lines are most likely either
2557 to be mistakes, or people trying to play silly games.
2558
255959. The size of a message, as listed by "-bp" or in the Exim monitor window,
2560 was being incorrectly given as 18 bytes larger than it should have been.
2561 This is a VOB (very old bug).
2562
256360. This may never have affected anything current, but just in case it has:
2564 When the local host is found other than at the start of a list of hosts,
2565 the local host, those with the same MX, and any that follow, are discarded.
2566 When the list in question was part of a longer list of hosts, the following
2567 hosts (not currently being processed) were also being discarded. This no
2568 longer happens. I'm not sure if this situation could ever has previously
2569 arisen.
2570
257161. Added the "/MX" feature to lists of hosts in the manualroute and query
2572 program routers.
2573
257462. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2575 header. This is something that is recommended in a new Internet Draft, and
2576 is something that is documented as being done by Sendmail. There are two
2577 possible values. For messages generated by the autoreply transport, Exim
2578 adds:
2579
2580 Auto-Submitted: auto-replied
2581
2582 whereas for all other generated messages (e.g. bounces) it adds
2583
2584 Auto-Submitted: auto-generated
2585
258663. The "personal" condition in filters now includes a test for the
2587 Auto-Submitted: header. If it contains the string "auto-" the message it
2588 not considered personal.
2589
259064. Added rcpt_include_affixes as a generic transport option.
2591
259265. Added queue_only_override (default true).
2593
259466. Added the syslog_duplication option.
2595
259667. If what should have been the first header line of a message consisted of
2597 a space followed by a colon, Exim was mis-interpreting it as a header line.
2598 It isn't of course - it is syntactically invalid and should therefore be
2599 treated as the start of the message body. The misbehaviour could have
2600 caused a number of strange effects, including loss of data in subsequent
2601 header lines, and spool format errors.
2602
260368. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2604 client host had authenticated. This control can now be exercised by an ACL
2605 for more flexibility.
2606
260769. By default, callouts do not happen when testing with -bh. There is now a
2608 variant, -bhc, which does actually run the callout code, including
2609 consulting and updating the callout cache.
2610
261170. Added support for saslauthd authentication, courtesy of Alexander
2612 Sabourenkov.
2613
261471. If statvfs() failed on the spool or log directories while checking their
2615 size for availability, Exim confusingly gave the error "space shortage".
2616 Furthermore, in debugging mode it crashed with a floating point exception.
2617 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2618 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2619 serious problem, Exim now writes to the main and panic logs when this
2620 happens, with details of the failure. It then refuses to accept the
2621 incoming message, giving the message "spool directory problem" or "log
2622 directory problem" with a 421 code for SMTP messages.
2623
262472. When Exim is about to re-exec itself, it ensures that the file descriptors
2625 0, 1, and 2 exist, because some OS complain for execs without them (see
2626 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2627 descriptors. However, the code omitted to check that the open succeeded,
2628 causing mysterious errors if for some reason the permissions on /dev/null
2629 got screwed. Now Exim writes a message to the main and panic logs, and
2630 bombs out if it can't open /dev/null.
2631
263273. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2633 interact so that it is all more flexible. It is supposed to remain
2634 backwards compatible. Also added extra_local_interfaces.
2635
263674. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2637 to bomb out with an assertion failure - to the client this appears as a
2638 connection drop. This problem occurs in the part of the code that was taken
2639 from the Samba project. Fortunately, the assertion is in a very simple
2640 function, so I have fixed this by reproducing the function inline in the
2641 one place where it is called, and arranging for authentication to fail
2642 instead of killing the process with assert().
2643
264475. The SPA client code was not working when the server requested OEM rather
2645 than Unicode encoding.
2646
264776. Added code to make require_files with a specific uid setting more usable in
2648 the case where statting the file as root fails - usually a non-root-mounted
2649 NFS file system. When this happens and the failure is EACCES, Exim now
2650 forks a subprocess and does the per-uid checking as the relevant uid.
2651
265277. Added process_log_path.
2653
265478. If log_file_path was not explicitly set, a setting of check_log_space or
2655 check_log_inodes was ignored.
2656
265779. If a space check for the spool or log partitions fails, the incident is now
2658 logged. Of course, in the latter case the data may get lost...
2659
266080. Added the %p formatting code to string_format() so that it can be used to
2661 print addresses in debug_print(). Adjusted all the address printing in the
2662 debugging in store.c to use %p rather than %d.
2663
266481. There was a concern that a line of code in smtp_in.c could overflow a
2665 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2666 initially expressed, the concern was not well-founded, because trailing
2667 spaces are removed early. However, if the trailing spaces were followed by
2668 a NULL, they did not get removed, so the overflow was possible. Two fixes
2669 were applied:
2670
2671 (a) I re-wrote the offending code in a cleaner fashion.
2672 (b) If an incoming SMTP command contains a NULL character, it is rejected
2673 as invalid.
2674
267582. When Exim changes uid/gid to the Exim user at daemon start time, it now
2676 runs initgroups(), so that if the Exim user is in any additional groups,
2677 they will be used during message reception.
2678
2679
2680Exim version 4.20
2681-----------------
2682
2683The change log for 4.20 and earlier releases has been archived.
2684
2685****