1. Preserve underlying transport filter timeout message if a pipe
[exim.git] / doc / doc-txt / ChangeLog
CommitLineData
22c3b60b 1$Cambridge: exim/doc/doc-txt/ChangeLog,v 1.135 2005/05/10 11:13:09 ph10 Exp $
495ae4b0
PH
2
3Change log file for Exim from version 4.21
4-------------------------------------------
5
bbe902f0 6
e5a9dba6
PH
7Exim version 4.52
8-----------------
9
10TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
11
22c3b60b
PH
12PH/01 When a transport filter timed out in a pipe delivery, and the pipe
13 command itself ended in error, the underlying message about the transport
14 filter timeout was being overwritten with the pipe command error. Now the
15 underlying error message should be appended to the second error message.
16
e5a9dba6 17
7982096b
PH
18Exim version 4.51
19-----------------
20
1a46a8c5
PH
21TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
22 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
23
2f079f46 24TK/02 Fix ACL "control" statement not being available in MIME ACL.
1a46a8c5
PH
25
26TK/03 Fix ACL "regex" condition not being available in MIME ACL.
27
28PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
29 to test Sieve filters that use "vacation".
30
31PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
32 that changes the way the GnuTLS parameters are stored in the cache file.
33 The new format can be generated externally. For backward compatibility,
34 if the data in the cache doesn't make sense, Exim assumes it has read an
35 old-format file, and it generates new data and writes a new file. This
36 means that you can't go back to an older release without removing the
37 file.
38
39PH/03 A redirect router that has both "unseen" and "one_time" set does not
40 work if there are any delivery delays because "one_time" forces the
41 parent to be marked "delivered", so its unseen clone is never tried
42 again. For this reason, Exim now forbids the simultaneous setting of
43 these two options.
44
45PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
46 redirected to themselves ("homonym" addresses). Read the long ChangeLog
47 entry if you want to know the details. The fix, however, neglected to
48 consider the case when local delivery batching is involved. The test for
49 "previously delivered" was not happening when checking to see if an
50 address could be batched with a previous (undelivered) one; under
51 certain circumstances this could lead to multiple deliveries to the same
c2c19e9d 52 address.
1a46a8c5
PH
53
54PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
55 in its include files, and this causes problems building Exim.
56
57PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
58 header_syntax) but Exim was just ignoring anything given after a slash.
59 In particular, this caused confusion with an attempt to use "verify =
60 reverse_host_lookup/defer_ok". An error is now given when options are
61 supplied for verify items that do not have them. (Maybe reverse_host_
62 lookup should have a defer_ok option, but that's a different point.)
63
64PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
65 defined by RFC 821) to 2048, because there were problems with some AUTH
66 commands, and RFC 1869 says the size should be increased for extended
67 SMTP commands that take arguments.
68
69PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
70 Finch).
71
72PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
73 "unknown" error; now it says that the functionality isn't in the binary.
8d67ada3 74
49c2d5ea
PH
75PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
76 an address' error message when a string expansion fails (syntax or
f331f3b6
PH
77 whatever). Otherwise the password may appear in the log. Following change
78 PH/42 below, there is no longer a chance of it appearing in a bounce
79 message.
49c2d5ea 80
bf759a8b
PH
81PH/11 Installed exipick version 20050225.0 from John Jetmore.
82
83364d30
PH
83PH/12 If the last host in a fallback_hosts list was multihomed, only the first
84 of its addresses was ever tried. (Bugzilla bug #2.)
85
7999bbd7
PH
86PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
87 the result incorrectly in the debug output. (It correctly added a newline
88 to what was transported.)
89
7dbf77c9
PH
90TF/01 Added $received_time.
91
74e0617f
PH
92PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
93 commented out examples of how to interface to a virus scanner and to
94 SpamAssassin. Also added commented examples of av_scanner and
95 spamd_address settings.
96
2f079f46
PH
97PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
98 and controls are allowed in which ACLs. There were a couple of minor
99 errors. Some of the entries in the conditions table (which is a table of
100 where they are NOT allowed) were getting very unwieldy; rewrote them as a
101 negation of where the condition IS allowed.
102
8c841523
PH
103PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
104
7766a4f0
PH
105PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
106 header file does not have a version number, so I've had to invent a new
107 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
108 API. The code is untested by me (my Linux distribution still has 0.3.2 of
109 radiusclient), but it was contributed by a Radius user.
110
8b417f2c
PH
111PH/18 Installed Lars Mainka's patch for the support of CRL collections in
112 files or directories, for OpenSSL.
113
901f42cb
PH
114PH/19 When an Exim process that is running as root has to create an Exim log
115 file, it does so in a subprocess that runs as exim:exim so as to get the
116 ownership right at creation (otherwise, other Exim processes might see
117 the file with the wrong ownership). There was no test for failure of this
118 fork() call, which would lead to the process getting stuck as it waited
119 for a non-existent subprocess. Forks do occasionally fail when resources
120 run out. I reviewed all the other calls to fork(); they all seem to check
121 for failure.
122
f9b9210e
PH
123PH/20 When checking for unexpected SMTP input at connect time (before writing
124 the banner), Exim was not dealing correctly with a non-positive return
125 from the read() function. If the client had disconnected by this time,
126 the result was a log entry for a synchronization error with an empty
127 string after "input=" when read() returned zero. If read() returned -1
128 (an event I could not check), uninitialized data bytes were printed.
129 There were reports of junk text (parts of files, etc) appearing after
130 "input=".
131
54cdb463
PH
132PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
133
cf00dad6
PH
134PH/22 Added support for macro redefinition, and (re)definition in between
135 driver and ACL definitions.
136
acb1b346
PH
137PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
138 forgetting to use the resulting value; it was using the unexpanded value.
139
c5ddb310
PH
140PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
141 hadn't been configured. The fix is from Juergen Kreileder, who
142 understands it better than I do:
143
144 "Here's what I see happening with three configured cyrus_sasl
145 authenticators configured (plain, login, cram-md5):
146
147 On startup auth_cyrus_sasl_init() gets called for each of these.
148 This means three calls to sasl_listmech() without a specified mech_list.
149 => SASL tests which mechs of all available mechs actually work
150 => three warnings about OTP not working
151 => the returned list contains: plain, login, cram-md5, digest-md5, ...
152
153 With the patch, sasl_listmech() also gets called three times. But now
154 SASL's mech_list option is set to the server_mech specified in the the
155 authenticator. Or in other words, the answer from sasl_listmech()
156 gets limited to just the mech you're testing for (which is different
157 for each call.)
158 => the return list contains just 'plain' or 'login', 'cram-md5' or
159 nothing depending on the value of ob->server_mech.
160
161 I've just tested the patch: Authentication still works fine,
162 unavailable mechs specified in the exim configuration are still
163 caught, and the auth.log warnings about OTP are gone."
164
31619da6
PH
165PH/25 When debugging is enabled, the contents of the command line are added
166 to the debugging output, even when log_selector=+arguments is not
167 specified.
168
bebaf0fc
PH
169PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
170 answer is "GNU", and only if the return is "GNU/something" is the answer
171 "Linux".
172
475fe28a
PH
173PH/27 $acl_verify_message is now set immediately after the failure of a
174 verification in an ACL, and so is available in subsequent modifiers. In
175 particular, the message can be preserved by coding like this:
176
177 warn !verify = sender
178 set acl_m0 = $acl_verify_message
179
180 Previously, $acl_verify_message was set only while expanding "message"
181 and "log_message" when a very denied access.
182
7e8bec7a
PH
183PH/28 Modified OS/os.c-Linux with
184
185 -#ifndef OS_LOAD_AVERAGE
186 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
187
188 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
189 nomenclature these days.)
190
e4a89c47
PH
191PH/29 Installed patch from the Sieve maintainer that adds the options
192 sieve_useraddress and sieve_subaddress to the redirect router.
193
5ca2a9a1
PH
194PH/30 In these circumstances:
195 . Two addresses routed to the same list of hosts;
196 . First host does not offer TLS;
197 . First host accepts first address;
198 . First host gives temporary error to second address;
199 . Second host offers TLS and a TLS session is established;
200 . Second host accepts second address.
201 Exim incorrectly logged both deliveries with the TLS parameters (cipher
202 and peerdn, if requested) that were in fact used only for the second
203 address.
7e8bec7a 204
c688b954
PH
205PH/31 When doing a callout as part of verifying an address, Exim was not paying
206 attention to any local part prefix or suffix that was matched by the
207 router that accepted the address. It now behaves in the same way as it
208 does for delivery: the affixes are removed from the local part unless
209 rcpt_include_affixes is set on the transport.
210
fed77020
PH
211PH/32 Add the sender address, as F=<...>, to the log line when logging a
212 timeout during the DATA phase of an incoming message.
213
7fe1560f
PH
214PH/33 Sieve envelope tests were broken for match types other than :is. I have
215 applied a patch sanctioned by the Sieve maintainer.
c688b954 216
ebb6e6d5
PH
217PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
218 the uid or gid is negative. A case of a negative gid caused this to be
219 noticed. The fix allows for either to be negative.
220
9c4e8f60
PH
221PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
222 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
223 overlooked.
224
225PH/36 The change PH/12 above was broken. Fixed it.
226
d7174846
PH
227PH/37 Exim used to check for duplicate addresses in the middle of routing, on
228 the grounds that routing the same address twice would always produce the
229 same answer. This might have been true once, but it is certainly no
230 longer true now. Routing a child address may depend on the previous
231 routing that produced that child. Some complicated redirection strategies
232 went wrong when messages had multiple recipients, and made Exim's
233 behaviour dependent on the order in which the addresses were given.
234
235 I have moved the duplicate checking until after the routing is complete.
236 Exim scans the addresses that are assigned to local and remote
237 transports, and removes any duplicates. This means that more work will be
238 done, as duplicates will always all be routed, but duplicates are
239 presumably rare, so I don't expect this is of any significance.
240
241 For deliveries to pipes, files, and autoreplies, the duplicate checking
242 still happens during the routing process, since they are not going to be
243 routed further.
244
cfe75fc3
PH
245PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
246 It corrects a timeout issue with spamd. This is Ian's comment: "The
247 background is that sometimes spamd either never reads data from a
248 connection it has accepted, or it never writes response data. The exiscan
249 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
250 blindly assumes that writes won't block so it may never time out."
251
be22d70e
PH
252PH/39 Allow G after quota size as well as K and M.
253
0612b098
PH
254PH/40 The value set for $authenticated_id in an authenticator may not contain
255 binary zeroes or newlines because the value is written to log lines and
256 to spool files. There was no check on this. Now the value is run through
257 the string_printing() function so that such characters are converted to
258 printable escape sequences.
259
2e0c1448
PH
260PH/41 $message_linecount is a new variable that contains the total number of
261 lines in the message. Compare $body_linecount, which is the count for the
262 body only.
263
447d236c
PH
264PH/42 Exim no longer gives details of delivery errors for specific addresses in
265 bounce and delay warning messages, except in certain special cases, which
266 are as follows:
267
268 (a) An SMTP error message from a remote host;
269 (b) A message specified in a :fail: redirection;
270 (c) A message specified in a "fail" command in a system filter;
271 (d) A message specified in a FAIL return from the queryprogram router;
272 (e) A message specified by the cannot_route_message router option.
273
274 In these cases only, Exim does include the error details in bounce and
275 warning messages. There are also a few cases where bland messages such
276 as "unrouteable address" or "local delivery error" are given.
277
d20976dc
PH
278PH/43 $value is now also set for the "else" part of a ${run expansion.
279
f656d135
PH
280PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
281 being worked on, but at least Exim now implements the latest version to
282 play with."
283
2e2a30b4
PH
284PH/45 In a pipe transport, although a timeout while waiting for the pipe
285 process to complete was treated as a delivery failure, a timeout while
286 writing the message to the pipe was logged, but erroneously treated as a
287 successful delivery. Such timeouts include transport filter timeouts. For
288 consistency with the overall process timeout, these timeouts are now
289 treated as errors, giving rise to delivery failures by default. However,
290 there is now a new Boolean option for the pipe transport called
291 timeout_defer, which, if set TRUE, converts the failures into defers for
292 both kinds of timeout. A transport filter timeout is now identified in
293 the log output.
294
9176e9f0
PH
295PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
296 systems where "make" and "gmake" are different, calling "gmake" at top
297 level broke things. I've arranged for the value of $(MAKE) to be passed
298 from the Makefile to this script so that it can call the same version of
299 "make".
300
7982096b 301
bbe902f0
PH
302A note about Exim versions 4.44 and 4.50
303----------------------------------------
304
305Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
306changes of various kinds. As a consequence, a big documentation update was
307needed. This delayed the release for rather longer than seemed good, especially
308in the light of a couple of (minor) security issues. Therefore, the changes
309that fixed bugs were backported into 4.43, to create a 4.44 maintenance
310release. So 4.44 and 4.50 are in effect two different branches that both start
311from 4.43.
312
313I have left the 4.50 change log unchanged; it contains all the changes since
3144.43. The change log for 4.44 is below; many of its items are identical to
315those for 4.50. This seems to be the most sensible way to preserve the
316historical information.
317
318
f7b63901 319Exim version 4.50
495ae4b0
PH
320-----------------
321
5fe762f6
PH
322 1. Minor wording change to the doc/README.SIEVE file.
323
139059f6 324 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
5fe762f6 325 computation of the current number of files was incorrect.
495ae4b0 326
7086e875
PH
327 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
328 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
329 place.
330
35af9f61
PH
331 4. Give more explanation in the error message when the command for a transport
332 filter fails to execute.
333
b668c215
PH
334 5. There are several places where Exim runs a non-Exim command in a
335 subprocess. The SIGUSR1 signal should be disabled for these processes. This
336 was being done only for the command run by the queryprogram router. It is
337 now done for all such subprocesses. The other cases are: ${run, transport
338 filters, and the commands run by the lmtp and pipe transports.
339
a494b1e1
PH
340 6. Added CONFIGURE_GROUP build-time option.
341
342 7. Some older OS have a limit of 256 on the maximum number of file
343 descriptors. Exim was using setrlimit() to set 1000 as a large value
344 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
345 systems. I've change it so that if it can't get 1000, it tries for 256.
35edf2ff 346
c5fcb476
PH
347 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
348 was an oversight, and furthermore, ever since the addition of extra
349 controls (e.g. 4.43/32), the checks on when to allow different forms of
350 "control" were broken. There should now be diagnostics for all cases when a
351 control that does not make sense is encountered.
352
69358f02
PH
353 9. Added the /retain_sender option to "control=submission".
354
5be20824
PH
35510. $recipients is now available in the predata ACL (oversight).
356
eb2c0248
PH
35711. Tidy the search cache before the fork to do a delivery from a message
358 received from the command line. Otherwise the child will trigger a lookup
359 failure and thereby defer the delivery if it tries to use (for example) a
360 cached ldap connection that the parent has called unbind on.
361
2a3eea10
PH
36212. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
363 of $address_data from the recipient verification was clobbered by the
364 sender verification.
365
36613. The value of address_data from a sender verification is now available in
367 $sender_address_data in subsequent conditions in the ACL statement.
368
23c7ff99
PH
36914. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
370
4deaf07d
PH
37115. Added a new option "connect=<time>" to callout options, to set a different
372 connection timeout.
373
926e1192
PH
37416. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
375 was its contents. (It was OK if the option was not defined at all.)
376
650edc6f
PH
37717. A "Completed" log line is now written for messages that are removed from
378 the spool by the -Mrm option.
379
2c7db3f5
PH
38018. New variables $sender_verify_failure and $recipient_verify_failure contain
381 information about exactly what failed.
382
3d235903
PH
38319. Added -dd to debug only the daemon process.
384
7c7ad977
PH
38520. Incorporated Michael Haardt's patch to ldap.c for improving the way it
386 handles timeouts, both on the server side and network timeouts. Renamed the
387 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
388
981756db
PH
38921. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
390 It is now set to "smtps".
391
d4eb88df
PH
39222. $host_address is now set to the target address during the checking of
393 ignore_target_hosts.
394
39523. When checking ignore_target_hosts for an ipliteral router, no host name was
396 being passed; this would have caused $sender_host_name to have been used if
397 matching the list had actually called for a host name (not very likely,
398 since this list is usually IP addresses). A host name is now passed as
399 "[x.x.x.x]".
400
7d468ab8
PH
40124. Changed the calls that set up the SIGCHLD handler in the daemon to use the
402 code that specifies a non-restarting handler (typically sigaction() in
403 modern systems) in an attempt to fix a rare and obscure crash bug.
404
40525. Narrowed the window for a race in the daemon that could cause it to ignore
406 SIGCHLD signals. This is not a major problem, because they are used only to
407 wake it up if nothing else does.
408
62c0818f
PH
40926. A malformed maildirsize file could cause Exim to calculate negative values
410 for the mailbox size or file count. Odd effects could occur as a result.
411 The maildirsize information is now recalculated if the size or filecount
412 end up negative.
413
26034054
PH
41427. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
415 support for a long time. Removed HAVE_SYS_VFS_H.
416
af66f652
PH
41728. Installed the latest version of exipick from John Jetmore.
418
90af77f4
PH
41929. In an address list, if the pattern was not a regular expression, an empty
420 subject address (from a bounce message) matched only if the pattern was an
421 empty string. Non-empty patterns were not even tested. This was the wrong
422 because it is perfectly reasonable to use an empty address as part of a
423 database query. An empty address is now tested by patterns that are
424 lookups. However, all the other forms of pattern expect the subject to
425 contain a local part and a domain, and therefore, for them, an empty
426 address still always fails if the pattern is not itself empty.
427
d8ef3577
PH
42830. Exim went into a mad DNS loop when attempting to do a callout where the
429 host was specified on an smtp transport, and looking it up yielded more
430 than one IP address.
431
5cb8cbc6
PH
43231. Re-factored the code for checking spool and log partition space into a
433 function that finds that data and another that does the check. The former
434 is then used to implement four new variables: $spool_space, $log_space,
435 $spool_inodes, and $log_inodes.
436
14702f5b
PH
43732. The RFC2047 encoding function was originally intended for short strings
438 such as real names; it was not keeping to the 75-character limit for
439 encoded words that the RFC imposes. It now respects the limit, and
440 generates multiple encoded words if necessary. To be on the safe side, I
441 have increased the buffer size for the ${rfc2047: expansion operator from
442 1024 to 2048 bytes.
443
063b1e99
PH
44433. It is now permitted to omit both strings after an "if" condition; if the
445 condition is true, the result is "true". As before, when the second string
446 is omitted, a false condition yields an empty string. This makes it less
447 cumbersome to write custom ACL and router conditions.
448
652e1b65
PH
44934. Failure to deliver a bounce message always caused it to be frozen, even if
450 there was an errors_to setting on the router. The errors_to setting is now
451 respected.
452
6f0c9a4f
PH
45335. If an IPv6 address is given for -bh or -bhc, it is now converted to the
454 canonical form (fully expanded) before being placed in
455 $sender_host_address.
456
33397d19
PH
45736. The table in the code that translates DNS record types into text (T_A to
458 "A" for instance) was missing entries for NS and CNAME. It is just possible
459 that this could have caused confusion if both these types were looked up
460 for the same domain, because the text type is used as part of Exim's
461 per-process caching. But the chance of anyone hitting this buglet seems
462 very small.
463
7bb56e1f
PH
46437. The dnsdb lookup has been extended in a number of ways.
465
466 (1) There is a new type, "zns", which walks up the domain tree until it
467 finds some nameserver records. It should be used with care.
468
ea3bc19b
PH
469 (2) There is a new type, "mxh", which is like "mx" except that it returns
470 just the host names, not the priorities.
471
472 (3) It is now possible to give a list of domains (or IP addresses) to be
ff4dbb19
PH
473 looked up. The behaviour when one of the lookups defers can be
474 controlled by a keyword.
7bb56e1f 475
ea3bc19b 476 (4) It is now possible to specify the separator character for use when
7bb56e1f 477 multiple records are returned.
33397d19 478
0bcb2a0e
PH
47938. The dnslists ACL condition has been extended: it is now possible to supply
480 a list of IP addresses and/or domains to be looked up in a particular DNS
481 domain.
482
2ac0e484
PH
48339. Added log_selector=+queue_time_overall.
484
4e1fde53
PH
48540. When running the queue in the test harness, wait just a tad after forking a
486 delivery process, to get repeatability of debugging output.
487
de365ded
PH
48841. Include certificate and key file names in error message when GnuTLS fails
489 to set them up, because the GnuTLS error message doesn't include the name
490 of the failing file when there is a problem reading it.
491
f05da2e8
PH
49242. Allow both -bf and -bF in the same test run.
493
d6453af2
PH
49443. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
495
f7b63901
PH
49644. The "Exiscan patch" is now merged into the mainline Exim source.
497
49845. Sometimes the final signoff response after QUIT could fail to get
499 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
500 < 0 before doing a fflush(). This bug looks as though it goes back to the
501 introduction of TLS in release 3.20, but "sometimes" must have been rare
502 because the tests only now provoked it.
503
a444213a
PH
50446. Reset the locale to "C" after calling embedded Perl, in case it was changed
505 (this can affect the format of dates).
506
0ec020ea
PH
50747. exim_tidydb, when checking for the continued existence of a message for
508 which it has found a message-specific retry record, was not finding
509 messages that were in split spool directories. Consequently, it was
510 deleting retry records that should have stayed in existence.
511
b1206957
PH
51248. Steve fixed some bugs in eximstats.
513
51449. The SPA authentication driver was not abandoning authentication and moving
515 on to the next authenticator when an expansion was forced to fail,
516 contradicting the general specification for all authenticators. Instead it
517 was generating a temporary error. It now behaves as specified.
518
26dd5a95
PH
51950. The default ordering of permitted cipher suites for GnuTLS was pessimal
520 (the order specifies the preference for clients). The order is now AES256,
521 AES128, 3DES, ARCFOUR128.
522
343b2385
PH
52351. Small patch to Sieve code - explicitly set From: when generating an
524 autoreply.
525
1c5466b9
PH
52652. Exim crashed if a remote delivery caused a very long error message to be
527 recorded - for instance if somebody sent an entire SpamAssassin report back
528 as a large number of 550 error lines. This bug was coincidentally fixed by
529 increasing the size of one of Exim's internal buffers (big_buffer) that
530 happened as part of the Exiscan merge. However, to be on the safe side, I
531 have made the code more robust (and fixed the comments that describe what
532 is going on).
533
55ee9ee3
PH
53453. Now that there can be additional text after "Completed" in log lines (if
535 the queue_time_overall log selector is set), a one-byte patch to exigrep
536 was needed to allow it to recognize "Completed" as not the last thing in
537 the line.
538
d38f8232
PH
53954. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
540 patch that reportedly fixes this has been added. I am not expert enough to
541 create a test for it. This is what the patch creator wrote:
542
543 "I found a little strange behaviour of ldap code when working with
544 Windows 2003 AD Domain, where users was placed in more than one
545 Organization Units. When I tried to give exim partial DN, the exit code
546 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
547 But simultaneously result of request was absolutely normal ldap result,
548 so I produce this patch..."
549
3295e65b
PH
550 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
551 REFERENCE, so I have modified the code to exclude the patch when that macro
552 is not defined.
553
7102e136
PH
55455. Some experimental protocols are using DNS PTR records for new purposes. The
555 keys for these records are domain names, not reversed IP addresses. The
b975ba52
PH
556 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
557 leaves it alone. Component reversal etc. now happens only for IP addresses.
ea3a6f44 558 CAN-2005-0021
7102e136 559
3ca0ba97
PH
56056. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
561
c2bcbe20
PH
56257. Double the size of the debug message buffer (to 2048) so that more of very
563 long debug lines gets shown.
564
18ce445d
PH
56558. The exicyclog utility now does better if the number of log files to keep
566 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
567
1f5b4c3d
PH
56859. Two changes related to the smtp_active_hostname option:
569
570 (1) $smtp_active_hostname is now available as a variable.
571 (2) The default for smtp_banner uses $smtp_active_hostname instead
572 of $primary_hostname.
573
b975ba52
PH
57460. The host_aton() function is supposed to be passed a string that is known
575 to be a valid IP address. However, in the case of IPv6 addresses, it was
576 not checking this. This is a hostage to fortune. Exim now panics and dies
577 if the condition is not met. A case was found where this could be provoked
85b87bc2
PH
578 from a dnsdb PTR lookup with an IPv6 address that had more than 8
579 components; fortuitously, this particular loophole had already been fixed
580 by change 4.50/55 above.
581
582 If there are any other similar loopholes, the new check in host_aton()
583 itself should stop them being exploited. The report I received stated that
584 data on the command line could provoke the exploit when Exim was running as
585 exim, but did not say which command line option was involved. All I could
586 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
587 running as the user.
ea3a6f44 588 CAN-2005-0021
85b87bc2
PH
589
59061. There was a buffer overflow vulnerability in the SPA authentication code
591 (which came originally from the Samba project). I have added a test to the
592 spa_base64_to_bits() function which I hope fixes it.
ea3a6f44 593 CAN-2005-0022
b975ba52 594
17ffcae7
PH
59562. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
596 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
597
d95f9fdb
PH
59863. The daemon start-up calls getloadavg() while still root for those OS that
599 need the first call to be done as root, but it missed one case: when
600 deliver_queue_load_max is set with deliver_drop_privilege. This is
601 necessary for the benefit of the queue runner, because there is no re-exec
602 when deliver_drop_privilege is set.
603
86b8287f
PH
60464. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
605 This has been fixed.
606
60dc5e56
PH
60765. Caching of lookup data for "hosts =" ACL conditions, when a named host list
608 was in use, was not putting the data itself into the right store pool;
609 consequently, it could be overwritten for a subsequent message in the same
610 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
611 the caching.)
612
533244af
PH
61366. Added hosts_max_try_hardlimit to the smtp transport, default 50.
614
a5a28604
PH
61567. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
616 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
617 the function were treating the return as a boolean value, which happened to
618 work because 0=false and not-0=true, but is not correct code.
619
7e634d24
PH
62068. The host_aton() function was not handling scoped IPv6 addresses (those
621 with, for example, "%eth0" on the end) correctly.
622
3e11c26b
PH
62369. Fixed some compiler warnings in acl.c for the bitmaps specified with
624 negated items (that is, ~something) in unsigned ints. Some compilers
625 apparently mutter when there is no cast.
626
6729cf78
PH
62770. If an address verification called from an ACL failed, and did not produce a
628 user-specific message (i.e. there was only a "system" message), nothing was
629 put in $acl_verify_message. In this situation, it now puts the system
630 message there.
631
00f00ca5
PH
63271. Change 4.23/11 added synchronization checking at the start of an SMTP
633 session; change 4.31/43 added the unwanted input to the log line - except
634 that it did not do this in the start of session case. It now does.
635
c9bdd01c
PH
63672. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
637 This is wrong and can cause the other end to generate a synchronization
638 error if it is another Exim or anything else that does the synchronization
639 check. A QUIT command is no longer sent after a timeout.
640
d43194df
PH
64173. $host_lookup_deferred has been added, to make it easier to detect DEFERs
642 during host lookups.
643
fe5b5d0b
PH
64474. The defer_ok option of callout verification was not working if it was used
645 when verifying addresses in header lines, that is, for this case:
646
647 verify = header_sender/callout=defer_ok
648
76a2d7ba
PH
64975. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
650 those file descriptors could be used for SMTP connections. If anything
651 wrote to stderr (the example that came up was "warn" in embedded Perl), it
652 could be sent to the SMTP client, causing chaos. The daemon now opens
653 stdin, stdout, and stderr to /dev/null when it puts itself into the
654 background.
655
65676. Arrange for output from Perl's "warn" command to be written to Exim's main
657 log by default. The user can override this with suitable Perl magic.
658
04f7d5b9
PH
65977. The use of log_message on a "discard" ACL verb, which is supposed to add to
660 the log message when discard triggers, was not working for the DATA ACL or
661 for the non-SMTP ACL.
662
bc60667e
PH
66378. Error message wording change in sieve.c.
664
bb6e88ff
PH
66579. If smtp_accept_max_per_host was set, the number of connections could be
666 restricted to fewer than expected, because the daemon was trying to set up
667 a new connection before checking whether the processes handling previous
668 connections had finished. The check for completed processes is now done
669 earlier. On busy systems, this bug wouldn't be noticed because something
670 else would have woken the daemon, and it would have reaped the completed
671 process earlier.
672
1e70f85b
PH
67380. If a message was submitted locally by a user whose login name contained one
674 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
675 It caused a spool format error. I have fixed the spool reading code. A
676 related problem was that the "from" clause in the Received: line became
677 illegal because of the space(s). It is now covered by ${quote_local_part.
678
67981. Included the latest eximstats from Steve (adds average sizes to HTML Top
680 tables).
681
4e01f9d6
PH
68282. Updated OS/Makefile-AIX as per message from Mike Meredith.
683
1ee1cef2
PH
68483. Patch from Sieve maintainer to fix unterminated string problem in
685 "vacation" handling.
686
6e2b4ccc
PH
68784. Some minor changes to the Linux configuration files to help with other
688 OS variants using glibc.
689
8e669ac1
PH
69085. One more patch for Sieve to update vacation handling to latest spec.
691
495ae4b0 692
bbe902f0
PH
693----------------------------------------------------
694See the note above about the 4.44 and 4.50 releases.
695----------------------------------------------------
696
697
698Exim version 4.44
699-----------------
700
701 1. Change 4.43/35 introduced a bug that caused file counts to be
702 incorrectly computed when quota_filecount was set in an appendfile
703 transport
704
705 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
706 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
707 place.
708
709 3. Give more explanation in the error message when the command for a transport
710 filter fails to execute.
711
712 4. There are several places where Exim runs a non-Exim command in a
713 subprocess. The SIGUSR1 signal should be disabled for these processes. This
714 was being done only for the command run by the queryprogram router. It is
715 now done for all such subprocesses. The other cases are: ${run, transport
716 filters, and the commands run by the lmtp and pipe transports.
717
718 5. Some older OS have a limit of 256 on the maximum number of file
719 descriptors. Exim was using setrlimit() to set 1000 as a large value
720 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
721 systems. I've change it so that if it can't get 1000, it tries for 256.
722
723 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
724 was an oversight, and furthermore, ever since the addition of extra
725 controls (e.g. 4.43/32), the checks on when to allow different forms of
726 "control" were broken. There should now be diagnostics for all cases when a
727 control that does not make sense is encountered.
728
729 7. $recipients is now available in the predata ACL (oversight).
730
731 8. Tidy the search cache before the fork to do a delivery from a message
732 received from the command line. Otherwise the child will trigger a lookup
733 failure and thereby defer the delivery if it tries to use (for example) a
734 cached ldap connection that the parent has called unbind on.
735
736 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
737 of $address_data from the recipient verification was clobbered by the
738 sender verification.
739
74010. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
741 was its contents. (It was OK if the option was not defined at all.)
742
74311. A "Completed" log line is now written for messages that are removed from
744 the spool by the -Mrm option.
745
74612. $host_address is now set to the target address during the checking of
747 ignore_target_hosts.
748
74913. When checking ignore_target_hosts for an ipliteral router, no host name was
750 being passed; this would have caused $sender_host_name to have been used if
751 matching the list had actually called for a host name (not very likely,
752 since this list is usually IP addresses). A host name is now passed as
753 "[x.x.x.x]".
754
75514. Changed the calls that set up the SIGCHLD handler in the daemon to use the
756 code that specifies a non-restarting handler (typically sigaction() in
757 modern systems) in an attempt to fix a rare and obscure crash bug.
758
75915. Narrowed the window for a race in the daemon that could cause it to ignore
760 SIGCHLD signals. This is not a major problem, because they are used only to
761 wake it up if nothing else does.
762
76316. A malformed maildirsize file could cause Exim to calculate negative values
764 for the mailbox size or file count. Odd effects could occur as a result.
765 The maildirsize information is now recalculated if the size or filecount
766 end up negative.
767
76817. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
769 support for a long time. Removed HAVE_SYS_VFS_H.
770
ea3a6f44 77118. Updated exipick to current release from John Jetmore.
bbe902f0
PH
772
77319. Allow an empty sender to be matched against a lookup in an address list.
774 Previously the only cases considered were a regular expression, or an
775 empty pattern.
776
77720. Exim went into a mad DNS lookup loop when doing a callout where the
778 host was specified on the transport, if the DNS lookup yielded more than
779 one IP address.
780
ea3a6f44
NM
78121. The RFC2047 encoding function was originally intended for short strings
782 such as real names; it was not keeping to the 75-character limit for
783 encoded words that the RFC imposes. It now respects the limit, and
784 generates multiple encoded words if necessary. To be on the safe side, I
785 have increased the buffer size for the ${rfc2047: expansion operator from
786 1024 to 2048 bytes.
bbe902f0 787
ea3a6f44
NM
78822. Failure to deliver a bounce message always caused it to be frozen, even if
789 there was an errors_to setting on the router. The errors_to setting is now
790 respected.
bbe902f0
PH
791
79223. If an IPv6 address is given for -bh or -bhc, it is now converted to the
793 canonical form (fully expanded) before being placed in
794 $sender_host_address.
795
79624. Updated eximstats to version 1.33
797
ea3a6f44
NM
79825. Include certificate and key file names in error message when GnuTLS fails
799 to set them up, because the GnuTLS error message doesn't include the name
800 of the failing file when there is a problem reading it.
bbe902f0
PH
801
80226. Expand error message when OpenSSL has problems setting up cert/key files.
ea3a6f44 803 As per change 25.
bbe902f0 804
ea3a6f44
NM
80527. Reset the locale to "C" after calling embedded Perl, in case it was changed
806 (this can affect the format of dates).
bbe902f0 807
ea3a6f44
NM
80828. exim_tidydb, when checking for the continued existence of a message for
809 which it has found a message-specific retry record, was not finding
810 messages that were in split spool directories. Consequently, it was
811 deleting retry records that should have stayed in existence.
bbe902f0
PH
812
81329. eximstats updated to version 1.35
814 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
815 1.35 - bugfix such that pie charts by volume are generated correctly
816
ea3a6f44
NM
81730. The SPA authentication driver was not abandoning authentication and moving
818 on to the next authenticator when an expansion was forced to fail,
819 contradicting the general specification for all authenticators. Instead it
820 was generating a temporary error. It now behaves as specified.
bbe902f0 821
ea3a6f44
NM
82231. The default ordering of permitted cipher suites for GnuTLS was pessimal
823 (the order specifies the preference for clients). The order is now AES256,
824 AES128, 3DES, ARCFOUR128.
bbe902f0 825
ea3a6f44
NM
82631. Small patch to Sieve code - explicitly set From: when generating an
827 autoreply.
bbe902f0 828
ea3a6f44
NM
82932. Exim crashed if a remote delivery caused a very long error message to be
830 recorded - for instance if somebody sent an entire SpamAssassin report back
831 as a large number of 550 error lines. This bug was coincidentally fixed by
832 increasing the size of one of Exim's internal buffers (big_buffer) that
833 happened as part of the Exiscan merge. However, to be on the safe side, I
834 have made the code more robust (and fixed the comments that describe what
835 is going on).
bbe902f0 836
ea3a6f44
NM
83733. Some experimental protocols are using DNS PTR records for new purposes. The
838 keys for these records are domain names, not reversed IP addresses. The
839 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
840 leaves it alone. Component reversal etc. now happens only for IP addresses.
bbe902f0
PH
841 CAN-2005-0021
842
ea3a6f44
NM
84334. The host_aton() function is supposed to be passed a string that is known
844 to be a valid IP address. However, in the case of IPv6 addresses, it was
845 not checking this. This is a hostage to fortune. Exim now panics and dies
846 if the condition is not met. A case was found where this could be provoked
847 from a dnsdb PTR lookup with an IPv6 address that had more than 8
848 components; fortuitously, this particular loophole had already been fixed
849 by change 4.50/55 or 4.44/33 above.
850
851 If there are any other similar loopholes, the new check in host_aton()
852 itself should stop them being exploited. The report I received stated that
853 data on the command line could provoke the exploit when Exim was running as
854 exim, but did not say which command line option was involved. All I could
855 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
856 running as the user.
bbe902f0
PH
857 CAN-2005-0021
858
ea3a6f44
NM
85935. There was a buffer overflow vulnerability in the SPA authentication code
860 (which came originally from the Samba project). I have added a test to the
861 spa_base64_to_bits() function which I hope fixes it.
bbe902f0
PH
862 CAN-2005-0022
863
ea3a6f44
NM
86436. The daemon start-up calls getloadavg() while still root for those OS that
865 need the first call to be done as root, but it missed one case: when
866 deliver_queue_load_max is set with deliver_drop_privilege. This is
867 necessary for the benefit of the queue runner, because there is no re-exec
868 when deliver_drop_privilege is set.
bbe902f0 869
ea3a6f44
NM
87037. Caching of lookup data for "hosts =" ACL conditions, when a named host list
871 was in use, was not putting the data itself into the right store pool;
872 consequently, it could be overwritten for a subsequent message in the same
873 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
874 the caching.)
bbe902f0 875
ea3a6f44
NM
87638. Sometimes the final signoff response after QUIT could fail to get
877 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
878 < 0 before doing a fflush(). This bug looks as though it goes back to the
879 introduction of TLS in release 3.20, but "sometimes" must have been rare
880 because the tests only now provoked it.
bbe902f0
PH
881
882
495ae4b0
PH
883Exim version 4.43
884-----------------
885
886 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
887 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
888 Now it returns an integer. A number of calls were still expecting a T/F
889 return. Fortuitously, in all cases, the tests worked in OK situations,
890 which is the norm. However, things would have gone wrong on any write
891 failures on the smtp file descriptor. This function is used when sending
892 messages over SMTP and also when doing verify callouts.
893
894 2. When Exim is called to do synchronous delivery of a locally submitted
895 message (the -odf or -odi options), it no longer closes stderr before doing
896 the delivery.
897
898 3. Implemented the mua_wrapper option.
899
900 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
901
902 5. Implemented the functions header_remove(), header_testname(),
903 header_add_at_position(), and receive_remove_recipient(), and exported them
904 to local_scan().
905
906 6. If an ACL "warn" statement specified the addition of headers, Exim already
907 inserted X-ACL-Warn: at the start if there was no header name. However, it
908 was not making this test for the second and subsequent header lines if
909 there were newlines in the string. This meant that an invalid header could
910 be inserted if Exim was badly configured.
911
912 7. Allow an ACL "warn" statement to add header lines at the start or after all
913 the Received: headers, as well as at the end.
914
915 8. Added the rcpt_4xx retry error code.
916
917 9. Added postmaster_mailfrom=xxx to callout verification option.
918
91910. Added mailfrom=xxxx to the callout verification option, for verify=
920 header_sender only.
921
92211. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
923 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
924
92512. Inserted some casts to stop certain compilers warning when using pointer
926 differences as field lengths or precisions in printf-type calls (mostly
927 affecting debugging statements).
928
92913. Added optional readline() support for -be (dynamically loaded).
930
93114. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
932 same clock tick as a message's arrival, so that its received time was the
933 same as the "first fail" time on the retry record, and that message
934 remained on the queue past the ultimate address timeout, every queue runner
935 would try a delivery (because it was past the ultimate address timeout) but
936 after another failure, the ultimate address timeout, which should have then
937 bounced the address, did not kick in. This was a "< instead of <=" error;
938 in most cases the first failure would have been in the next clock tick
939 after the received time, and all would be well.
940
94115. The special items beginning with @ in domain lists (e.g. @mx_any) were not
942 being recognized when the domain list was tested by the match_domain
943 condition in an expansion string.
944
94516. Added the ${str2b64: operator.
946
94717. Exim was always calling setrlimit() to set a large limit for the number of
948 processes, without checking whether the existing limit was already
949 adequate. (It did check for the limit on file descriptors.) Furthermore,
950 errors from getrlimit() and setrlimit() were being ignored. Now they are
951 logged to the main and panic logs, but Exim does carry on, to try to do its
952 job under whatever limits there are.
953
95418. Imported PCRE 5.0.
955
95619. Trivial typo in log message " temporarily refused connection" (the leading
957 space).
958
95920. If the log selector return_path_on_delivery was set and an address was
960 redirected to /dev/null, the delivery process crashed because it assumed
961 that a return path would always be set for a "successful" delivery. In this
962 case, the whole delivery is bypassed as an optimization, and therefore no
963 return path is set.
964
96521. Internal re-arrangement: the function for sending a challenge and reading
966 a response while authentication was assuming a zero-terminated challenge
967 string. It's now changed to take a pointer and a length, to allow for
968 binary data in such strings.
969
97022. Added the cyrus_sasl authenticator (code supplied by MBM).
971
97223. Exim was not respecting finduser_retries when seeking the login of the
973 uid under which it was called; it was always trying 10 times. (The default
974 setting of finduser_retries is zero.) Also, it was sleeping after the final
975 failure, which is pointless.
976
97724. Implemented tls_on_connect_ports.
978
97925. Implemented acl_smtp_predata.
980
98126. If the domain in control=submission is set empty, Exim assumes that the
982 authenticated id is a complete email address when it generates From: or
983 Sender: header lines.
984
98527. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
986 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
987 chown and chgrp in /bin and hostname in /usr/bin.
988
98928. Exim was keeping the "process log" file open after each use, just as it
990 does for the main log. This opens the possibility of it remaining open for
991 long periods when the USR1 signal hits a daemon. Occasional processlog
992 errors were reported, that could have been caused by this. Anyway, it seems
993 much more sensible not to leave this file open at all, so that is what now
994 happens.
995
99629. The long-running daemon process does not normally write to the log once it
997 has entered its main loop, and it closes the log before doing so. This is
998 so that log files can straightforwardly be renamed and moved. However,
999 there are a couple of unusual error situations where the daemon does write
1000 log entries, and I had neglected to close the log afterwards.
1001
100230. The text of an SMTP error response that was received during a remote
1003 delivery was being truncated at 512 bytes. This is too short for some of
1004 the long messages that one sometimes sees. I've increased the limit to
1005 1024.
1006
100731. It is now possible to make retry rules that apply only when a message has a
1008 specific sender, in particular, an empty sender.
1009
101032. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1011 it possible to be selective about when SMTP synchronization is enforced.
1012
101333. Added "control = caseful_local_part" and "control = "caselower_local_part".
1014
101532. Implemented hosts_connection_nolog.
1016
101733. Added an ACL for QUIT.
1018
101934. Setting "delay_warning=" to disable warnings was not working; it gave a
1020 syntax error.
1021
102235. Added mailbox_size and mailbox_filecount to appendfile.
1023
102436. Added control = no_multiline_responses to ACLs.
1025
102637. There was a bug in the logic of the code that waits for the clock to tick
1027 in the case where the clock went backwards by a substantial amount such
1028 that the microsecond fraction of "now" was more than the microsecond
1029 fraction of "then" (but the whole seconds number was less).
1030
103138. Added support for the libradius Radius client library this is found on
1032 FreeBSD (previously only the radiusclient library was supported).
1033
1034
1035Exim version 4.42
1036-----------------
1037
1038 1. When certain lookups returned multiple values in the form name=value, the
1039 quoting of the values was not always being done properly. Specifically:
1040 (a) If the value started with a double quote, but contained no whitespace,
1041 it was not quoted.
1042 (b) If the value contained whitespace other than a space character (i.e.
1043 tabs or newlines or carriage returns) it was not quoted.
1044 This fix has been applied to the mysql and pgsql lookups by writing a
1045 separate quoting function and calling it from the lookup code. The fix
1046 should probably also be applied to nisplus, ibase and oracle lookups, but
1047 since I cannot test any of those, I have not disturbed their existing code.
1048
1049 2. A hit in the callout cache for a specific address caused a log line with no
1050 reason for rejecting RCPT. Now it says "Previous (cached) callout
1051 verification failure".
1052
1053 3. There was an off-by-one bug in the queryprogram router. An over-long
1054 return line was truncated at 256 instead of 255 characters, thereby
1055 overflowing its buffer with the terminating zero. As well as fixing this, I
1056 have increased the buffer size to 1024 (and made a note to document this).
1057
1058 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1059 when Exim is waiting for an SMTP response from a remote server, Exim
1060 restarts its select() call on the socket, thereby resetting its timeout.
1061 This is not a problem when such interrupts are rare. Somebody set up a cron
1062 job to run exiwhat every 2 minutes, which is less than the normal select()
1063 timeout (5 or 10 minutes). This meant that the select() timeout never
1064 kicked in because it was always reset. I have fixed this by comparing the
1065 time when an interrupt arrives with the time at the start of the first call
1066 to select(). If more time than the timeout has elapsed, the interrupt is
1067 treated as a timeout.
1068
1069 5. Some internal re-factoring in preparation for the addition of Sieve
1070 extensions (by MH). In particular, the "personal" test is moved to a
1071 separate function, and given an option for scanning Cc: and Bcc: (which is
1072 not set for Exim filters).
1073
1074 6. When Exim created an email address using the login of the caller as the
1075 local part (e.g. when creating a From: or Sender: header line), it was not
1076 quoting the local part when it contained special characters such as @.
1077
1078 7. Installed new OpenBSD configuration files.
1079
1080 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1081 try to make them clearer.
1082
1083 9. Callout options, other than the timeout value, were being ignored when
1084 verifying sender addresses in header lines. For example, when using
1085
1086 verify = header_sender/callout=no_cache
1087
1088 the cache was (incorrectly) being used.
1089
109010. Added a missing instance of ${EXE} to the exim_install script; this affects
1091 only the Cygwin environment.
1092
109311. When return_path_on_delivery was set as a log selector, if different remote
1094 addresses in the same message used different return paths and parallel
1095 remote delivery occurred, the wrong values would sometimes be logged.
1096 (Whenever a remote delivery process finished, the return path value from
1097 the most recently started remote delivery process was logged.)
1098
109912. RFC 3848 specifies standard names for the "with" phrase in Received: header
1100 lines when AUTH and/or TLS are in use. This is the "received protocol"
1101 field. Exim used to use "asmtp" for authenticated SMTP, without any
1102 indication (in the protocol name) for TLS use. Now it follows the RFC and
1103 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1104 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1105 names appear in log lines as well as in Received: header lines.
1106
110713. Installed MH's patches for Sieve to add the "copy" and "vacation"
1108 extensions, and comparison tests, and to fix some bugs.
1109
111014. Changes to the "personal" filter test:
1111
1112 (1) The test was buggy in that it was just doing the equivalent of
1113 "contains" tests on header lines. For example, if a user's address was
1114 anne@some.where, the "personal" test would incorrectly be true for
1115
1116 To: susanne@some.where
1117
1118 This test is now done by extracting each address from the header in turn,
1119 and checking the entire address. Other tests that are part of "personal"
1120 are now done using regular expressions (for example, to check local parts
1121 of addresses in From: header lines).
1122
1123 (2) The list of non-personal local parts in From: addresses has been
1124 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1125 taken from the Sieve specification recommendations.
1126
1127 (3) If the message contains any header line starting with "List-" it is
1128 treated as non-personal.
1129
1130 (4) The test for "circular" in the Subject: header line has been removed
1131 because it now seems ill-conceived.
1132
113315. Minor typos in src/EDITME comments corrected.
1134
113516. Installed latest exipick from John Jetmore.
1136
113717. If headers_add on a router specified a text string that was too long for
1138 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1139 of string_sprintf() is now avoided.
1140
114118. $message_body_size was not set (it was always zero) when running the DATA
1142 ACL and the local_scan() function.
1143
114419. For the "mail" command in an Exim filter, no default was being set for
1145 the once_repeat time, causing a random time value to be used if "once" was
1146 specified. (If the value happened to be <= 0, no repeat happened.) The
1147 default is now 0s, meaning "never repeat". The "vacation" command was OK
1148 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1149 (I found it when inspecting the code).
1150
115120. There is now an overall timeout for performing a callout verification. It
1152 defaults to 4 times the callout timeout, which applies to individual SMTP
1153 commands during the callout. The overall timeout applies when there is more
1154 than one host that can be tried. The timeout is checked before trying the
1155 next host. This prevents very long delays if there are a large number of
1156 hosts and all are timing out (e.g. when the network connections are timing
1157 out). The value of the overall timeout can be changed by specifying an
1158 additional sub-option for "callout", called "maxwait". For example:
1159
1160 verify = sender/callout=5s,maxwait=20s
1161
116221. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1163 to the end before writing, but this should make it even safer).
1164
116522. Exim was forgetting that it had advertised PIPELINING for the second and
1166 subsequent messages on an SMTP connection. It was also not resetting its
1167 memory on STARTTLS and an internal HELO.
1168
116923. When Exim logs an SMTP synchronization error within a session, it now
1170 records whether PIPELINING has been advertised or not.
1171
117224. Added 3 instances of "(long int)" casts to time_t variables that were being
1173 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1174 rather than long int.
1175
117625. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1177
117826. Added the never_mail option to autoreply.
1179
1180
1181Exim version 4.41
1182-----------------
1183
1184 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1185 crash if the getsockname() call failed; this can happen if a connection is
1186 closed very soon after it is established. The problem was simply in the
1187 order in which certain operations were done, causing Exim to try to write
1188 to the SMTP stream before it had set up the file descriptor. The bug has
1189 been fixed by making things happen in the correct order.
1190
1191
1192Exim version 4.40
1193-----------------
1194
1195 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1196 before the connection was closed, thus losing the rejection response.
1197
1198 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1199 some early Solaris releases, but causes trouble in current releases where
1200 socklen_t is defined.
1201
1202 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1203 always exist.
1204
1205 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1206 configured.
1207
1208 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1209 that releasing the top of it at the end releases what was used for sub-
1210 expansions (unless the block got too big). However, discard this block if
1211 the first thing is a variable or header, so that we can use its block when
1212 it is dynamic (useful for very large $message_headers, for example).
1213
1214 6. Lookups now cache *every* query, not just the most recent. A new, separate
1215 store pool is used for this. It can be recovered when all lookup caches are
1216 flushed. Lookups now release memory at the end of their result strings.
1217 This has involved some general refactoring of the lookup sources.
1218
1219 7. Some code has been added to the store_xxx() functions to reduce the amount
1220 of flapping under certain conditions.
1221
1222 8. log_incoming_interface used to affect only the <= reception log lines. Now
1223 it causes the local interface and port to be added to several more SMTP log
1224 lines, for example "SMTP connection from", and rejection lines.
1225
1226 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1227
122810. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1229
123011. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1231 could be overwritten at the end of the current message (or the start of a
1232 new message if it was set in a HELO ACL). The value is now preserved for
1233 the duration of the SMTP connection.
1234
123512. If a transport had a headers_rewrite setting, and a matching header line
1236 contained an unqualified address, that address was qualified, even if it
1237 did not match any rewriting rules. The underlying bug was that the values
1238 of the flags that permit the existence of unqualified sender and recipient
1239 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1240 non-local messages, and by -bnq for local messages) were not being
1241 preserved with the message after it was received.
1242
124313. When Exim was logging an SMTP synchronization error, it could sometimes log
1244 "next input=" as part of the text comprising the host identity instead of
1245 the correct text. The code was using the same buffer for two different
1246 strings. However, depending on which order the printing function evaluated
1247 its arguments, the bug did not always show up. Under Linux, for example, my
1248 test suite worked just fine.
1249
125014. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1251 doesn't work with some older versions of Perl. It has been changed to "my",
1252 which in any case is probably the better facility to use.
1253
125415. A really picky compiler found some instances of statements for creating
1255 error messages that either had too many or two few arguments for the format
1256 string.
1257
125816. The size of the buffer for calls to the DNS resolver has been increased
1259 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1260 for addresses that have a lot of PTR records. This alleviates a problem; it
1261 does not fully solve it.
1262
126317. A dnsdb lookup for PTR records that receives more data than will fit in the
1264 buffer now truncates the list and logs the incident, which is the same
1265 action as happens when Exim is looking up a host name and its aliases.
1266 Previously in this situation something unpredictable would happen;
1267 sometimes it was "internal error: store_reset failed".
1268
126918. If a server dropped the connection unexpectedly when an Exim client was
1270 using GnuTLS and trying to read a response, the client delivery process
1271 crashed while trying to generate an error log message.
1272
127319. If a "warn" verb in an ACL added multiple headers to a message in a single
1274 string, for example:
1275
1276 warn message = H1: something\nH2: something
1277
1278 the text was added as a single header line from Exim's point of view
1279 though it ended up OK in the delivered message. However, searching for the
1280 second and subsequent header lines using $h_h2: did not work. This has been
1281 fixed. Similarly, if a system filter added multiple headers in this way,
1282 the routers could not see them.
1283
128420. Expanded the error message when iplsearch is called with an invalid key to
1285 suggest using net-iplsearch in a host list.
1286
128721. When running tests using -bh, any delays imposed by "delay" modifiers in
1288 ACLs are no longer actually imposed (and a message to that effect is
1289 output).
1290
129122. If a "gecos" field in a passwd entry contained escaped characters, in
1292 particular, if it contained a \" sequence, Exim got it wrong when building
1293 a From: or a Sender: header from that name. A second bug also caused
1294 incorrect handling when an unquoted " was present following a character
1295 that needed quoting.
1296
129723. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1298 was not being matched caselessly.
1299
130024. Arranged for all hyphens in the exim.8 source to be escaped with
1301 backslashes.
1302
130325. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1304 properly. Recipient callout cache records were still being keyed to include
1305 the sender, even when use_sender was set false. This led to far more
1306 callouts that were necessary. The sender is no longer included in the key
1307 when use_sender is false.
1308
130926. Added "control = submission" modifier to ACLs.
1310
131127. Added the ${base62d: operator to decode base 62 numbers.
1312
131328. dnsdb lookups can now access SRV records.
1314
131529. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1316 the configuration file.
1317
131830. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1319 (-v) mode. This makes the output for a verbose queue run more intelligible.
1320
132131. Added a use_postmaster feature to recipient callouts.
1322
132332. Added the $body_zerocount variable, containing the number of binary zero
1324 bytes in the message body.
1325
132633. The time of last modification of the "new" subdirectory is now used as the
1327 "mailbox time last read" when there is a quota error for a maildir
1328 delivery.
1329
133034. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1331
133235. Added +ignore_unknown as a special item in host lists.
1333
133436. Code for decoding IPv6 addresses in host lists is now included, even if
1335 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1336 address was recognized as an IP address, but was then not correctly decoded
1337 into binary, causing unexpected and incorrect effects when compared with
1338 another IP address.
1339
1340
1341Exim version 4.34
1342-----------------
1343
1344 1. Very minor rewording of debugging text in manualroute to say "list of
1345 hosts" instead of "hostlist".
1346
1347 2. If verify=header_syntax was set, and a header line with an unqualified
1348 address (no domain) and a large number of spaces between the end of the
1349 name and the colon was received, the reception process suffered a buffer
1350 overflow, and (when I tested it) crashed. This was caused by some obsolete
1351 code that should have been removed. The fix is to remove it!
1352
1353 3. When running in the test harness, delay a bit after writing a bounce
1354 message to get a bit more predictability in the log output.
1355
1356 4. Added a call to search_tidyup() just before forking a reception process. In
1357 theory, someone could use a lookup in the expansion of smtp_accept_max_
1358 per_host which, without the tidyup, could leave open a database connection.
1359
1360 5. Added the variables $recipient_data and $sender_data which get set from a
1361 lookup success in an ACL "recipients" or "senders" condition, or a router
1362 "senders" option, similar to $domain_data and $local_part_data.
1363
1364 6. Moved the writing of debug_print from before to after the "senders" test
1365 for routers.
1366
1367 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1368 problems for message scanning, either using a data ACL, or using
1369 local_scan() because the Received: header was not generated till after they
1370 were called (in order to set the time as the time of reception completion).
1371 I have revised the way this works. The header is now generated after the
1372 body is received, but before the ACL or local_scan() are called. After they
1373 are run, the timestamp in the header is updated.
1374
1375
1376Exim version 4.33
1377-----------------
1378
1379 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1380 before starting a queue runner without re-exec. This happened only when
1381 deliver_drop_privilege was set or when the Exim user was set to root. The
1382 effect of the bug was that timeouts during subsequent deliveries caused
1383 crashes instead of being properly handled. The handler is now left at its
1384 default (and expected) setting.
1385
1386 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1387 message, again when deliver_drop_privilege is set or Exim is run as root.
1388 The bug described in (1) was not present in this case, but the tidying up
1389 of the other signals was missing. I have made the two cases consistent.
1390
1391 3. The ignore_target_hosts setting on a manualroute router was being ignored
1392 for hosts that were looked up using the /MX notation.
1393
1394 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1395 in domain lists.
1396
1397 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1398 operated on the sender address. After changing the $sender_address to <>
1399 for the sender address verify, Exim was re-instated it as the original
1400 (before rewriting) address, but remembering that it had rewritten it, so it
1401 wasn't rewriting it again. This bug also had the effect of breaking the
1402 sender address verification caching when the sender address was rewritten.
1403
1404 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1405 This has been changed so that if the ip literal address matches
1406 ignore_target_hosts, the router declines.
1407
1408 7. Added expansion conditions match_domain, match_address, and match_local_
1409 part (NOT match_host).
1410
1411 8. The placeholder for the Received: header didn't have a length field set.
1412
1413 9. Added code to Exim itself and to exim_lock to test for a specific race
1414 condition that could lead to file corruption when using MBX delivery. The
1415 issue is with the lockfile that is created in /tmp. If this file is removed
1416 after a process has opened it but before that process has acquired a lock,
1417 there is the potential for a second process to recreate the file and also
1418 acquire a lock. This could lead to two Exim processes writing to the file
1419 at the same time. The added code performs the same test as UW imapd; it
1420 checks after acquiring the lock that its file descriptor still refers to
1421 the same named file.
1422
142310. The buffer for building added header lines was of fixed size, 8192 bytes.
1424 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1425 when Exim is built.
1426
142711. Added the smtp_active_hostname option. If used, this will typically be made
1428 to depend on the incoming interface address. Because $interface_address is
1429 not set up until the daemon has forked a reception process, error responses
1430 that can happen earlier (such as "too many connections") no longer contain
1431 a host name.
1432
143312. If an expansion in a condition on a "warn" statement fails because a lookup
1434 defers, the "warn" statement is abandoned, and the next ACL statement is
1435 processed. Previously this caused the whole ACL to be aborted.
1436
143713. Added the iplsearch lookup type.
1438
143914. Added ident_timeout as a log selector.
1440
144115. Added tls_certificate_verified as a log selector.
1442
144316. Added a global option tls_require_ciphers (compare the smtp transport
1444 option of the same name). This controls incoming TLS connections.
1445
144617. I finally figured out how to make tls_require_ciphers do a similar thing
1447 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1448 before starting the TLS session.
1449
145018. Tabs are now shown as \t in -bP output.
1451
145219. If the log selector return_path_on_delivery was set, Exim crashed when
1453 bouncing a message because it had too many Received: header lines.
1454
145520. If two routers both had headers_remove settings, and the first one included
1456 a superfluous trailing colon, the final name in the first list and the
1457 first name in the second list were incorrectly joined into one item (with a
1458 colon in the middle).
1459
1460
1461Exim version 4.32
1462-----------------
1463
1464 1. Added -C and -D options to the exinext utility, mainly to make it easier
1465 to include in the automated testing, but these could be helpful when
1466 multiple configurations are in use.
1467
1468 2. The exinext utility was not formatting the output nicely when there was
1469 an alternate port involved in the retry record key, nor when there was a
1470 message id as well (for retries that were specific to a specific message
1471 and a specific host). It was also confused by IPv6 addresses, because of
1472 the additional colons they contain. I have fixed the IPv4 problem, and
1473 patched it up to do a reasonable job for IPv6.
1474
1475 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1476 delivery, the log line now contains "pipelined" if PIPELINING was used.
1477
1478 4. An SMTP transport process used to panic and die if the bind() call to set
1479 an explicit outgoing interface failed. This has been changed; it is now
1480 treated in the same way as a connect() failure.
1481
1482 5. A reference to $sender_host_name in the part of a conditional expansion
1483 that was being skipped was still causing a DNS lookup. This no longer
1484 occurs.
1485
1486 6. The def: expansion condition was not recognizing references to header lines
1487 that used bh_ and bheader_.
1488
1489 7. Added the _cache feature to named lists.
1490
1491 8. The code for checking quota_filecount in the appendfile transport was
1492 allowing one more file than it should have been.
1493
1494 9. For compatibility with Sendmail, the command line option
1495
1496 -prval:sval
1497
1498 is equivalent to
1499
1500 -oMr rval -oMs sval
1501
1502 and sets the incoming protocol and host name (for trusted callers). The
1503 host name and its colon can be omitted when only the protocol is to be set.
1504 Note the Exim already has two private options, -pd and -ps, that refer to
1505 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1506 "s", but I don't think that's a major issue.
1507
150810. A number of refactoring changes to the code, none of which should affect
1509 Exim's behaviour:
1510
1511 (a) The number of logging options was getting close to filling up the
1512 32-bit word that was used as a bit map. I have split them into two classes:
1513 those that are passed in the argument to log_write(), and those that are
1514 only ever tested independently outside of that function. These are now in
1515 separate 32-bit words, so there is plenty of room for expansion again.
1516 There is no change in the user interface or the logging behaviour.
1517
1518 (b) When building, for example, log lines, the code previously used a
1519 macro that called string_cat() twice, in order to add two strings. This is
1520 not really sufficiently general. Furthermore, there was one instance where
1521 it was actually wrong because one of the argument was used twice, and in
1522 one call a function was used. (As it happened, calling the function twice
1523 did not affect the overall behaviour.) The macro has been replaced by a
1524 function that can join an arbitrary number of extra strings onto a growing
1525 string.
1526
1527 (c) The code for expansion conditions now uses a table and a binary chop
1528 instead of a serial search (which was left over from when there were very
1529 few conditions). Also, it now recognizes conditions like "pam" even when
1530 the relevant support is not compiled in: a suitably worded error message is
1531 given if an attempt is made to use such a condition.
1532
153311. Added ${time_interval:xxxxx}.
1534
153512. A bug was causing one of the ddress fields not to be passed back correctly
1536 from remote delivery subprocesses. The field in question was not being
1537 subsequently used, so this caused to problems in practice.
1538
153913. Added new log selectors queue_time and deliver_time.
1540
154114. Might have fixed a bug in maildirsizefile handling that threw up
1542 "unexpected character" debug warnings, and recalculated the data
1543 unnecessarily. In any case, I expanded the warning message to give more
1544 information.
1545
154615. Added the message "Restricted characters in address" to the statements in
1547 the default ACL that block characters like @ and % in local parts.
1548
154916. Change 71 for release 4.31 proved to be much less benign that I imagined.
1550 Three changes have been made:
1551
1552 (a) There was a serious bug; a negative response to MAIL caused the whole
1553 recipient domain to be cached as invalid, thereby blocking all messages
1554 to all local parts at the same domain, from all senders. This bug has
1555 been fixed. The domain is no longer cached after a negative response to
1556 MAIL if the sender used is not empty.
1557
1558 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1559 been restored.
1560
1561 (c) A new callout option, "use_sender" has been added for people who want
1562 the modified behaviour.
1563
1564
1565Exim version 4.31
1566-----------------
1567
1568 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1569 Larry Rosenman.
1570
1571 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1572 indeed breaks things for older releases.
1573
1574 3. Added additional logging to the case where there is a problem reading data
1575 from a filter that is running in a subprocess using a pipe, in order to
1576 try to track down a specific problem.
1577
1578 4. Testing facility fudge: when running in the test harness and attempting
1579 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1580 getting "No route to host". Convert this to a timeout.
1581
1582 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1583 warning.
1584
1585 6. Some OS don't have socklen_t but use size_t instead. This affects the
1586 fifth argument of getsockopt() amongst other things. This is now
1587 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1588 can be set for individual OS. I have set it for SunOS5, OSF1, and
1589 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1590 some earlier ones do not.
1591
1592 7. Change 4.30/15 was not doing the test caselessly.
1593
1594 8. The standard form for an IPv6 address literal was being rejected by address
1595 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1596 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1597 this, as well as the form without the "IPv6" on the front (but only when
1598 address literals are enabled, of course).
1599
1600 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1601
160210. Exim crashed if a message with an empty sender address specified by -f
1603 encountered a router with an errors_to setting. This could be provoked only
1604 by a command such as
1605
1606 exim -f "" ...
1607
1608 where an empty string was supplied; "<>" did not hit this bug.
1609
161011. Installed PCRE release 4.5.
1611
161212. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1613 remained set. It is now erased.
1614
161513. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1616 times from message ids (which are base 36 rather than the normal 62).
1617
161814. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1619 were being counted as actual protocol errors, and logged if the log
1620 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1621 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1622 and DATA following a set of rejected RCPTs do not count as protocol errors.
1623 In other words, Exim assumes they were pipelined, though this may not
1624 actually be the case. Of course, in all cases the client gets an
1625 appropriate error code.
1626
162715. If a lookup fails in an ACL condition, a message about the failure may
1628 be available; it is used if testing the ACL cannot continue, because most
1629 such messages specify what the cause of the deferral is. However, some
1630 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1631 that caused an old message to be retained and used if a later statement
1632 caused a defer, replacing the real cause of the deferral.
1633
163416. If an IP address had so many PTR records that the DNS lookup buffer
1635 was not large enough to hold them, Exim could crash while trying to process
1636 the truncated data. It now detects and logs this case.
1637
163817. Further to 4.21/58, another change has been made: if (and only if) the
1639 first line of a message (the first header line) ends with CRLF, a bare LF
1640 in a subsequent header line has a space inserted after it, so as not to
1641 terminate the header.
1642
164318. Refactoring: tidied an ugly bit of code in appendfile that copied data
1644 unnecessarily, used atoi() instead of strtol(), and didn't check the
1645 termination when getting file sizes from file names by regex.
1646
164719. Completely re-implemented the support for maildirsize files, in the light
1648 of a number of problems with the previous contributed implementation
1649 (4.30/29). In particular:
1650
1651 . If the quota is zero, the maildirsize file is maintained, but no quota is
1652 imposed.
1653
1654 . If the maildir directory does not exist, it is created before any attempt
1655 to write a maildirsize file.
1656
1657 . The quota value in the file is just a cache; if the quota is changed in
1658 the transport, the new value overrides.
1659
1660 . A regular expression is available for excluding directories from the
1661 count.
1662
166320. The autoreply transport checks the characters in options that define the
1664 message's headers; it allows continued headers, but it was checking with
1665 isspace() after an embedded newline instead of explicitly looking for a
1666 space or a tab.
1667
166821. If all the "regular" hosts to which an address was routed had passed their
1669 expiry times, and had not reached their retry times, the address was
1670 bounced, even if fallback hosts were defined. Now Exim should go on to try
1671 the fallback hosts.
1672
167322. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1674 equivalent code in the SMTP transport. Some hosts send humungous responses
1675 to HELO/EHLO, more than 1024 it seems.
1676
167723. Refactoring: code in filter.c used (void *) for "any old type" but this
1678 gives compiler warnings in some environments. I've now done it "properly",
1679 using a union.
1680
168124. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1682 (because of problems with the built-in one) was declared to return uschar *
1683 instead of char *, causing compiler failure.
1684
168525. Fixed a file descriptor leak when processing alias/forward files.
1686
168726. Fixed a minor format string issue in dbfn.c.
1688
168927. Typo in exim.c: ("dmbnz" for "dbmnz").
1690
169128. If a filter file refered to $h_xxx or $message_headers, and the headers
1692 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1693 become corrupted.
1694
169529. When a sender address is verified, it is cached, to save repeating the test
1696 when there is more than one recipient in a message. However, when the
1697 verification involves a callout, it is possible for different callout
1698 options to be set for different recipients. It is too complicated to keep
1699 track of this in the cache, so now Exim always runs a verification when a
1700 callout is required, relying on the callout cache for the optimization.
1701 The overhead is duplication of the address routing, but this should not be
1702 too great.
1703
170430. Fixed a bug in callout caching. If a RCPT command caused the sender address
1705 to be verified with callout=postmaster, and the main callout worked but the
1706 postmaster check failed, the verification correctly failed. However, if a
1707 subsequent RCPT command asked for sender verification *without* the
1708 postmaster check, incorrect caching caused this verification also to fail,
1709 incorrectly.
1710
171131. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1712 it was not caching the DNS options (qualify_single, search_parents) that
1713 were used when the lookup failed. A subsequent lookup with different
1714 options therefore always gave the same answer, though there were cases
1715 where it should not have. (Example: a "domains = !$mx_any" option on a
1716 dnslookup router: the "domains" option is always processed without any
1717 widening, but the router might have qualify_single set.) Now Exim uses the
1718 cached value only when the same options are set.
1719
172032. Added John Jetmore's "exipick" utility to the distribution.
1721
172233. GnuTLS: When an attempt to start a TLS session fails for any reason other
1723 than a timeout (e.g. a certificate is required, and is not provided), an
1724 Exim server now closes the connection immediately. Previously it waited for
1725 the client to close - but if the client is SSL, it seems that they each
1726 wait for each other, leading to a delay before one of them times out.
1727
172834: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1729 maintained 0.8.x compatibility because I don't think many are using it, and
1730 it is clearly obsolete.
1731
173235. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1733 transport.
1734
173536. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1736 client certificate was expired. A simple patch fixes this, though I don't
1737 understand the full logic of why the verify callback is called multiple
1738 times.
1739
174037. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1741 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1742 which causes problems with some clients (such as the Certicom SSL Plus
1743 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1744 disables the coutermeasure allowing Eudora to connect."
1745
174638. Exim was not checking that a write() to a log file succeeded. This could
1747 lead to Bad Things if a log got too big, in particular if it hit a file
1748 size limit. Exim now panics and dies if it cannot write to a log file, just
1749 as it does if it cannot open a log file.
1750
175139. Modified OS/Makefile-Linux so that it now contains
1752
1753 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1754
1755 The two -D definitions ensure that Exim is compiled with large file
1756 support, which makes it possible to handle log files that are bigger than
1757 2^31.
1758
175940. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1760 instance) a domain was checked against a named list that involved a lookup,
1761 causing $domain_data to be set, then another domain was checked against the
1762 same list, then the first domain was re-checked, the value of $domain_data
1763 after the final check could be wrong. In particular, if the second check
1764 failed, it could be set empty. This bug probably also applied to
1765 $localpart_data.
1766
176741. The strip_trailing_dot option was not being applied to the address given
1768 with the -f command-line option.
1769
177042. The code for reading a message's header from the spool was incrementing
1771 $received_count, but never initializing it. This meant that the value was
1772 incorrect (doubled) while delivering a message in the same process in which
1773 it was received. In the most common configuration of Exim, this never
1774 happens - a fresh exec is done - but it can happen when
1775 deliver_drop_privilege is set.
1776
177743. When Exim logs an SMTP synchronization error - client data sent too soon -
1778 it now includes up to 150 characters of the unexpected data in the log
1779 line.
1780
178144. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1782 and building data strings. The size of both of these buffers was 10 000
1783 bytes - far larger than anybody would *ever* want, thought I. Needless to
1784 say, somebody hit the limit. I have increased the maximum line length to
1785 20 000 and the maximum data length of concatenated lines to 100 000. I have
1786 also fixed two bugs, because there was no checking on these buffers. Tsk,
1787 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1788 buffer is too small.
1789
179045. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1791 lsearch lookups. Now it does.
1792
179346. When parsing a route_list item in a manualroute router, a fixed-length
1794 buffer was used for the list of hosts. I made this 1024 bytes long,
1795 thinking that nobody would ever have a list of hosts that long. Wrong.
1796 Somebody had a whole pile of complicated expansion conditions, and the
1797 string was silently truncated, leading to an expansion error. It turns out
1798 that it is easier to change to an unlimited length (owing to other changes
1799 that have happened since this code was originally written) than to build
1800 structure for giving a limitation error. The length of the item that
1801 expands into the list of hosts is now unlimited.
1802
180347. The lsearch lookup could not handle data where the length of text line was
1804 more than 4095 characters. Such lines were truncated, leading to shortened
1805 data being returned. It should now handle lines of any length.
1806
180748. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1808 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1809 ACL").
1810
181149. Cosmetic tidy to scripts like exicyclog that are generated by globally
1812 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1813 no longer happens in comment lines. A list of replacements is now placed
1814 at the head of all of the source files, except those whose only change is
1815 to replace PERL_COMMAND in the very first #! line.
1816
181750. Replaced the slow insertion sort in queue.c, for sorting the list of
1818 messages on the queue, with a bottom-up merge sort, using code contributed
1819 by Michael Haardt. This should make operations like -bp somewhat faster on
1820 large queues. It won't affect queue runners, except when queue_run_in_order
1821 is set.
1822
182351. Installed eximstats 1.31 in the distribution.
1824
182552. Added support for SRV lookups to the dnslookup router.
1826
182753. If an ACL referred to $message_body or $message_body_end, the value was not
1828 reset for any messages that followed in the same SMTP session.
1829
183054. The store-handling optimization for building very long strings was not
1831 differentiating between the different store pools. I don't think this
1832 actually made any difference in practice, but I've tidied it.
1833
183455. While running the routers to verify a sender address, $sender_address
1835 was still set to the sender address. This is wrong, because when routing to
1836 send a bounce to the sender, it would be empty. Therefore, I have changed
1837 it so that, while verifying a sender address, $sender_address is set to <>.
1838 (There is no change to what happens when verifying a recipient address.)
1839
184056. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1841 target A or AAAA records (if not already returned) without resetting the
1842 qualify_single or search_parents options of the DNS resolver. These are
1843 inappropriate in this case because the targets of MX and SRV records must
1844 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1845 target that, when qualified, matched something in the local domain. These
1846 two options are now turned off when doing these lookups.
1847
184857. It seems that at least some releases of Reiserfs (which does not have the
1849 concept of a fixed number of inodes) returns zero and not -1 for the
1850 number of available inodes. This interacted badly with check_spool_inodes,
1851 which assumed that -1 was the "no such thing" setting. What I have done is
1852 to check that the total number of inodes is greater than zero before doing
1853 the test of how many are available.
1854
185558. When a "warn" ACL statement has a log_message modifier, the message is
1856 remembered, and not repeated. This is to avoid a lot of repetition when a
1857 message has many recipients that cause the same warning to be written.
1858 Howewer, Exim was preserving the list of already written lines for an
1859 entire SMTP session, which doesn't seem right. The memory is now reset if a
1860 new message is started.
1861
186259. The "rewrite" debugging flag was not showing the result of rewriting in the
1863 debugging output unless log_rewrite was also set.
1864
186560. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1866 of (int)(handle) when we know that handle contains (void *)(-1).
1867
186861. The Exim daemon panic-logs an error return when it closes the incoming
1869 connection. However "connection reset by peer" seems to be common, and
1870 isn't really an error worthy of noting specially, so that particular error
1871 is no long logged.
1872
187362. When Exim is trying to find all the local interfaces, it used to panic and
1874 die if the ioctl to get the interface flags failed. However, it seems that
1875 on at least one OS (Solaris 9) it is possible to have an interface that is
1876 included in the list of interfaces, but for which you get a failure error
1877 for this call. This happens when the interface is not "plumbed" into a
1878 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1879 failure of the "get flags" call assumes that the interface is down.
1880
188163. Added a ${eval10: operator, which assumes all numbers are decimal. This
1882 makes life easier for people who are doing arithmetic on fields extracted
1883 from dates, where you often get leading zeros that should not be
1884 interpreted as octal.
1885
188664. Added qualify_domain to the redirect router, to override the global
1887 setting.
1888
188965. If a pathologically long header line contained very many addresses (the
1890 report of this problem mentioned 10 000) and each of them was rewritten,
1891 Exim could use up a very large amount of memory. (It kept on making new
1892 copies of the header line as it rewrote, and never released the old ones.)
1893 At the expense of a bit more processing, the header rewriting function has
1894 been changed so that it no longer eats memory in this way.
1895
189666. The generation of the Received: header has been moved from the time that a
1897 message starts to be received, to the time that it finishes. The timestamp
1898 in the Received: header should now be very close to that of the <= log
1899 line. There are two side-effects of this change:
1900
1901 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1902 logged header lines no longer include the local Received: line, because
1903 it has not yet been created. The same applies to a copy of the message
1904 that is returned to a non-SMTP sender when a message is rejected.
1905
1906 (b) When a filter file is tested using -bf, no additional Received: header
1907 is added to the test message. After some thought, I decided that this
1908 is a bug fix.
1909
1910 This change does not affect the value of $received_for. It is still set
1911 after address rewriting, but before local_scan() is called.
1912
191367. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1914
191568. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1916 gave an unhelpful panic error message, and a defer error. I have managed to
1917 change this behaviour so that it now rejects any supplied certificate,
1918 which seems right, as the list of acceptable certificates is empty.
1919
192069. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1921 gave an unhelpful defer error. I have not managed to make this reject any
1922 supplied certificates, but the error message it gives is "no certificate
1923 supplied", which is not helpful.
1924
192570. exigrep's output now also includes lines that are not associated with any
1926 message, but which match the given pattern. Implemented by a patch from
1927 Martin Sluka, which also tidied up the Perl a bit.
1928
192971. Recipient callout verification, like sender verification, was using <> in
1930 the MAIL FROM command. This isn't really the right thing, since the actual
1931 sender may affect whether the remote host accepts the recipient or not. I
1932 have changed it to use the actual sender in the callout; this means that
1933 the cache record is now keyed on a recipient/sender pair, not just the
1934 recipient address. There doesn't seem to be a real danger of callout loops,
1935 since a callout by the remote host to check the sender would use <>.
1936 [SEE ABOVE: changed after hitting problems.]
1937
193872. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1939 temporary errors. However, in the case of such a code being given after
1940 the end of a data transmission (i.e. after ".") Exim was failing to write
1941 a retry record for the message. (Yes, there was some broken host that was
1942 actually sending 8xx at this point.)
1943
194473. An unknown lookup type in a host list could cause Exim to panic-die when
1945 the list was checked. (An example that provoked this was putting <; in the
1946 middle of a list instead of at the start.) If this happened during a DATA
1947 ACL check, a -D file could be left lying around. This kind of configuration
1948 error no longer causes Exim to die; instead it causes a defer errror. The
1949 incident is still logged to the main and panic logs.
1950
195174. Buglet left over from Exim 3 conversion. The message "too many messages
1952 in one connection" was written to the rejectlog but not the mainlog, except
1953 when address rewriting (yes!) was being logged.
1954
195575. Added write_rejectlog option.
1956
195776. When a system filter was run not as root (that is, when system_filter_user
1958 was set), the values of the $n variables were not being returned to the
1959 main process; thus, they were not subsequently available in the $sn
1960 variables.
1961
196277. Added +return_path_on_delivery log selector.
1963
196478. A connection timeout was being treated differently from recipients deferred
1965 when testing hosts_max_try with a message that was older than the host's
1966 retry timeout. (The host should not be counted, thus allowing all hosts to
1967 be tried at least once before bouncing.) This may have been the cause of an
1968 occasionally reported bug whereby a message would remain on the queue
1969 longer than the retry timeout, but would be bounced if a delivery was
1970 forced. I say "may" because I never totally pinned down the problem;
1971 setting up timeout/retry tests is difficult. See also the next item.
1972
197379. The ultimate address timeout was not being applied to errors that involved
1974 a combination of host plus message (for example, a timeout on a MAIL
1975 command). When an address resolved to a number of possible hosts, and they
1976 were not all tried for each delivery (e.g. because of hosts_max_try), a
1977 message could remain on the queue longer than the retry timeout.
1978
197980. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1980 Haardt.
1981
198281. Fixed an obscure SMTP outgoing bug which required at least the following
1983 conditions: (a) there was another message waiting for the same server;
1984 (b) the server returned 5xx to all RCPT commands in the first message so
1985 that the message was not completed; (c) the server dropped the connection
1986 or gave a negative response to the RSET that Exim sends to abort the
1987 transaction. The observed case was a dropped connection after DATA that had
1988 been sent in pipelining mode. That is, the server had advertised PIPELINING
1989 but was not implementing it correctly. The effect of the bug was incorrect
1990 behaviour, such as trying another host, and this could lead to a crash.
1991
1992
1993Exim version 4.30
1994-----------------
1995
1996 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1997 and daemon.c were passed as pointers to ints; they should have been
1998 pointers to socklen_t variables (which are typically unsigned ints).
1999
2000 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2001 fixed.
2002
2003 3. Fixed a really odd bug that affected only the testing scheme; patching a
2004 certain fixed string in the binary changed the value of another string that
2005 happened to be identical to the end of the original first string.
2006
2007 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2008 name", it returns that address as the IP address. On some operating
2009 systems (e.g. Solaris), it also passes back the IP address string as the
2010 "host name". However, on others (e.g. Linux), it passes back an empty
2011 string. Exim wasn't checking for this, and was changing the host name to an
2012 empty string, assuming it had been canonicized.
2013
2014 5. Although rare, it is permitted to have more than one PTR record for a given
2015 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2016 all the names associated with an address, because they do in Solaris.
2017 However, it seems that they do not in Linux for data that comes from the
2018 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2019 I found this out when I moved to a new Linux workstation and tried to run
2020 the Exim test suite.
2021
2022 To get round this problem I have changed the code so that it now does its
2023 own call to the DNS to look up PTR records when searching for a host name.
2024 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2025 addresses that are only in /etc/hosts are still found.
2026
2027 This behaviour is, however, controlled by an option called host_lookup_
2028 order, which defaults to "bydns:byaddr". If people want to use the other
2029 order, or indeed, just use one or the other means of lookup, they can
2030 specify it in this variable.
2031
2032 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2033 some operating systems, this comes back from gethostbyaddr() as an empty
2034 string, and this is what Exim used to test for. However, it seems that in
2035 other systems, "." is yielded. Exim now tests for this case too.
2036
2037 7. The values of check_spool_space and check_log_space are now held internally
2038 as a number of kilobytes instead of an absolute number of bytes. If a
2039 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2040 kilobyte. This means that much larger values can be stored.
2041
2042 8. Exim monitor: an attempt to get the action menu when not actually pointing
2043 at a message produces an empty menu entitled "No message selected". This
2044 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2045 no entries in it ("Shell widget menu has zero width and/or height"). So I
2046 have added a single, blank menu entry in this case.
2047
2048 9. Added ${quote_local_part.
2049
205010. MIME decoding is now applied to the contents of Subject: header lines when
2051 they are logged.
2052
205311. Now that a reference to $sender_host_address automatically causes a reverse
2054 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2055 host lookup before query-style lookups in lists that might use this
2056 variable. This has therefore been abolished, and the "net-" prefix is no
2057 longer necessary for query-style lookups.
2058
205912. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2060 have been a typo for LFLAGS, so it has been changed.
2061
206213. The install script calls Exim with "-C /dev/null" in order to find the
2063 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2064 to be output. Howeve, since Exim outputs its version number before the
2065 error, it didn't break the script. It just looked ugly. I fixed this by
2066 always allowing "-C /dev/null" if the caller is root.
2067
206814. Ignore overlarge ACL variable number when reading spool file - insurance
2069 against a later release with more variables having written the file.
2070
207115. The standard form for an IPv6 address literal was being rejected by EHLO.
2072 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2073 this, as well as the form without the "IPv6" on the front.
2074
207516. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2076 OS/Makefile-Darwin file.
2077
207817. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2079 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2080
208118. After change 4.21/52, "%ld" was used to format the contents of the $inode
2082 variable. However, some OS use ints for inodes. I've added cast to long int
2083 to get rid of the compiler warning.
2084
208519. I had forgotten to lock out "/../" in configuration file names when
2086 ALT_CONFIG_PREFIX was set.
2087
208820. Routers used for verification do not need to specify transports. However,
2089 if such a router generated a host list, and callout was configured, Exim
2090 crashed, because it could not find a port number from the (non-existent)
2091 transport. It now assumes port 25 in this circumstance.
2092
209321. Added the -t option to exigrep.
2094
209522. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2096 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2097 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2098 them is set, LOOKUP_LSEARCH is forced.
2099
210023. "exim -bV" now outputs a list of lookups that are included in the binary.
2101
210224. Added sender and host information to the "rejected by local_scan()" log
2103 line; previously there was no indication of these.
2104
210525. Added .include_if_exists.
2106
210726. Change 3.952/11 added an explicit directory sync on top of a file sync for
2108 Linux. It turns out that not all file systems support this. Apparently some
2109 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2110 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2111 not supported on the file descriptor, is now ignored when Exim is trying to
2112 sync a directory. This applies only to Linux.
2113
211427. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2115
211628. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2117 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2118 this.
2119
212029. Added support for maildirsize files from supplied patch (modified a bit).
2121
212230. The use of :fail: followed by an empty string could lead Exim to respond to
2123 sender verification failures with (e.g.):
2124
2125 550 Verification failed for <xxx>
2126 550 Sender verify failed
2127
2128 where the first response line was missing the '-' that indicates it is not
2129 the final line of the response.
2130
213131. The loop for finding the name of the user that called Exim had a hardwired
2132 limit of 10; it now uses the value of finduser_retries, which is used for
2133 all other user lookups.
2134
213532. Added $received_count variable, available in data and not_smtp ACLs, and at
2136 delivery time.
2137
213833. Exim was neglecting to zero errno before one call of strtol() when
2139 expanding a string and expecting an integer value. On some systems this
2140 resulted in spurious "integer overflow" errors. Also, it was casting the
2141 result into an int without checking.
2142
214334. Testing for a connection timeout using "timeout_connect" in the retry rules
2144 did not work. The code looks as if it has *never* worked, though it appears
2145 to have been documented since at least releast 1.62. I have made it work.
2146
214735. The "timeout_DNS" error in retry rules, also documented since at least
2148 1.62, also never worked. As it isn't clear exactly what this means, and
2149 clearly it isn't a major issue, I have abolished the feature by treating it
2150 as "timeout", and writing a warning to the main and panic logs.
2151
215236. The display of retry rules for -brt wasn't always showing the error code
2153 correctly.
2154
215537. Added new error conditions to retry rules: timeout_A, timeout_MX,
2156 timeout_connect_A, timeout_connect_MX.
2157
215838. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2159 to the empty sender.
2160
216139. The daemon was not analysing the content of -oX till after it had closed
2162 stderr and disconnected from the controlling terminal. This meant that any
2163 syntax errors were only noted on the panic log, and the return code from
2164 the command was 0. By re-arranging the code a little, I've made the
2165 decoding happen first, so such errors now appear on stderr, and the return
2166 code is 1. However, the actual setting up of the sockets still happens in
2167 the disconnected process, so errors there are still only recorded on the
2168 panic log.
2169
217040. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2171 connections (as happens on some IP stacks) was logged at start up time as
2172 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2173 it from "IPv6 and IPv4", which means that two separate sockets are being
2174 used.
2175
217641. The debug output for gethostbyname2() or getipnodebyname() failures now
2177 says whether AF_INET or AF_INET6 was passed as an argument.
2178
217942. Exiwhat output was messed up when time zones were included in log
2180 timestamps.
2181
218243. Exiwhat now gives more information about the daemon's listening ports,
2183 and whether -tls-on-connect was used.
2184
218544. The "port" option of the smtp transport is now expanded.
2186
218745. A "message" modifier in a "warn" statement in a non-message ACL was being
2188 silently ignored. Now an error message is written to the main and panic
2189 logs.
2190
219146. There's a new ACL modifier called "logwrite" which writes to a log file
2192 as soon as it is encountered.
2193
219447. Added $local_user_uid and $local_user_gid at routing time.
2195
219648. Exim crashed when trying to verify a sender address that was being
2197 rewritten to "<>".
2198
219949. Exim was recognizing only a space character after ".include". It now also
2200 recognizes a tab character.
2201
220250. Fixed several bugs in the Perl script that creates the exim.8 man page by
2203 extracting the relevant information from the specification. The man page no
2204 longer contains scrambled data for the -d option, and I've added a section
2205 at the front about calling Exim under different names.
2206
220751. Added "extra_headers" argument to the "mail" command in filter files.
2208
220952. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2210 crash.
2211
221253. Installed eximstats 1.29.
2213
221454. Added transport_filter_timeout as a generic transport option.
2215
221655. Exim no longer adds an empty Bcc: header to messages that have no To: or
2217 Cc: header lines. This was required by RFC 822, but it not required by RFC
2218 2822.
2219
222056. Exim used to add From:, Date:, and Message-Id: header lines to any
2221 incoming messages that did not have them. Now it does so only if the
2222 message originates locally, that is, if there is no associated remote host
2223 address. When Resent- header lines are present, this applies to the Resent-
2224 lines rather than the non-Resent- lines.
2225
222657. Drop incoming SMTP connection after too many syntax or protocol errors. The
2227 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2228
222958. Messages for configuration errors now include the name of the main
2230 configuration file - useful now that there may be more than one file in a
2231 list (.included file names were always shown).
2232
223359. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2234 for those rare installations that do not start the daemon as root or run it
2235 setuid root. I've cut out the call to initgroups() if the daemon is not
2236 root at that time.
2237
223860. The Exim user and group can now be bound into the binary as text strings
2239 that are looked up at the start of Exim's processing.
2240
224161. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2242
224362. Added $mailstore_basename variable.
2244
224563. Installed patch to sieve.c from Michael Haardt.
2246
224764. When Exim failed to open the panic log after failing to open the main log,
2248 the original message it was trying to log was written to stderr and debug
2249 output, but if they were not available (the usual case in production), it
2250 was lost. Now it is written to syslog before the two lines that record the
2251 failures to open the logs.
2252
225365. Users' Exim filters run in subprocesses under the user's uid. It is
2254 possible for a "deliver" command or an alias in a "personal" command to
2255 provoke an address rewrite. If logging of address rewriting is configured,
2256 this fails because the process is not running as root or exim. There may be
2257 a better way of dealing with this, but for the moment (because 4.30 needs
2258 to be released), I have disabled address rewrite logging when running a
2259 filter in a non-root, non-exim process.
2260
2261
2262Exim version 4.24
2263-----------------
2264
2265 1. The buildconfig auxiliary program wasn't quoting the value set for
2266 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2267 not defined. This bug was masked in 4.22 by the effect that was fixed in
2268 change 4.23/1.
2269
2270 2. Some messages that were rejected after a message id was allocated were
2271 shown as "incomplete" by exigrep. It no longer does this for messages that
2272 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2273
2274 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2275 have allow_domain_literals set, the ID did not get logged in the <= line.
2276 Domain literals are now always recognized in Message-ID: header lines.
2277
2278 4. The first argument for a ${extract expansion item is the key name or field
2279 number. Leading and trailing spaces in this item were not being ignored,
2280 causing some misleading effects.
2281
2282 5. When deliver_drop_privilege was set, single queue runner processes started
2283 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2284 same command in the process it spins off) were not dropping privilege.
2285
2286 6. When the daemon running as "exim" started a queue runner, it always
2287 re-executed Exim in the spun-off process. This is a waste of effort when
2288 deliver_drop_privilege is set. The new process now just calls the
2289 queue-runner function directly.
2290
2291
2292Exim version 4.23
2293-----------------
2294
2295 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2296 HEADERS_CHARSET.
2297
2298 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2299 ignored. Though the use of -oP was forcing the writing of a pid file, it
2300 was always written to the default place.
2301
2302 3. If the message "no IP address found for host xxxx" is generated during
2303 incoming verification, it is now followed by identification of the incoming
2304 connection (so you can more easily find what provoked it).
2305
2306 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2307
2308 5. Added some features to "harden" Exim a bit more against certain attacks:
2309
2310 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2311 be put in Local/Makefile. This is like the never_users runtime option,
2312 but it cannot be overridden. The default setting is "root".
2313
2314 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2315 prefix string with which any file named in a -C command line option
2316 must start.
2317
2318 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2319 is retained for -C and -D only if the caller of Exim is root. Without
2320 it, the exim user may also use -C and -D and retain privilege.
2321
2322 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2323 command line option is disabled.
2324
2325 6. Macro names set by the -D option must start with an upper case letter, just
2326 like macro names defined in the configuration file.
2327
2328 7. Added "dereference=" facility to LDAP.
2329
2330 8. Two instances of the typo "uknown" in the source files are fixed.
2331
2332 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2333 the Configure-Makefile script screwed up while processing it.
2334
233510. Incorporated PCRE 4.4.
2336
233711. The SMTP synchronization check was not operating right at the start of an
2338 SMTP session. For example, it could not catch a HELO sent before the client
2339 waited for the greeting. There is now a check for outstanding input at the
2340 point when the greeting is written. Because of the duplex, asynchronous
2341 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2342 way, but not yet received, when the check is performed.
2343
234412. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2345 on TCP/IP sockets, because this apparently causes some broken clients to
2346 timeout.
2347
234813. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2349 unchanged) from the Cygwin maintainer.
2350
235114. The code for -bV that shows what is in the binary showed "mbx" when maildir
2352 was supported instead of testing for mbx. Effectively a typo.
2353
235415. The spa authenticator server code was not checking that the input it
2355 received was valid base64.
2356
235716. The debug output line for the "set" modifier in ACLs was not showing the
2358 name of the variable that was being set.
2359
236017. Code tidy: the variable type "vtype_string" was never used. Removed it.
2361
236218. Previously, a reference to $sender_host_name did not cause a DNS reverse
2363 lookup on its own. Something else was needed to trigger the lookup. For
2364 example, a match in host_lookup or the need for a host name in a host list.
2365 Now, if $sender_host_name is referenced and the host name has not yet been
2366 looked up, a lookup is performed. If the lookup fails, the variable remains
2367 empty, and $host_lookup_failed is set to "1".
2368
236919. Added "eqi" as a case-independent comparison operator.
2370
237120. The saslauthd authentication condition could segfault if neither service
2372 nor realm was specified.
2373
237421. If an overflowing value such as "2048M" was set for message_size_limit, the
2375 error message that was logged was misleading, and incoming SMTP
2376 connections were dropped. The message is now more accurate, and temporary
2377 errors are given to SMTP connections.
2378
237922. In some error situations (such as 21 above) Exim rejects all SMTP commands
2380 (except RSET) with a 421 error, until QUIT is received. However, it was
2381 failing to send a response to QUIT.
2382
238323. The HELO ACL was being run before the code for helo_try_verify_hosts,
2384 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2385 ACL is now run after the helo_try_verify_hosts code.
2386
238724. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2388 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2389 case-independent, so other case variants are also recognized). Apparently
2390 some systems use these upper case variants.
2391
239225. If more than two messages were waiting for the same host, and a transport
2393 filter was specified for the transport, Exim sent two messages over the
2394 same TCP/IP connection, and then failed with "socket operation on non-
2395 socket" when it tried to send the third.
2396
239726. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2398
239927. The extern definition of crypt16() in expand.c was not being excluded when
2400 the OS had its own crypt16() function.
2401
240228. Added bounce_return_body as a new option, and bounce_return_size_limit
2403 as a preferred synonym for return_size_limit, both as an option and as an
2404 expansion variable.
2405
240629. Added LIBS=-liconv to OS/Makefile-OSF1.
2407
240830. Changed the default configuration ACL to relax the local part checking rule
2409 for addresses that are not in any local domains. For these addresses,
2410 slashes and pipe symbols are allowed within local parts, but the sequence
2411 /../ is explicitly forbidden.
2412
241331. SPA server authentication was not clearing the challenge buffer before
2414 using it.
2415
241632. log_message in a "warn" ACL statement was writing to the reject log as
2417 well as to the main log, which contradicts the documentation and doesn't
2418 seem right (because no rejection is happening). So I have stopped it.
2419
242033. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2421 However, I am unable to do any testing of this.
2422
242334. Fixed an infelicity in the appendfile transport. When checking directories
2424 for a mailbox, to see if any needed to be created, it was accidentally
2425 using path names with one or more superfluous leading slashes; tracing
2426 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2427
242835. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2429 contents are added to the log line that is written for every discarded
2430 recipient. (Previously a log_message setting was ignored.)
2431
243236. The ${quote: operator now quotes the string if it is empty.
2433
243437. The install script runs exim in order to find its version number. If for
2435 some reason other than non-existence or emptiness, which it checks, it
2436 could not run './exim', it was installing it with an empty version number,
2437 i.e. as "exim-". This error state is now caught, and the installation is
2438 aborted.
2439
244038. An argument was missing from the function that creates an error message
2441 when Exim fails to connect to the socket for saslauthd authentication.
2442 This could cause Exim to crash, or give a corrupted message.
2443
244439. Added isip, isip4, and isip6 to ${if conditions.
2445
244640. The ACL variables $acl_xx are now saved with the message, and can be
2447 accessed later in routers, transports, and filters.
2448
244941. The new lookup type nwildlsearch is like wildlsearch, except that the key
2450 strings in the file are not string-expanded.
2451
245242. If a MAIL command specified a SIZE value that was too large to fit into an
2453 int variable, the check against message_size_limit failed. Such values are
2454 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2455 day this will have to be increased, but I don't think I want to be around
2456 when emails are that large.
2457
2458
2459
2460Exim version 4.22
2461-----------------
2462
2463 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2464 iconv() is not standard in FreeBSD.
2465
2466 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2467 IPv6 enabled. The observed symptom was a segmentation fault on return from
2468 the function os_common_find_running_interfaces() in src/os.c.
2469
2470 3. In the check_special_case() function in daemon.c I had used "errno" as an
2471 argument name, which causes warnings on some systems. This was basically a
2472 typo, since it was named "eno" in the comments!
2473
2474 4. The code that waits for the clock to tick (at a resolution of some fraction
2475 of a second) so as to ensure message-id uniqueness was always waiting for
2476 at least one whole tick, when it could have waited for less. [This is
2477 almost certainly not relevant at current processor speeds, where it is
2478 unlikely to ever wait at all. But we try to future-proof.]
2479
2480 5. The function that sleeps for a time interval that includes fractions of a
2481 second contained a race. It did not block SIGALRM between setting the
2482 timer, and suspending (a couple of lines later). If the interval was short
2483 and the sigsuspend() was delayed until after it had expired, the suspension
2484 never ended. On busy systems this could lead to processes getting stuck for
2485 ever.
2486
2487 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2488 process, before it forks any delivery processes. The open lookup caching
2489 mechanism meant that the open file or database connection was passed into
2490 the delivery process. The problem was that delivery processes always tidy
2491 up cached lookup data. This could cause a problem for the next delivery
2492 process started by the queue runner, because the external queue runner
2493 process does not know about the closure. So the next delivery process
2494 still has data in the lookup cache. In the case of a file lookup, there was
2495 no problem because closing a file descriptor in a subprocess doesn't affect
2496 the parent. However, if the lookup was caching a connection to a database,
2497 the connection was closed, and the second delivery process was likely to
2498 see errors such as "PGSQL: query failed: server closed the connection
2499 unexpectedly". The problem has been fixed by closing all cached lookups
2500 in a queue runner before running a delivery process.
2501
2502 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2503 seem to have the "const" qualifier which it has on other OS. I've
2504 parameterised it.
2505
2506 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2507 *of the same type* (client or server) with the same public name that an
2508 error should be diagnosed.
2509
2510 9. When Exim looked up a host name for an IP address, but failed to find the
2511 original IP address when looking up the host name (a safety check), it
2512 output the message "<ip address> does not match any IP for NULL", which was
2513 confusing, to say the least. The bug was that the host name should have
2514 appeared instead of "NULL".
2515
251610. Since release 3.03, if Exim is called by a uid other than root or the Exim
2517 user that is built into the binary, and the -C or -D options is used, root
2518 privilege is dropped before the configuration file is read. In addition,
2519 logging is switched to stderr instead of the normal log files. If the
2520 configuration then re-defines the Exim user, the unprivileged environment
2521 is probably not what is expected, so Exim logs a panic warning message (but
2522 proceeds).
2523
2524 However, if deliver_drop_privilege is set, the unprivileged state may well
2525 be exactly what is intended, so the warning has been cut out in that case,
2526 and Exim is allowed to try to write to its normal log files.
2527
2528
2529Exim version 4.21
2530-----------------
2531
2532 1. smtp_return_error_details was not giving details for temporary sender
2533 or receiver verification errors.
2534
2535 2. Diagnose a configuration error if two authenticators have the same public
2536 name.
2537
2538 3. Exim used not to create the message log file for a message until the first
2539 delivery attempt. This could be confusing when incoming messages were held
2540 for policy or load reasons. The message log file is now created at the time
2541 the message is received, and an initial "Received" line is written to it.
2542
2543 4. The automatically generated man page for command line options had a minor
2544 bug that caused no ill effects; however, a more serious problem was that
2545 the procedure for building the man page automatically didn't always
2546 operate. Consequently, release 4.20 contains an out-of-date version. This
2547 shouldn't happen again.
2548
2549 5. When building Exim with embedded Perl support, the script that builds the
2550 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2551 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2552
2553 6. The freeze_tell option was not being used for messages that were frozen on
2554 arrival, either by an ACL or by local_scan().
2555
2556 7. Added the smtp_incomplete_transaction log selector.
2557
2558 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2559 was accepting AUTH without a new EHLO.
2560
2561 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2562 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2563 RFC.
2564
256510. Logging of TCP/IP connections (when configured) now happens in the main
2566 daemon process instead of the child process, so that the TCP/IP connection
2567 count is more accurate (but it can never be perfect).
2568
256911. The use of "drop" in a nested ACL was not being handled correctly in the
2570 outer ACL. Now, if condition failure induced by the nested "drop" causes
2571 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2572 or "require"), the connection is dropped.
2573
257412. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2575 that yield "discard" can now be used with an "accept" or a "discard" verb,
2576 but an error is generated for any others (because I can't see a useful way
2577 to define what should happen).
2578
257913. When an ACL is read dynamically from a file (or anywhere else), the lines
2580 are now processed in the same way as lines in the Exim configuration file.
2581 In particular, continuation lines are supported.
2582
258314. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2584
258515. Added -ti meaning -t -i.
2586
258716. Check for letters, digits, hyphens, and dots in the names of dnslist
2588 domains, and warn by logging if others are found.
2589
259017. At least on BSD, alignment is not guarenteed for the array of ifreq's
2591 returned from GIFCONF when Exim is trying to find the list of interfaces on
2592 a host. The code in os.c has been modified to copy each ifreq to an aligned
2593 structure in all cases.
2594
2595 Also, in some cases, the returned ifreq's were being copied to a 'struct
2596 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2597 means the last couple of bytes of an IPv6 address could be chopped if the
2598 ifreq contained only a normal sockaddr (14 bytes storage).
2599
260018. Named domain lists were not supported in the hosts_treat_as_local option.
2601 An entry such as +xxxx was not recognized, and was treated as a literal
2602 domain name.
2603
260419. Ensure that header lines added by a DATA ACL are included in the reject log
2605 if the ACL subsequently rejects the message.
2606
260720. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2608 MD5 (which is deprecated).
2609
261021. When testing a filter file using -bf, Exim was writing a message when it
2611 took the sender from a "From " line in the message, but it was not doing so
2612 when it took $return_path from a Return-Path: header line. It now does.
2613
261422. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2615 with a valid header line field name (a series of printing characters
2616 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2617
261823. Changed "disc" in the source to "disk" to conform to the documentation and
2619 the book and for uniformity.
2620
262124. Ignore Sendmail's -Ooption=value command line item.
2622
262325. When execve() failed while trying to run a command in a pipe transport,
2624 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2625 could be confused with a return value of 69 from the command itself. This
2626 has been changed to 127, the value the shell returns if it is asked to run
2627 a non-existent command. The wording for the related log line suggests a
2628 non-existent command as the problem.
2629
263026. If received_header_text expands to an empty string, do not add a Received:
2631 header line to the message. (Well, it adds a token one on the spool, but
2632 marks it "old" so that it doesn't get used or transmitted.)
2633
263427. Installed eximstats 1.28 (addition of -nt option).
2635
263628. There was no check for failure on the call to getsockname() in the daemon
2637 code. This can fail if there is a shortage of resources on the system, with
2638 ENOMEM, for example. A temporary error is now given on failure.
2639
264029. Contrary to the C standard, it seems that in some environments, the
2641 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2642 program. Exim now does this explicitly; it affects the formatting of
2643 timestamps using strftime().
2644
264530. If exiqsumm was given junk data, it threw up some uninitialized variable
2646 complaints. I've now initialized all the variables, to avoid this.
2647
264832. Header lines added by a system filter were not being "seen" during
2649 transport-time rewrites.
2650
265133. The info_callback() function passed to OpenSSL is set up with type void
2652 (*)(SSL *, int, int), as described somewhere. However, when calling the
2653 function (actually a macro) that sets it up, the type void(*)() is
2654 expected. I've put in a cast to prevent warnings from picky compilers.
2655
265634. If a DNS black list lookup found a CNAME record, but there were no A
2657 records associated with the domain it pointed at, Exim crashed.
2658
265935. If a DNS black list lookup returned more than one A record, Exim ignored
2660 all but the first. It now scans all returned addresses if a particular IP
2661 value is being sought. In this situation, the contents of the
2662 $dnslist_value variable are a list of all the addresses, separated by a
2663 comma and a space.
2664
266536. Tightened up the rules for host name lookups using reverse DNS. Exim used
2666 to accept a host name and all its aliases if the forward lookup for any of
2667 them yielded the IP address of the incoming connection. Now it accepts only
2668 those names whose forward lookup yields the correct IP address. Any other
2669 names are discarded. This closes a loophole whereby a rogue DNS
2670 administrator could create reverse DNS records to break through a
2671 wildcarded host restriction in an ACL.
2672
267337. If a user filter or a system filter that ran in a subprocess used any of
2674 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2675 the wrong values were passed to the pipe command ($thisaddress had the
2676 value of $0, $0 had the value of $1, etc). This bug was introduced by
2677 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2678
267938. Improved the line breaking for long SMTP error messages from ACLs.
2680 Previously, if there was no break point between 40 and 75 characters, Exim
2681 left the rest of the message alone. Two changes have been made: (a) I've
2682 reduced the minimum length to 35 characters; (b) if it can't find a break
2683 point between 35 and 75 characters, it looks ahead and uses the first one
2684 that it finds. This may give the occasional overlong line, but at least the
2685 remaining text gets split now.
2686
268739. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2688 file descriptors might be low, and that setting 1000 would always raise it.
2689 It turns out that in some environments, the limit is already over 1000 and
2690 that lowering it causes trouble. So now Exim takes care not to decrease it.
2691
269240. When delivering a message, the value of $return_path is set to $sender_
2693 address at the start of routing (routers may change the value). By an
2694 oversight, this default was not being set up when an address was tested by
2695 -bt or -bv, which affected the outcome if any router or filter referred to
2696 $return_path.
2697
269841. The idea of the "warn" ACL verb is that it adds a header or writes to the
2699 log only when "message" or "log_message" are set. However, if one of the
2700 conditions was an address verification, or a call to a nested ACL, the
2701 messages generated by the underlying test were being passed through. This
2702 no longer happens. The underlying message is available in $acl_verify_
2703 message for both "message" and "log_message" expansions, so it can be
2704 passed through if needed.
2705
270642. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2707 new expansion $bh_ to give the encoded byte string without charset
2708 translation. Translation happens only if iconv() is available; HAVE_ICONV
2709 indicates this at build time. HEADERS_CHARSET gives the charset to
2710 translate to; headers_charset can change it in the configuration, and
2711 "headers charset" can change it in an individual filter file.
2712
271343. Now that we have a default RFC 2047 charset (see above), the code in Exim
2714 that creates RFC 2047 encoded "words" labels them as that charset instead
2715 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2716 expansion operator; (ii) when Exim creates a From: line for a local
2717 message; (iii) when a header line is rewritten to include a "phrase" part.
2718
271944. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2720 buggy, causing it to skip the first lines of messages whose message ID
2721 ended in 'D'. This would not have bitten before Exim release 4.14, because
2722 message IDs were unlikely to end in 'D' before then. The effect was to have
2723 incorrect size information for certain domains.
2724
272545. #include "config.h" was missing at the start of the crypt16.c module. This
2726 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2727 noticed.
2728
272946. If there was a timeout during a "random" callout check, Exim treated it as
2730 a failure of the random address, and carried on sending RSET and the real
2731 address. If the delay was just some slowness somewhere, the response to the
2732 original RCPT would be taken as a response to RSET and so on, causing
2733 mayhem of various kinds.
2734
273547. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2736 when I implemented it. It didn't allow for the fact that some option values
2737 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2738 the right test for positive values.
2739
274048. Domain names in DNS records are case-independent. Exim always looks them up
2741 in lower case. Some resolvers return domain names in exactly the case they
2742 appear in the zone file, that is, they may contain uppercase letters. Not
2743 all resolvers do this - some return always lower case. Exim was treating a
2744 change of case by a resolver as a change of domain, similar to a widening
2745 of a domain abbreviation. This triggered its re-routing code and so it was
2746 trying to route what was effectively the same domain again. This normally
2747 caused routing to fail (because the router wouldn't handle the domain
2748 twice). Now Exim checks for this case specially, and just changes the
2749 casing of the domain that it ultimately uses when it transmits the message
2750 envelope.
2751
275249. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2753 module.
2754
275550. If a filter generated a file delivery with a non-absolute name (possible if
2756 no home directory exists for the router), the forbid_file option was not
2757 forbidding it.
2758
275951. Added '&' feature to dnslists, to provide bit mask matching in addition to
2760 the existing equality matching.
2761
276252. Exim was using ints instead of ino_t variables in some places where it was
2763 dealing with inode numbers.
2764
276553. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2766 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2767 called TMPDIR, and if it finds it is different, it changes its value.
2768
276954. The smtp_printf() function is now made available to local_scan() so
2770 additional output lines can be written before returning. There is also an
2771 smtp_fflush() function to enable the detection of a dropped connection.
2772 The variables smtp_input and smtp_batched_input are exported to
2773 local_scan().
2774
277555. Changed the default runtime configuration: the message "Unknown user"
2776 has been removed from the ACL, and instead placed on the localuser router,
2777 using the cannot_route_message feature. This means that any verification
2778 failures that generate their own messages won't get overridden. Similarly,
2779 the "Unrouteable address" message that was in the ACL for unverifiable
2780 relay addresses has also been removed.
2781
278256. Added hosts_avoid_esmtp to the smtp transport.
2783
278457. The exicyclog script was not checking for the esoteric option
2785 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2786 will work only if exicyclog is run under the appropriate euid.
2787
278858. Following a discussion on the list, the rules by which Exim recognises line
2789 endings on incoming messages have been changed. The -dropcr and drop_cr
2790 options are now no-ops, retained only for backwards compatibility. The
2791 following line terminators are recognized: LF CRLF CR. However, special
2792 processing applies to CR:
2793
2794 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2795 nor a local message in the state where . is a terminator.
2796
2797 (ii) If a bare CR is encountered in a header line, an extra space is added
2798 after the line terminator so as not to end the header. The reasoning
2799 behind this is that bare CRs in header lines are most likely either
2800 to be mistakes, or people trying to play silly games.
2801
280259. The size of a message, as listed by "-bp" or in the Exim monitor window,
2803 was being incorrectly given as 18 bytes larger than it should have been.
2804 This is a VOB (very old bug).
2805
280660. This may never have affected anything current, but just in case it has:
2807 When the local host is found other than at the start of a list of hosts,
2808 the local host, those with the same MX, and any that follow, are discarded.
2809 When the list in question was part of a longer list of hosts, the following
2810 hosts (not currently being processed) were also being discarded. This no
2811 longer happens. I'm not sure if this situation could ever has previously
2812 arisen.
2813
281461. Added the "/MX" feature to lists of hosts in the manualroute and query
2815 program routers.
2816
281762. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2818 header. This is something that is recommended in a new Internet Draft, and
2819 is something that is documented as being done by Sendmail. There are two
2820 possible values. For messages generated by the autoreply transport, Exim
2821 adds:
2822
2823 Auto-Submitted: auto-replied
2824
2825 whereas for all other generated messages (e.g. bounces) it adds
2826
2827 Auto-Submitted: auto-generated
2828
282963. The "personal" condition in filters now includes a test for the
2830 Auto-Submitted: header. If it contains the string "auto-" the message it
2831 not considered personal.
2832
283364. Added rcpt_include_affixes as a generic transport option.
2834
283565. Added queue_only_override (default true).
2836
283766. Added the syslog_duplication option.
2838
283967. If what should have been the first header line of a message consisted of
2840 a space followed by a colon, Exim was mis-interpreting it as a header line.
2841 It isn't of course - it is syntactically invalid and should therefore be
2842 treated as the start of the message body. The misbehaviour could have
2843 caused a number of strange effects, including loss of data in subsequent
2844 header lines, and spool format errors.
2845
284668. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2847 client host had authenticated. This control can now be exercised by an ACL
2848 for more flexibility.
2849
285069. By default, callouts do not happen when testing with -bh. There is now a
2851 variant, -bhc, which does actually run the callout code, including
2852 consulting and updating the callout cache.
2853
285470. Added support for saslauthd authentication, courtesy of Alexander
2855 Sabourenkov.
2856
285771. If statvfs() failed on the spool or log directories while checking their
2858 size for availability, Exim confusingly gave the error "space shortage".
2859 Furthermore, in debugging mode it crashed with a floating point exception.
2860 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2861 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2862 serious problem, Exim now writes to the main and panic logs when this
2863 happens, with details of the failure. It then refuses to accept the
2864 incoming message, giving the message "spool directory problem" or "log
2865 directory problem" with a 421 code for SMTP messages.
2866
286772. When Exim is about to re-exec itself, it ensures that the file descriptors
2868 0, 1, and 2 exist, because some OS complain for execs without them (see
2869 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2870 descriptors. However, the code omitted to check that the open succeeded,
2871 causing mysterious errors if for some reason the permissions on /dev/null
2872 got screwed. Now Exim writes a message to the main and panic logs, and
2873 bombs out if it can't open /dev/null.
2874
287573. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2876 interact so that it is all more flexible. It is supposed to remain
2877 backwards compatible. Also added extra_local_interfaces.
2878
287974. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2880 to bomb out with an assertion failure - to the client this appears as a
2881 connection drop. This problem occurs in the part of the code that was taken
2882 from the Samba project. Fortunately, the assertion is in a very simple
2883 function, so I have fixed this by reproducing the function inline in the
2884 one place where it is called, and arranging for authentication to fail
2885 instead of killing the process with assert().
2886
288775. The SPA client code was not working when the server requested OEM rather
2888 than Unicode encoding.
2889
289076. Added code to make require_files with a specific uid setting more usable in
2891 the case where statting the file as root fails - usually a non-root-mounted
2892 NFS file system. When this happens and the failure is EACCES, Exim now
2893 forks a subprocess and does the per-uid checking as the relevant uid.
2894
289577. Added process_log_path.
2896
289778. If log_file_path was not explicitly set, a setting of check_log_space or
2898 check_log_inodes was ignored.
2899
290079. If a space check for the spool or log partitions fails, the incident is now
2901 logged. Of course, in the latter case the data may get lost...
2902
290380. Added the %p formatting code to string_format() so that it can be used to
2904 print addresses in debug_print(). Adjusted all the address printing in the
2905 debugging in store.c to use %p rather than %d.
2906
290781. There was a concern that a line of code in smtp_in.c could overflow a
2908 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2909 initially expressed, the concern was not well-founded, because trailing
2910 spaces are removed early. However, if the trailing spaces were followed by
2911 a NULL, they did not get removed, so the overflow was possible. Two fixes
2912 were applied:
2913
2914 (a) I re-wrote the offending code in a cleaner fashion.
2915 (b) If an incoming SMTP command contains a NULL character, it is rejected
2916 as invalid.
2917
291882. When Exim changes uid/gid to the Exim user at daemon start time, it now
2919 runs initgroups(), so that if the Exim user is in any additional groups,
2920 they will be used during message reception.
2921
2922
2923Exim version 4.20
2924-----------------
2925
2926The change log for 4.20 and earlier releases has been archived.
2927
2928****