Add simple SPF lookup method (armored in EXPERIMENTAL_SPF)
[exim.git] / doc / doc-txt / ChangeLog
CommitLineData
958541e9 1$Cambridge: exim/doc/doc-txt/ChangeLog,v 1.144 2005/05/24 14:56:26 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
06a9b4b5
PH
17TK/01 Fix poll() being unavailable on Mac OSX 10.2.
18
c1ac6996
PH
19PH/02 Reduce the amount of output that "make" produces by default. Full output
20 can still be requested.
21
9c7a242c
PH
22PH/03 The warning log line about a condition test deferring for a "warn" verb
23 was being output only once per connection, rather than after each
24 occurrence (because it was using the same function as for successful
25 "warn" verbs). This seems wrong, so I have changed it.
26
87ba3f5f
PH
27TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
28 it should not have, which might have caused a crash in the right
29 circumstances, but probably never did.
30
31PH/04 Installed a modified version of Tony Finch's patch to make submission
32 mode fix the return path as well as the Sender: header line, and to
33 add a /name= option so that you can make the user's friendly name appear
34 in the header line.
35
29aba418
TF
36TF/03 Added the control = fakedefer ACL modifier.
37
870f6ba8
TF
38TF/04 Added the ratelimit ACL condition. See NewStuff for details.
39
11d337a4
TK
40TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
41
42TK/03 Merged latest SRS patch from Miles Wilton.
43
415c8f3b
PH
44PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
45 with the definition in sysexits.h (which is #included earlier).
46 Fortunately, Exim does not actually use EX_OK. The code used to try to
47 preserve the sysexits.h value, by assumimg that macro definitions were
48 scanned for macro replacements. I have been disabused of this notion,
49 so now the code just undefines EX_OK before #including unistd.h.
11d337a4 50
958541e9
PH
51PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
52 in the smtp transport. When a block could not be written in a single
53 write() function, the timeout was being re-applied to each part-write.
54 This seems wrong - if the receiver was accepting one byte at a time it
55 would take for ever. The timeout is now adjusted when this happens. It
56 doesn't have to be particularly precise.
57
e5a9dba6 58
7982096b
PH
59Exim version 4.51
60-----------------
61
1a46a8c5
PH
62TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
63 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
64
2f079f46 65TK/02 Fix ACL "control" statement not being available in MIME ACL.
1a46a8c5
PH
66
67TK/03 Fix ACL "regex" condition not being available in MIME ACL.
68
69PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
70 to test Sieve filters that use "vacation".
71
72PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
73 that changes the way the GnuTLS parameters are stored in the cache file.
74 The new format can be generated externally. For backward compatibility,
75 if the data in the cache doesn't make sense, Exim assumes it has read an
76 old-format file, and it generates new data and writes a new file. This
77 means that you can't go back to an older release without removing the
78 file.
79
80PH/03 A redirect router that has both "unseen" and "one_time" set does not
81 work if there are any delivery delays because "one_time" forces the
82 parent to be marked "delivered", so its unseen clone is never tried
83 again. For this reason, Exim now forbids the simultaneous setting of
84 these two options.
85
86PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
87 redirected to themselves ("homonym" addresses). Read the long ChangeLog
88 entry if you want to know the details. The fix, however, neglected to
89 consider the case when local delivery batching is involved. The test for
90 "previously delivered" was not happening when checking to see if an
91 address could be batched with a previous (undelivered) one; under
92 certain circumstances this could lead to multiple deliveries to the same
c2c19e9d 93 address.
1a46a8c5
PH
94
95PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
96 in its include files, and this causes problems building Exim.
97
98PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
99 header_syntax) but Exim was just ignoring anything given after a slash.
100 In particular, this caused confusion with an attempt to use "verify =
101 reverse_host_lookup/defer_ok". An error is now given when options are
102 supplied for verify items that do not have them. (Maybe reverse_host_
103 lookup should have a defer_ok option, but that's a different point.)
104
105PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
106 defined by RFC 821) to 2048, because there were problems with some AUTH
107 commands, and RFC 1869 says the size should be increased for extended
108 SMTP commands that take arguments.
109
110PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
111 Finch).
112
113PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
114 "unknown" error; now it says that the functionality isn't in the binary.
8d67ada3 115
49c2d5ea
PH
116PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
117 an address' error message when a string expansion fails (syntax or
f331f3b6
PH
118 whatever). Otherwise the password may appear in the log. Following change
119 PH/42 below, there is no longer a chance of it appearing in a bounce
120 message.
49c2d5ea 121
bf759a8b
PH
122PH/11 Installed exipick version 20050225.0 from John Jetmore.
123
83364d30
PH
124PH/12 If the last host in a fallback_hosts list was multihomed, only the first
125 of its addresses was ever tried. (Bugzilla bug #2.)
126
7999bbd7
PH
127PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
128 the result incorrectly in the debug output. (It correctly added a newline
129 to what was transported.)
130
7dbf77c9
PH
131TF/01 Added $received_time.
132
74e0617f
PH
133PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
134 commented out examples of how to interface to a virus scanner and to
135 SpamAssassin. Also added commented examples of av_scanner and
136 spamd_address settings.
137
2f079f46
PH
138PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
139 and controls are allowed in which ACLs. There were a couple of minor
140 errors. Some of the entries in the conditions table (which is a table of
141 where they are NOT allowed) were getting very unwieldy; rewrote them as a
142 negation of where the condition IS allowed.
143
8c841523
PH
144PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
145
7766a4f0
PH
146PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
147 header file does not have a version number, so I've had to invent a new
148 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
149 API. The code is untested by me (my Linux distribution still has 0.3.2 of
150 radiusclient), but it was contributed by a Radius user.
151
8b417f2c
PH
152PH/18 Installed Lars Mainka's patch for the support of CRL collections in
153 files or directories, for OpenSSL.
154
901f42cb
PH
155PH/19 When an Exim process that is running as root has to create an Exim log
156 file, it does so in a subprocess that runs as exim:exim so as to get the
157 ownership right at creation (otherwise, other Exim processes might see
158 the file with the wrong ownership). There was no test for failure of this
159 fork() call, which would lead to the process getting stuck as it waited
160 for a non-existent subprocess. Forks do occasionally fail when resources
161 run out. I reviewed all the other calls to fork(); they all seem to check
162 for failure.
163
f9b9210e
PH
164PH/20 When checking for unexpected SMTP input at connect time (before writing
165 the banner), Exim was not dealing correctly with a non-positive return
166 from the read() function. If the client had disconnected by this time,
167 the result was a log entry for a synchronization error with an empty
168 string after "input=" when read() returned zero. If read() returned -1
169 (an event I could not check), uninitialized data bytes were printed.
170 There were reports of junk text (parts of files, etc) appearing after
171 "input=".
172
54cdb463
PH
173PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
174
cf00dad6
PH
175PH/22 Added support for macro redefinition, and (re)definition in between
176 driver and ACL definitions.
177
acb1b346
PH
178PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
179 forgetting to use the resulting value; it was using the unexpanded value.
180
c5ddb310
PH
181PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
182 hadn't been configured. The fix is from Juergen Kreileder, who
183 understands it better than I do:
184
185 "Here's what I see happening with three configured cyrus_sasl
186 authenticators configured (plain, login, cram-md5):
187
188 On startup auth_cyrus_sasl_init() gets called for each of these.
189 This means three calls to sasl_listmech() without a specified mech_list.
190 => SASL tests which mechs of all available mechs actually work
191 => three warnings about OTP not working
192 => the returned list contains: plain, login, cram-md5, digest-md5, ...
193
194 With the patch, sasl_listmech() also gets called three times. But now
195 SASL's mech_list option is set to the server_mech specified in the the
196 authenticator. Or in other words, the answer from sasl_listmech()
197 gets limited to just the mech you're testing for (which is different
198 for each call.)
199 => the return list contains just 'plain' or 'login', 'cram-md5' or
200 nothing depending on the value of ob->server_mech.
201
202 I've just tested the patch: Authentication still works fine,
203 unavailable mechs specified in the exim configuration are still
204 caught, and the auth.log warnings about OTP are gone."
205
31619da6
PH
206PH/25 When debugging is enabled, the contents of the command line are added
207 to the debugging output, even when log_selector=+arguments is not
208 specified.
209
bebaf0fc
PH
210PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
211 answer is "GNU", and only if the return is "GNU/something" is the answer
212 "Linux".
213
475fe28a
PH
214PH/27 $acl_verify_message is now set immediately after the failure of a
215 verification in an ACL, and so is available in subsequent modifiers. In
216 particular, the message can be preserved by coding like this:
217
218 warn !verify = sender
219 set acl_m0 = $acl_verify_message
220
221 Previously, $acl_verify_message was set only while expanding "message"
222 and "log_message" when a very denied access.
223
7e8bec7a
PH
224PH/28 Modified OS/os.c-Linux with
225
226 -#ifndef OS_LOAD_AVERAGE
227 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
228
229 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
230 nomenclature these days.)
231
e4a89c47
PH
232PH/29 Installed patch from the Sieve maintainer that adds the options
233 sieve_useraddress and sieve_subaddress to the redirect router.
234
5ca2a9a1
PH
235PH/30 In these circumstances:
236 . Two addresses routed to the same list of hosts;
237 . First host does not offer TLS;
238 . First host accepts first address;
239 . First host gives temporary error to second address;
240 . Second host offers TLS and a TLS session is established;
241 . Second host accepts second address.
242 Exim incorrectly logged both deliveries with the TLS parameters (cipher
243 and peerdn, if requested) that were in fact used only for the second
244 address.
7e8bec7a 245
c688b954
PH
246PH/31 When doing a callout as part of verifying an address, Exim was not paying
247 attention to any local part prefix or suffix that was matched by the
248 router that accepted the address. It now behaves in the same way as it
249 does for delivery: the affixes are removed from the local part unless
250 rcpt_include_affixes is set on the transport.
251
fed77020
PH
252PH/32 Add the sender address, as F=<...>, to the log line when logging a
253 timeout during the DATA phase of an incoming message.
254
7fe1560f
PH
255PH/33 Sieve envelope tests were broken for match types other than :is. I have
256 applied a patch sanctioned by the Sieve maintainer.
c688b954 257
ebb6e6d5
PH
258PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
259 the uid or gid is negative. A case of a negative gid caused this to be
260 noticed. The fix allows for either to be negative.
261
9c4e8f60
PH
262PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
263 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
264 overlooked.
265
266PH/36 The change PH/12 above was broken. Fixed it.
267
d7174846
PH
268PH/37 Exim used to check for duplicate addresses in the middle of routing, on
269 the grounds that routing the same address twice would always produce the
270 same answer. This might have been true once, but it is certainly no
271 longer true now. Routing a child address may depend on the previous
272 routing that produced that child. Some complicated redirection strategies
273 went wrong when messages had multiple recipients, and made Exim's
274 behaviour dependent on the order in which the addresses were given.
275
276 I have moved the duplicate checking until after the routing is complete.
277 Exim scans the addresses that are assigned to local and remote
278 transports, and removes any duplicates. This means that more work will be
279 done, as duplicates will always all be routed, but duplicates are
280 presumably rare, so I don't expect this is of any significance.
281
282 For deliveries to pipes, files, and autoreplies, the duplicate checking
283 still happens during the routing process, since they are not going to be
284 routed further.
285
cfe75fc3
PH
286PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
287 It corrects a timeout issue with spamd. This is Ian's comment: "The
288 background is that sometimes spamd either never reads data from a
289 connection it has accepted, or it never writes response data. The exiscan
290 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
291 blindly assumes that writes won't block so it may never time out."
292
be22d70e
PH
293PH/39 Allow G after quota size as well as K and M.
294
0612b098
PH
295PH/40 The value set for $authenticated_id in an authenticator may not contain
296 binary zeroes or newlines because the value is written to log lines and
297 to spool files. There was no check on this. Now the value is run through
298 the string_printing() function so that such characters are converted to
299 printable escape sequences.
300
2e0c1448
PH
301PH/41 $message_linecount is a new variable that contains the total number of
302 lines in the message. Compare $body_linecount, which is the count for the
303 body only.
304
447d236c
PH
305PH/42 Exim no longer gives details of delivery errors for specific addresses in
306 bounce and delay warning messages, except in certain special cases, which
307 are as follows:
308
309 (a) An SMTP error message from a remote host;
310 (b) A message specified in a :fail: redirection;
311 (c) A message specified in a "fail" command in a system filter;
312 (d) A message specified in a FAIL return from the queryprogram router;
313 (e) A message specified by the cannot_route_message router option.
314
315 In these cases only, Exim does include the error details in bounce and
316 warning messages. There are also a few cases where bland messages such
317 as "unrouteable address" or "local delivery error" are given.
318
d20976dc
PH
319PH/43 $value is now also set for the "else" part of a ${run expansion.
320
f656d135
PH
321PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
322 being worked on, but at least Exim now implements the latest version to
323 play with."
324
2e2a30b4
PH
325PH/45 In a pipe transport, although a timeout while waiting for the pipe
326 process to complete was treated as a delivery failure, a timeout while
327 writing the message to the pipe was logged, but erroneously treated as a
328 successful delivery. Such timeouts include transport filter timeouts. For
329 consistency with the overall process timeout, these timeouts are now
330 treated as errors, giving rise to delivery failures by default. However,
331 there is now a new Boolean option for the pipe transport called
332 timeout_defer, which, if set TRUE, converts the failures into defers for
333 both kinds of timeout. A transport filter timeout is now identified in
334 the log output.
335
9176e9f0
PH
336PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
337 systems where "make" and "gmake" are different, calling "gmake" at top
338 level broke things. I've arranged for the value of $(MAKE) to be passed
339 from the Makefile to this script so that it can call the same version of
340 "make".
341
7982096b 342
bbe902f0
PH
343A note about Exim versions 4.44 and 4.50
344----------------------------------------
345
346Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
347changes of various kinds. As a consequence, a big documentation update was
348needed. This delayed the release for rather longer than seemed good, especially
349in the light of a couple of (minor) security issues. Therefore, the changes
350that fixed bugs were backported into 4.43, to create a 4.44 maintenance
351release. So 4.44 and 4.50 are in effect two different branches that both start
352from 4.43.
353
354I have left the 4.50 change log unchanged; it contains all the changes since
3554.43. The change log for 4.44 is below; many of its items are identical to
356those for 4.50. This seems to be the most sensible way to preserve the
357historical information.
358
359
f7b63901 360Exim version 4.50
495ae4b0
PH
361-----------------
362
5fe762f6
PH
363 1. Minor wording change to the doc/README.SIEVE file.
364
139059f6 365 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
5fe762f6 366 computation of the current number of files was incorrect.
495ae4b0 367
7086e875
PH
368 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
369 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
370 place.
371
35af9f61
PH
372 4. Give more explanation in the error message when the command for a transport
373 filter fails to execute.
374
b668c215
PH
375 5. There are several places where Exim runs a non-Exim command in a
376 subprocess. The SIGUSR1 signal should be disabled for these processes. This
377 was being done only for the command run by the queryprogram router. It is
378 now done for all such subprocesses. The other cases are: ${run, transport
379 filters, and the commands run by the lmtp and pipe transports.
380
a494b1e1
PH
381 6. Added CONFIGURE_GROUP build-time option.
382
383 7. Some older OS have a limit of 256 on the maximum number of file
384 descriptors. Exim was using setrlimit() to set 1000 as a large value
385 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
386 systems. I've change it so that if it can't get 1000, it tries for 256.
35edf2ff 387
c5fcb476
PH
388 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
389 was an oversight, and furthermore, ever since the addition of extra
390 controls (e.g. 4.43/32), the checks on when to allow different forms of
391 "control" were broken. There should now be diagnostics for all cases when a
392 control that does not make sense is encountered.
393
69358f02
PH
394 9. Added the /retain_sender option to "control=submission".
395
5be20824
PH
39610. $recipients is now available in the predata ACL (oversight).
397
eb2c0248
PH
39811. Tidy the search cache before the fork to do a delivery from a message
399 received from the command line. Otherwise the child will trigger a lookup
400 failure and thereby defer the delivery if it tries to use (for example) a
401 cached ldap connection that the parent has called unbind on.
402
2a3eea10
PH
40312. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
404 of $address_data from the recipient verification was clobbered by the
405 sender verification.
406
40713. The value of address_data from a sender verification is now available in
408 $sender_address_data in subsequent conditions in the ACL statement.
409
23c7ff99
PH
41014. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
411
4deaf07d
PH
41215. Added a new option "connect=<time>" to callout options, to set a different
413 connection timeout.
414
926e1192
PH
41516. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
416 was its contents. (It was OK if the option was not defined at all.)
417
650edc6f
PH
41817. A "Completed" log line is now written for messages that are removed from
419 the spool by the -Mrm option.
420
2c7db3f5
PH
42118. New variables $sender_verify_failure and $recipient_verify_failure contain
422 information about exactly what failed.
423
3d235903
PH
42419. Added -dd to debug only the daemon process.
425
7c7ad977
PH
42620. Incorporated Michael Haardt's patch to ldap.c for improving the way it
427 handles timeouts, both on the server side and network timeouts. Renamed the
428 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
429
981756db
PH
43021. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
431 It is now set to "smtps".
432
d4eb88df
PH
43322. $host_address is now set to the target address during the checking of
434 ignore_target_hosts.
435
43623. When checking ignore_target_hosts for an ipliteral router, no host name was
437 being passed; this would have caused $sender_host_name to have been used if
438 matching the list had actually called for a host name (not very likely,
439 since this list is usually IP addresses). A host name is now passed as
440 "[x.x.x.x]".
441
7d468ab8
PH
44224. Changed the calls that set up the SIGCHLD handler in the daemon to use the
443 code that specifies a non-restarting handler (typically sigaction() in
444 modern systems) in an attempt to fix a rare and obscure crash bug.
445
44625. Narrowed the window for a race in the daemon that could cause it to ignore
447 SIGCHLD signals. This is not a major problem, because they are used only to
448 wake it up if nothing else does.
449
62c0818f
PH
45026. A malformed maildirsize file could cause Exim to calculate negative values
451 for the mailbox size or file count. Odd effects could occur as a result.
452 The maildirsize information is now recalculated if the size or filecount
453 end up negative.
454
26034054
PH
45527. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
456 support for a long time. Removed HAVE_SYS_VFS_H.
457
af66f652
PH
45828. Installed the latest version of exipick from John Jetmore.
459
90af77f4
PH
46029. In an address list, if the pattern was not a regular expression, an empty
461 subject address (from a bounce message) matched only if the pattern was an
462 empty string. Non-empty patterns were not even tested. This was the wrong
463 because it is perfectly reasonable to use an empty address as part of a
464 database query. An empty address is now tested by patterns that are
465 lookups. However, all the other forms of pattern expect the subject to
466 contain a local part and a domain, and therefore, for them, an empty
467 address still always fails if the pattern is not itself empty.
468
d8ef3577
PH
46930. Exim went into a mad DNS loop when attempting to do a callout where the
470 host was specified on an smtp transport, and looking it up yielded more
471 than one IP address.
472
5cb8cbc6
PH
47331. Re-factored the code for checking spool and log partition space into a
474 function that finds that data and another that does the check. The former
475 is then used to implement four new variables: $spool_space, $log_space,
476 $spool_inodes, and $log_inodes.
477
14702f5b
PH
47832. The RFC2047 encoding function was originally intended for short strings
479 such as real names; it was not keeping to the 75-character limit for
480 encoded words that the RFC imposes. It now respects the limit, and
481 generates multiple encoded words if necessary. To be on the safe side, I
482 have increased the buffer size for the ${rfc2047: expansion operator from
483 1024 to 2048 bytes.
484
063b1e99
PH
48533. It is now permitted to omit both strings after an "if" condition; if the
486 condition is true, the result is "true". As before, when the second string
487 is omitted, a false condition yields an empty string. This makes it less
488 cumbersome to write custom ACL and router conditions.
489
652e1b65
PH
49034. Failure to deliver a bounce message always caused it to be frozen, even if
491 there was an errors_to setting on the router. The errors_to setting is now
492 respected.
493
6f0c9a4f
PH
49435. If an IPv6 address is given for -bh or -bhc, it is now converted to the
495 canonical form (fully expanded) before being placed in
496 $sender_host_address.
497
33397d19
PH
49836. The table in the code that translates DNS record types into text (T_A to
499 "A" for instance) was missing entries for NS and CNAME. It is just possible
500 that this could have caused confusion if both these types were looked up
501 for the same domain, because the text type is used as part of Exim's
502 per-process caching. But the chance of anyone hitting this buglet seems
503 very small.
504
7bb56e1f
PH
50537. The dnsdb lookup has been extended in a number of ways.
506
507 (1) There is a new type, "zns", which walks up the domain tree until it
508 finds some nameserver records. It should be used with care.
509
ea3bc19b
PH
510 (2) There is a new type, "mxh", which is like "mx" except that it returns
511 just the host names, not the priorities.
512
513 (3) It is now possible to give a list of domains (or IP addresses) to be
ff4dbb19
PH
514 looked up. The behaviour when one of the lookups defers can be
515 controlled by a keyword.
7bb56e1f 516
ea3bc19b 517 (4) It is now possible to specify the separator character for use when
7bb56e1f 518 multiple records are returned.
33397d19 519
0bcb2a0e
PH
52038. The dnslists ACL condition has been extended: it is now possible to supply
521 a list of IP addresses and/or domains to be looked up in a particular DNS
522 domain.
523
2ac0e484
PH
52439. Added log_selector=+queue_time_overall.
525
4e1fde53
PH
52640. When running the queue in the test harness, wait just a tad after forking a
527 delivery process, to get repeatability of debugging output.
528
de365ded
PH
52941. Include certificate and key file names in error message when GnuTLS fails
530 to set them up, because the GnuTLS error message doesn't include the name
531 of the failing file when there is a problem reading it.
532
f05da2e8
PH
53342. Allow both -bf and -bF in the same test run.
534
d6453af2
PH
53543. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
536
f7b63901
PH
53744. The "Exiscan patch" is now merged into the mainline Exim source.
538
53945. Sometimes the final signoff response after QUIT could fail to get
540 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
541 < 0 before doing a fflush(). This bug looks as though it goes back to the
542 introduction of TLS in release 3.20, but "sometimes" must have been rare
543 because the tests only now provoked it.
544
a444213a
PH
54546. Reset the locale to "C" after calling embedded Perl, in case it was changed
546 (this can affect the format of dates).
547
0ec020ea
PH
54847. exim_tidydb, when checking for the continued existence of a message for
549 which it has found a message-specific retry record, was not finding
550 messages that were in split spool directories. Consequently, it was
551 deleting retry records that should have stayed in existence.
552
b1206957
PH
55348. Steve fixed some bugs in eximstats.
554
55549. The SPA authentication driver was not abandoning authentication and moving
556 on to the next authenticator when an expansion was forced to fail,
557 contradicting the general specification for all authenticators. Instead it
558 was generating a temporary error. It now behaves as specified.
559
26dd5a95
PH
56050. The default ordering of permitted cipher suites for GnuTLS was pessimal
561 (the order specifies the preference for clients). The order is now AES256,
562 AES128, 3DES, ARCFOUR128.
563
343b2385
PH
56451. Small patch to Sieve code - explicitly set From: when generating an
565 autoreply.
566
1c5466b9
PH
56752. Exim crashed if a remote delivery caused a very long error message to be
568 recorded - for instance if somebody sent an entire SpamAssassin report back
569 as a large number of 550 error lines. This bug was coincidentally fixed by
570 increasing the size of one of Exim's internal buffers (big_buffer) that
571 happened as part of the Exiscan merge. However, to be on the safe side, I
572 have made the code more robust (and fixed the comments that describe what
573 is going on).
574
55ee9ee3
PH
57553. Now that there can be additional text after "Completed" in log lines (if
576 the queue_time_overall log selector is set), a one-byte patch to exigrep
577 was needed to allow it to recognize "Completed" as not the last thing in
578 the line.
579
d38f8232
PH
58054. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
581 patch that reportedly fixes this has been added. I am not expert enough to
582 create a test for it. This is what the patch creator wrote:
583
584 "I found a little strange behaviour of ldap code when working with
585 Windows 2003 AD Domain, where users was placed in more than one
586 Organization Units. When I tried to give exim partial DN, the exit code
587 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
588 But simultaneously result of request was absolutely normal ldap result,
589 so I produce this patch..."
590
3295e65b
PH
591 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
592 REFERENCE, so I have modified the code to exclude the patch when that macro
593 is not defined.
594
7102e136
PH
59555. Some experimental protocols are using DNS PTR records for new purposes. The
596 keys for these records are domain names, not reversed IP addresses. The
b975ba52
PH
597 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
598 leaves it alone. Component reversal etc. now happens only for IP addresses.
ea3a6f44 599 CAN-2005-0021
7102e136 600
3ca0ba97
PH
60156. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
602
c2bcbe20
PH
60357. Double the size of the debug message buffer (to 2048) so that more of very
604 long debug lines gets shown.
605
18ce445d
PH
60658. The exicyclog utility now does better if the number of log files to keep
607 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
608
1f5b4c3d
PH
60959. Two changes related to the smtp_active_hostname option:
610
611 (1) $smtp_active_hostname is now available as a variable.
612 (2) The default for smtp_banner uses $smtp_active_hostname instead
613 of $primary_hostname.
614
b975ba52
PH
61560. The host_aton() function is supposed to be passed a string that is known
616 to be a valid IP address. However, in the case of IPv6 addresses, it was
617 not checking this. This is a hostage to fortune. Exim now panics and dies
618 if the condition is not met. A case was found where this could be provoked
85b87bc2
PH
619 from a dnsdb PTR lookup with an IPv6 address that had more than 8
620 components; fortuitously, this particular loophole had already been fixed
621 by change 4.50/55 above.
622
623 If there are any other similar loopholes, the new check in host_aton()
624 itself should stop them being exploited. The report I received stated that
625 data on the command line could provoke the exploit when Exim was running as
626 exim, but did not say which command line option was involved. All I could
627 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
628 running as the user.
ea3a6f44 629 CAN-2005-0021
85b87bc2
PH
630
63161. There was a buffer overflow vulnerability in the SPA authentication code
632 (which came originally from the Samba project). I have added a test to the
633 spa_base64_to_bits() function which I hope fixes it.
ea3a6f44 634 CAN-2005-0022
b975ba52 635
17ffcae7
PH
63662. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
637 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
638
d95f9fdb
PH
63963. The daemon start-up calls getloadavg() while still root for those OS that
640 need the first call to be done as root, but it missed one case: when
641 deliver_queue_load_max is set with deliver_drop_privilege. This is
642 necessary for the benefit of the queue runner, because there is no re-exec
643 when deliver_drop_privilege is set.
644
86b8287f
PH
64564. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
646 This has been fixed.
647
60dc5e56
PH
64865. Caching of lookup data for "hosts =" ACL conditions, when a named host list
649 was in use, was not putting the data itself into the right store pool;
650 consequently, it could be overwritten for a subsequent message in the same
651 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
652 the caching.)
653
533244af
PH
65466. Added hosts_max_try_hardlimit to the smtp transport, default 50.
655
a5a28604
PH
65667. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
657 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
658 the function were treating the return as a boolean value, which happened to
659 work because 0=false and not-0=true, but is not correct code.
660
7e634d24
PH
66168. The host_aton() function was not handling scoped IPv6 addresses (those
662 with, for example, "%eth0" on the end) correctly.
663
3e11c26b
PH
66469. Fixed some compiler warnings in acl.c for the bitmaps specified with
665 negated items (that is, ~something) in unsigned ints. Some compilers
666 apparently mutter when there is no cast.
667
6729cf78
PH
66870. If an address verification called from an ACL failed, and did not produce a
669 user-specific message (i.e. there was only a "system" message), nothing was
670 put in $acl_verify_message. In this situation, it now puts the system
671 message there.
672
00f00ca5
PH
67371. Change 4.23/11 added synchronization checking at the start of an SMTP
674 session; change 4.31/43 added the unwanted input to the log line - except
675 that it did not do this in the start of session case. It now does.
676
c9bdd01c
PH
67772. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
678 This is wrong and can cause the other end to generate a synchronization
679 error if it is another Exim or anything else that does the synchronization
680 check. A QUIT command is no longer sent after a timeout.
681
d43194df
PH
68273. $host_lookup_deferred has been added, to make it easier to detect DEFERs
683 during host lookups.
684
fe5b5d0b
PH
68574. The defer_ok option of callout verification was not working if it was used
686 when verifying addresses in header lines, that is, for this case:
687
688 verify = header_sender/callout=defer_ok
689
76a2d7ba
PH
69075. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
691 those file descriptors could be used for SMTP connections. If anything
692 wrote to stderr (the example that came up was "warn" in embedded Perl), it
693 could be sent to the SMTP client, causing chaos. The daemon now opens
694 stdin, stdout, and stderr to /dev/null when it puts itself into the
695 background.
696
69776. Arrange for output from Perl's "warn" command to be written to Exim's main
698 log by default. The user can override this with suitable Perl magic.
699
04f7d5b9
PH
70077. The use of log_message on a "discard" ACL verb, which is supposed to add to
701 the log message when discard triggers, was not working for the DATA ACL or
702 for the non-SMTP ACL.
703
bc60667e
PH
70478. Error message wording change in sieve.c.
705
bb6e88ff
PH
70679. If smtp_accept_max_per_host was set, the number of connections could be
707 restricted to fewer than expected, because the daemon was trying to set up
708 a new connection before checking whether the processes handling previous
709 connections had finished. The check for completed processes is now done
710 earlier. On busy systems, this bug wouldn't be noticed because something
711 else would have woken the daemon, and it would have reaped the completed
712 process earlier.
713
1e70f85b
PH
71480. If a message was submitted locally by a user whose login name contained one
715 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
716 It caused a spool format error. I have fixed the spool reading code. A
717 related problem was that the "from" clause in the Received: line became
718 illegal because of the space(s). It is now covered by ${quote_local_part.
719
72081. Included the latest eximstats from Steve (adds average sizes to HTML Top
721 tables).
722
4e01f9d6
PH
72382. Updated OS/Makefile-AIX as per message from Mike Meredith.
724
1ee1cef2
PH
72583. Patch from Sieve maintainer to fix unterminated string problem in
726 "vacation" handling.
727
6e2b4ccc
PH
72884. Some minor changes to the Linux configuration files to help with other
729 OS variants using glibc.
730
8e669ac1
PH
73185. One more patch for Sieve to update vacation handling to latest spec.
732
495ae4b0 733
bbe902f0
PH
734----------------------------------------------------
735See the note above about the 4.44 and 4.50 releases.
736----------------------------------------------------
737
738
739Exim version 4.44
740-----------------
741
742 1. Change 4.43/35 introduced a bug that caused file counts to be
743 incorrectly computed when quota_filecount was set in an appendfile
744 transport
745
746 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
747 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
748 place.
749
750 3. Give more explanation in the error message when the command for a transport
751 filter fails to execute.
752
753 4. There are several places where Exim runs a non-Exim command in a
754 subprocess. The SIGUSR1 signal should be disabled for these processes. This
755 was being done only for the command run by the queryprogram router. It is
756 now done for all such subprocesses. The other cases are: ${run, transport
757 filters, and the commands run by the lmtp and pipe transports.
758
759 5. Some older OS have a limit of 256 on the maximum number of file
760 descriptors. Exim was using setrlimit() to set 1000 as a large value
761 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
762 systems. I've change it so that if it can't get 1000, it tries for 256.
763
764 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
765 was an oversight, and furthermore, ever since the addition of extra
766 controls (e.g. 4.43/32), the checks on when to allow different forms of
767 "control" were broken. There should now be diagnostics for all cases when a
768 control that does not make sense is encountered.
769
770 7. $recipients is now available in the predata ACL (oversight).
771
772 8. Tidy the search cache before the fork to do a delivery from a message
773 received from the command line. Otherwise the child will trigger a lookup
774 failure and thereby defer the delivery if it tries to use (for example) a
775 cached ldap connection that the parent has called unbind on.
776
777 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
778 of $address_data from the recipient verification was clobbered by the
779 sender verification.
780
78110. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
782 was its contents. (It was OK if the option was not defined at all.)
783
78411. A "Completed" log line is now written for messages that are removed from
785 the spool by the -Mrm option.
786
78712. $host_address is now set to the target address during the checking of
788 ignore_target_hosts.
789
79013. When checking ignore_target_hosts for an ipliteral router, no host name was
791 being passed; this would have caused $sender_host_name to have been used if
792 matching the list had actually called for a host name (not very likely,
793 since this list is usually IP addresses). A host name is now passed as
794 "[x.x.x.x]".
795
79614. Changed the calls that set up the SIGCHLD handler in the daemon to use the
797 code that specifies a non-restarting handler (typically sigaction() in
798 modern systems) in an attempt to fix a rare and obscure crash bug.
799
80015. Narrowed the window for a race in the daemon that could cause it to ignore
801 SIGCHLD signals. This is not a major problem, because they are used only to
802 wake it up if nothing else does.
803
80416. A malformed maildirsize file could cause Exim to calculate negative values
805 for the mailbox size or file count. Odd effects could occur as a result.
806 The maildirsize information is now recalculated if the size or filecount
807 end up negative.
808
80917. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
810 support for a long time. Removed HAVE_SYS_VFS_H.
811
ea3a6f44 81218. Updated exipick to current release from John Jetmore.
bbe902f0
PH
813
81419. Allow an empty sender to be matched against a lookup in an address list.
815 Previously the only cases considered were a regular expression, or an
816 empty pattern.
817
81820. Exim went into a mad DNS lookup loop when doing a callout where the
819 host was specified on the transport, if the DNS lookup yielded more than
820 one IP address.
821
ea3a6f44
NM
82221. The RFC2047 encoding function was originally intended for short strings
823 such as real names; it was not keeping to the 75-character limit for
824 encoded words that the RFC imposes. It now respects the limit, and
825 generates multiple encoded words if necessary. To be on the safe side, I
826 have increased the buffer size for the ${rfc2047: expansion operator from
827 1024 to 2048 bytes.
bbe902f0 828
ea3a6f44
NM
82922. Failure to deliver a bounce message always caused it to be frozen, even if
830 there was an errors_to setting on the router. The errors_to setting is now
831 respected.
bbe902f0
PH
832
83323. If an IPv6 address is given for -bh or -bhc, it is now converted to the
834 canonical form (fully expanded) before being placed in
835 $sender_host_address.
836
83724. Updated eximstats to version 1.33
838
ea3a6f44
NM
83925. Include certificate and key file names in error message when GnuTLS fails
840 to set them up, because the GnuTLS error message doesn't include the name
841 of the failing file when there is a problem reading it.
bbe902f0
PH
842
84326. Expand error message when OpenSSL has problems setting up cert/key files.
ea3a6f44 844 As per change 25.
bbe902f0 845
ea3a6f44
NM
84627. Reset the locale to "C" after calling embedded Perl, in case it was changed
847 (this can affect the format of dates).
bbe902f0 848
ea3a6f44
NM
84928. exim_tidydb, when checking for the continued existence of a message for
850 which it has found a message-specific retry record, was not finding
851 messages that were in split spool directories. Consequently, it was
852 deleting retry records that should have stayed in existence.
bbe902f0
PH
853
85429. eximstats updated to version 1.35
855 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
856 1.35 - bugfix such that pie charts by volume are generated correctly
857
ea3a6f44
NM
85830. The SPA authentication driver was not abandoning authentication and moving
859 on to the next authenticator when an expansion was forced to fail,
860 contradicting the general specification for all authenticators. Instead it
861 was generating a temporary error. It now behaves as specified.
bbe902f0 862
ea3a6f44
NM
86331. The default ordering of permitted cipher suites for GnuTLS was pessimal
864 (the order specifies the preference for clients). The order is now AES256,
865 AES128, 3DES, ARCFOUR128.
bbe902f0 866
ea3a6f44
NM
86731. Small patch to Sieve code - explicitly set From: when generating an
868 autoreply.
bbe902f0 869
ea3a6f44
NM
87032. Exim crashed if a remote delivery caused a very long error message to be
871 recorded - for instance if somebody sent an entire SpamAssassin report back
872 as a large number of 550 error lines. This bug was coincidentally fixed by
873 increasing the size of one of Exim's internal buffers (big_buffer) that
874 happened as part of the Exiscan merge. However, to be on the safe side, I
875 have made the code more robust (and fixed the comments that describe what
876 is going on).
bbe902f0 877
ea3a6f44
NM
87833. Some experimental protocols are using DNS PTR records for new purposes. The
879 keys for these records are domain names, not reversed IP addresses. The
880 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
881 leaves it alone. Component reversal etc. now happens only for IP addresses.
bbe902f0
PH
882 CAN-2005-0021
883
ea3a6f44
NM
88434. The host_aton() function is supposed to be passed a string that is known
885 to be a valid IP address. However, in the case of IPv6 addresses, it was
886 not checking this. This is a hostage to fortune. Exim now panics and dies
887 if the condition is not met. A case was found where this could be provoked
888 from a dnsdb PTR lookup with an IPv6 address that had more than 8
889 components; fortuitously, this particular loophole had already been fixed
890 by change 4.50/55 or 4.44/33 above.
891
892 If there are any other similar loopholes, the new check in host_aton()
893 itself should stop them being exploited. The report I received stated that
894 data on the command line could provoke the exploit when Exim was running as
895 exim, but did not say which command line option was involved. All I could
896 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
897 running as the user.
bbe902f0
PH
898 CAN-2005-0021
899
ea3a6f44
NM
90035. There was a buffer overflow vulnerability in the SPA authentication code
901 (which came originally from the Samba project). I have added a test to the
902 spa_base64_to_bits() function which I hope fixes it.
bbe902f0
PH
903 CAN-2005-0022
904
ea3a6f44
NM
90536. The daemon start-up calls getloadavg() while still root for those OS that
906 need the first call to be done as root, but it missed one case: when
907 deliver_queue_load_max is set with deliver_drop_privilege. This is
908 necessary for the benefit of the queue runner, because there is no re-exec
909 when deliver_drop_privilege is set.
bbe902f0 910
ea3a6f44
NM
91137. Caching of lookup data for "hosts =" ACL conditions, when a named host list
912 was in use, was not putting the data itself into the right store pool;
913 consequently, it could be overwritten for a subsequent message in the same
914 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
915 the caching.)
bbe902f0 916
ea3a6f44
NM
91738. Sometimes the final signoff response after QUIT could fail to get
918 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
919 < 0 before doing a fflush(). This bug looks as though it goes back to the
920 introduction of TLS in release 3.20, but "sometimes" must have been rare
921 because the tests only now provoked it.
bbe902f0
PH
922
923
495ae4b0
PH
924Exim version 4.43
925-----------------
926
927 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
928 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
929 Now it returns an integer. A number of calls were still expecting a T/F
930 return. Fortuitously, in all cases, the tests worked in OK situations,
931 which is the norm. However, things would have gone wrong on any write
932 failures on the smtp file descriptor. This function is used when sending
933 messages over SMTP and also when doing verify callouts.
934
935 2. When Exim is called to do synchronous delivery of a locally submitted
936 message (the -odf or -odi options), it no longer closes stderr before doing
937 the delivery.
938
939 3. Implemented the mua_wrapper option.
940
941 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
942
943 5. Implemented the functions header_remove(), header_testname(),
944 header_add_at_position(), and receive_remove_recipient(), and exported them
945 to local_scan().
946
947 6. If an ACL "warn" statement specified the addition of headers, Exim already
948 inserted X-ACL-Warn: at the start if there was no header name. However, it
949 was not making this test for the second and subsequent header lines if
950 there were newlines in the string. This meant that an invalid header could
951 be inserted if Exim was badly configured.
952
953 7. Allow an ACL "warn" statement to add header lines at the start or after all
954 the Received: headers, as well as at the end.
955
956 8. Added the rcpt_4xx retry error code.
957
958 9. Added postmaster_mailfrom=xxx to callout verification option.
959
96010. Added mailfrom=xxxx to the callout verification option, for verify=
961 header_sender only.
962
96311. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
964 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
965
96612. Inserted some casts to stop certain compilers warning when using pointer
967 differences as field lengths or precisions in printf-type calls (mostly
968 affecting debugging statements).
969
97013. Added optional readline() support for -be (dynamically loaded).
971
97214. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
973 same clock tick as a message's arrival, so that its received time was the
974 same as the "first fail" time on the retry record, and that message
975 remained on the queue past the ultimate address timeout, every queue runner
976 would try a delivery (because it was past the ultimate address timeout) but
977 after another failure, the ultimate address timeout, which should have then
978 bounced the address, did not kick in. This was a "< instead of <=" error;
979 in most cases the first failure would have been in the next clock tick
980 after the received time, and all would be well.
981
98215. The special items beginning with @ in domain lists (e.g. @mx_any) were not
983 being recognized when the domain list was tested by the match_domain
984 condition in an expansion string.
985
98616. Added the ${str2b64: operator.
987
98817. Exim was always calling setrlimit() to set a large limit for the number of
989 processes, without checking whether the existing limit was already
990 adequate. (It did check for the limit on file descriptors.) Furthermore,
991 errors from getrlimit() and setrlimit() were being ignored. Now they are
992 logged to the main and panic logs, but Exim does carry on, to try to do its
993 job under whatever limits there are.
994
99518. Imported PCRE 5.0.
996
99719. Trivial typo in log message " temporarily refused connection" (the leading
998 space).
999
100020. If the log selector return_path_on_delivery was set and an address was
1001 redirected to /dev/null, the delivery process crashed because it assumed
1002 that a return path would always be set for a "successful" delivery. In this
1003 case, the whole delivery is bypassed as an optimization, and therefore no
1004 return path is set.
1005
100621. Internal re-arrangement: the function for sending a challenge and reading
1007 a response while authentication was assuming a zero-terminated challenge
1008 string. It's now changed to take a pointer and a length, to allow for
1009 binary data in such strings.
1010
101122. Added the cyrus_sasl authenticator (code supplied by MBM).
1012
101323. Exim was not respecting finduser_retries when seeking the login of the
1014 uid under which it was called; it was always trying 10 times. (The default
1015 setting of finduser_retries is zero.) Also, it was sleeping after the final
1016 failure, which is pointless.
1017
101824. Implemented tls_on_connect_ports.
1019
102025. Implemented acl_smtp_predata.
1021
102226. If the domain in control=submission is set empty, Exim assumes that the
1023 authenticated id is a complete email address when it generates From: or
1024 Sender: header lines.
1025
102627. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1027 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1028 chown and chgrp in /bin and hostname in /usr/bin.
1029
103028. Exim was keeping the "process log" file open after each use, just as it
1031 does for the main log. This opens the possibility of it remaining open for
1032 long periods when the USR1 signal hits a daemon. Occasional processlog
1033 errors were reported, that could have been caused by this. Anyway, it seems
1034 much more sensible not to leave this file open at all, so that is what now
1035 happens.
1036
103729. The long-running daemon process does not normally write to the log once it
1038 has entered its main loop, and it closes the log before doing so. This is
1039 so that log files can straightforwardly be renamed and moved. However,
1040 there are a couple of unusual error situations where the daemon does write
1041 log entries, and I had neglected to close the log afterwards.
1042
104330. The text of an SMTP error response that was received during a remote
1044 delivery was being truncated at 512 bytes. This is too short for some of
1045 the long messages that one sometimes sees. I've increased the limit to
1046 1024.
1047
104831. It is now possible to make retry rules that apply only when a message has a
1049 specific sender, in particular, an empty sender.
1050
105132. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1052 it possible to be selective about when SMTP synchronization is enforced.
1053
105433. Added "control = caseful_local_part" and "control = "caselower_local_part".
1055
105632. Implemented hosts_connection_nolog.
1057
105833. Added an ACL for QUIT.
1059
106034. Setting "delay_warning=" to disable warnings was not working; it gave a
1061 syntax error.
1062
106335. Added mailbox_size and mailbox_filecount to appendfile.
1064
106536. Added control = no_multiline_responses to ACLs.
1066
106737. There was a bug in the logic of the code that waits for the clock to tick
1068 in the case where the clock went backwards by a substantial amount such
1069 that the microsecond fraction of "now" was more than the microsecond
1070 fraction of "then" (but the whole seconds number was less).
1071
107238. Added support for the libradius Radius client library this is found on
1073 FreeBSD (previously only the radiusclient library was supported).
1074
1075
1076Exim version 4.42
1077-----------------
1078
1079 1. When certain lookups returned multiple values in the form name=value, the
1080 quoting of the values was not always being done properly. Specifically:
1081 (a) If the value started with a double quote, but contained no whitespace,
1082 it was not quoted.
1083 (b) If the value contained whitespace other than a space character (i.e.
1084 tabs or newlines or carriage returns) it was not quoted.
1085 This fix has been applied to the mysql and pgsql lookups by writing a
1086 separate quoting function and calling it from the lookup code. The fix
1087 should probably also be applied to nisplus, ibase and oracle lookups, but
1088 since I cannot test any of those, I have not disturbed their existing code.
1089
1090 2. A hit in the callout cache for a specific address caused a log line with no
1091 reason for rejecting RCPT. Now it says "Previous (cached) callout
1092 verification failure".
1093
1094 3. There was an off-by-one bug in the queryprogram router. An over-long
1095 return line was truncated at 256 instead of 255 characters, thereby
1096 overflowing its buffer with the terminating zero. As well as fixing this, I
1097 have increased the buffer size to 1024 (and made a note to document this).
1098
1099 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1100 when Exim is waiting for an SMTP response from a remote server, Exim
1101 restarts its select() call on the socket, thereby resetting its timeout.
1102 This is not a problem when such interrupts are rare. Somebody set up a cron
1103 job to run exiwhat every 2 minutes, which is less than the normal select()
1104 timeout (5 or 10 minutes). This meant that the select() timeout never
1105 kicked in because it was always reset. I have fixed this by comparing the
1106 time when an interrupt arrives with the time at the start of the first call
1107 to select(). If more time than the timeout has elapsed, the interrupt is
1108 treated as a timeout.
1109
1110 5. Some internal re-factoring in preparation for the addition of Sieve
1111 extensions (by MH). In particular, the "personal" test is moved to a
1112 separate function, and given an option for scanning Cc: and Bcc: (which is
1113 not set for Exim filters).
1114
1115 6. When Exim created an email address using the login of the caller as the
1116 local part (e.g. when creating a From: or Sender: header line), it was not
1117 quoting the local part when it contained special characters such as @.
1118
1119 7. Installed new OpenBSD configuration files.
1120
1121 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1122 try to make them clearer.
1123
1124 9. Callout options, other than the timeout value, were being ignored when
1125 verifying sender addresses in header lines. For example, when using
1126
1127 verify = header_sender/callout=no_cache
1128
1129 the cache was (incorrectly) being used.
1130
113110. Added a missing instance of ${EXE} to the exim_install script; this affects
1132 only the Cygwin environment.
1133
113411. When return_path_on_delivery was set as a log selector, if different remote
1135 addresses in the same message used different return paths and parallel
1136 remote delivery occurred, the wrong values would sometimes be logged.
1137 (Whenever a remote delivery process finished, the return path value from
1138 the most recently started remote delivery process was logged.)
1139
114012. RFC 3848 specifies standard names for the "with" phrase in Received: header
1141 lines when AUTH and/or TLS are in use. This is the "received protocol"
1142 field. Exim used to use "asmtp" for authenticated SMTP, without any
1143 indication (in the protocol name) for TLS use. Now it follows the RFC and
1144 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1145 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1146 names appear in log lines as well as in Received: header lines.
1147
114813. Installed MH's patches for Sieve to add the "copy" and "vacation"
1149 extensions, and comparison tests, and to fix some bugs.
1150
115114. Changes to the "personal" filter test:
1152
1153 (1) The test was buggy in that it was just doing the equivalent of
1154 "contains" tests on header lines. For example, if a user's address was
1155 anne@some.where, the "personal" test would incorrectly be true for
1156
1157 To: susanne@some.where
1158
1159 This test is now done by extracting each address from the header in turn,
1160 and checking the entire address. Other tests that are part of "personal"
1161 are now done using regular expressions (for example, to check local parts
1162 of addresses in From: header lines).
1163
1164 (2) The list of non-personal local parts in From: addresses has been
1165 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1166 taken from the Sieve specification recommendations.
1167
1168 (3) If the message contains any header line starting with "List-" it is
1169 treated as non-personal.
1170
1171 (4) The test for "circular" in the Subject: header line has been removed
1172 because it now seems ill-conceived.
1173
117415. Minor typos in src/EDITME comments corrected.
1175
117616. Installed latest exipick from John Jetmore.
1177
117817. If headers_add on a router specified a text string that was too long for
1179 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1180 of string_sprintf() is now avoided.
1181
118218. $message_body_size was not set (it was always zero) when running the DATA
1183 ACL and the local_scan() function.
1184
118519. For the "mail" command in an Exim filter, no default was being set for
1186 the once_repeat time, causing a random time value to be used if "once" was
1187 specified. (If the value happened to be <= 0, no repeat happened.) The
1188 default is now 0s, meaning "never repeat". The "vacation" command was OK
1189 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1190 (I found it when inspecting the code).
1191
119220. There is now an overall timeout for performing a callout verification. It
1193 defaults to 4 times the callout timeout, which applies to individual SMTP
1194 commands during the callout. The overall timeout applies when there is more
1195 than one host that can be tried. The timeout is checked before trying the
1196 next host. This prevents very long delays if there are a large number of
1197 hosts and all are timing out (e.g. when the network connections are timing
1198 out). The value of the overall timeout can be changed by specifying an
1199 additional sub-option for "callout", called "maxwait". For example:
1200
1201 verify = sender/callout=5s,maxwait=20s
1202
120321. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1204 to the end before writing, but this should make it even safer).
1205
120622. Exim was forgetting that it had advertised PIPELINING for the second and
1207 subsequent messages on an SMTP connection. It was also not resetting its
1208 memory on STARTTLS and an internal HELO.
1209
121023. When Exim logs an SMTP synchronization error within a session, it now
1211 records whether PIPELINING has been advertised or not.
1212
121324. Added 3 instances of "(long int)" casts to time_t variables that were being
1214 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1215 rather than long int.
1216
121725. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1218
121926. Added the never_mail option to autoreply.
1220
1221
1222Exim version 4.41
1223-----------------
1224
1225 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1226 crash if the getsockname() call failed; this can happen if a connection is
1227 closed very soon after it is established. The problem was simply in the
1228 order in which certain operations were done, causing Exim to try to write
1229 to the SMTP stream before it had set up the file descriptor. The bug has
1230 been fixed by making things happen in the correct order.
1231
1232
1233Exim version 4.40
1234-----------------
1235
1236 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1237 before the connection was closed, thus losing the rejection response.
1238
1239 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1240 some early Solaris releases, but causes trouble in current releases where
1241 socklen_t is defined.
1242
1243 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1244 always exist.
1245
1246 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1247 configured.
1248
1249 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1250 that releasing the top of it at the end releases what was used for sub-
1251 expansions (unless the block got too big). However, discard this block if
1252 the first thing is a variable or header, so that we can use its block when
1253 it is dynamic (useful for very large $message_headers, for example).
1254
1255 6. Lookups now cache *every* query, not just the most recent. A new, separate
1256 store pool is used for this. It can be recovered when all lookup caches are
1257 flushed. Lookups now release memory at the end of their result strings.
1258 This has involved some general refactoring of the lookup sources.
1259
1260 7. Some code has been added to the store_xxx() functions to reduce the amount
1261 of flapping under certain conditions.
1262
1263 8. log_incoming_interface used to affect only the <= reception log lines. Now
1264 it causes the local interface and port to be added to several more SMTP log
1265 lines, for example "SMTP connection from", and rejection lines.
1266
1267 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1268
126910. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1270
127111. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1272 could be overwritten at the end of the current message (or the start of a
1273 new message if it was set in a HELO ACL). The value is now preserved for
1274 the duration of the SMTP connection.
1275
127612. If a transport had a headers_rewrite setting, and a matching header line
1277 contained an unqualified address, that address was qualified, even if it
1278 did not match any rewriting rules. The underlying bug was that the values
1279 of the flags that permit the existence of unqualified sender and recipient
1280 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1281 non-local messages, and by -bnq for local messages) were not being
1282 preserved with the message after it was received.
1283
128413. When Exim was logging an SMTP synchronization error, it could sometimes log
1285 "next input=" as part of the text comprising the host identity instead of
1286 the correct text. The code was using the same buffer for two different
1287 strings. However, depending on which order the printing function evaluated
1288 its arguments, the bug did not always show up. Under Linux, for example, my
1289 test suite worked just fine.
1290
129114. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1292 doesn't work with some older versions of Perl. It has been changed to "my",
1293 which in any case is probably the better facility to use.
1294
129515. A really picky compiler found some instances of statements for creating
1296 error messages that either had too many or two few arguments for the format
1297 string.
1298
129916. The size of the buffer for calls to the DNS resolver has been increased
1300 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1301 for addresses that have a lot of PTR records. This alleviates a problem; it
1302 does not fully solve it.
1303
130417. A dnsdb lookup for PTR records that receives more data than will fit in the
1305 buffer now truncates the list and logs the incident, which is the same
1306 action as happens when Exim is looking up a host name and its aliases.
1307 Previously in this situation something unpredictable would happen;
1308 sometimes it was "internal error: store_reset failed".
1309
131018. If a server dropped the connection unexpectedly when an Exim client was
1311 using GnuTLS and trying to read a response, the client delivery process
1312 crashed while trying to generate an error log message.
1313
131419. If a "warn" verb in an ACL added multiple headers to a message in a single
1315 string, for example:
1316
1317 warn message = H1: something\nH2: something
1318
1319 the text was added as a single header line from Exim's point of view
1320 though it ended up OK in the delivered message. However, searching for the
1321 second and subsequent header lines using $h_h2: did not work. This has been
1322 fixed. Similarly, if a system filter added multiple headers in this way,
1323 the routers could not see them.
1324
132520. Expanded the error message when iplsearch is called with an invalid key to
1326 suggest using net-iplsearch in a host list.
1327
132821. When running tests using -bh, any delays imposed by "delay" modifiers in
1329 ACLs are no longer actually imposed (and a message to that effect is
1330 output).
1331
133222. If a "gecos" field in a passwd entry contained escaped characters, in
1333 particular, if it contained a \" sequence, Exim got it wrong when building
1334 a From: or a Sender: header from that name. A second bug also caused
1335 incorrect handling when an unquoted " was present following a character
1336 that needed quoting.
1337
133823. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1339 was not being matched caselessly.
1340
134124. Arranged for all hyphens in the exim.8 source to be escaped with
1342 backslashes.
1343
134425. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1345 properly. Recipient callout cache records were still being keyed to include
1346 the sender, even when use_sender was set false. This led to far more
1347 callouts that were necessary. The sender is no longer included in the key
1348 when use_sender is false.
1349
135026. Added "control = submission" modifier to ACLs.
1351
135227. Added the ${base62d: operator to decode base 62 numbers.
1353
135428. dnsdb lookups can now access SRV records.
1355
135629. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1357 the configuration file.
1358
135930. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1360 (-v) mode. This makes the output for a verbose queue run more intelligible.
1361
136231. Added a use_postmaster feature to recipient callouts.
1363
136432. Added the $body_zerocount variable, containing the number of binary zero
1365 bytes in the message body.
1366
136733. The time of last modification of the "new" subdirectory is now used as the
1368 "mailbox time last read" when there is a quota error for a maildir
1369 delivery.
1370
137134. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1372
137335. Added +ignore_unknown as a special item in host lists.
1374
137536. Code for decoding IPv6 addresses in host lists is now included, even if
1376 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1377 address was recognized as an IP address, but was then not correctly decoded
1378 into binary, causing unexpected and incorrect effects when compared with
1379 another IP address.
1380
1381
1382Exim version 4.34
1383-----------------
1384
1385 1. Very minor rewording of debugging text in manualroute to say "list of
1386 hosts" instead of "hostlist".
1387
1388 2. If verify=header_syntax was set, and a header line with an unqualified
1389 address (no domain) and a large number of spaces between the end of the
1390 name and the colon was received, the reception process suffered a buffer
1391 overflow, and (when I tested it) crashed. This was caused by some obsolete
1392 code that should have been removed. The fix is to remove it!
1393
1394 3. When running in the test harness, delay a bit after writing a bounce
1395 message to get a bit more predictability in the log output.
1396
1397 4. Added a call to search_tidyup() just before forking a reception process. In
1398 theory, someone could use a lookup in the expansion of smtp_accept_max_
1399 per_host which, without the tidyup, could leave open a database connection.
1400
1401 5. Added the variables $recipient_data and $sender_data which get set from a
1402 lookup success in an ACL "recipients" or "senders" condition, or a router
1403 "senders" option, similar to $domain_data and $local_part_data.
1404
1405 6. Moved the writing of debug_print from before to after the "senders" test
1406 for routers.
1407
1408 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1409 problems for message scanning, either using a data ACL, or using
1410 local_scan() because the Received: header was not generated till after they
1411 were called (in order to set the time as the time of reception completion).
1412 I have revised the way this works. The header is now generated after the
1413 body is received, but before the ACL or local_scan() are called. After they
1414 are run, the timestamp in the header is updated.
1415
1416
1417Exim version 4.33
1418-----------------
1419
1420 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1421 before starting a queue runner without re-exec. This happened only when
1422 deliver_drop_privilege was set or when the Exim user was set to root. The
1423 effect of the bug was that timeouts during subsequent deliveries caused
1424 crashes instead of being properly handled. The handler is now left at its
1425 default (and expected) setting.
1426
1427 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1428 message, again when deliver_drop_privilege is set or Exim is run as root.
1429 The bug described in (1) was not present in this case, but the tidying up
1430 of the other signals was missing. I have made the two cases consistent.
1431
1432 3. The ignore_target_hosts setting on a manualroute router was being ignored
1433 for hosts that were looked up using the /MX notation.
1434
1435 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1436 in domain lists.
1437
1438 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1439 operated on the sender address. After changing the $sender_address to <>
1440 for the sender address verify, Exim was re-instated it as the original
1441 (before rewriting) address, but remembering that it had rewritten it, so it
1442 wasn't rewriting it again. This bug also had the effect of breaking the
1443 sender address verification caching when the sender address was rewritten.
1444
1445 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1446 This has been changed so that if the ip literal address matches
1447 ignore_target_hosts, the router declines.
1448
1449 7. Added expansion conditions match_domain, match_address, and match_local_
1450 part (NOT match_host).
1451
1452 8. The placeholder for the Received: header didn't have a length field set.
1453
1454 9. Added code to Exim itself and to exim_lock to test for a specific race
1455 condition that could lead to file corruption when using MBX delivery. The
1456 issue is with the lockfile that is created in /tmp. If this file is removed
1457 after a process has opened it but before that process has acquired a lock,
1458 there is the potential for a second process to recreate the file and also
1459 acquire a lock. This could lead to two Exim processes writing to the file
1460 at the same time. The added code performs the same test as UW imapd; it
1461 checks after acquiring the lock that its file descriptor still refers to
1462 the same named file.
1463
146410. The buffer for building added header lines was of fixed size, 8192 bytes.
1465 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1466 when Exim is built.
1467
146811. Added the smtp_active_hostname option. If used, this will typically be made
1469 to depend on the incoming interface address. Because $interface_address is
1470 not set up until the daemon has forked a reception process, error responses
1471 that can happen earlier (such as "too many connections") no longer contain
1472 a host name.
1473
147412. If an expansion in a condition on a "warn" statement fails because a lookup
1475 defers, the "warn" statement is abandoned, and the next ACL statement is
1476 processed. Previously this caused the whole ACL to be aborted.
1477
147813. Added the iplsearch lookup type.
1479
148014. Added ident_timeout as a log selector.
1481
148215. Added tls_certificate_verified as a log selector.
1483
148416. Added a global option tls_require_ciphers (compare the smtp transport
1485 option of the same name). This controls incoming TLS connections.
1486
148717. I finally figured out how to make tls_require_ciphers do a similar thing
1488 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1489 before starting the TLS session.
1490
149118. Tabs are now shown as \t in -bP output.
1492
149319. If the log selector return_path_on_delivery was set, Exim crashed when
1494 bouncing a message because it had too many Received: header lines.
1495
149620. If two routers both had headers_remove settings, and the first one included
1497 a superfluous trailing colon, the final name in the first list and the
1498 first name in the second list were incorrectly joined into one item (with a
1499 colon in the middle).
1500
1501
1502Exim version 4.32
1503-----------------
1504
1505 1. Added -C and -D options to the exinext utility, mainly to make it easier
1506 to include in the automated testing, but these could be helpful when
1507 multiple configurations are in use.
1508
1509 2. The exinext utility was not formatting the output nicely when there was
1510 an alternate port involved in the retry record key, nor when there was a
1511 message id as well (for retries that were specific to a specific message
1512 and a specific host). It was also confused by IPv6 addresses, because of
1513 the additional colons they contain. I have fixed the IPv4 problem, and
1514 patched it up to do a reasonable job for IPv6.
1515
1516 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1517 delivery, the log line now contains "pipelined" if PIPELINING was used.
1518
1519 4. An SMTP transport process used to panic and die if the bind() call to set
1520 an explicit outgoing interface failed. This has been changed; it is now
1521 treated in the same way as a connect() failure.
1522
1523 5. A reference to $sender_host_name in the part of a conditional expansion
1524 that was being skipped was still causing a DNS lookup. This no longer
1525 occurs.
1526
1527 6. The def: expansion condition was not recognizing references to header lines
1528 that used bh_ and bheader_.
1529
1530 7. Added the _cache feature to named lists.
1531
1532 8. The code for checking quota_filecount in the appendfile transport was
1533 allowing one more file than it should have been.
1534
1535 9. For compatibility with Sendmail, the command line option
1536
1537 -prval:sval
1538
1539 is equivalent to
1540
1541 -oMr rval -oMs sval
1542
1543 and sets the incoming protocol and host name (for trusted callers). The
1544 host name and its colon can be omitted when only the protocol is to be set.
1545 Note the Exim already has two private options, -pd and -ps, that refer to
1546 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1547 "s", but I don't think that's a major issue.
1548
154910. A number of refactoring changes to the code, none of which should affect
1550 Exim's behaviour:
1551
1552 (a) The number of logging options was getting close to filling up the
1553 32-bit word that was used as a bit map. I have split them into two classes:
1554 those that are passed in the argument to log_write(), and those that are
1555 only ever tested independently outside of that function. These are now in
1556 separate 32-bit words, so there is plenty of room for expansion again.
1557 There is no change in the user interface or the logging behaviour.
1558
1559 (b) When building, for example, log lines, the code previously used a
1560 macro that called string_cat() twice, in order to add two strings. This is
1561 not really sufficiently general. Furthermore, there was one instance where
1562 it was actually wrong because one of the argument was used twice, and in
1563 one call a function was used. (As it happened, calling the function twice
1564 did not affect the overall behaviour.) The macro has been replaced by a
1565 function that can join an arbitrary number of extra strings onto a growing
1566 string.
1567
1568 (c) The code for expansion conditions now uses a table and a binary chop
1569 instead of a serial search (which was left over from when there were very
1570 few conditions). Also, it now recognizes conditions like "pam" even when
1571 the relevant support is not compiled in: a suitably worded error message is
1572 given if an attempt is made to use such a condition.
1573
157411. Added ${time_interval:xxxxx}.
1575
157612. A bug was causing one of the ddress fields not to be passed back correctly
1577 from remote delivery subprocesses. The field in question was not being
1578 subsequently used, so this caused to problems in practice.
1579
158013. Added new log selectors queue_time and deliver_time.
1581
158214. Might have fixed a bug in maildirsizefile handling that threw up
1583 "unexpected character" debug warnings, and recalculated the data
1584 unnecessarily. In any case, I expanded the warning message to give more
1585 information.
1586
158715. Added the message "Restricted characters in address" to the statements in
1588 the default ACL that block characters like @ and % in local parts.
1589
159016. Change 71 for release 4.31 proved to be much less benign that I imagined.
1591 Three changes have been made:
1592
1593 (a) There was a serious bug; a negative response to MAIL caused the whole
1594 recipient domain to be cached as invalid, thereby blocking all messages
1595 to all local parts at the same domain, from all senders. This bug has
1596 been fixed. The domain is no longer cached after a negative response to
1597 MAIL if the sender used is not empty.
1598
1599 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1600 been restored.
1601
1602 (c) A new callout option, "use_sender" has been added for people who want
1603 the modified behaviour.
1604
1605
1606Exim version 4.31
1607-----------------
1608
1609 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1610 Larry Rosenman.
1611
1612 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1613 indeed breaks things for older releases.
1614
1615 3. Added additional logging to the case where there is a problem reading data
1616 from a filter that is running in a subprocess using a pipe, in order to
1617 try to track down a specific problem.
1618
1619 4. Testing facility fudge: when running in the test harness and attempting
1620 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1621 getting "No route to host". Convert this to a timeout.
1622
1623 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1624 warning.
1625
1626 6. Some OS don't have socklen_t but use size_t instead. This affects the
1627 fifth argument of getsockopt() amongst other things. This is now
1628 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1629 can be set for individual OS. I have set it for SunOS5, OSF1, and
1630 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1631 some earlier ones do not.
1632
1633 7. Change 4.30/15 was not doing the test caselessly.
1634
1635 8. The standard form for an IPv6 address literal was being rejected by address
1636 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1637 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1638 this, as well as the form without the "IPv6" on the front (but only when
1639 address literals are enabled, of course).
1640
1641 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1642
164310. Exim crashed if a message with an empty sender address specified by -f
1644 encountered a router with an errors_to setting. This could be provoked only
1645 by a command such as
1646
1647 exim -f "" ...
1648
1649 where an empty string was supplied; "<>" did not hit this bug.
1650
165111. Installed PCRE release 4.5.
1652
165312. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1654 remained set. It is now erased.
1655
165613. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1657 times from message ids (which are base 36 rather than the normal 62).
1658
165914. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1660 were being counted as actual protocol errors, and logged if the log
1661 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1662 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1663 and DATA following a set of rejected RCPTs do not count as protocol errors.
1664 In other words, Exim assumes they were pipelined, though this may not
1665 actually be the case. Of course, in all cases the client gets an
1666 appropriate error code.
1667
166815. If a lookup fails in an ACL condition, a message about the failure may
1669 be available; it is used if testing the ACL cannot continue, because most
1670 such messages specify what the cause of the deferral is. However, some
1671 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1672 that caused an old message to be retained and used if a later statement
1673 caused a defer, replacing the real cause of the deferral.
1674
167516. If an IP address had so many PTR records that the DNS lookup buffer
1676 was not large enough to hold them, Exim could crash while trying to process
1677 the truncated data. It now detects and logs this case.
1678
167917. Further to 4.21/58, another change has been made: if (and only if) the
1680 first line of a message (the first header line) ends with CRLF, a bare LF
1681 in a subsequent header line has a space inserted after it, so as not to
1682 terminate the header.
1683
168418. Refactoring: tidied an ugly bit of code in appendfile that copied data
1685 unnecessarily, used atoi() instead of strtol(), and didn't check the
1686 termination when getting file sizes from file names by regex.
1687
168819. Completely re-implemented the support for maildirsize files, in the light
1689 of a number of problems with the previous contributed implementation
1690 (4.30/29). In particular:
1691
1692 . If the quota is zero, the maildirsize file is maintained, but no quota is
1693 imposed.
1694
1695 . If the maildir directory does not exist, it is created before any attempt
1696 to write a maildirsize file.
1697
1698 . The quota value in the file is just a cache; if the quota is changed in
1699 the transport, the new value overrides.
1700
1701 . A regular expression is available for excluding directories from the
1702 count.
1703
170420. The autoreply transport checks the characters in options that define the
1705 message's headers; it allows continued headers, but it was checking with
1706 isspace() after an embedded newline instead of explicitly looking for a
1707 space or a tab.
1708
170921. If all the "regular" hosts to which an address was routed had passed their
1710 expiry times, and had not reached their retry times, the address was
1711 bounced, even if fallback hosts were defined. Now Exim should go on to try
1712 the fallback hosts.
1713
171422. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1715 equivalent code in the SMTP transport. Some hosts send humungous responses
1716 to HELO/EHLO, more than 1024 it seems.
1717
171823. Refactoring: code in filter.c used (void *) for "any old type" but this
1719 gives compiler warnings in some environments. I've now done it "properly",
1720 using a union.
1721
172224. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1723 (because of problems with the built-in one) was declared to return uschar *
1724 instead of char *, causing compiler failure.
1725
172625. Fixed a file descriptor leak when processing alias/forward files.
1727
172826. Fixed a minor format string issue in dbfn.c.
1729
173027. Typo in exim.c: ("dmbnz" for "dbmnz").
1731
173228. If a filter file refered to $h_xxx or $message_headers, and the headers
1733 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1734 become corrupted.
1735
173629. When a sender address is verified, it is cached, to save repeating the test
1737 when there is more than one recipient in a message. However, when the
1738 verification involves a callout, it is possible for different callout
1739 options to be set for different recipients. It is too complicated to keep
1740 track of this in the cache, so now Exim always runs a verification when a
1741 callout is required, relying on the callout cache for the optimization.
1742 The overhead is duplication of the address routing, but this should not be
1743 too great.
1744
174530. Fixed a bug in callout caching. If a RCPT command caused the sender address
1746 to be verified with callout=postmaster, and the main callout worked but the
1747 postmaster check failed, the verification correctly failed. However, if a
1748 subsequent RCPT command asked for sender verification *without* the
1749 postmaster check, incorrect caching caused this verification also to fail,
1750 incorrectly.
1751
175231. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1753 it was not caching the DNS options (qualify_single, search_parents) that
1754 were used when the lookup failed. A subsequent lookup with different
1755 options therefore always gave the same answer, though there were cases
1756 where it should not have. (Example: a "domains = !$mx_any" option on a
1757 dnslookup router: the "domains" option is always processed without any
1758 widening, but the router might have qualify_single set.) Now Exim uses the
1759 cached value only when the same options are set.
1760
176132. Added John Jetmore's "exipick" utility to the distribution.
1762
176333. GnuTLS: When an attempt to start a TLS session fails for any reason other
1764 than a timeout (e.g. a certificate is required, and is not provided), an
1765 Exim server now closes the connection immediately. Previously it waited for
1766 the client to close - but if the client is SSL, it seems that they each
1767 wait for each other, leading to a delay before one of them times out.
1768
176934: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1770 maintained 0.8.x compatibility because I don't think many are using it, and
1771 it is clearly obsolete.
1772
177335. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1774 transport.
1775
177636. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1777 client certificate was expired. A simple patch fixes this, though I don't
1778 understand the full logic of why the verify callback is called multiple
1779 times.
1780
178137. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1782 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1783 which causes problems with some clients (such as the Certicom SSL Plus
1784 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1785 disables the coutermeasure allowing Eudora to connect."
1786
178738. Exim was not checking that a write() to a log file succeeded. This could
1788 lead to Bad Things if a log got too big, in particular if it hit a file
1789 size limit. Exim now panics and dies if it cannot write to a log file, just
1790 as it does if it cannot open a log file.
1791
179239. Modified OS/Makefile-Linux so that it now contains
1793
1794 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1795
1796 The two -D definitions ensure that Exim is compiled with large file
1797 support, which makes it possible to handle log files that are bigger than
1798 2^31.
1799
180040. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1801 instance) a domain was checked against a named list that involved a lookup,
1802 causing $domain_data to be set, then another domain was checked against the
1803 same list, then the first domain was re-checked, the value of $domain_data
1804 after the final check could be wrong. In particular, if the second check
1805 failed, it could be set empty. This bug probably also applied to
1806 $localpart_data.
1807
180841. The strip_trailing_dot option was not being applied to the address given
1809 with the -f command-line option.
1810
181142. The code for reading a message's header from the spool was incrementing
1812 $received_count, but never initializing it. This meant that the value was
1813 incorrect (doubled) while delivering a message in the same process in which
1814 it was received. In the most common configuration of Exim, this never
1815 happens - a fresh exec is done - but it can happen when
1816 deliver_drop_privilege is set.
1817
181843. When Exim logs an SMTP synchronization error - client data sent too soon -
1819 it now includes up to 150 characters of the unexpected data in the log
1820 line.
1821
182244. The exim_dbmbuild utility uses fixed size buffers for reading input lines
1823 and building data strings. The size of both of these buffers was 10 000
1824 bytes - far larger than anybody would *ever* want, thought I. Needless to
1825 say, somebody hit the limit. I have increased the maximum line length to
1826 20 000 and the maximum data length of concatenated lines to 100 000. I have
1827 also fixed two bugs, because there was no checking on these buffers. Tsk,
1828 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
1829 buffer is too small.
1830
183145. The exim_dbmbuild utility did not support quoted keys, as Exim does in
1832 lsearch lookups. Now it does.
1833
183446. When parsing a route_list item in a manualroute router, a fixed-length
1835 buffer was used for the list of hosts. I made this 1024 bytes long,
1836 thinking that nobody would ever have a list of hosts that long. Wrong.
1837 Somebody had a whole pile of complicated expansion conditions, and the
1838 string was silently truncated, leading to an expansion error. It turns out
1839 that it is easier to change to an unlimited length (owing to other changes
1840 that have happened since this code was originally written) than to build
1841 structure for giving a limitation error. The length of the item that
1842 expands into the list of hosts is now unlimited.
1843
184447. The lsearch lookup could not handle data where the length of text line was
1845 more than 4095 characters. Such lines were truncated, leading to shortened
1846 data being returned. It should now handle lines of any length.
1847
184848. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
1849 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
1850 ACL").
1851
185249. Cosmetic tidy to scripts like exicyclog that are generated by globally
1853 replacing strings such as BIN_DIRECTORY in a source file: the replacement
1854 no longer happens in comment lines. A list of replacements is now placed
1855 at the head of all of the source files, except those whose only change is
1856 to replace PERL_COMMAND in the very first #! line.
1857
185850. Replaced the slow insertion sort in queue.c, for sorting the list of
1859 messages on the queue, with a bottom-up merge sort, using code contributed
1860 by Michael Haardt. This should make operations like -bp somewhat faster on
1861 large queues. It won't affect queue runners, except when queue_run_in_order
1862 is set.
1863
186451. Installed eximstats 1.31 in the distribution.
1865
186652. Added support for SRV lookups to the dnslookup router.
1867
186853. If an ACL referred to $message_body or $message_body_end, the value was not
1869 reset for any messages that followed in the same SMTP session.
1870
187154. The store-handling optimization for building very long strings was not
1872 differentiating between the different store pools. I don't think this
1873 actually made any difference in practice, but I've tidied it.
1874
187555. While running the routers to verify a sender address, $sender_address
1876 was still set to the sender address. This is wrong, because when routing to
1877 send a bounce to the sender, it would be empty. Therefore, I have changed
1878 it so that, while verifying a sender address, $sender_address is set to <>.
1879 (There is no change to what happens when verifying a recipient address.)
1880
188156. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1882 target A or AAAA records (if not already returned) without resetting the
1883 qualify_single or search_parents options of the DNS resolver. These are
1884 inappropriate in this case because the targets of MX and SRV records must
1885 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1886 target that, when qualified, matched something in the local domain. These
1887 two options are now turned off when doing these lookups.
1888
188957. It seems that at least some releases of Reiserfs (which does not have the
1890 concept of a fixed number of inodes) returns zero and not -1 for the
1891 number of available inodes. This interacted badly with check_spool_inodes,
1892 which assumed that -1 was the "no such thing" setting. What I have done is
1893 to check that the total number of inodes is greater than zero before doing
1894 the test of how many are available.
1895
189658. When a "warn" ACL statement has a log_message modifier, the message is
1897 remembered, and not repeated. This is to avoid a lot of repetition when a
1898 message has many recipients that cause the same warning to be written.
1899 Howewer, Exim was preserving the list of already written lines for an
1900 entire SMTP session, which doesn't seem right. The memory is now reset if a
1901 new message is started.
1902
190359. The "rewrite" debugging flag was not showing the result of rewriting in the
1904 debugging output unless log_rewrite was also set.
1905
190660. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1907 of (int)(handle) when we know that handle contains (void *)(-1).
1908
190961. The Exim daemon panic-logs an error return when it closes the incoming
1910 connection. However "connection reset by peer" seems to be common, and
1911 isn't really an error worthy of noting specially, so that particular error
1912 is no long logged.
1913
191462. When Exim is trying to find all the local interfaces, it used to panic and
1915 die if the ioctl to get the interface flags failed. However, it seems that
1916 on at least one OS (Solaris 9) it is possible to have an interface that is
1917 included in the list of interfaces, but for which you get a failure error
1918 for this call. This happens when the interface is not "plumbed" into a
1919 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1920 failure of the "get flags" call assumes that the interface is down.
1921
192263. Added a ${eval10: operator, which assumes all numbers are decimal. This
1923 makes life easier for people who are doing arithmetic on fields extracted
1924 from dates, where you often get leading zeros that should not be
1925 interpreted as octal.
1926
192764. Added qualify_domain to the redirect router, to override the global
1928 setting.
1929
193065. If a pathologically long header line contained very many addresses (the
1931 report of this problem mentioned 10 000) and each of them was rewritten,
1932 Exim could use up a very large amount of memory. (It kept on making new
1933 copies of the header line as it rewrote, and never released the old ones.)
1934 At the expense of a bit more processing, the header rewriting function has
1935 been changed so that it no longer eats memory in this way.
1936
193766. The generation of the Received: header has been moved from the time that a
1938 message starts to be received, to the time that it finishes. The timestamp
1939 in the Received: header should now be very close to that of the <= log
1940 line. There are two side-effects of this change:
1941
1942 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1943 logged header lines no longer include the local Received: line, because
1944 it has not yet been created. The same applies to a copy of the message
1945 that is returned to a non-SMTP sender when a message is rejected.
1946
1947 (b) When a filter file is tested using -bf, no additional Received: header
1948 is added to the test message. After some thought, I decided that this
1949 is a bug fix.
1950
1951 This change does not affect the value of $received_for. It is still set
1952 after address rewriting, but before local_scan() is called.
1953
195467. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1955
195668. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1957 gave an unhelpful panic error message, and a defer error. I have managed to
1958 change this behaviour so that it now rejects any supplied certificate,
1959 which seems right, as the list of acceptable certificates is empty.
1960
196169. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1962 gave an unhelpful defer error. I have not managed to make this reject any
1963 supplied certificates, but the error message it gives is "no certificate
1964 supplied", which is not helpful.
1965
196670. exigrep's output now also includes lines that are not associated with any
1967 message, but which match the given pattern. Implemented by a patch from
1968 Martin Sluka, which also tidied up the Perl a bit.
1969
197071. Recipient callout verification, like sender verification, was using <> in
1971 the MAIL FROM command. This isn't really the right thing, since the actual
1972 sender may affect whether the remote host accepts the recipient or not. I
1973 have changed it to use the actual sender in the callout; this means that
1974 the cache record is now keyed on a recipient/sender pair, not just the
1975 recipient address. There doesn't seem to be a real danger of callout loops,
1976 since a callout by the remote host to check the sender would use <>.
1977 [SEE ABOVE: changed after hitting problems.]
1978
197972. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1980 temporary errors. However, in the case of such a code being given after
1981 the end of a data transmission (i.e. after ".") Exim was failing to write
1982 a retry record for the message. (Yes, there was some broken host that was
1983 actually sending 8xx at this point.)
1984
198573. An unknown lookup type in a host list could cause Exim to panic-die when
1986 the list was checked. (An example that provoked this was putting <; in the
1987 middle of a list instead of at the start.) If this happened during a DATA
1988 ACL check, a -D file could be left lying around. This kind of configuration
1989 error no longer causes Exim to die; instead it causes a defer errror. The
1990 incident is still logged to the main and panic logs.
1991
199274. Buglet left over from Exim 3 conversion. The message "too many messages
1993 in one connection" was written to the rejectlog but not the mainlog, except
1994 when address rewriting (yes!) was being logged.
1995
199675. Added write_rejectlog option.
1997
199876. When a system filter was run not as root (that is, when system_filter_user
1999 was set), the values of the $n variables were not being returned to the
2000 main process; thus, they were not subsequently available in the $sn
2001 variables.
2002
200377. Added +return_path_on_delivery log selector.
2004
200578. A connection timeout was being treated differently from recipients deferred
2006 when testing hosts_max_try with a message that was older than the host's
2007 retry timeout. (The host should not be counted, thus allowing all hosts to
2008 be tried at least once before bouncing.) This may have been the cause of an
2009 occasionally reported bug whereby a message would remain on the queue
2010 longer than the retry timeout, but would be bounced if a delivery was
2011 forced. I say "may" because I never totally pinned down the problem;
2012 setting up timeout/retry tests is difficult. See also the next item.
2013
201479. The ultimate address timeout was not being applied to errors that involved
2015 a combination of host plus message (for example, a timeout on a MAIL
2016 command). When an address resolved to a number of possible hosts, and they
2017 were not all tried for each delivery (e.g. because of hosts_max_try), a
2018 message could remain on the queue longer than the retry timeout.
2019
202080. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2021 Haardt.
2022
202381. Fixed an obscure SMTP outgoing bug which required at least the following
2024 conditions: (a) there was another message waiting for the same server;
2025 (b) the server returned 5xx to all RCPT commands in the first message so
2026 that the message was not completed; (c) the server dropped the connection
2027 or gave a negative response to the RSET that Exim sends to abort the
2028 transaction. The observed case was a dropped connection after DATA that had
2029 been sent in pipelining mode. That is, the server had advertised PIPELINING
2030 but was not implementing it correctly. The effect of the bug was incorrect
2031 behaviour, such as trying another host, and this could lead to a crash.
2032
2033
2034Exim version 4.30
2035-----------------
2036
2037 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2038 and daemon.c were passed as pointers to ints; they should have been
2039 pointers to socklen_t variables (which are typically unsigned ints).
2040
2041 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2042 fixed.
2043
2044 3. Fixed a really odd bug that affected only the testing scheme; patching a
2045 certain fixed string in the binary changed the value of another string that
2046 happened to be identical to the end of the original first string.
2047
2048 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2049 name", it returns that address as the IP address. On some operating
2050 systems (e.g. Solaris), it also passes back the IP address string as the
2051 "host name". However, on others (e.g. Linux), it passes back an empty
2052 string. Exim wasn't checking for this, and was changing the host name to an
2053 empty string, assuming it had been canonicized.
2054
2055 5. Although rare, it is permitted to have more than one PTR record for a given
2056 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2057 all the names associated with an address, because they do in Solaris.
2058 However, it seems that they do not in Linux for data that comes from the
2059 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2060 I found this out when I moved to a new Linux workstation and tried to run
2061 the Exim test suite.
2062
2063 To get round this problem I have changed the code so that it now does its
2064 own call to the DNS to look up PTR records when searching for a host name.
2065 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2066 addresses that are only in /etc/hosts are still found.
2067
2068 This behaviour is, however, controlled by an option called host_lookup_
2069 order, which defaults to "bydns:byaddr". If people want to use the other
2070 order, or indeed, just use one or the other means of lookup, they can
2071 specify it in this variable.
2072
2073 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2074 some operating systems, this comes back from gethostbyaddr() as an empty
2075 string, and this is what Exim used to test for. However, it seems that in
2076 other systems, "." is yielded. Exim now tests for this case too.
2077
2078 7. The values of check_spool_space and check_log_space are now held internally
2079 as a number of kilobytes instead of an absolute number of bytes. If a
2080 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2081 kilobyte. This means that much larger values can be stored.
2082
2083 8. Exim monitor: an attempt to get the action menu when not actually pointing
2084 at a message produces an empty menu entitled "No message selected". This
2085 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2086 no entries in it ("Shell widget menu has zero width and/or height"). So I
2087 have added a single, blank menu entry in this case.
2088
2089 9. Added ${quote_local_part.
2090
209110. MIME decoding is now applied to the contents of Subject: header lines when
2092 they are logged.
2093
209411. Now that a reference to $sender_host_address automatically causes a reverse
2095 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2096 host lookup before query-style lookups in lists that might use this
2097 variable. This has therefore been abolished, and the "net-" prefix is no
2098 longer necessary for query-style lookups.
2099
210012. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2101 have been a typo for LFLAGS, so it has been changed.
2102
210313. The install script calls Exim with "-C /dev/null" in order to find the
2104 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2105 to be output. Howeve, since Exim outputs its version number before the
2106 error, it didn't break the script. It just looked ugly. I fixed this by
2107 always allowing "-C /dev/null" if the caller is root.
2108
210914. Ignore overlarge ACL variable number when reading spool file - insurance
2110 against a later release with more variables having written the file.
2111
211215. The standard form for an IPv6 address literal was being rejected by EHLO.
2113 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2114 this, as well as the form without the "IPv6" on the front.
2115
211616. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2117 OS/Makefile-Darwin file.
2118
211917. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2120 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2121
212218. After change 4.21/52, "%ld" was used to format the contents of the $inode
2123 variable. However, some OS use ints for inodes. I've added cast to long int
2124 to get rid of the compiler warning.
2125
212619. I had forgotten to lock out "/../" in configuration file names when
2127 ALT_CONFIG_PREFIX was set.
2128
212920. Routers used for verification do not need to specify transports. However,
2130 if such a router generated a host list, and callout was configured, Exim
2131 crashed, because it could not find a port number from the (non-existent)
2132 transport. It now assumes port 25 in this circumstance.
2133
213421. Added the -t option to exigrep.
2135
213622. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2137 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2138 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2139 them is set, LOOKUP_LSEARCH is forced.
2140
214123. "exim -bV" now outputs a list of lookups that are included in the binary.
2142
214324. Added sender and host information to the "rejected by local_scan()" log
2144 line; previously there was no indication of these.
2145
214625. Added .include_if_exists.
2147
214826. Change 3.952/11 added an explicit directory sync on top of a file sync for
2149 Linux. It turns out that not all file systems support this. Apparently some
2150 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2151 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2152 not supported on the file descriptor, is now ignored when Exim is trying to
2153 sync a directory. This applies only to Linux.
2154
215527. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2156
215728. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2158 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2159 this.
2160
216129. Added support for maildirsize files from supplied patch (modified a bit).
2162
216330. The use of :fail: followed by an empty string could lead Exim to respond to
2164 sender verification failures with (e.g.):
2165
2166 550 Verification failed for <xxx>
2167 550 Sender verify failed
2168
2169 where the first response line was missing the '-' that indicates it is not
2170 the final line of the response.
2171
217231. The loop for finding the name of the user that called Exim had a hardwired
2173 limit of 10; it now uses the value of finduser_retries, which is used for
2174 all other user lookups.
2175
217632. Added $received_count variable, available in data and not_smtp ACLs, and at
2177 delivery time.
2178
217933. Exim was neglecting to zero errno before one call of strtol() when
2180 expanding a string and expecting an integer value. On some systems this
2181 resulted in spurious "integer overflow" errors. Also, it was casting the
2182 result into an int without checking.
2183
218434. Testing for a connection timeout using "timeout_connect" in the retry rules
2185 did not work. The code looks as if it has *never* worked, though it appears
2186 to have been documented since at least releast 1.62. I have made it work.
2187
218835. The "timeout_DNS" error in retry rules, also documented since at least
2189 1.62, also never worked. As it isn't clear exactly what this means, and
2190 clearly it isn't a major issue, I have abolished the feature by treating it
2191 as "timeout", and writing a warning to the main and panic logs.
2192
219336. The display of retry rules for -brt wasn't always showing the error code
2194 correctly.
2195
219637. Added new error conditions to retry rules: timeout_A, timeout_MX,
2197 timeout_connect_A, timeout_connect_MX.
2198
219938. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2200 to the empty sender.
2201
220239. The daemon was not analysing the content of -oX till after it had closed
2203 stderr and disconnected from the controlling terminal. This meant that any
2204 syntax errors were only noted on the panic log, and the return code from
2205 the command was 0. By re-arranging the code a little, I've made the
2206 decoding happen first, so such errors now appear on stderr, and the return
2207 code is 1. However, the actual setting up of the sockets still happens in
2208 the disconnected process, so errors there are still only recorded on the
2209 panic log.
2210
221140. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2212 connections (as happens on some IP stacks) was logged at start up time as
2213 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2214 it from "IPv6 and IPv4", which means that two separate sockets are being
2215 used.
2216
221741. The debug output for gethostbyname2() or getipnodebyname() failures now
2218 says whether AF_INET or AF_INET6 was passed as an argument.
2219
222042. Exiwhat output was messed up when time zones were included in log
2221 timestamps.
2222
222343. Exiwhat now gives more information about the daemon's listening ports,
2224 and whether -tls-on-connect was used.
2225
222644. The "port" option of the smtp transport is now expanded.
2227
222845. A "message" modifier in a "warn" statement in a non-message ACL was being
2229 silently ignored. Now an error message is written to the main and panic
2230 logs.
2231
223246. There's a new ACL modifier called "logwrite" which writes to a log file
2233 as soon as it is encountered.
2234
223547. Added $local_user_uid and $local_user_gid at routing time.
2236
223748. Exim crashed when trying to verify a sender address that was being
2238 rewritten to "<>".
2239
224049. Exim was recognizing only a space character after ".include". It now also
2241 recognizes a tab character.
2242
224350. Fixed several bugs in the Perl script that creates the exim.8 man page by
2244 extracting the relevant information from the specification. The man page no
2245 longer contains scrambled data for the -d option, and I've added a section
2246 at the front about calling Exim under different names.
2247
224851. Added "extra_headers" argument to the "mail" command in filter files.
2249
225052. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2251 crash.
2252
225353. Installed eximstats 1.29.
2254
225554. Added transport_filter_timeout as a generic transport option.
2256
225755. Exim no longer adds an empty Bcc: header to messages that have no To: or
2258 Cc: header lines. This was required by RFC 822, but it not required by RFC
2259 2822.
2260
226156. Exim used to add From:, Date:, and Message-Id: header lines to any
2262 incoming messages that did not have them. Now it does so only if the
2263 message originates locally, that is, if there is no associated remote host
2264 address. When Resent- header lines are present, this applies to the Resent-
2265 lines rather than the non-Resent- lines.
2266
226757. Drop incoming SMTP connection after too many syntax or protocol errors. The
2268 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2269
227058. Messages for configuration errors now include the name of the main
2271 configuration file - useful now that there may be more than one file in a
2272 list (.included file names were always shown).
2273
227459. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2275 for those rare installations that do not start the daemon as root or run it
2276 setuid root. I've cut out the call to initgroups() if the daemon is not
2277 root at that time.
2278
227960. The Exim user and group can now be bound into the binary as text strings
2280 that are looked up at the start of Exim's processing.
2281
228261. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2283
228462. Added $mailstore_basename variable.
2285
228663. Installed patch to sieve.c from Michael Haardt.
2287
228864. When Exim failed to open the panic log after failing to open the main log,
2289 the original message it was trying to log was written to stderr and debug
2290 output, but if they were not available (the usual case in production), it
2291 was lost. Now it is written to syslog before the two lines that record the
2292 failures to open the logs.
2293
229465. Users' Exim filters run in subprocesses under the user's uid. It is
2295 possible for a "deliver" command or an alias in a "personal" command to
2296 provoke an address rewrite. If logging of address rewriting is configured,
2297 this fails because the process is not running as root or exim. There may be
2298 a better way of dealing with this, but for the moment (because 4.30 needs
2299 to be released), I have disabled address rewrite logging when running a
2300 filter in a non-root, non-exim process.
2301
2302
2303Exim version 4.24
2304-----------------
2305
2306 1. The buildconfig auxiliary program wasn't quoting the value set for
2307 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2308 not defined. This bug was masked in 4.22 by the effect that was fixed in
2309 change 4.23/1.
2310
2311 2. Some messages that were rejected after a message id was allocated were
2312 shown as "incomplete" by exigrep. It no longer does this for messages that
2313 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2314
2315 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2316 have allow_domain_literals set, the ID did not get logged in the <= line.
2317 Domain literals are now always recognized in Message-ID: header lines.
2318
2319 4. The first argument for a ${extract expansion item is the key name or field
2320 number. Leading and trailing spaces in this item were not being ignored,
2321 causing some misleading effects.
2322
2323 5. When deliver_drop_privilege was set, single queue runner processes started
2324 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2325 same command in the process it spins off) were not dropping privilege.
2326
2327 6. When the daemon running as "exim" started a queue runner, it always
2328 re-executed Exim in the spun-off process. This is a waste of effort when
2329 deliver_drop_privilege is set. The new process now just calls the
2330 queue-runner function directly.
2331
2332
2333Exim version 4.23
2334-----------------
2335
2336 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2337 HEADERS_CHARSET.
2338
2339 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2340 ignored. Though the use of -oP was forcing the writing of a pid file, it
2341 was always written to the default place.
2342
2343 3. If the message "no IP address found for host xxxx" is generated during
2344 incoming verification, it is now followed by identification of the incoming
2345 connection (so you can more easily find what provoked it).
2346
2347 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2348
2349 5. Added some features to "harden" Exim a bit more against certain attacks:
2350
2351 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2352 be put in Local/Makefile. This is like the never_users runtime option,
2353 but it cannot be overridden. The default setting is "root".
2354
2355 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2356 prefix string with which any file named in a -C command line option
2357 must start.
2358
2359 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2360 is retained for -C and -D only if the caller of Exim is root. Without
2361 it, the exim user may also use -C and -D and retain privilege.
2362
2363 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2364 command line option is disabled.
2365
2366 6. Macro names set by the -D option must start with an upper case letter, just
2367 like macro names defined in the configuration file.
2368
2369 7. Added "dereference=" facility to LDAP.
2370
2371 8. Two instances of the typo "uknown" in the source files are fixed.
2372
2373 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2374 the Configure-Makefile script screwed up while processing it.
2375
237610. Incorporated PCRE 4.4.
2377
237811. The SMTP synchronization check was not operating right at the start of an
2379 SMTP session. For example, it could not catch a HELO sent before the client
2380 waited for the greeting. There is now a check for outstanding input at the
2381 point when the greeting is written. Because of the duplex, asynchronous
2382 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2383 way, but not yet received, when the check is performed.
2384
238512. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2386 on TCP/IP sockets, because this apparently causes some broken clients to
2387 timeout.
2388
238913. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2390 unchanged) from the Cygwin maintainer.
2391
239214. The code for -bV that shows what is in the binary showed "mbx" when maildir
2393 was supported instead of testing for mbx. Effectively a typo.
2394
239515. The spa authenticator server code was not checking that the input it
2396 received was valid base64.
2397
239816. The debug output line for the "set" modifier in ACLs was not showing the
2399 name of the variable that was being set.
2400
240117. Code tidy: the variable type "vtype_string" was never used. Removed it.
2402
240318. Previously, a reference to $sender_host_name did not cause a DNS reverse
2404 lookup on its own. Something else was needed to trigger the lookup. For
2405 example, a match in host_lookup or the need for a host name in a host list.
2406 Now, if $sender_host_name is referenced and the host name has not yet been
2407 looked up, a lookup is performed. If the lookup fails, the variable remains
2408 empty, and $host_lookup_failed is set to "1".
2409
241019. Added "eqi" as a case-independent comparison operator.
2411
241220. The saslauthd authentication condition could segfault if neither service
2413 nor realm was specified.
2414
241521. If an overflowing value such as "2048M" was set for message_size_limit, the
2416 error message that was logged was misleading, and incoming SMTP
2417 connections were dropped. The message is now more accurate, and temporary
2418 errors are given to SMTP connections.
2419
242022. In some error situations (such as 21 above) Exim rejects all SMTP commands
2421 (except RSET) with a 421 error, until QUIT is received. However, it was
2422 failing to send a response to QUIT.
2423
242423. The HELO ACL was being run before the code for helo_try_verify_hosts,
2425 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2426 ACL is now run after the helo_try_verify_hosts code.
2427
242824. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2429 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2430 case-independent, so other case variants are also recognized). Apparently
2431 some systems use these upper case variants.
2432
243325. If more than two messages were waiting for the same host, and a transport
2434 filter was specified for the transport, Exim sent two messages over the
2435 same TCP/IP connection, and then failed with "socket operation on non-
2436 socket" when it tried to send the third.
2437
243826. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2439
244027. The extern definition of crypt16() in expand.c was not being excluded when
2441 the OS had its own crypt16() function.
2442
244328. Added bounce_return_body as a new option, and bounce_return_size_limit
2444 as a preferred synonym for return_size_limit, both as an option and as an
2445 expansion variable.
2446
244729. Added LIBS=-liconv to OS/Makefile-OSF1.
2448
244930. Changed the default configuration ACL to relax the local part checking rule
2450 for addresses that are not in any local domains. For these addresses,
2451 slashes and pipe symbols are allowed within local parts, but the sequence
2452 /../ is explicitly forbidden.
2453
245431. SPA server authentication was not clearing the challenge buffer before
2455 using it.
2456
245732. log_message in a "warn" ACL statement was writing to the reject log as
2458 well as to the main log, which contradicts the documentation and doesn't
2459 seem right (because no rejection is happening). So I have stopped it.
2460
246133. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2462 However, I am unable to do any testing of this.
2463
246434. Fixed an infelicity in the appendfile transport. When checking directories
2465 for a mailbox, to see if any needed to be created, it was accidentally
2466 using path names with one or more superfluous leading slashes; tracing
2467 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2468
246935. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2470 contents are added to the log line that is written for every discarded
2471 recipient. (Previously a log_message setting was ignored.)
2472
247336. The ${quote: operator now quotes the string if it is empty.
2474
247537. The install script runs exim in order to find its version number. If for
2476 some reason other than non-existence or emptiness, which it checks, it
2477 could not run './exim', it was installing it with an empty version number,
2478 i.e. as "exim-". This error state is now caught, and the installation is
2479 aborted.
2480
248138. An argument was missing from the function that creates an error message
2482 when Exim fails to connect to the socket for saslauthd authentication.
2483 This could cause Exim to crash, or give a corrupted message.
2484
248539. Added isip, isip4, and isip6 to ${if conditions.
2486
248740. The ACL variables $acl_xx are now saved with the message, and can be
2488 accessed later in routers, transports, and filters.
2489
249041. The new lookup type nwildlsearch is like wildlsearch, except that the key
2491 strings in the file are not string-expanded.
2492
249342. If a MAIL command specified a SIZE value that was too large to fit into an
2494 int variable, the check against message_size_limit failed. Such values are
2495 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2496 day this will have to be increased, but I don't think I want to be around
2497 when emails are that large.
2498
2499
2500
2501Exim version 4.22
2502-----------------
2503
2504 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2505 iconv() is not standard in FreeBSD.
2506
2507 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2508 IPv6 enabled. The observed symptom was a segmentation fault on return from
2509 the function os_common_find_running_interfaces() in src/os.c.
2510
2511 3. In the check_special_case() function in daemon.c I had used "errno" as an
2512 argument name, which causes warnings on some systems. This was basically a
2513 typo, since it was named "eno" in the comments!
2514
2515 4. The code that waits for the clock to tick (at a resolution of some fraction
2516 of a second) so as to ensure message-id uniqueness was always waiting for
2517 at least one whole tick, when it could have waited for less. [This is
2518 almost certainly not relevant at current processor speeds, where it is
2519 unlikely to ever wait at all. But we try to future-proof.]
2520
2521 5. The function that sleeps for a time interval that includes fractions of a
2522 second contained a race. It did not block SIGALRM between setting the
2523 timer, and suspending (a couple of lines later). If the interval was short
2524 and the sigsuspend() was delayed until after it had expired, the suspension
2525 never ended. On busy systems this could lead to processes getting stuck for
2526 ever.
2527
2528 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2529 process, before it forks any delivery processes. The open lookup caching
2530 mechanism meant that the open file or database connection was passed into
2531 the delivery process. The problem was that delivery processes always tidy
2532 up cached lookup data. This could cause a problem for the next delivery
2533 process started by the queue runner, because the external queue runner
2534 process does not know about the closure. So the next delivery process
2535 still has data in the lookup cache. In the case of a file lookup, there was
2536 no problem because closing a file descriptor in a subprocess doesn't affect
2537 the parent. However, if the lookup was caching a connection to a database,
2538 the connection was closed, and the second delivery process was likely to
2539 see errors such as "PGSQL: query failed: server closed the connection
2540 unexpectedly". The problem has been fixed by closing all cached lookups
2541 in a queue runner before running a delivery process.
2542
2543 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2544 seem to have the "const" qualifier which it has on other OS. I've
2545 parameterised it.
2546
2547 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2548 *of the same type* (client or server) with the same public name that an
2549 error should be diagnosed.
2550
2551 9. When Exim looked up a host name for an IP address, but failed to find the
2552 original IP address when looking up the host name (a safety check), it
2553 output the message "<ip address> does not match any IP for NULL", which was
2554 confusing, to say the least. The bug was that the host name should have
2555 appeared instead of "NULL".
2556
255710. Since release 3.03, if Exim is called by a uid other than root or the Exim
2558 user that is built into the binary, and the -C or -D options is used, root
2559 privilege is dropped before the configuration file is read. In addition,
2560 logging is switched to stderr instead of the normal log files. If the
2561 configuration then re-defines the Exim user, the unprivileged environment
2562 is probably not what is expected, so Exim logs a panic warning message (but
2563 proceeds).
2564
2565 However, if deliver_drop_privilege is set, the unprivileged state may well
2566 be exactly what is intended, so the warning has been cut out in that case,
2567 and Exim is allowed to try to write to its normal log files.
2568
2569
2570Exim version 4.21
2571-----------------
2572
2573 1. smtp_return_error_details was not giving details for temporary sender
2574 or receiver verification errors.
2575
2576 2. Diagnose a configuration error if two authenticators have the same public
2577 name.
2578
2579 3. Exim used not to create the message log file for a message until the first
2580 delivery attempt. This could be confusing when incoming messages were held
2581 for policy or load reasons. The message log file is now created at the time
2582 the message is received, and an initial "Received" line is written to it.
2583
2584 4. The automatically generated man page for command line options had a minor
2585 bug that caused no ill effects; however, a more serious problem was that
2586 the procedure for building the man page automatically didn't always
2587 operate. Consequently, release 4.20 contains an out-of-date version. This
2588 shouldn't happen again.
2589
2590 5. When building Exim with embedded Perl support, the script that builds the
2591 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2592 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2593
2594 6. The freeze_tell option was not being used for messages that were frozen on
2595 arrival, either by an ACL or by local_scan().
2596
2597 7. Added the smtp_incomplete_transaction log selector.
2598
2599 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2600 was accepting AUTH without a new EHLO.
2601
2602 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2603 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2604 RFC.
2605
260610. Logging of TCP/IP connections (when configured) now happens in the main
2607 daemon process instead of the child process, so that the TCP/IP connection
2608 count is more accurate (but it can never be perfect).
2609
261011. The use of "drop" in a nested ACL was not being handled correctly in the
2611 outer ACL. Now, if condition failure induced by the nested "drop" causes
2612 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2613 or "require"), the connection is dropped.
2614
261512. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2616 that yield "discard" can now be used with an "accept" or a "discard" verb,
2617 but an error is generated for any others (because I can't see a useful way
2618 to define what should happen).
2619
262013. When an ACL is read dynamically from a file (or anywhere else), the lines
2621 are now processed in the same way as lines in the Exim configuration file.
2622 In particular, continuation lines are supported.
2623
262414. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2625
262615. Added -ti meaning -t -i.
2627
262816. Check for letters, digits, hyphens, and dots in the names of dnslist
2629 domains, and warn by logging if others are found.
2630
263117. At least on BSD, alignment is not guarenteed for the array of ifreq's
2632 returned from GIFCONF when Exim is trying to find the list of interfaces on
2633 a host. The code in os.c has been modified to copy each ifreq to an aligned
2634 structure in all cases.
2635
2636 Also, in some cases, the returned ifreq's were being copied to a 'struct
2637 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2638 means the last couple of bytes of an IPv6 address could be chopped if the
2639 ifreq contained only a normal sockaddr (14 bytes storage).
2640
264118. Named domain lists were not supported in the hosts_treat_as_local option.
2642 An entry such as +xxxx was not recognized, and was treated as a literal
2643 domain name.
2644
264519. Ensure that header lines added by a DATA ACL are included in the reject log
2646 if the ACL subsequently rejects the message.
2647
264820. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2649 MD5 (which is deprecated).
2650
265121. When testing a filter file using -bf, Exim was writing a message when it
2652 took the sender from a "From " line in the message, but it was not doing so
2653 when it took $return_path from a Return-Path: header line. It now does.
2654
265522. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2656 with a valid header line field name (a series of printing characters
2657 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2658
265923. Changed "disc" in the source to "disk" to conform to the documentation and
2660 the book and for uniformity.
2661
266224. Ignore Sendmail's -Ooption=value command line item.
2663
266425. When execve() failed while trying to run a command in a pipe transport,
2665 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2666 could be confused with a return value of 69 from the command itself. This
2667 has been changed to 127, the value the shell returns if it is asked to run
2668 a non-existent command. The wording for the related log line suggests a
2669 non-existent command as the problem.
2670
267126. If received_header_text expands to an empty string, do not add a Received:
2672 header line to the message. (Well, it adds a token one on the spool, but
2673 marks it "old" so that it doesn't get used or transmitted.)
2674
267527. Installed eximstats 1.28 (addition of -nt option).
2676
267728. There was no check for failure on the call to getsockname() in the daemon
2678 code. This can fail if there is a shortage of resources on the system, with
2679 ENOMEM, for example. A temporary error is now given on failure.
2680
268129. Contrary to the C standard, it seems that in some environments, the
2682 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2683 program. Exim now does this explicitly; it affects the formatting of
2684 timestamps using strftime().
2685
268630. If exiqsumm was given junk data, it threw up some uninitialized variable
2687 complaints. I've now initialized all the variables, to avoid this.
2688
268932. Header lines added by a system filter were not being "seen" during
2690 transport-time rewrites.
2691
269233. The info_callback() function passed to OpenSSL is set up with type void
2693 (*)(SSL *, int, int), as described somewhere. However, when calling the
2694 function (actually a macro) that sets it up, the type void(*)() is
2695 expected. I've put in a cast to prevent warnings from picky compilers.
2696
269734. If a DNS black list lookup found a CNAME record, but there were no A
2698 records associated with the domain it pointed at, Exim crashed.
2699
270035. If a DNS black list lookup returned more than one A record, Exim ignored
2701 all but the first. It now scans all returned addresses if a particular IP
2702 value is being sought. In this situation, the contents of the
2703 $dnslist_value variable are a list of all the addresses, separated by a
2704 comma and a space.
2705
270636. Tightened up the rules for host name lookups using reverse DNS. Exim used
2707 to accept a host name and all its aliases if the forward lookup for any of
2708 them yielded the IP address of the incoming connection. Now it accepts only
2709 those names whose forward lookup yields the correct IP address. Any other
2710 names are discarded. This closes a loophole whereby a rogue DNS
2711 administrator could create reverse DNS records to break through a
2712 wildcarded host restriction in an ACL.
2713
271437. If a user filter or a system filter that ran in a subprocess used any of
2715 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2716 the wrong values were passed to the pipe command ($thisaddress had the
2717 value of $0, $0 had the value of $1, etc). This bug was introduced by
2718 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2719
272038. Improved the line breaking for long SMTP error messages from ACLs.
2721 Previously, if there was no break point between 40 and 75 characters, Exim
2722 left the rest of the message alone. Two changes have been made: (a) I've
2723 reduced the minimum length to 35 characters; (b) if it can't find a break
2724 point between 35 and 75 characters, it looks ahead and uses the first one
2725 that it finds. This may give the occasional overlong line, but at least the
2726 remaining text gets split now.
2727
272839. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2729 file descriptors might be low, and that setting 1000 would always raise it.
2730 It turns out that in some environments, the limit is already over 1000 and
2731 that lowering it causes trouble. So now Exim takes care not to decrease it.
2732
273340. When delivering a message, the value of $return_path is set to $sender_
2734 address at the start of routing (routers may change the value). By an
2735 oversight, this default was not being set up when an address was tested by
2736 -bt or -bv, which affected the outcome if any router or filter referred to
2737 $return_path.
2738
273941. The idea of the "warn" ACL verb is that it adds a header or writes to the
2740 log only when "message" or "log_message" are set. However, if one of the
2741 conditions was an address verification, or a call to a nested ACL, the
2742 messages generated by the underlying test were being passed through. This
2743 no longer happens. The underlying message is available in $acl_verify_
2744 message for both "message" and "log_message" expansions, so it can be
2745 passed through if needed.
2746
274742. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2748 new expansion $bh_ to give the encoded byte string without charset
2749 translation. Translation happens only if iconv() is available; HAVE_ICONV
2750 indicates this at build time. HEADERS_CHARSET gives the charset to
2751 translate to; headers_charset can change it in the configuration, and
2752 "headers charset" can change it in an individual filter file.
2753
275443. Now that we have a default RFC 2047 charset (see above), the code in Exim
2755 that creates RFC 2047 encoded "words" labels them as that charset instead
2756 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2757 expansion operator; (ii) when Exim creates a From: line for a local
2758 message; (iii) when a header line is rewritten to include a "phrase" part.
2759
276044. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2761 buggy, causing it to skip the first lines of messages whose message ID
2762 ended in 'D'. This would not have bitten before Exim release 4.14, because
2763 message IDs were unlikely to end in 'D' before then. The effect was to have
2764 incorrect size information for certain domains.
2765
276645. #include "config.h" was missing at the start of the crypt16.c module. This
2767 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2768 noticed.
2769
277046. If there was a timeout during a "random" callout check, Exim treated it as
2771 a failure of the random address, and carried on sending RSET and the real
2772 address. If the delay was just some slowness somewhere, the response to the
2773 original RCPT would be taken as a response to RSET and so on, causing
2774 mayhem of various kinds.
2775
277647. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2777 when I implemented it. It didn't allow for the fact that some option values
2778 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2779 the right test for positive values.
2780
278148. Domain names in DNS records are case-independent. Exim always looks them up
2782 in lower case. Some resolvers return domain names in exactly the case they
2783 appear in the zone file, that is, they may contain uppercase letters. Not
2784 all resolvers do this - some return always lower case. Exim was treating a
2785 change of case by a resolver as a change of domain, similar to a widening
2786 of a domain abbreviation. This triggered its re-routing code and so it was
2787 trying to route what was effectively the same domain again. This normally
2788 caused routing to fail (because the router wouldn't handle the domain
2789 twice). Now Exim checks for this case specially, and just changes the
2790 casing of the domain that it ultimately uses when it transmits the message
2791 envelope.
2792
279349. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2794 module.
2795
279650. If a filter generated a file delivery with a non-absolute name (possible if
2797 no home directory exists for the router), the forbid_file option was not
2798 forbidding it.
2799
280051. Added '&' feature to dnslists, to provide bit mask matching in addition to
2801 the existing equality matching.
2802
280352. Exim was using ints instead of ino_t variables in some places where it was
2804 dealing with inode numbers.
2805
280653. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2807 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2808 called TMPDIR, and if it finds it is different, it changes its value.
2809
281054. The smtp_printf() function is now made available to local_scan() so
2811 additional output lines can be written before returning. There is also an
2812 smtp_fflush() function to enable the detection of a dropped connection.
2813 The variables smtp_input and smtp_batched_input are exported to
2814 local_scan().
2815
281655. Changed the default runtime configuration: the message "Unknown user"
2817 has been removed from the ACL, and instead placed on the localuser router,
2818 using the cannot_route_message feature. This means that any verification
2819 failures that generate their own messages won't get overridden. Similarly,
2820 the "Unrouteable address" message that was in the ACL for unverifiable
2821 relay addresses has also been removed.
2822
282356. Added hosts_avoid_esmtp to the smtp transport.
2824
282557. The exicyclog script was not checking for the esoteric option
2826 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
2827 will work only if exicyclog is run under the appropriate euid.
2828
282958. Following a discussion on the list, the rules by which Exim recognises line
2830 endings on incoming messages have been changed. The -dropcr and drop_cr
2831 options are now no-ops, retained only for backwards compatibility. The
2832 following line terminators are recognized: LF CRLF CR. However, special
2833 processing applies to CR:
2834
2835 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
2836 nor a local message in the state where . is a terminator.
2837
2838 (ii) If a bare CR is encountered in a header line, an extra space is added
2839 after the line terminator so as not to end the header. The reasoning
2840 behind this is that bare CRs in header lines are most likely either
2841 to be mistakes, or people trying to play silly games.
2842
284359. The size of a message, as listed by "-bp" or in the Exim monitor window,
2844 was being incorrectly given as 18 bytes larger than it should have been.
2845 This is a VOB (very old bug).
2846
284760. This may never have affected anything current, but just in case it has:
2848 When the local host is found other than at the start of a list of hosts,
2849 the local host, those with the same MX, and any that follow, are discarded.
2850 When the list in question was part of a longer list of hosts, the following
2851 hosts (not currently being processed) were also being discarded. This no
2852 longer happens. I'm not sure if this situation could ever has previously
2853 arisen.
2854
285561. Added the "/MX" feature to lists of hosts in the manualroute and query
2856 program routers.
2857
285862. Whenever Exim generates a new message, it now adds an Auto-Submitted:
2859 header. This is something that is recommended in a new Internet Draft, and
2860 is something that is documented as being done by Sendmail. There are two
2861 possible values. For messages generated by the autoreply transport, Exim
2862 adds:
2863
2864 Auto-Submitted: auto-replied
2865
2866 whereas for all other generated messages (e.g. bounces) it adds
2867
2868 Auto-Submitted: auto-generated
2869
287063. The "personal" condition in filters now includes a test for the
2871 Auto-Submitted: header. If it contains the string "auto-" the message it
2872 not considered personal.
2873
287464. Added rcpt_include_affixes as a generic transport option.
2875
287665. Added queue_only_override (default true).
2877
287866. Added the syslog_duplication option.
2879
288067. If what should have been the first header line of a message consisted of
2881 a space followed by a colon, Exim was mis-interpreting it as a header line.
2882 It isn't of course - it is syntactically invalid and should therefore be
2883 treated as the start of the message body. The misbehaviour could have
2884 caused a number of strange effects, including loss of data in subsequent
2885 header lines, and spool format errors.
2886
288768. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2888 client host had authenticated. This control can now be exercised by an ACL
2889 for more flexibility.
2890
289169. By default, callouts do not happen when testing with -bh. There is now a
2892 variant, -bhc, which does actually run the callout code, including
2893 consulting and updating the callout cache.
2894
289570. Added support for saslauthd authentication, courtesy of Alexander
2896 Sabourenkov.
2897
289871. If statvfs() failed on the spool or log directories while checking their
2899 size for availability, Exim confusingly gave the error "space shortage".
2900 Furthermore, in debugging mode it crashed with a floating point exception.
2901 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2902 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2903 serious problem, Exim now writes to the main and panic logs when this
2904 happens, with details of the failure. It then refuses to accept the
2905 incoming message, giving the message "spool directory problem" or "log
2906 directory problem" with a 421 code for SMTP messages.
2907
290872. When Exim is about to re-exec itself, it ensures that the file descriptors
2909 0, 1, and 2 exist, because some OS complain for execs without them (see
2910 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2911 descriptors. However, the code omitted to check that the open succeeded,
2912 causing mysterious errors if for some reason the permissions on /dev/null
2913 got screwed. Now Exim writes a message to the main and panic logs, and
2914 bombs out if it can't open /dev/null.
2915
291673. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2917 interact so that it is all more flexible. It is supposed to remain
2918 backwards compatible. Also added extra_local_interfaces.
2919
292074. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2921 to bomb out with an assertion failure - to the client this appears as a
2922 connection drop. This problem occurs in the part of the code that was taken
2923 from the Samba project. Fortunately, the assertion is in a very simple
2924 function, so I have fixed this by reproducing the function inline in the
2925 one place where it is called, and arranging for authentication to fail
2926 instead of killing the process with assert().
2927
292875. The SPA client code was not working when the server requested OEM rather
2929 than Unicode encoding.
2930
293176. Added code to make require_files with a specific uid setting more usable in
2932 the case where statting the file as root fails - usually a non-root-mounted
2933 NFS file system. When this happens and the failure is EACCES, Exim now
2934 forks a subprocess and does the per-uid checking as the relevant uid.
2935
293677. Added process_log_path.
2937
293878. If log_file_path was not explicitly set, a setting of check_log_space or
2939 check_log_inodes was ignored.
2940
294179. If a space check for the spool or log partitions fails, the incident is now
2942 logged. Of course, in the latter case the data may get lost...
2943
294480. Added the %p formatting code to string_format() so that it can be used to
2945 print addresses in debug_print(). Adjusted all the address printing in the
2946 debugging in store.c to use %p rather than %d.
2947
294881. There was a concern that a line of code in smtp_in.c could overflow a
2949 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2950 initially expressed, the concern was not well-founded, because trailing
2951 spaces are removed early. However, if the trailing spaces were followed by
2952 a NULL, they did not get removed, so the overflow was possible. Two fixes
2953 were applied:
2954
2955 (a) I re-wrote the offending code in a cleaner fashion.
2956 (b) If an incoming SMTP command contains a NULL character, it is rejected
2957 as invalid.
2958
295982. When Exim changes uid/gid to the Exim user at daemon start time, it now
2960 runs initgroups(), so that if the Exim user is in any additional groups,
2961 they will be used during message reception.
2962
2963
2964Exim version 4.20
2965-----------------
2966
2967The change log for 4.20 and earlier releases has been archived.
2968
2969****