(1) $address_data in an ACL was clobbered by a sender verify that
[exim.git] / doc / doc-txt / ChangeLog
CommitLineData
2a3eea10 1$Cambridge: exim/doc/doc-txt/ChangeLog,v 1.13 2004/10/19 13:40:39 ph10 Exp $
495ae4b0
PH
2
3Change log file for Exim from version 4.21
4-------------------------------------------
5
6
7Exim version 4.44
8-----------------
9
5fe762f6
PH
10 1. Minor wording change to the doc/README.SIEVE file.
11
139059f6 12 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
5fe762f6 13 computation of the current number of files was incorrect.
495ae4b0 14
7086e875
PH
15 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
16 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
17 place.
18
35af9f61
PH
19 4. Give more explanation in the error message when the command for a transport
20 filter fails to execute.
21
b668c215
PH
22 5. There are several places where Exim runs a non-Exim command in a
23 subprocess. The SIGUSR1 signal should be disabled for these processes. This
24 was being done only for the command run by the queryprogram router. It is
25 now done for all such subprocesses. The other cases are: ${run, transport
26 filters, and the commands run by the lmtp and pipe transports.
27
a494b1e1
PH
28 6. Added CONFIGURE_GROUP build-time option.
29
30 7. Some older OS have a limit of 256 on the maximum number of file
31 descriptors. Exim was using setrlimit() to set 1000 as a large value
32 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
33 systems. I've change it so that if it can't get 1000, it tries for 256.
35edf2ff 34
c5fcb476
PH
35 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
36 was an oversight, and furthermore, ever since the addition of extra
37 controls (e.g. 4.43/32), the checks on when to allow different forms of
38 "control" were broken. There should now be diagnostics for all cases when a
39 control that does not make sense is encountered.
40
69358f02
PH
41 9. Added the /retain_sender option to "control=submission".
42
5be20824
PH
4310. $recipients is now available in the predata ACL (oversight).
44
eb2c0248
PH
4511. Tidy the search cache before the fork to do a delivery from a message
46 received from the command line. Otherwise the child will trigger a lookup
47 failure and thereby defer the delivery if it tries to use (for example) a
48 cached ldap connection that the parent has called unbind on.
49
2a3eea10
PH
5012. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
51 of $address_data from the recipient verification was clobbered by the
52 sender verification.
53
5413. The value of address_data from a sender verification is now available in
55 $sender_address_data in subsequent conditions in the ACL statement.
56
495ae4b0
PH
57
58Exim version 4.43
59-----------------
60
61 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
62 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
63 Now it returns an integer. A number of calls were still expecting a T/F
64 return. Fortuitously, in all cases, the tests worked in OK situations,
65 which is the norm. However, things would have gone wrong on any write
66 failures on the smtp file descriptor. This function is used when sending
67 messages over SMTP and also when doing verify callouts.
68
69 2. When Exim is called to do synchronous delivery of a locally submitted
70 message (the -odf or -odi options), it no longer closes stderr before doing
71 the delivery.
72
73 3. Implemented the mua_wrapper option.
74
75 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
76
77 5. Implemented the functions header_remove(), header_testname(),
78 header_add_at_position(), and receive_remove_recipient(), and exported them
79 to local_scan().
80
81 6. If an ACL "warn" statement specified the addition of headers, Exim already
82 inserted X-ACL-Warn: at the start if there was no header name. However, it
83 was not making this test for the second and subsequent header lines if
84 there were newlines in the string. This meant that an invalid header could
85 be inserted if Exim was badly configured.
86
87 7. Allow an ACL "warn" statement to add header lines at the start or after all
88 the Received: headers, as well as at the end.
89
90 8. Added the rcpt_4xx retry error code.
91
92 9. Added postmaster_mailfrom=xxx to callout verification option.
93
9410. Added mailfrom=xxxx to the callout verification option, for verify=
95 header_sender only.
96
9711. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
98 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
99
10012. Inserted some casts to stop certain compilers warning when using pointer
101 differences as field lengths or precisions in printf-type calls (mostly
102 affecting debugging statements).
103
10413. Added optional readline() support for -be (dynamically loaded).
105
10614. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
107 same clock tick as a message's arrival, so that its received time was the
108 same as the "first fail" time on the retry record, and that message
109 remained on the queue past the ultimate address timeout, every queue runner
110 would try a delivery (because it was past the ultimate address timeout) but
111 after another failure, the ultimate address timeout, which should have then
112 bounced the address, did not kick in. This was a "< instead of <=" error;
113 in most cases the first failure would have been in the next clock tick
114 after the received time, and all would be well.
115
11615. The special items beginning with @ in domain lists (e.g. @mx_any) were not
117 being recognized when the domain list was tested by the match_domain
118 condition in an expansion string.
119
12016. Added the ${str2b64: operator.
121
12217. Exim was always calling setrlimit() to set a large limit for the number of
123 processes, without checking whether the existing limit was already
124 adequate. (It did check for the limit on file descriptors.) Furthermore,
125 errors from getrlimit() and setrlimit() were being ignored. Now they are
126 logged to the main and panic logs, but Exim does carry on, to try to do its
127 job under whatever limits there are.
128
12918. Imported PCRE 5.0.
130
13119. Trivial typo in log message " temporarily refused connection" (the leading
132 space).
133
13420. If the log selector return_path_on_delivery was set and an address was
135 redirected to /dev/null, the delivery process crashed because it assumed
136 that a return path would always be set for a "successful" delivery. In this
137 case, the whole delivery is bypassed as an optimization, and therefore no
138 return path is set.
139
14021. Internal re-arrangement: the function for sending a challenge and reading
141 a response while authentication was assuming a zero-terminated challenge
142 string. It's now changed to take a pointer and a length, to allow for
143 binary data in such strings.
144
14522. Added the cyrus_sasl authenticator (code supplied by MBM).
146
14723. Exim was not respecting finduser_retries when seeking the login of the
148 uid under which it was called; it was always trying 10 times. (The default
149 setting of finduser_retries is zero.) Also, it was sleeping after the final
150 failure, which is pointless.
151
15224. Implemented tls_on_connect_ports.
153
15425. Implemented acl_smtp_predata.
155
15626. If the domain in control=submission is set empty, Exim assumes that the
157 authenticated id is a complete email address when it generates From: or
158 Sender: header lines.
159
16027. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
161 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
162 chown and chgrp in /bin and hostname in /usr/bin.
163
16428. Exim was keeping the "process log" file open after each use, just as it
165 does for the main log. This opens the possibility of it remaining open for
166 long periods when the USR1 signal hits a daemon. Occasional processlog
167 errors were reported, that could have been caused by this. Anyway, it seems
168 much more sensible not to leave this file open at all, so that is what now
169 happens.
170
17129. The long-running daemon process does not normally write to the log once it
172 has entered its main loop, and it closes the log before doing so. This is
173 so that log files can straightforwardly be renamed and moved. However,
174 there are a couple of unusual error situations where the daemon does write
175 log entries, and I had neglected to close the log afterwards.
176
17730. The text of an SMTP error response that was received during a remote
178 delivery was being truncated at 512 bytes. This is too short for some of
179 the long messages that one sometimes sees. I've increased the limit to
180 1024.
181
18231. It is now possible to make retry rules that apply only when a message has a
183 specific sender, in particular, an empty sender.
184
18532. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
186 it possible to be selective about when SMTP synchronization is enforced.
187
18833. Added "control = caseful_local_part" and "control = "caselower_local_part".
189
19032. Implemented hosts_connection_nolog.
191
19233. Added an ACL for QUIT.
193
19434. Setting "delay_warning=" to disable warnings was not working; it gave a
195 syntax error.
196
19735. Added mailbox_size and mailbox_filecount to appendfile.
198
19936. Added control = no_multiline_responses to ACLs.
200
20137. There was a bug in the logic of the code that waits for the clock to tick
202 in the case where the clock went backwards by a substantial amount such
203 that the microsecond fraction of "now" was more than the microsecond
204 fraction of "then" (but the whole seconds number was less).
205
20638. Added support for the libradius Radius client library this is found on
207 FreeBSD (previously only the radiusclient library was supported).
208
209
210Exim version 4.42
211-----------------
212
213 1. When certain lookups returned multiple values in the form name=value, the
214 quoting of the values was not always being done properly. Specifically:
215 (a) If the value started with a double quote, but contained no whitespace,
216 it was not quoted.
217 (b) If the value contained whitespace other than a space character (i.e.
218 tabs or newlines or carriage returns) it was not quoted.
219 This fix has been applied to the mysql and pgsql lookups by writing a
220 separate quoting function and calling it from the lookup code. The fix
221 should probably also be applied to nisplus, ibase and oracle lookups, but
222 since I cannot test any of those, I have not disturbed their existing code.
223
224 2. A hit in the callout cache for a specific address caused a log line with no
225 reason for rejecting RCPT. Now it says "Previous (cached) callout
226 verification failure".
227
228 3. There was an off-by-one bug in the queryprogram router. An over-long
229 return line was truncated at 256 instead of 255 characters, thereby
230 overflowing its buffer with the terminating zero. As well as fixing this, I
231 have increased the buffer size to 1024 (and made a note to document this).
232
233 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
234 when Exim is waiting for an SMTP response from a remote server, Exim
235 restarts its select() call on the socket, thereby resetting its timeout.
236 This is not a problem when such interrupts are rare. Somebody set up a cron
237 job to run exiwhat every 2 minutes, which is less than the normal select()
238 timeout (5 or 10 minutes). This meant that the select() timeout never
239 kicked in because it was always reset. I have fixed this by comparing the
240 time when an interrupt arrives with the time at the start of the first call
241 to select(). If more time than the timeout has elapsed, the interrupt is
242 treated as a timeout.
243
244 5. Some internal re-factoring in preparation for the addition of Sieve
245 extensions (by MH). In particular, the "personal" test is moved to a
246 separate function, and given an option for scanning Cc: and Bcc: (which is
247 not set for Exim filters).
248
249 6. When Exim created an email address using the login of the caller as the
250 local part (e.g. when creating a From: or Sender: header line), it was not
251 quoting the local part when it contained special characters such as @.
252
253 7. Installed new OpenBSD configuration files.
254
255 8. Reworded some messages for syntax errors in "and" and "or" conditions to
256 try to make them clearer.
257
258 9. Callout options, other than the timeout value, were being ignored when
259 verifying sender addresses in header lines. For example, when using
260
261 verify = header_sender/callout=no_cache
262
263 the cache was (incorrectly) being used.
264
26510. Added a missing instance of ${EXE} to the exim_install script; this affects
266 only the Cygwin environment.
267
26811. When return_path_on_delivery was set as a log selector, if different remote
269 addresses in the same message used different return paths and parallel
270 remote delivery occurred, the wrong values would sometimes be logged.
271 (Whenever a remote delivery process finished, the return path value from
272 the most recently started remote delivery process was logged.)
273
27412. RFC 3848 specifies standard names for the "with" phrase in Received: header
275 lines when AUTH and/or TLS are in use. This is the "received protocol"
276 field. Exim used to use "asmtp" for authenticated SMTP, without any
277 indication (in the protocol name) for TLS use. Now it follows the RFC and
278 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
279 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
280 names appear in log lines as well as in Received: header lines.
281
28213. Installed MH's patches for Sieve to add the "copy" and "vacation"
283 extensions, and comparison tests, and to fix some bugs.
284
28514. Changes to the "personal" filter test:
286
287 (1) The test was buggy in that it was just doing the equivalent of
288 "contains" tests on header lines. For example, if a user's address was
289 anne@some.where, the "personal" test would incorrectly be true for
290
291 To: susanne@some.where
292
293 This test is now done by extracting each address from the header in turn,
294 and checking the entire address. Other tests that are part of "personal"
295 are now done using regular expressions (for example, to check local parts
296 of addresses in From: header lines).
297
298 (2) The list of non-personal local parts in From: addresses has been
299 extended to include "listserv", "majordomo", "*-request", and "owner-*",
300 taken from the Sieve specification recommendations.
301
302 (3) If the message contains any header line starting with "List-" it is
303 treated as non-personal.
304
305 (4) The test for "circular" in the Subject: header line has been removed
306 because it now seems ill-conceived.
307
30815. Minor typos in src/EDITME comments corrected.
309
31016. Installed latest exipick from John Jetmore.
311
31217. If headers_add on a router specified a text string that was too long for
313 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
314 of string_sprintf() is now avoided.
315
31618. $message_body_size was not set (it was always zero) when running the DATA
317 ACL and the local_scan() function.
318
31919. For the "mail" command in an Exim filter, no default was being set for
320 the once_repeat time, causing a random time value to be used if "once" was
321 specified. (If the value happened to be <= 0, no repeat happened.) The
322 default is now 0s, meaning "never repeat". The "vacation" command was OK
323 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
324 (I found it when inspecting the code).
325
32620. There is now an overall timeout for performing a callout verification. It
327 defaults to 4 times the callout timeout, which applies to individual SMTP
328 commands during the callout. The overall timeout applies when there is more
329 than one host that can be tried. The timeout is checked before trying the
330 next host. This prevents very long delays if there are a large number of
331 hosts and all are timing out (e.g. when the network connections are timing
332 out). The value of the overall timeout can be changed by specifying an
333 additional sub-option for "callout", called "maxwait". For example:
334
335 verify = sender/callout=5s,maxwait=20s
336
33721. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
338 to the end before writing, but this should make it even safer).
339
34022. Exim was forgetting that it had advertised PIPELINING for the second and
341 subsequent messages on an SMTP connection. It was also not resetting its
342 memory on STARTTLS and an internal HELO.
343
34423. When Exim logs an SMTP synchronization error within a session, it now
345 records whether PIPELINING has been advertised or not.
346
34724. Added 3 instances of "(long int)" casts to time_t variables that were being
348 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
349 rather than long int.
350
35125. Installed the latest Cygwin configuration files from the Cygwin maintainer.
352
35326. Added the never_mail option to autoreply.
354
355
356Exim version 4.41
357-----------------
358
359 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
360 crash if the getsockname() call failed; this can happen if a connection is
361 closed very soon after it is established. The problem was simply in the
362 order in which certain operations were done, causing Exim to try to write
363 to the SMTP stream before it had set up the file descriptor. The bug has
364 been fixed by making things happen in the correct order.
365
366
367Exim version 4.40
368-----------------
369
370 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
371 before the connection was closed, thus losing the rejection response.
372
373 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
374 some early Solaris releases, but causes trouble in current releases where
375 socklen_t is defined.
376
377 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
378 always exist.
379
380 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
381 configured.
382
383 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
384 that releasing the top of it at the end releases what was used for sub-
385 expansions (unless the block got too big). However, discard this block if
386 the first thing is a variable or header, so that we can use its block when
387 it is dynamic (useful for very large $message_headers, for example).
388
389 6. Lookups now cache *every* query, not just the most recent. A new, separate
390 store pool is used for this. It can be recovered when all lookup caches are
391 flushed. Lookups now release memory at the end of their result strings.
392 This has involved some general refactoring of the lookup sources.
393
394 7. Some code has been added to the store_xxx() functions to reduce the amount
395 of flapping under certain conditions.
396
397 8. log_incoming_interface used to affect only the <= reception log lines. Now
398 it causes the local interface and port to be added to several more SMTP log
399 lines, for example "SMTP connection from", and rejection lines.
400
401 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
402
40310. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
404
40511. If $host_data was set by virtue of a hosts lookup in an ACL, its value
406 could be overwritten at the end of the current message (or the start of a
407 new message if it was set in a HELO ACL). The value is now preserved for
408 the duration of the SMTP connection.
409
41012. If a transport had a headers_rewrite setting, and a matching header line
411 contained an unqualified address, that address was qualified, even if it
412 did not match any rewriting rules. The underlying bug was that the values
413 of the flags that permit the existence of unqualified sender and recipient
414 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
415 non-local messages, and by -bnq for local messages) were not being
416 preserved with the message after it was received.
417
41813. When Exim was logging an SMTP synchronization error, it could sometimes log
419 "next input=" as part of the text comprising the host identity instead of
420 the correct text. The code was using the same buffer for two different
421 strings. However, depending on which order the printing function evaluated
422 its arguments, the bug did not always show up. Under Linux, for example, my
423 test suite worked just fine.
424
42514. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
426 doesn't work with some older versions of Perl. It has been changed to "my",
427 which in any case is probably the better facility to use.
428
42915. A really picky compiler found some instances of statements for creating
430 error messages that either had too many or two few arguments for the format
431 string.
432
43316. The size of the buffer for calls to the DNS resolver has been increased
434 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
435 for addresses that have a lot of PTR records. This alleviates a problem; it
436 does not fully solve it.
437
43817. A dnsdb lookup for PTR records that receives more data than will fit in the
439 buffer now truncates the list and logs the incident, which is the same
440 action as happens when Exim is looking up a host name and its aliases.
441 Previously in this situation something unpredictable would happen;
442 sometimes it was "internal error: store_reset failed".
443
44418. If a server dropped the connection unexpectedly when an Exim client was
445 using GnuTLS and trying to read a response, the client delivery process
446 crashed while trying to generate an error log message.
447
44819. If a "warn" verb in an ACL added multiple headers to a message in a single
449 string, for example:
450
451 warn message = H1: something\nH2: something
452
453 the text was added as a single header line from Exim's point of view
454 though it ended up OK in the delivered message. However, searching for the
455 second and subsequent header lines using $h_h2: did not work. This has been
456 fixed. Similarly, if a system filter added multiple headers in this way,
457 the routers could not see them.
458
45920. Expanded the error message when iplsearch is called with an invalid key to
460 suggest using net-iplsearch in a host list.
461
46221. When running tests using -bh, any delays imposed by "delay" modifiers in
463 ACLs are no longer actually imposed (and a message to that effect is
464 output).
465
46622. If a "gecos" field in a passwd entry contained escaped characters, in
467 particular, if it contained a \" sequence, Exim got it wrong when building
468 a From: or a Sender: header from that name. A second bug also caused
469 incorrect handling when an unquoted " was present following a character
470 that needed quoting.
471
47223. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
473 was not being matched caselessly.
474
47524. Arranged for all hyphens in the exim.8 source to be escaped with
476 backslashes.
477
47825. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
479 properly. Recipient callout cache records were still being keyed to include
480 the sender, even when use_sender was set false. This led to far more
481 callouts that were necessary. The sender is no longer included in the key
482 when use_sender is false.
483
48426. Added "control = submission" modifier to ACLs.
485
48627. Added the ${base62d: operator to decode base 62 numbers.
487
48828. dnsdb lookups can now access SRV records.
489
49029. CONFIGURE_OWNER can be set at build time to define an alternative owner for
491 the configuration file.
492
49330. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
494 (-v) mode. This makes the output for a verbose queue run more intelligible.
495
49631. Added a use_postmaster feature to recipient callouts.
497
49832. Added the $body_zerocount variable, containing the number of binary zero
499 bytes in the message body.
500
50133. The time of last modification of the "new" subdirectory is now used as the
502 "mailbox time last read" when there is a quota error for a maildir
503 delivery.
504
50534. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
506
50735. Added +ignore_unknown as a special item in host lists.
508
50936. Code for decoding IPv6 addresses in host lists is now included, even if
510 IPv6 support is not being compiled. This fixes a bug in which an IPv6
511 address was recognized as an IP address, but was then not correctly decoded
512 into binary, causing unexpected and incorrect effects when compared with
513 another IP address.
514
515
516Exim version 4.34
517-----------------
518
519 1. Very minor rewording of debugging text in manualroute to say "list of
520 hosts" instead of "hostlist".
521
522 2. If verify=header_syntax was set, and a header line with an unqualified
523 address (no domain) and a large number of spaces between the end of the
524 name and the colon was received, the reception process suffered a buffer
525 overflow, and (when I tested it) crashed. This was caused by some obsolete
526 code that should have been removed. The fix is to remove it!
527
528 3. When running in the test harness, delay a bit after writing a bounce
529 message to get a bit more predictability in the log output.
530
531 4. Added a call to search_tidyup() just before forking a reception process. In
532 theory, someone could use a lookup in the expansion of smtp_accept_max_
533 per_host which, without the tidyup, could leave open a database connection.
534
535 5. Added the variables $recipient_data and $sender_data which get set from a
536 lookup success in an ACL "recipients" or "senders" condition, or a router
537 "senders" option, similar to $domain_data and $local_part_data.
538
539 6. Moved the writing of debug_print from before to after the "senders" test
540 for routers.
541
542 7. Change 4.31/66 (moving the time when the Received: is generated) caused
543 problems for message scanning, either using a data ACL, or using
544 local_scan() because the Received: header was not generated till after they
545 were called (in order to set the time as the time of reception completion).
546 I have revised the way this works. The header is now generated after the
547 body is received, but before the ACL or local_scan() are called. After they
548 are run, the timestamp in the header is updated.
549
550
551Exim version 4.33
552-----------------
553
554 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
555 before starting a queue runner without re-exec. This happened only when
556 deliver_drop_privilege was set or when the Exim user was set to root. The
557 effect of the bug was that timeouts during subsequent deliveries caused
558 crashes instead of being properly handled. The handler is now left at its
559 default (and expected) setting.
560
561 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
562 message, again when deliver_drop_privilege is set or Exim is run as root.
563 The bug described in (1) was not present in this case, but the tidying up
564 of the other signals was missing. I have made the two cases consistent.
565
566 3. The ignore_target_hosts setting on a manualroute router was being ignored
567 for hosts that were looked up using the /MX notation.
568
569 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
570 in domain lists.
571
572 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
573 operated on the sender address. After changing the $sender_address to <>
574 for the sender address verify, Exim was re-instated it as the original
575 (before rewriting) address, but remembering that it had rewritten it, so it
576 wasn't rewriting it again. This bug also had the effect of breaking the
577 sender address verification caching when the sender address was rewritten.
578
579 6. The ignore_target_hosts option was being ignored by the ipliteral router.
580 This has been changed so that if the ip literal address matches
581 ignore_target_hosts, the router declines.
582
583 7. Added expansion conditions match_domain, match_address, and match_local_
584 part (NOT match_host).
585
586 8. The placeholder for the Received: header didn't have a length field set.
587
588 9. Added code to Exim itself and to exim_lock to test for a specific race
589 condition that could lead to file corruption when using MBX delivery. The
590 issue is with the lockfile that is created in /tmp. If this file is removed
591 after a process has opened it but before that process has acquired a lock,
592 there is the potential for a second process to recreate the file and also
593 acquire a lock. This could lead to two Exim processes writing to the file
594 at the same time. The added code performs the same test as UW imapd; it
595 checks after acquiring the lock that its file descriptor still refers to
596 the same named file.
597
59810. The buffer for building added header lines was of fixed size, 8192 bytes.
599 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
600 when Exim is built.
601
60211. Added the smtp_active_hostname option. If used, this will typically be made
603 to depend on the incoming interface address. Because $interface_address is
604 not set up until the daemon has forked a reception process, error responses
605 that can happen earlier (such as "too many connections") no longer contain
606 a host name.
607
60812. If an expansion in a condition on a "warn" statement fails because a lookup
609 defers, the "warn" statement is abandoned, and the next ACL statement is
610 processed. Previously this caused the whole ACL to be aborted.
611
61213. Added the iplsearch lookup type.
613
61414. Added ident_timeout as a log selector.
615
61615. Added tls_certificate_verified as a log selector.
617
61816. Added a global option tls_require_ciphers (compare the smtp transport
619 option of the same name). This controls incoming TLS connections.
620
62117. I finally figured out how to make tls_require_ciphers do a similar thing
622 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
623 before starting the TLS session.
624
62518. Tabs are now shown as \t in -bP output.
626
62719. If the log selector return_path_on_delivery was set, Exim crashed when
628 bouncing a message because it had too many Received: header lines.
629
63020. If two routers both had headers_remove settings, and the first one included
631 a superfluous trailing colon, the final name in the first list and the
632 first name in the second list were incorrectly joined into one item (with a
633 colon in the middle).
634
635
636Exim version 4.32
637-----------------
638
639 1. Added -C and -D options to the exinext utility, mainly to make it easier
640 to include in the automated testing, but these could be helpful when
641 multiple configurations are in use.
642
643 2. The exinext utility was not formatting the output nicely when there was
644 an alternate port involved in the retry record key, nor when there was a
645 message id as well (for retries that were specific to a specific message
646 and a specific host). It was also confused by IPv6 addresses, because of
647 the additional colons they contain. I have fixed the IPv4 problem, and
648 patched it up to do a reasonable job for IPv6.
649
650 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
651 delivery, the log line now contains "pipelined" if PIPELINING was used.
652
653 4. An SMTP transport process used to panic and die if the bind() call to set
654 an explicit outgoing interface failed. This has been changed; it is now
655 treated in the same way as a connect() failure.
656
657 5. A reference to $sender_host_name in the part of a conditional expansion
658 that was being skipped was still causing a DNS lookup. This no longer
659 occurs.
660
661 6. The def: expansion condition was not recognizing references to header lines
662 that used bh_ and bheader_.
663
664 7. Added the _cache feature to named lists.
665
666 8. The code for checking quota_filecount in the appendfile transport was
667 allowing one more file than it should have been.
668
669 9. For compatibility with Sendmail, the command line option
670
671 -prval:sval
672
673 is equivalent to
674
675 -oMr rval -oMs sval
676
677 and sets the incoming protocol and host name (for trusted callers). The
678 host name and its colon can be omitted when only the protocol is to be set.
679 Note the Exim already has two private options, -pd and -ps, that refer to
680 embedded Perl. It is therefore impossible to set a protocol value of "d" or
681 "s", but I don't think that's a major issue.
682
68310. A number of refactoring changes to the code, none of which should affect
684 Exim's behaviour:
685
686 (a) The number of logging options was getting close to filling up the
687 32-bit word that was used as a bit map. I have split them into two classes:
688 those that are passed in the argument to log_write(), and those that are
689 only ever tested independently outside of that function. These are now in
690 separate 32-bit words, so there is plenty of room for expansion again.
691 There is no change in the user interface or the logging behaviour.
692
693 (b) When building, for example, log lines, the code previously used a
694 macro that called string_cat() twice, in order to add two strings. This is
695 not really sufficiently general. Furthermore, there was one instance where
696 it was actually wrong because one of the argument was used twice, and in
697 one call a function was used. (As it happened, calling the function twice
698 did not affect the overall behaviour.) The macro has been replaced by a
699 function that can join an arbitrary number of extra strings onto a growing
700 string.
701
702 (c) The code for expansion conditions now uses a table and a binary chop
703 instead of a serial search (which was left over from when there were very
704 few conditions). Also, it now recognizes conditions like "pam" even when
705 the relevant support is not compiled in: a suitably worded error message is
706 given if an attempt is made to use such a condition.
707
70811. Added ${time_interval:xxxxx}.
709
71012. A bug was causing one of the ddress fields not to be passed back correctly
711 from remote delivery subprocesses. The field in question was not being
712 subsequently used, so this caused to problems in practice.
713
71413. Added new log selectors queue_time and deliver_time.
715
71614. Might have fixed a bug in maildirsizefile handling that threw up
717 "unexpected character" debug warnings, and recalculated the data
718 unnecessarily. In any case, I expanded the warning message to give more
719 information.
720
72115. Added the message "Restricted characters in address" to the statements in
722 the default ACL that block characters like @ and % in local parts.
723
72416. Change 71 for release 4.31 proved to be much less benign that I imagined.
725 Three changes have been made:
726
727 (a) There was a serious bug; a negative response to MAIL caused the whole
728 recipient domain to be cached as invalid, thereby blocking all messages
729 to all local parts at the same domain, from all senders. This bug has
730 been fixed. The domain is no longer cached after a negative response to
731 MAIL if the sender used is not empty.
732
733 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
734 been restored.
735
736 (c) A new callout option, "use_sender" has been added for people who want
737 the modified behaviour.
738
739
740Exim version 4.31
741-----------------
742
743 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
744 Larry Rosenman.
745
746 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
747 indeed breaks things for older releases.
748
749 3. Added additional logging to the case where there is a problem reading data
750 from a filter that is running in a subprocess using a pipe, in order to
751 try to track down a specific problem.
752
753 4. Testing facility fudge: when running in the test harness and attempting
754 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
755 getting "No route to host". Convert this to a timeout.
756
757 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
758 warning.
759
760 6. Some OS don't have socklen_t but use size_t instead. This affects the
761 fifth argument of getsockopt() amongst other things. This is now
762 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
763 can be set for individual OS. I have set it for SunOS5, OSF1, and
764 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
765 some earlier ones do not.
766
767 7. Change 4.30/15 was not doing the test caselessly.
768
769 8. The standard form for an IPv6 address literal was being rejected by address
770 parsing in, for example, MAIL and RCPT commands. An example of this kind of
771 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
772 this, as well as the form without the "IPv6" on the front (but only when
773 address literals are enabled, of course).
774
775 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
776
77710. Exim crashed if a message with an empty sender address specified by -f
778 encountered a router with an errors_to setting. This could be provoked only
779 by a command such as
780
781 exim -f "" ...
782
783 where an empty string was supplied; "<>" did not hit this bug.
784
78511. Installed PCRE release 4.5.
786
78712. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
788 remained set. It is now erased.
789
79013. exiqgrep wasn't working on MacOS X because it didn't correctly compute
791 times from message ids (which are base 36 rather than the normal 62).
792
79314. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
794 were being counted as actual protocol errors, and logged if the log
795 selector +smtp_protocol_error was set. One cannot be perfect in this test,
796 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
797 and DATA following a set of rejected RCPTs do not count as protocol errors.
798 In other words, Exim assumes they were pipelined, though this may not
799 actually be the case. Of course, in all cases the client gets an
800 appropriate error code.
801
80215. If a lookup fails in an ACL condition, a message about the failure may
803 be available; it is used if testing the ACL cannot continue, because most
804 such messages specify what the cause of the deferral is. However, some
805 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
806 that caused an old message to be retained and used if a later statement
807 caused a defer, replacing the real cause of the deferral.
808
80916. If an IP address had so many PTR records that the DNS lookup buffer
810 was not large enough to hold them, Exim could crash while trying to process
811 the truncated data. It now detects and logs this case.
812
81317. Further to 4.21/58, another change has been made: if (and only if) the
814 first line of a message (the first header line) ends with CRLF, a bare LF
815 in a subsequent header line has a space inserted after it, so as not to
816 terminate the header.
817
81818. Refactoring: tidied an ugly bit of code in appendfile that copied data
819 unnecessarily, used atoi() instead of strtol(), and didn't check the
820 termination when getting file sizes from file names by regex.
821
82219. Completely re-implemented the support for maildirsize files, in the light
823 of a number of problems with the previous contributed implementation
824 (4.30/29). In particular:
825
826 . If the quota is zero, the maildirsize file is maintained, but no quota is
827 imposed.
828
829 . If the maildir directory does not exist, it is created before any attempt
830 to write a maildirsize file.
831
832 . The quota value in the file is just a cache; if the quota is changed in
833 the transport, the new value overrides.
834
835 . A regular expression is available for excluding directories from the
836 count.
837
83820. The autoreply transport checks the characters in options that define the
839 message's headers; it allows continued headers, but it was checking with
840 isspace() after an embedded newline instead of explicitly looking for a
841 space or a tab.
842
84321. If all the "regular" hosts to which an address was routed had passed their
844 expiry times, and had not reached their retry times, the address was
845 bounced, even if fallback hosts were defined. Now Exim should go on to try
846 the fallback hosts.
847
84822. Increased buffer sizes in the callout code from 1024 to 4096 to match the
849 equivalent code in the SMTP transport. Some hosts send humungous responses
850 to HELO/EHLO, more than 1024 it seems.
851
85223. Refactoring: code in filter.c used (void *) for "any old type" but this
853 gives compiler warnings in some environments. I've now done it "properly",
854 using a union.
855
85624. The replacement for inet_ntoa() that is used with gcc on IRIX systems
857 (because of problems with the built-in one) was declared to return uschar *
858 instead of char *, causing compiler failure.
859
86025. Fixed a file descriptor leak when processing alias/forward files.
861
86226. Fixed a minor format string issue in dbfn.c.
863
86427. Typo in exim.c: ("dmbnz" for "dbmnz").
865
86628. If a filter file refered to $h_xxx or $message_headers, and the headers
867 contained RFC 2047 "words", Exim's memory could, under certain conditions,
868 become corrupted.
869
87029. When a sender address is verified, it is cached, to save repeating the test
871 when there is more than one recipient in a message. However, when the
872 verification involves a callout, it is possible for different callout
873 options to be set for different recipients. It is too complicated to keep
874 track of this in the cache, so now Exim always runs a verification when a
875 callout is required, relying on the callout cache for the optimization.
876 The overhead is duplication of the address routing, but this should not be
877 too great.
878
87930. Fixed a bug in callout caching. If a RCPT command caused the sender address
880 to be verified with callout=postmaster, and the main callout worked but the
881 postmaster check failed, the verification correctly failed. However, if a
882 subsequent RCPT command asked for sender verification *without* the
883 postmaster check, incorrect caching caused this verification also to fail,
884 incorrectly.
885
88631. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
887 it was not caching the DNS options (qualify_single, search_parents) that
888 were used when the lookup failed. A subsequent lookup with different
889 options therefore always gave the same answer, though there were cases
890 where it should not have. (Example: a "domains = !$mx_any" option on a
891 dnslookup router: the "domains" option is always processed without any
892 widening, but the router might have qualify_single set.) Now Exim uses the
893 cached value only when the same options are set.
894
89532. Added John Jetmore's "exipick" utility to the distribution.
896
89733. GnuTLS: When an attempt to start a TLS session fails for any reason other
898 than a timeout (e.g. a certificate is required, and is not provided), an
899 Exim server now closes the connection immediately. Previously it waited for
900 the client to close - but if the client is SSL, it seems that they each
901 wait for each other, leading to a delay before one of them times out.
902
90334: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
904 maintained 0.8.x compatibility because I don't think many are using it, and
905 it is clearly obsolete.
906
90735. Added TLS support for CRLs: a tls_crl global option and one for the smtp
908 transport.
909
91036. OpenSSL: $tls_certificate_verified was being set to 1 even if the
911 client certificate was expired. A simple patch fixes this, though I don't
912 understand the full logic of why the verify callback is called multiple
913 times.
914
91537. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
916 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
917 which causes problems with some clients (such as the Certicom SSL Plus
918 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
919 disables the coutermeasure allowing Eudora to connect."
920
92138. Exim was not checking that a write() to a log file succeeded. This could
922 lead to Bad Things if a log got too big, in particular if it hit a file
923 size limit. Exim now panics and dies if it cannot write to a log file, just
924 as it does if it cannot open a log file.
925
92639. Modified OS/Makefile-Linux so that it now contains
927
928 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
929
930 The two -D definitions ensure that Exim is compiled with large file
931 support, which makes it possible to handle log files that are bigger than
932 2^31.
933
93440. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
935 instance) a domain was checked against a named list that involved a lookup,
936 causing $domain_data to be set, then another domain was checked against the
937 same list, then the first domain was re-checked, the value of $domain_data
938 after the final check could be wrong. In particular, if the second check
939 failed, it could be set empty. This bug probably also applied to
940 $localpart_data.
941
94241. The strip_trailing_dot option was not being applied to the address given
943 with the -f command-line option.
944
94542. The code for reading a message's header from the spool was incrementing
946 $received_count, but never initializing it. This meant that the value was
947 incorrect (doubled) while delivering a message in the same process in which
948 it was received. In the most common configuration of Exim, this never
949 happens - a fresh exec is done - but it can happen when
950 deliver_drop_privilege is set.
951
95243. When Exim logs an SMTP synchronization error - client data sent too soon -
953 it now includes up to 150 characters of the unexpected data in the log
954 line.
955
95644. The exim_dbmbuild utility uses fixed size buffers for reading input lines
957 and building data strings. The size of both of these buffers was 10 000
958 bytes - far larger than anybody would *ever* want, thought I. Needless to
959 say, somebody hit the limit. I have increased the maximum line length to
960 20 000 and the maximum data length of concatenated lines to 100 000. I have
961 also fixed two bugs, because there was no checking on these buffers. Tsk,
962 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
963 buffer is too small.
964
96545. The exim_dbmbuild utility did not support quoted keys, as Exim does in
966 lsearch lookups. Now it does.
967
96846. When parsing a route_list item in a manualroute router, a fixed-length
969 buffer was used for the list of hosts. I made this 1024 bytes long,
970 thinking that nobody would ever have a list of hosts that long. Wrong.
971 Somebody had a whole pile of complicated expansion conditions, and the
972 string was silently truncated, leading to an expansion error. It turns out
973 that it is easier to change to an unlimited length (owing to other changes
974 that have happened since this code was originally written) than to build
975 structure for giving a limitation error. The length of the item that
976 expands into the list of hosts is now unlimited.
977
97847. The lsearch lookup could not handle data where the length of text line was
979 more than 4095 characters. Such lines were truncated, leading to shortened
980 data being returned. It should now handle lines of any length.
981
98248. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
983 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
984 ACL").
985
98649. Cosmetic tidy to scripts like exicyclog that are generated by globally
987 replacing strings such as BIN_DIRECTORY in a source file: the replacement
988 no longer happens in comment lines. A list of replacements is now placed
989 at the head of all of the source files, except those whose only change is
990 to replace PERL_COMMAND in the very first #! line.
991
99250. Replaced the slow insertion sort in queue.c, for sorting the list of
993 messages on the queue, with a bottom-up merge sort, using code contributed
994 by Michael Haardt. This should make operations like -bp somewhat faster on
995 large queues. It won't affect queue runners, except when queue_run_in_order
996 is set.
997
99851. Installed eximstats 1.31 in the distribution.
999
100052. Added support for SRV lookups to the dnslookup router.
1001
100253. If an ACL referred to $message_body or $message_body_end, the value was not
1003 reset for any messages that followed in the same SMTP session.
1004
100554. The store-handling optimization for building very long strings was not
1006 differentiating between the different store pools. I don't think this
1007 actually made any difference in practice, but I've tidied it.
1008
100955. While running the routers to verify a sender address, $sender_address
1010 was still set to the sender address. This is wrong, because when routing to
1011 send a bounce to the sender, it would be empty. Therefore, I have changed
1012 it so that, while verifying a sender address, $sender_address is set to <>.
1013 (There is no change to what happens when verifying a recipient address.)
1014
101556. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
1016 target A or AAAA records (if not already returned) without resetting the
1017 qualify_single or search_parents options of the DNS resolver. These are
1018 inappropriate in this case because the targets of MX and SRV records must
1019 be FQDNs. A broken DNS record could cause trouble if it happened to have a
1020 target that, when qualified, matched something in the local domain. These
1021 two options are now turned off when doing these lookups.
1022
102357. It seems that at least some releases of Reiserfs (which does not have the
1024 concept of a fixed number of inodes) returns zero and not -1 for the
1025 number of available inodes. This interacted badly with check_spool_inodes,
1026 which assumed that -1 was the "no such thing" setting. What I have done is
1027 to check that the total number of inodes is greater than zero before doing
1028 the test of how many are available.
1029
103058. When a "warn" ACL statement has a log_message modifier, the message is
1031 remembered, and not repeated. This is to avoid a lot of repetition when a
1032 message has many recipients that cause the same warning to be written.
1033 Howewer, Exim was preserving the list of already written lines for an
1034 entire SMTP session, which doesn't seem right. The memory is now reset if a
1035 new message is started.
1036
103759. The "rewrite" debugging flag was not showing the result of rewriting in the
1038 debugging output unless log_rewrite was also set.
1039
104060. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
1041 of (int)(handle) when we know that handle contains (void *)(-1).
1042
104361. The Exim daemon panic-logs an error return when it closes the incoming
1044 connection. However "connection reset by peer" seems to be common, and
1045 isn't really an error worthy of noting specially, so that particular error
1046 is no long logged.
1047
104862. When Exim is trying to find all the local interfaces, it used to panic and
1049 die if the ioctl to get the interface flags failed. However, it seems that
1050 on at least one OS (Solaris 9) it is possible to have an interface that is
1051 included in the list of interfaces, but for which you get a failure error
1052 for this call. This happens when the interface is not "plumbed" into a
1053 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
1054 failure of the "get flags" call assumes that the interface is down.
1055
105663. Added a ${eval10: operator, which assumes all numbers are decimal. This
1057 makes life easier for people who are doing arithmetic on fields extracted
1058 from dates, where you often get leading zeros that should not be
1059 interpreted as octal.
1060
106164. Added qualify_domain to the redirect router, to override the global
1062 setting.
1063
106465. If a pathologically long header line contained very many addresses (the
1065 report of this problem mentioned 10 000) and each of them was rewritten,
1066 Exim could use up a very large amount of memory. (It kept on making new
1067 copies of the header line as it rewrote, and never released the old ones.)
1068 At the expense of a bit more processing, the header rewriting function has
1069 been changed so that it no longer eats memory in this way.
1070
107166. The generation of the Received: header has been moved from the time that a
1072 message starts to be received, to the time that it finishes. The timestamp
1073 in the Received: header should now be very close to that of the <= log
1074 line. There are two side-effects of this change:
1075
1076 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
1077 logged header lines no longer include the local Received: line, because
1078 it has not yet been created. The same applies to a copy of the message
1079 that is returned to a non-SMTP sender when a message is rejected.
1080
1081 (b) When a filter file is tested using -bf, no additional Received: header
1082 is added to the test message. After some thought, I decided that this
1083 is a bug fix.
1084
1085 This change does not affect the value of $received_for. It is still set
1086 after address rewriting, but before local_scan() is called.
1087
108867. Installed the latest Cygwin-specific files from the Cygwin maintainer.
1089
109068. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
1091 gave an unhelpful panic error message, and a defer error. I have managed to
1092 change this behaviour so that it now rejects any supplied certificate,
1093 which seems right, as the list of acceptable certificates is empty.
1094
109569. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
1096 gave an unhelpful defer error. I have not managed to make this reject any
1097 supplied certificates, but the error message it gives is "no certificate
1098 supplied", which is not helpful.
1099
110070. exigrep's output now also includes lines that are not associated with any
1101 message, but which match the given pattern. Implemented by a patch from
1102 Martin Sluka, which also tidied up the Perl a bit.
1103
110471. Recipient callout verification, like sender verification, was using <> in
1105 the MAIL FROM command. This isn't really the right thing, since the actual
1106 sender may affect whether the remote host accepts the recipient or not. I
1107 have changed it to use the actual sender in the callout; this means that
1108 the cache record is now keyed on a recipient/sender pair, not just the
1109 recipient address. There doesn't seem to be a real danger of callout loops,
1110 since a callout by the remote host to check the sender would use <>.
1111 [SEE ABOVE: changed after hitting problems.]
1112
111372. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
1114 temporary errors. However, in the case of such a code being given after
1115 the end of a data transmission (i.e. after ".") Exim was failing to write
1116 a retry record for the message. (Yes, there was some broken host that was
1117 actually sending 8xx at this point.)
1118
111973. An unknown lookup type in a host list could cause Exim to panic-die when
1120 the list was checked. (An example that provoked this was putting <; in the
1121 middle of a list instead of at the start.) If this happened during a DATA
1122 ACL check, a -D file could be left lying around. This kind of configuration
1123 error no longer causes Exim to die; instead it causes a defer errror. The
1124 incident is still logged to the main and panic logs.
1125
112674. Buglet left over from Exim 3 conversion. The message "too many messages
1127 in one connection" was written to the rejectlog but not the mainlog, except
1128 when address rewriting (yes!) was being logged.
1129
113075. Added write_rejectlog option.
1131
113276. When a system filter was run not as root (that is, when system_filter_user
1133 was set), the values of the $n variables were not being returned to the
1134 main process; thus, they were not subsequently available in the $sn
1135 variables.
1136
113777. Added +return_path_on_delivery log selector.
1138
113978. A connection timeout was being treated differently from recipients deferred
1140 when testing hosts_max_try with a message that was older than the host's
1141 retry timeout. (The host should not be counted, thus allowing all hosts to
1142 be tried at least once before bouncing.) This may have been the cause of an
1143 occasionally reported bug whereby a message would remain on the queue
1144 longer than the retry timeout, but would be bounced if a delivery was
1145 forced. I say "may" because I never totally pinned down the problem;
1146 setting up timeout/retry tests is difficult. See also the next item.
1147
114879. The ultimate address timeout was not being applied to errors that involved
1149 a combination of host plus message (for example, a timeout on a MAIL
1150 command). When an address resolved to a number of possible hosts, and they
1151 were not all tried for each delivery (e.g. because of hosts_max_try), a
1152 message could remain on the queue longer than the retry timeout.
1153
115480. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
1155 Haardt.
1156
115781. Fixed an obscure SMTP outgoing bug which required at least the following
1158 conditions: (a) there was another message waiting for the same server;
1159 (b) the server returned 5xx to all RCPT commands in the first message so
1160 that the message was not completed; (c) the server dropped the connection
1161 or gave a negative response to the RSET that Exim sends to abort the
1162 transaction. The observed case was a dropped connection after DATA that had
1163 been sent in pipelining mode. That is, the server had advertised PIPELINING
1164 but was not implementing it correctly. The effect of the bug was incorrect
1165 behaviour, such as trying another host, and this could lead to a crash.
1166
1167
1168Exim version 4.30
1169-----------------
1170
1171 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
1172 and daemon.c were passed as pointers to ints; they should have been
1173 pointers to socklen_t variables (which are typically unsigned ints).
1174
1175 2. Some signed/unsigned type warnings in the os.c file for Linux have been
1176 fixed.
1177
1178 3. Fixed a really odd bug that affected only the testing scheme; patching a
1179 certain fixed string in the binary changed the value of another string that
1180 happened to be identical to the end of the original first string.
1181
1182 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
1183 name", it returns that address as the IP address. On some operating
1184 systems (e.g. Solaris), it also passes back the IP address string as the
1185 "host name". However, on others (e.g. Linux), it passes back an empty
1186 string. Exim wasn't checking for this, and was changing the host name to an
1187 empty string, assuming it had been canonicized.
1188
1189 5. Although rare, it is permitted to have more than one PTR record for a given
1190 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
1191 all the names associated with an address, because they do in Solaris.
1192 However, it seems that they do not in Linux for data that comes from the
1193 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
1194 I found this out when I moved to a new Linux workstation and tried to run
1195 the Exim test suite.
1196
1197 To get round this problem I have changed the code so that it now does its
1198 own call to the DNS to look up PTR records when searching for a host name.
1199 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
1200 addresses that are only in /etc/hosts are still found.
1201
1202 This behaviour is, however, controlled by an option called host_lookup_
1203 order, which defaults to "bydns:byaddr". If people want to use the other
1204 order, or indeed, just use one or the other means of lookup, they can
1205 specify it in this variable.
1206
1207 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
1208 some operating systems, this comes back from gethostbyaddr() as an empty
1209 string, and this is what Exim used to test for. However, it seems that in
1210 other systems, "." is yielded. Exim now tests for this case too.
1211
1212 7. The values of check_spool_space and check_log_space are now held internally
1213 as a number of kilobytes instead of an absolute number of bytes. If a
1214 numbers is specified without 'K' or 'M', it is rounded up to the nearest
1215 kilobyte. This means that much larger values can be stored.
1216
1217 8. Exim monitor: an attempt to get the action menu when not actually pointing
1218 at a message produces an empty menu entitled "No message selected". This
1219 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
1220 no entries in it ("Shell widget menu has zero width and/or height"). So I
1221 have added a single, blank menu entry in this case.
1222
1223 9. Added ${quote_local_part.
1224
122510. MIME decoding is now applied to the contents of Subject: header lines when
1226 they are logged.
1227
122811. Now that a reference to $sender_host_address automatically causes a reverse
1229 lookup to occur if necessary (4.13/18), there is no need to arrange for a
1230 host lookup before query-style lookups in lists that might use this
1231 variable. This has therefore been abolished, and the "net-" prefix is no
1232 longer necessary for query-style lookups.
1233
123412. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
1235 have been a typo for LFLAGS, so it has been changed.
1236
123713. The install script calls Exim with "-C /dev/null" in order to find the
1238 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
1239 to be output. Howeve, since Exim outputs its version number before the
1240 error, it didn't break the script. It just looked ugly. I fixed this by
1241 always allowing "-C /dev/null" if the caller is root.
1242
124314. Ignore overlarge ACL variable number when reading spool file - insurance
1244 against a later release with more variables having written the file.
1245
124615. The standard form for an IPv6 address literal was being rejected by EHLO.
1247 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1248 this, as well as the form without the "IPv6" on the front.
1249
125016. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
1251 OS/Makefile-Darwin file.
1252
125317. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
1254 only to LDAP libraries that do not have LDAP_OPT_DEREF.
1255
125618. After change 4.21/52, "%ld" was used to format the contents of the $inode
1257 variable. However, some OS use ints for inodes. I've added cast to long int
1258 to get rid of the compiler warning.
1259
126019. I had forgotten to lock out "/../" in configuration file names when
1261 ALT_CONFIG_PREFIX was set.
1262
126320. Routers used for verification do not need to specify transports. However,
1264 if such a router generated a host list, and callout was configured, Exim
1265 crashed, because it could not find a port number from the (non-existent)
1266 transport. It now assumes port 25 in this circumstance.
1267
126821. Added the -t option to exigrep.
1269
127022. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
1271 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
1272 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
1273 them is set, LOOKUP_LSEARCH is forced.
1274
127523. "exim -bV" now outputs a list of lookups that are included in the binary.
1276
127724. Added sender and host information to the "rejected by local_scan()" log
1278 line; previously there was no indication of these.
1279
128025. Added .include_if_exists.
1281
128226. Change 3.952/11 added an explicit directory sync on top of a file sync for
1283 Linux. It turns out that not all file systems support this. Apparently some
1284 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
1285 do it.) To cope with this, the error EINVAL, which means that sync-ing is
1286 not supported on the file descriptor, is now ignored when Exim is trying to
1287 sync a directory. This applies only to Linux.
1288
128927. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
1290
129128. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
1292 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
1293 this.
1294
129529. Added support for maildirsize files from supplied patch (modified a bit).
1296
129730. The use of :fail: followed by an empty string could lead Exim to respond to
1298 sender verification failures with (e.g.):
1299
1300 550 Verification failed for <xxx>
1301 550 Sender verify failed
1302
1303 where the first response line was missing the '-' that indicates it is not
1304 the final line of the response.
1305
130631. The loop for finding the name of the user that called Exim had a hardwired
1307 limit of 10; it now uses the value of finduser_retries, which is used for
1308 all other user lookups.
1309
131032. Added $received_count variable, available in data and not_smtp ACLs, and at
1311 delivery time.
1312
131333. Exim was neglecting to zero errno before one call of strtol() when
1314 expanding a string and expecting an integer value. On some systems this
1315 resulted in spurious "integer overflow" errors. Also, it was casting the
1316 result into an int without checking.
1317
131834. Testing for a connection timeout using "timeout_connect" in the retry rules
1319 did not work. The code looks as if it has *never* worked, though it appears
1320 to have been documented since at least releast 1.62. I have made it work.
1321
132235. The "timeout_DNS" error in retry rules, also documented since at least
1323 1.62, also never worked. As it isn't clear exactly what this means, and
1324 clearly it isn't a major issue, I have abolished the feature by treating it
1325 as "timeout", and writing a warning to the main and panic logs.
1326
132736. The display of retry rules for -brt wasn't always showing the error code
1328 correctly.
1329
133037. Added new error conditions to retry rules: timeout_A, timeout_MX,
1331 timeout_connect_A, timeout_connect_MX.
1332
133338. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
1334 to the empty sender.
1335
133639. The daemon was not analysing the content of -oX till after it had closed
1337 stderr and disconnected from the controlling terminal. This meant that any
1338 syntax errors were only noted on the panic log, and the return code from
1339 the command was 0. By re-arranging the code a little, I've made the
1340 decoding happen first, so such errors now appear on stderr, and the return
1341 code is 1. However, the actual setting up of the sockets still happens in
1342 the disconnected process, so errors there are still only recorded on the
1343 panic log.
1344
134540. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
1346 connections (as happens on some IP stacks) was logged at start up time as
1347 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
1348 it from "IPv6 and IPv4", which means that two separate sockets are being
1349 used.
1350
135141. The debug output for gethostbyname2() or getipnodebyname() failures now
1352 says whether AF_INET or AF_INET6 was passed as an argument.
1353
135442. Exiwhat output was messed up when time zones were included in log
1355 timestamps.
1356
135743. Exiwhat now gives more information about the daemon's listening ports,
1358 and whether -tls-on-connect was used.
1359
136044. The "port" option of the smtp transport is now expanded.
1361
136245. A "message" modifier in a "warn" statement in a non-message ACL was being
1363 silently ignored. Now an error message is written to the main and panic
1364 logs.
1365
136646. There's a new ACL modifier called "logwrite" which writes to a log file
1367 as soon as it is encountered.
1368
136947. Added $local_user_uid and $local_user_gid at routing time.
1370
137148. Exim crashed when trying to verify a sender address that was being
1372 rewritten to "<>".
1373
137449. Exim was recognizing only a space character after ".include". It now also
1375 recognizes a tab character.
1376
137750. Fixed several bugs in the Perl script that creates the exim.8 man page by
1378 extracting the relevant information from the specification. The man page no
1379 longer contains scrambled data for the -d option, and I've added a section
1380 at the front about calling Exim under different names.
1381
138251. Added "extra_headers" argument to the "mail" command in filter files.
1383
138452. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
1385 crash.
1386
138753. Installed eximstats 1.29.
1388
138954. Added transport_filter_timeout as a generic transport option.
1390
139155. Exim no longer adds an empty Bcc: header to messages that have no To: or
1392 Cc: header lines. This was required by RFC 822, but it not required by RFC
1393 2822.
1394
139556. Exim used to add From:, Date:, and Message-Id: header lines to any
1396 incoming messages that did not have them. Now it does so only if the
1397 message originates locally, that is, if there is no associated remote host
1398 address. When Resent- header lines are present, this applies to the Resent-
1399 lines rather than the non-Resent- lines.
1400
140157. Drop incoming SMTP connection after too many syntax or protocol errors. The
1402 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
1403
140458. Messages for configuration errors now include the name of the main
1405 configuration file - useful now that there may be more than one file in a
1406 list (.included file names were always shown).
1407
140859. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
1409 for those rare installations that do not start the daemon as root or run it
1410 setuid root. I've cut out the call to initgroups() if the daemon is not
1411 root at that time.
1412
141360. The Exim user and group can now be bound into the binary as text strings
1414 that are looked up at the start of Exim's processing.
1415
141661. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
1417
141862. Added $mailstore_basename variable.
1419
142063. Installed patch to sieve.c from Michael Haardt.
1421
142264. When Exim failed to open the panic log after failing to open the main log,
1423 the original message it was trying to log was written to stderr and debug
1424 output, but if they were not available (the usual case in production), it
1425 was lost. Now it is written to syslog before the two lines that record the
1426 failures to open the logs.
1427
142865. Users' Exim filters run in subprocesses under the user's uid. It is
1429 possible for a "deliver" command or an alias in a "personal" command to
1430 provoke an address rewrite. If logging of address rewriting is configured,
1431 this fails because the process is not running as root or exim. There may be
1432 a better way of dealing with this, but for the moment (because 4.30 needs
1433 to be released), I have disabled address rewrite logging when running a
1434 filter in a non-root, non-exim process.
1435
1436
1437Exim version 4.24
1438-----------------
1439
1440 1. The buildconfig auxiliary program wasn't quoting the value set for
1441 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
1442 not defined. This bug was masked in 4.22 by the effect that was fixed in
1443 change 4.23/1.
1444
1445 2. Some messages that were rejected after a message id was allocated were
1446 shown as "incomplete" by exigrep. It no longer does this for messages that
1447 are rejected by local_scan() or the DATA or non-SMTP ACLs.
1448
1449 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
1450 have allow_domain_literals set, the ID did not get logged in the <= line.
1451 Domain literals are now always recognized in Message-ID: header lines.
1452
1453 4. The first argument for a ${extract expansion item is the key name or field
1454 number. Leading and trailing spaces in this item were not being ignored,
1455 causing some misleading effects.
1456
1457 5. When deliver_drop_privilege was set, single queue runner processes started
1458 manually (i.e. by the command "exim -q") or by the daemon (which uses the
1459 same command in the process it spins off) were not dropping privilege.
1460
1461 6. When the daemon running as "exim" started a queue runner, it always
1462 re-executed Exim in the spun-off process. This is a waste of effort when
1463 deliver_drop_privilege is set. The new process now just calls the
1464 queue-runner function directly.
1465
1466
1467Exim version 4.23
1468-----------------
1469
1470 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
1471 HEADERS_CHARSET.
1472
1473 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
1474 ignored. Though the use of -oP was forcing the writing of a pid file, it
1475 was always written to the default place.
1476
1477 3. If the message "no IP address found for host xxxx" is generated during
1478 incoming verification, it is now followed by identification of the incoming
1479 connection (so you can more easily find what provoked it).
1480
1481 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
1482
1483 5. Added some features to "harden" Exim a bit more against certain attacks:
1484
1485 (a) There is now a build-time option called FIXED_NEVER_USERS that can
1486 be put in Local/Makefile. This is like the never_users runtime option,
1487 but it cannot be overridden. The default setting is "root".
1488
1489 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
1490 prefix string with which any file named in a -C command line option
1491 must start.
1492
1493 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
1494 is retained for -C and -D only if the caller of Exim is root. Without
1495 it, the exim user may also use -C and -D and retain privilege.
1496
1497 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
1498 command line option is disabled.
1499
1500 6. Macro names set by the -D option must start with an upper case letter, just
1501 like macro names defined in the configuration file.
1502
1503 7. Added "dereference=" facility to LDAP.
1504
1505 8. Two instances of the typo "uknown" in the source files are fixed.
1506
1507 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
1508 the Configure-Makefile script screwed up while processing it.
1509
151010. Incorporated PCRE 4.4.
1511
151211. The SMTP synchronization check was not operating right at the start of an
1513 SMTP session. For example, it could not catch a HELO sent before the client
1514 waited for the greeting. There is now a check for outstanding input at the
1515 point when the greeting is written. Because of the duplex, asynchronous
1516 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
1517 way, but not yet received, when the check is performed.
1518
151912. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
1520 on TCP/IP sockets, because this apparently causes some broken clients to
1521 timeout.
1522
152313. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
1524 unchanged) from the Cygwin maintainer.
1525
152614. The code for -bV that shows what is in the binary showed "mbx" when maildir
1527 was supported instead of testing for mbx. Effectively a typo.
1528
152915. The spa authenticator server code was not checking that the input it
1530 received was valid base64.
1531
153216. The debug output line for the "set" modifier in ACLs was not showing the
1533 name of the variable that was being set.
1534
153517. Code tidy: the variable type "vtype_string" was never used. Removed it.
1536
153718. Previously, a reference to $sender_host_name did not cause a DNS reverse
1538 lookup on its own. Something else was needed to trigger the lookup. For
1539 example, a match in host_lookup or the need for a host name in a host list.
1540 Now, if $sender_host_name is referenced and the host name has not yet been
1541 looked up, a lookup is performed. If the lookup fails, the variable remains
1542 empty, and $host_lookup_failed is set to "1".
1543
154419. Added "eqi" as a case-independent comparison operator.
1545
154620. The saslauthd authentication condition could segfault if neither service
1547 nor realm was specified.
1548
154921. If an overflowing value such as "2048M" was set for message_size_limit, the
1550 error message that was logged was misleading, and incoming SMTP
1551 connections were dropped. The message is now more accurate, and temporary
1552 errors are given to SMTP connections.
1553
155422. In some error situations (such as 21 above) Exim rejects all SMTP commands
1555 (except RSET) with a 421 error, until QUIT is received. However, it was
1556 failing to send a response to QUIT.
1557
155823. The HELO ACL was being run before the code for helo_try_verify_hosts,
1559 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
1560 ACL is now run after the helo_try_verify_hosts code.
1561
156224. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
1563 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
1564 case-independent, so other case variants are also recognized). Apparently
1565 some systems use these upper case variants.
1566
156725. If more than two messages were waiting for the same host, and a transport
1568 filter was specified for the transport, Exim sent two messages over the
1569 same TCP/IP connection, and then failed with "socket operation on non-
1570 socket" when it tried to send the third.
1571
157226. Added Exim::debug_write and Exim::log_write for embedded Perl use.
1573
157427. The extern definition of crypt16() in expand.c was not being excluded when
1575 the OS had its own crypt16() function.
1576
157728. Added bounce_return_body as a new option, and bounce_return_size_limit
1578 as a preferred synonym for return_size_limit, both as an option and as an
1579 expansion variable.
1580
158129. Added LIBS=-liconv to OS/Makefile-OSF1.
1582
158330. Changed the default configuration ACL to relax the local part checking rule
1584 for addresses that are not in any local domains. For these addresses,
1585 slashes and pipe symbols are allowed within local parts, but the sequence
1586 /../ is explicitly forbidden.
1587
158831. SPA server authentication was not clearing the challenge buffer before
1589 using it.
1590
159132. log_message in a "warn" ACL statement was writing to the reject log as
1592 well as to the main log, which contradicts the documentation and doesn't
1593 seem right (because no rejection is happening). So I have stopped it.
1594
159533. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
1596 However, I am unable to do any testing of this.
1597
159834. Fixed an infelicity in the appendfile transport. When checking directories
1599 for a mailbox, to see if any needed to be created, it was accidentally
1600 using path names with one or more superfluous leading slashes; tracing
1601 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
1602
160335. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
1604 contents are added to the log line that is written for every discarded
1605 recipient. (Previously a log_message setting was ignored.)
1606
160736. The ${quote: operator now quotes the string if it is empty.
1608
160937. The install script runs exim in order to find its version number. If for
1610 some reason other than non-existence or emptiness, which it checks, it
1611 could not run './exim', it was installing it with an empty version number,
1612 i.e. as "exim-". This error state is now caught, and the installation is
1613 aborted.
1614
161538. An argument was missing from the function that creates an error message
1616 when Exim fails to connect to the socket for saslauthd authentication.
1617 This could cause Exim to crash, or give a corrupted message.
1618
161939. Added isip, isip4, and isip6 to ${if conditions.
1620
162140. The ACL variables $acl_xx are now saved with the message, and can be
1622 accessed later in routers, transports, and filters.
1623
162441. The new lookup type nwildlsearch is like wildlsearch, except that the key
1625 strings in the file are not string-expanded.
1626
162742. If a MAIL command specified a SIZE value that was too large to fit into an
1628 int variable, the check against message_size_limit failed. Such values are
1629 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
1630 day this will have to be increased, but I don't think I want to be around
1631 when emails are that large.
1632
1633
1634
1635Exim version 4.22
1636-----------------
1637
1638 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
1639 iconv() is not standard in FreeBSD.
1640
1641 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
1642 IPv6 enabled. The observed symptom was a segmentation fault on return from
1643 the function os_common_find_running_interfaces() in src/os.c.
1644
1645 3. In the check_special_case() function in daemon.c I had used "errno" as an
1646 argument name, which causes warnings on some systems. This was basically a
1647 typo, since it was named "eno" in the comments!
1648
1649 4. The code that waits for the clock to tick (at a resolution of some fraction
1650 of a second) so as to ensure message-id uniqueness was always waiting for
1651 at least one whole tick, when it could have waited for less. [This is
1652 almost certainly not relevant at current processor speeds, where it is
1653 unlikely to ever wait at all. But we try to future-proof.]
1654
1655 5. The function that sleeps for a time interval that includes fractions of a
1656 second contained a race. It did not block SIGALRM between setting the
1657 timer, and suspending (a couple of lines later). If the interval was short
1658 and the sigsuspend() was delayed until after it had expired, the suspension
1659 never ended. On busy systems this could lead to processes getting stuck for
1660 ever.
1661
1662 6. Some uncommon configurations may cause a lookup to happen in a queue runner
1663 process, before it forks any delivery processes. The open lookup caching
1664 mechanism meant that the open file or database connection was passed into
1665 the delivery process. The problem was that delivery processes always tidy
1666 up cached lookup data. This could cause a problem for the next delivery
1667 process started by the queue runner, because the external queue runner
1668 process does not know about the closure. So the next delivery process
1669 still has data in the lookup cache. In the case of a file lookup, there was
1670 no problem because closing a file descriptor in a subprocess doesn't affect
1671 the parent. However, if the lookup was caching a connection to a database,
1672 the connection was closed, and the second delivery process was likely to
1673 see errors such as "PGSQL: query failed: server closed the connection
1674 unexpectedly". The problem has been fixed by closing all cached lookups
1675 in a queue runner before running a delivery process.
1676
1677 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
1678 seem to have the "const" qualifier which it has on other OS. I've
1679 parameterised it.
1680
1681 8. Change 4.21/2 was too strict. It is only if there are two authenticators
1682 *of the same type* (client or server) with the same public name that an
1683 error should be diagnosed.
1684
1685 9. When Exim looked up a host name for an IP address, but failed to find the
1686 original IP address when looking up the host name (a safety check), it
1687 output the message "<ip address> does not match any IP for NULL", which was
1688 confusing, to say the least. The bug was that the host name should have
1689 appeared instead of "NULL".
1690
169110. Since release 3.03, if Exim is called by a uid other than root or the Exim
1692 user that is built into the binary, and the -C or -D options is used, root
1693 privilege is dropped before the configuration file is read. In addition,
1694 logging is switched to stderr instead of the normal log files. If the
1695 configuration then re-defines the Exim user, the unprivileged environment
1696 is probably not what is expected, so Exim logs a panic warning message (but
1697 proceeds).
1698
1699 However, if deliver_drop_privilege is set, the unprivileged state may well
1700 be exactly what is intended, so the warning has been cut out in that case,
1701 and Exim is allowed to try to write to its normal log files.
1702
1703
1704Exim version 4.21
1705-----------------
1706
1707 1. smtp_return_error_details was not giving details for temporary sender
1708 or receiver verification errors.
1709
1710 2. Diagnose a configuration error if two authenticators have the same public
1711 name.
1712
1713 3. Exim used not to create the message log file for a message until the first
1714 delivery attempt. This could be confusing when incoming messages were held
1715 for policy or load reasons. The message log file is now created at the time
1716 the message is received, and an initial "Received" line is written to it.
1717
1718 4. The automatically generated man page for command line options had a minor
1719 bug that caused no ill effects; however, a more serious problem was that
1720 the procedure for building the man page automatically didn't always
1721 operate. Consequently, release 4.20 contains an out-of-date version. This
1722 shouldn't happen again.
1723
1724 5. When building Exim with embedded Perl support, the script that builds the
1725 Makefile was calling 'perl' to find its compile-time parameters, ignoring
1726 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
1727
1728 6. The freeze_tell option was not being used for messages that were frozen on
1729 arrival, either by an ACL or by local_scan().
1730
1731 7. Added the smtp_incomplete_transaction log selector.
1732
1733 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
1734 was accepting AUTH without a new EHLO.
1735
1736 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
1737 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
1738 RFC.
1739
174010. Logging of TCP/IP connections (when configured) now happens in the main
1741 daemon process instead of the child process, so that the TCP/IP connection
1742 count is more accurate (but it can never be perfect).
1743
174411. The use of "drop" in a nested ACL was not being handled correctly in the
1745 outer ACL. Now, if condition failure induced by the nested "drop" causes
1746 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
1747 or "require"), the connection is dropped.
1748
174912. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
1750 that yield "discard" can now be used with an "accept" or a "discard" verb,
1751 but an error is generated for any others (because I can't see a useful way
1752 to define what should happen).
1753
175413. When an ACL is read dynamically from a file (or anywhere else), the lines
1755 are now processed in the same way as lines in the Exim configuration file.
1756 In particular, continuation lines are supported.
1757
175814. Added the "dnslists = a.b.c!=n.n.n.n" feature.
1759
176015. Added -ti meaning -t -i.
1761
176216. Check for letters, digits, hyphens, and dots in the names of dnslist
1763 domains, and warn by logging if others are found.
1764
176517. At least on BSD, alignment is not guarenteed for the array of ifreq's
1766 returned from GIFCONF when Exim is trying to find the list of interfaces on
1767 a host. The code in os.c has been modified to copy each ifreq to an aligned
1768 structure in all cases.
1769
1770 Also, in some cases, the returned ifreq's were being copied to a 'struct
1771 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
1772 means the last couple of bytes of an IPv6 address could be chopped if the
1773 ifreq contained only a normal sockaddr (14 bytes storage).
1774
177518. Named domain lists were not supported in the hosts_treat_as_local option.
1776 An entry such as +xxxx was not recognized, and was treated as a literal
1777 domain name.
1778
177919. Ensure that header lines added by a DATA ACL are included in the reject log
1780 if the ACL subsequently rejects the message.
1781
178220. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
1783 MD5 (which is deprecated).
1784
178521. When testing a filter file using -bf, Exim was writing a message when it
1786 took the sender from a "From " line in the message, but it was not doing so
1787 when it took $return_path from a Return-Path: header line. It now does.
1788
178922. If the contents of a "message" modifier for a "warn" ACL verb do not begin
1790 with a valid header line field name (a series of printing characters
1791 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
1792
179323. Changed "disc" in the source to "disk" to conform to the documentation and
1794 the book and for uniformity.
1795
179624. Ignore Sendmail's -Ooption=value command line item.
1797
179825. When execve() failed while trying to run a command in a pipe transport,
1799 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
1800 could be confused with a return value of 69 from the command itself. This
1801 has been changed to 127, the value the shell returns if it is asked to run
1802 a non-existent command. The wording for the related log line suggests a
1803 non-existent command as the problem.
1804
180526. If received_header_text expands to an empty string, do not add a Received:
1806 header line to the message. (Well, it adds a token one on the spool, but
1807 marks it "old" so that it doesn't get used or transmitted.)
1808
180927. Installed eximstats 1.28 (addition of -nt option).
1810
181128. There was no check for failure on the call to getsockname() in the daemon
1812 code. This can fail if there is a shortage of resources on the system, with
1813 ENOMEM, for example. A temporary error is now given on failure.
1814
181529. Contrary to the C standard, it seems that in some environments, the
1816 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
1817 program. Exim now does this explicitly; it affects the formatting of
1818 timestamps using strftime().
1819
182030. If exiqsumm was given junk data, it threw up some uninitialized variable
1821 complaints. I've now initialized all the variables, to avoid this.
1822
182332. Header lines added by a system filter were not being "seen" during
1824 transport-time rewrites.
1825
182633. The info_callback() function passed to OpenSSL is set up with type void
1827 (*)(SSL *, int, int), as described somewhere. However, when calling the
1828 function (actually a macro) that sets it up, the type void(*)() is
1829 expected. I've put in a cast to prevent warnings from picky compilers.
1830
183134. If a DNS black list lookup found a CNAME record, but there were no A
1832 records associated with the domain it pointed at, Exim crashed.
1833
183435. If a DNS black list lookup returned more than one A record, Exim ignored
1835 all but the first. It now scans all returned addresses if a particular IP
1836 value is being sought. In this situation, the contents of the
1837 $dnslist_value variable are a list of all the addresses, separated by a
1838 comma and a space.
1839
184036. Tightened up the rules for host name lookups using reverse DNS. Exim used
1841 to accept a host name and all its aliases if the forward lookup for any of
1842 them yielded the IP address of the incoming connection. Now it accepts only
1843 those names whose forward lookup yields the correct IP address. Any other
1844 names are discarded. This closes a loophole whereby a rogue DNS
1845 administrator could create reverse DNS records to break through a
1846 wildcarded host restriction in an ACL.
1847
184837. If a user filter or a system filter that ran in a subprocess used any of
1849 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
1850 the wrong values were passed to the pipe command ($thisaddress had the
1851 value of $0, $0 had the value of $1, etc). This bug was introduced by
1852 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
1853
185438. Improved the line breaking for long SMTP error messages from ACLs.
1855 Previously, if there was no break point between 40 and 75 characters, Exim
1856 left the rest of the message alone. Two changes have been made: (a) I've
1857 reduced the minimum length to 35 characters; (b) if it can't find a break
1858 point between 35 and 75 characters, it looks ahead and uses the first one
1859 that it finds. This may give the occasional overlong line, but at least the
1860 remaining text gets split now.
1861
186239. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
1863 file descriptors might be low, and that setting 1000 would always raise it.
1864 It turns out that in some environments, the limit is already over 1000 and
1865 that lowering it causes trouble. So now Exim takes care not to decrease it.
1866
186740. When delivering a message, the value of $return_path is set to $sender_
1868 address at the start of routing (routers may change the value). By an
1869 oversight, this default was not being set up when an address was tested by
1870 -bt or -bv, which affected the outcome if any router or filter referred to
1871 $return_path.
1872
187341. The idea of the "warn" ACL verb is that it adds a header or writes to the
1874 log only when "message" or "log_message" are set. However, if one of the
1875 conditions was an address verification, or a call to a nested ACL, the
1876 messages generated by the underlying test were being passed through. This
1877 no longer happens. The underlying message is available in $acl_verify_
1878 message for both "message" and "log_message" expansions, so it can be
1879 passed through if needed.
1880
188142. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
1882 new expansion $bh_ to give the encoded byte string without charset
1883 translation. Translation happens only if iconv() is available; HAVE_ICONV
1884 indicates this at build time. HEADERS_CHARSET gives the charset to
1885 translate to; headers_charset can change it in the configuration, and
1886 "headers charset" can change it in an individual filter file.
1887
188843. Now that we have a default RFC 2047 charset (see above), the code in Exim
1889 that creates RFC 2047 encoded "words" labels them as that charset instead
1890 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
1891 expansion operator; (ii) when Exim creates a From: line for a local
1892 message; (iii) when a header line is rewritten to include a "phrase" part.
1893
189444. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
1895 buggy, causing it to skip the first lines of messages whose message ID
1896 ended in 'D'. This would not have bitten before Exim release 4.14, because
1897 message IDs were unlikely to end in 'D' before then. The effect was to have
1898 incorrect size information for certain domains.
1899
190045. #include "config.h" was missing at the start of the crypt16.c module. This
1901 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
1902 noticed.
1903
190446. If there was a timeout during a "random" callout check, Exim treated it as
1905 a failure of the random address, and carried on sending RSET and the real
1906 address. If the delay was just some slowness somewhere, the response to the
1907 original RCPT would be taken as a response to RSET and so on, causing
1908 mayhem of various kinds.
1909
191047. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
1911 when I implemented it. It didn't allow for the fact that some option values
1912 may legitimatetly be negative (e.g. size_addition), and it didn't even do
1913 the right test for positive values.
1914
191548. Domain names in DNS records are case-independent. Exim always looks them up
1916 in lower case. Some resolvers return domain names in exactly the case they
1917 appear in the zone file, that is, they may contain uppercase letters. Not
1918 all resolvers do this - some return always lower case. Exim was treating a
1919 change of case by a resolver as a change of domain, similar to a widening
1920 of a domain abbreviation. This triggered its re-routing code and so it was
1921 trying to route what was effectively the same domain again. This normally
1922 caused routing to fail (because the router wouldn't handle the domain
1923 twice). Now Exim checks for this case specially, and just changes the
1924 casing of the domain that it ultimately uses when it transmits the message
1925 envelope.
1926
192749. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
1928 module.
1929
193050. If a filter generated a file delivery with a non-absolute name (possible if
1931 no home directory exists for the router), the forbid_file option was not
1932 forbidding it.
1933
193451. Added '&' feature to dnslists, to provide bit mask matching in addition to
1935 the existing equality matching.
1936
193752. Exim was using ints instead of ino_t variables in some places where it was
1938 dealing with inode numbers.
1939
194053. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
1941 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
1942 called TMPDIR, and if it finds it is different, it changes its value.
1943
194454. The smtp_printf() function is now made available to local_scan() so
1945 additional output lines can be written before returning. There is also an
1946 smtp_fflush() function to enable the detection of a dropped connection.
1947 The variables smtp_input and smtp_batched_input are exported to
1948 local_scan().
1949
195055. Changed the default runtime configuration: the message "Unknown user"
1951 has been removed from the ACL, and instead placed on the localuser router,
1952 using the cannot_route_message feature. This means that any verification
1953 failures that generate their own messages won't get overridden. Similarly,
1954 the "Unrouteable address" message that was in the ACL for unverifiable
1955 relay addresses has also been removed.
1956
195756. Added hosts_avoid_esmtp to the smtp transport.
1958
195957. The exicyclog script was not checking for the esoteric option
1960 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
1961 will work only if exicyclog is run under the appropriate euid.
1962
196358. Following a discussion on the list, the rules by which Exim recognises line
1964 endings on incoming messages have been changed. The -dropcr and drop_cr
1965 options are now no-ops, retained only for backwards compatibility. The
1966 following line terminators are recognized: LF CRLF CR. However, special
1967 processing applies to CR:
1968
1969 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
1970 nor a local message in the state where . is a terminator.
1971
1972 (ii) If a bare CR is encountered in a header line, an extra space is added
1973 after the line terminator so as not to end the header. The reasoning
1974 behind this is that bare CRs in header lines are most likely either
1975 to be mistakes, or people trying to play silly games.
1976
197759. The size of a message, as listed by "-bp" or in the Exim monitor window,
1978 was being incorrectly given as 18 bytes larger than it should have been.
1979 This is a VOB (very old bug).
1980
198160. This may never have affected anything current, but just in case it has:
1982 When the local host is found other than at the start of a list of hosts,
1983 the local host, those with the same MX, and any that follow, are discarded.
1984 When the list in question was part of a longer list of hosts, the following
1985 hosts (not currently being processed) were also being discarded. This no
1986 longer happens. I'm not sure if this situation could ever has previously
1987 arisen.
1988
198961. Added the "/MX" feature to lists of hosts in the manualroute and query
1990 program routers.
1991
199262. Whenever Exim generates a new message, it now adds an Auto-Submitted:
1993 header. This is something that is recommended in a new Internet Draft, and
1994 is something that is documented as being done by Sendmail. There are two
1995 possible values. For messages generated by the autoreply transport, Exim
1996 adds:
1997
1998 Auto-Submitted: auto-replied
1999
2000 whereas for all other generated messages (e.g. bounces) it adds
2001
2002 Auto-Submitted: auto-generated
2003
200463. The "personal" condition in filters now includes a test for the
2005 Auto-Submitted: header. If it contains the string "auto-" the message it
2006 not considered personal.
2007
200864. Added rcpt_include_affixes as a generic transport option.
2009
201065. Added queue_only_override (default true).
2011
201266. Added the syslog_duplication option.
2013
201467. If what should have been the first header line of a message consisted of
2015 a space followed by a colon, Exim was mis-interpreting it as a header line.
2016 It isn't of course - it is syntactically invalid and should therefore be
2017 treated as the start of the message body. The misbehaviour could have
2018 caused a number of strange effects, including loss of data in subsequent
2019 header lines, and spool format errors.
2020
202168. Formerly, the AUTH parameter on a MAIL command was trusted only if the
2022 client host had authenticated. This control can now be exercised by an ACL
2023 for more flexibility.
2024
202569. By default, callouts do not happen when testing with -bh. There is now a
2026 variant, -bhc, which does actually run the callout code, including
2027 consulting and updating the callout cache.
2028
202970. Added support for saslauthd authentication, courtesy of Alexander
2030 Sabourenkov.
2031
203271. If statvfs() failed on the spool or log directories while checking their
2033 size for availability, Exim confusingly gave the error "space shortage".
2034 Furthermore, in debugging mode it crashed with a floating point exception.
2035 These checks are done if check_{spool,log}_{space,inodes} are set, and when
2036 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
2037 serious problem, Exim now writes to the main and panic logs when this
2038 happens, with details of the failure. It then refuses to accept the
2039 incoming message, giving the message "spool directory problem" or "log
2040 directory problem" with a 421 code for SMTP messages.
2041
204272. When Exim is about to re-exec itself, it ensures that the file descriptors
2043 0, 1, and 2 exist, because some OS complain for execs without them (see
2044 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
2045 descriptors. However, the code omitted to check that the open succeeded,
2046 causing mysterious errors if for some reason the permissions on /dev/null
2047 got screwed. Now Exim writes a message to the main and panic logs, and
2048 bombs out if it can't open /dev/null.
2049
205073. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
2051 interact so that it is all more flexible. It is supposed to remain
2052 backwards compatible. Also added extra_local_interfaces.
2053
205474. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
2055 to bomb out with an assertion failure - to the client this appears as a
2056 connection drop. This problem occurs in the part of the code that was taken
2057 from the Samba project. Fortunately, the assertion is in a very simple
2058 function, so I have fixed this by reproducing the function inline in the
2059 one place where it is called, and arranging for authentication to fail
2060 instead of killing the process with assert().
2061
206275. The SPA client code was not working when the server requested OEM rather
2063 than Unicode encoding.
2064
206576. Added code to make require_files with a specific uid setting more usable in
2066 the case where statting the file as root fails - usually a non-root-mounted
2067 NFS file system. When this happens and the failure is EACCES, Exim now
2068 forks a subprocess and does the per-uid checking as the relevant uid.
2069
207077. Added process_log_path.
2071
207278. If log_file_path was not explicitly set, a setting of check_log_space or
2073 check_log_inodes was ignored.
2074
207579. If a space check for the spool or log partitions fails, the incident is now
2076 logged. Of course, in the latter case the data may get lost...
2077
207880. Added the %p formatting code to string_format() so that it can be used to
2079 print addresses in debug_print(). Adjusted all the address printing in the
2080 debugging in store.c to use %p rather than %d.
2081
208281. There was a concern that a line of code in smtp_in.c could overflow a
2083 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
2084 initially expressed, the concern was not well-founded, because trailing
2085 spaces are removed early. However, if the trailing spaces were followed by
2086 a NULL, they did not get removed, so the overflow was possible. Two fixes
2087 were applied:
2088
2089 (a) I re-wrote the offending code in a cleaner fashion.
2090 (b) If an incoming SMTP command contains a NULL character, it is rejected
2091 as invalid.
2092
209382. When Exim changes uid/gid to the Exim user at daemon start time, it now
2094 runs initgroups(), so that if the Exim user is in any additional groups,
2095 they will be used during message reception.
2096
2097
2098Exim version 4.20
2099-----------------
2100
2101The change log for 4.20 and earlier releases has been archived.
2102
2103****