1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.351 2006/05/08 12:45:05 steve Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 SC/01 Use a glob alias rather than an array ref in eximstats generated
10 parser. This improves both readability and performance.
12 SC/02 Collect SpamAssassin and rejection statistics in eximstats.
13 Don't display local sender or destination tables in eximstats unless
14 there is data to show.
15 Added average volumes into the eximstats top table text output.
17 SC/03 Collect data on the number of addresses (recipients) as well
18 as the number of messages in eximstats.
25 TF/01 Fix the add_header change below (4.61 PH/55) which had a bug that (amongst
26 other effects) broke the use of negated acl sub-conditions.
28 PH/01 ${readsocket now supports Internet domain sockets (modified John Jetmore
31 PH/02 When tcp-wrappers is called from Exim, it returns only "deny" or "allow".
32 "Deny" causes Exim to reject the incoming connection with a 554 error.
33 Unfortunately, if there is a major crisis, such as a disk failure,
34 tcp-wrappers gives "deny", whereas what one would like would be some
35 kind of temporary error. A kludge has been added to help with this.
36 Before calling hosts_ctl(), errno is set zero. If the result is "deny", a
37 554 error is used if errno is still zero or contains ENOENT (which occurs
38 if either of the /etc/hosts.{allow,deny} files is missing). Otherwise, a
41 PH/03 Add -lutil to the default FreeBSD LIBS setting.
43 PH/04 Change PH/19 for 4.61 was too wide. It should not be applied to host
44 errors. Otherwise a message that provokes a temporary error (when other
45 messages do not) can cause a whole host to time out.
47 PH/05 Batch deliveries by appendfile and pipe transports did not work when the
48 addresses were routed directly to files or pipes from a redirect router.
49 File deliveries just didn't batch; pipe deliveries might have suffered
52 PH/06 A failure to get a lock for a hints database would erroneously always say
53 "Failed to get write lock", even when it was really a read lock.
55 PH/07 The appendfile transport was creating MBX lock files with a fixed mode
56 of 0600. This has been changed to use the value of the lockfile_mode
57 option (which defaults to 0600).
59 PH/08 Applied small patch from the Sieve maintainer.
61 PH/09 If maildir_quota_directory_regex was set to exclude (say) the .Trash
62 folder from quota calculations, a direct delivery into this folder messed
63 up the contents of the maildirsize file. This was because the regex was
64 used only to exclude .Trash (or whatever) when the size of the mailbox
65 was calculated. There was no check that a delivery was happening into an
66 excluded directory. This bug has been fixed by ignoring all quota
67 processing for deliveries into excluded directories.
69 PH/10 Added the maildirfolder_create_regex option to appendfile.
75 PH/01 The code for finding all the local interface addresses on a FreeBSD
76 system running IPv6 was broken. This may well have applied to all BSD
77 systems, as well as to others that have similar system calls. The broken
78 code found IPv4 interfaces correctly, but gave incorrect values for the
79 IPv6 interfaces. In particular, ::1 was not found. The effect in Exim was
80 that it would not match correctly against @[] and not recognize the IPv6
83 PH/02 The ipliteral router was not recognizing addresses of the form user@
84 [ipv6:....] because it didn't know about the "ipv6:" prefix.
86 PH/03 Added disable_ipv6.
88 PH/04 Changed $reply_address to use the raw form of the headers instead of the
89 decoded form, because it is most often used to construct To: headers
90 lines in autoreplies, and the decoded form may well be syntactically
91 invalid. However, $reply_address has leading white space removed, and all
92 newlines turned into spaces so that the autoreply transport does not
95 PH/05 If group was specified without a user on a router, and no group or user
96 was specified on a transport, the group from the router was ignored.
98 PH/06 Increased the number of ACL variables to 20 of each type, and arranged
99 for visible compile-time settings that can be used to change these
100 numbers, for those that want even more. Backwards compatibility with old
101 spool files has been maintained. However, going back to a previous Exim
102 release will lost any variables that are in spool files.
104 PH/07 Two small changes when running in the test harness: increase delay when
105 passing a TCP/IP connection to a new process, in case the original
106 process has to generate a bounce, and remove special handling of
107 127.0.0.2 (sic), which is no longer necessary.
109 PH/08 Changed debug output of dbfn_open() flags from numbers to names, so as to
110 be the same on different OS.
112 PH/09 Moved a debug statement in filter processing to avoid a race problem when
115 JJ/01 exipick: fixed bug where -b (brief) output option showed "Vars:"
116 whether --show-vars was specified or not
118 JJ/02 exipick: Added support for new ACL variable spool format introduced
121 PH/10 Fixed another bug related to PH/04 above: if an incoming message had a
122 syntactically invalid From: or Reply-to: line, and a filter used this to
123 generate an autoreply, and therefore failed to obtain an address for the
124 autoreply, Exim could try to deliver to a non-existent relative file
125 name, causing unrelated and misleading errors. What now happens is that
126 it logs this as a hard delivery error, but does not attempt to create a
129 PH/11 The exinext utility has a -C option for testing purposes, but although
130 the given file was scanned by exinext itself; it wasn't being passed on
131 when Exim was called.
133 PH/12 In the smtp transport, treat an explicit ECONNRESET error the same as
134 an end-of-file indication when reading a command response.
136 PH/13 Domain literals for IPv6 were not recognized unless IPv6 support was
137 compiled. In many other places in Exim, IPv6 addresses are always
138 recognized, so I have changed this. It also means that IPv4 domain
139 literals of the form [IPV4:n.n.n.n] are now always recognized.
141 PH/14 When a uid/gid is specified for the queryprogram router, it cannot be
142 used if the router is not running as root, for example, when verifying at
143 ACL time, or when using -bh. The debugging output from this situation was
144 non-existent - all you got was a failure to exec. I have made two
147 (a) Failures to set uid/gid, the current directory, or a process leader
148 in a subprocess such as that created by queryprogram now generate
149 suitable debugging ouput when -d is set.
151 (b) The queryprogram router detects when it is not running as root,
152 outputs suitable debugging information if -d is set, and then runs
153 the subprocess without attempting to change uid/gid.
155 PH/15 Minor change to Makefile for building test_host (undocumented testing
158 PH/16 As discussed on the list in Nov/Dec: Exim no longer looks at the
159 additional section of a DNS packet that returns MX or SRV records.
160 Instead, it always explicitly searches for A/AAAA records. This avoids
161 major problems that occur when a DNS server includes only records of one
162 type (A or AAAA) in an MX/SRV packet. A byproduct of this change has
163 fixed another bug: if SRV records were looked up and the corresponding
164 address records were *not* found in the additional section, the port
165 values from the SRV records were lost.
167 PH/17 If a delivery to a pipe, file, or autoreply was deferred, Exim was not
168 using the correct key (the original address) when searching the retry
169 rules in order to find which one to use for generating the retry hint.
171 PH/18 If quota_warn_message contains a From: header, Exim now refrains from
172 adding the default one. Similarly, if it contains a Reply-To: header, the
173 errors_reply_to option, if set, is not used.
175 PH/19 When calculating a retry time, Exim used to measure the "time since
176 failure" by looking at the "first failed" field in the retry record. Now
177 it does not use this if it is later than than the arrival time of the
178 message. Instead it uses the arrival time. This makes for better
179 behaviour in cases where some deliveries succeed, thus re-setting the
180 "first failed" field. An example is a quota failure for a huge message
181 when small messages continue to be delivered. Without this change, the
182 "time since failure" will always be short, possible causing more frequent
183 delivery attempts for the huge message than are intended.
184 [Note: This change was subsequently modified - see PH/04 for 4.62.]
186 PH/20 Added $auth1, $auth2, $auth3 to contain authentication data (as well as
187 $1, $2, $3) because the numerical variables can be reset during some
188 expansion items (e.g. "match"), thereby losing the authentication data.
190 PH/21 Make -bV show the size of off_t variables so that the test suite can
191 decide whether to run tests for quotas > 2G.
193 PH/22 Test the values given for quota, quota_filecount, quota_warn_threshold,
194 mailbox_size, and mailbox_filecount in the appendfile transport. If a
195 filecount value is greater than 2G or if a quota value is greater than 2G
196 on a system where the size of off_t is not greater than 4, a panic error
199 PH/23 When a malformed item such as 1.2.3/24 appears in a host list, it can
200 never match. The debug and -bh output now contains an explicit error
201 message indicating a malformed IPv4 address or mask.
203 PH/24 An host item such as 1.2.3.4/abc was being treated as the IP address
204 1.2.3.4 without a mask. Now it is not recognized as an IP address, and
207 PH/25 Do not write to syslog when running in the test harness. The only
208 occasion when this arises is a failure to open the main or panic logs
209 (for which there is an explicit test).
211 PH/26 Added the /no_tell option to "control=freeze".
213 PH/27 If a host name lookup failed very early in a connection, for example, if
214 the IP address matched host_lookup and the reverse lookup yielded a name
215 that did not have a forward lookup, an error message of the form "no IP
216 address found for host xxx.xxx.xxx (during SMTP connection from NULL)"
217 could be logged. Now it outputs the IP address instead of "NULL".
219 PH/28 An enabling patch from MH: add new function child_open_exim2() which
220 allows the sender and the authenticated sender to be set when
221 submitting a message from within Exim. Since child_open_exim() is
222 documented for local_scan(), the new function should be too.
224 PH/29 In GnuTLS, a forced expansion failure for tls_privatekey was not being
225 ignored. In both GnuTLS and OpenSSL, an expansion of tls_privatekey that
226 results in an empty string is now treated as unset.
228 PH/30 Fix eximon buffer overflow bug (Bugzilla #73).
230 PH/31 Added sender_verify_fail logging option.
232 PH/32 In November 2003, the code in Exim that added an empty Bcc: header when
233 needed by RFC 822 but not by RFC 2822 was commented out. I have now
234 tidied the source and removed it altogether.
236 PH/33 When a queue run was abandoned because the load average was too high, a
237 log line was always written; now it is written only if the queue_run log
238 selector is set. In addition, the log line for abandonment now contains
239 information about the queue run such as the pid. This is always present
240 in "start" and "stop" lines but was omitted from the "abandon" line.
242 PH/34 Omit spaces between a header name and the colon in the error message that
243 is given when verify = headers_syntax fails (if there are lots of them,
244 the message gets confusing).
246 PH/35 Change the default for dns_check_names_pattern to allow slashes within
247 names, as there are now some PTR records that contain slashes. This check
248 is only to protect against broken name servers that fall over on strange
249 characters, so the fact that it applies to all lookups doesn't matter.
251 PH/36 Now that the new test suite is complete, we can remove some of the
252 special code in Exim that was needed for the old test suite. For example,
253 sorting DNS records because real resolvers return them in an arbitrary
254 order. The new test suite's fake resolver always returns records in the
257 PH/37 When running in the test harness, use -odi for submitted messages (e.g.
258 bounces) except when queue_only is set, to avoid logging races between
259 the different processes.
261 PH/38 Panic-die if .include specifies a non-absolute path.
263 PH/39 A tweak to the "H" retry rule from its user.
265 JJ/03 exipick: Removed parentheses from 'next' and 'last' calls that specified
266 a label. They prevented compilation on older perls.
268 JJ/04 exipick: Refactored code to prevent implicit split to @_ which caused
269 a warning to be raised on newish perls.
271 JJ/05 exipick: Fixed bug where -bpc always showed a count of all messages
272 on queue. Changes to match documented behaviour of showing count of
273 messages matching specified criteria.
275 PH/40 Changed the default ident timeout from 30s to 5s.
277 PH/41 Added support for the use of login_cap features, on those BSD systems
278 that have them, for controlling the resources used by pipe deliveries.
280 PH/42 The content-scanning code uses fopen() to create files in which to put
281 message data. Previously it was not paying any attention to the mode of
282 the files. Exim runs with umask(0) because the rest of the code creates
283 files with open(), and sets the required mode explicitly. Thus, these
284 files were ending up world-writeable. This was not a big issue, because,
285 being within the spool directory, they were not world-accessible. I have
286 created a function called modefopen, which takes an additional mode
287 argument. It sets umask(777), creates the file, chmods it to the required
288 mode, then resets the umask. All the relevant calls to fopen() in the
289 content scanning code have been changed to use this function.
291 PH/43 If retry_interval_max is set greater than 24 hours, it is quietly reset
292 to 24 hours. This avoids potential overflow problems when processing G
293 and H retry rules. I suspect nobody ever tinkers with this value.
295 PH/44 Added STRIP_COMMAND=/usr/bin/strip to the FreeBSD Makefile.
297 PH/45 When the plaintext authenticator is running as a client, the server's
298 challenges are checked to ensure they are valid base64 strings. By
299 default, the authentication attempt is cancelled if an invalid string is
300 received. Setting client_ignore_invalid_base64 true ignores these errors.
301 The decoded challenge strings are now placed in $auth1, $auth2, etc. as
302 they are received. Thus, the responses can be made to depend on the
303 challenges. If an invalid string is ignored, an empty string is placed in
306 PH/46 Messages that are created by the autoreply transport now contains a
307 References: header, in accordance with RFCs 2822 and 3834.
309 PH/47 Added authenticated_sender_force to the smtp transport.
311 PH/48 The ${prvs expansion was broken on systems where time_t was long long.
313 PH/49 Installed latest patch from the Sieve maintainer.
315 PH/50 When an Exim quota was set without a file count quota, and mailbox_size
316 was also set, the appendfile transport was unnecessarily scanning a
317 directory of message files (e.g. for maildir delivery) to find the count
318 of files (along with the size), even though it did not need this
319 information. It now does the scan only if it needs to find either the
320 size of the count of files.
322 PH/51 Added ${time_eval: to convert Exim time strings into seconds.
324 PH/52 Two bugs concerned with error handling when the smtp transport is
327 (i) Exim was not creating retry information for temporary errors given
328 for individual recipients after the DATA command when the smtp transport
329 was used in LMTP mode. This meant that they could be retried too
330 frequently, and not timed out correctly.
332 (ii) Exim was setting the flag that allows error details to be returned
333 for LMTP errors on RCPT commands, but not for LMTP errors for individual
334 recipients that were returned after the DATA command.
336 PH/53 This is related to PH/52, but is more general: for any failing address,
337 when detailed error information was permitted to be returned to the
338 sender, but the error was temporary, then after the final timeout, only
339 "retry timeout exceeded" was returned. Now it returns the full error as
340 well as "retry timeout exceeded".
342 PH/54 Added control=allow_auth_unadvertised, as it seems there are clients that
343 do this, and (what is worse) MTAs that accept it.
345 PH/55 Added the add_header modified to ACLs. The use of "message" with "warn"
346 will now be deprecated.
348 PH/56 New os.c-cygwin from the Cygwin maintainer.
350 JJ/06 exipick: added --unsorted option to allow unsorted output in all output
351 formats (previously only available in exim formats via -bpr, -bpru,
352 and -bpra. Now also available in native and exiqgrep formats)
354 JJ/07 exipick: added --freeze and --thaw options to allow faster interaction
355 with very large, slow to parse queues
357 JJ/08 exipick: added ! as generic prefix to negate any criteria format
359 JJ/09 exipick: miscellaneous performance enhancements (~24% improvements)
361 PH/57 Tidies in SMTP dialogue display in debug output: (i) It was not showing
362 responses to authentication challenges, though it was showing the
363 challenges; (ii) I've removed the CR characters from the debug output for
366 PH/58 Allow for the insertion of a newline as well as a space when a string
367 is turned into more than one encoded-word during RFC 2047 encoding. The
368 Sieve code now uses this.
370 PH/59 Added the following errors that can be detected in retry rules: mail_4xx,
371 data_4xx, lost_connection, tls_required.
373 PH/60 When a VRFY deferred or FAILED, the log message rather than the user
374 message was being sent as an SMTP response.
376 PH/61 Add -l and -k options to exicyclog.
378 PH/62 When verifying, if an address was redirected to one new address, so that
379 verification continued, and the new address failed or deferred after
380 having set something in $address_data, the value of $address_data was not
381 passed back to the ACL. This was different to the case when no
382 redirection occurred. The value is now passed back in both cases.
384 PH/63 Changed the macro HAVE_LOGIN_CAP (see PH/41 for this release above) to
385 HAVE_SETCLASSRESOURCES because there are different APIs in use that all
386 use login_cap.h, so on its own it isn't the distinguishing feature. The
387 new name refers directly to the setclassresources() function.
389 PH/65 Added configuration files for NetBSD3.
391 PH/66 Updated OS/Makefile-HP-UX for gcc 4.1.0 with HP-UX 11.
393 PH/67 Fixed minor infelicity in the sorting of addresses to ensure that IPv6
394 is preferred over IPv4.
396 PH/68 The bounce_return_message and bounce_return_body options were not being
397 honoured for bounces generated during the reception of non-SMTP messages.
398 In particular, this applied to messages rejected by the ACL. This bug has
399 been fixed. However, if bounce_return_message is true and bounce_return_
400 body is false, the headers that are returned for a non-SMTP message
401 include only those that have been read before the error was detected.
402 (In the case of an ACL rejection, they have all been read.)
404 PH/69 The HTML version of the specification is now built in a directory called
405 spec_html instead of spec.html, because the latter looks like a path with
406 a MIME-type, and this confuses some software.
408 PH/70 Catch two compiler warnings in sieve.c.
410 PH/71 Fixed an obscure and subtle bug (thanks Alexander & Matthias). The
411 function verify_get_ident() calls ip_connect() to connect a socket, but
412 if the "connect()" function timed out, ip_connect() used to close the
413 socket. However, verify_get_ident() also closes the socket later, and in
414 between Exim writes to the log, which may get opened at this point. When
415 the socket was closed in ip_connect(), the log could get the same file
416 descriptor number as the socket. This naturally causes chaos. The fix is
417 not to close the socket in ip_connect(); the socket should be closed by
418 the function that creates it. There was only one place in the code where
419 this was missing, in the iplookup router, which I don't think anybody now
420 uses, but I've fixed it anyway.
422 PH/72 Make dns_again_means_nonexist apply to lookups using gethostbyname() as
423 well as to direct DNS lookups. Otherwise the handling of names in host
424 lists is inconsistent and therefore confusing.
430 PH/01 Two changes to the default runtime configuration:
432 (1) Move the checks for relay_from_hosts and authenticated clients from
433 after to before the (commented out) DNS black list checks.
435 (2) Add control=submission to the relay_from_hosts and authenticated
436 clients checks, on the grounds that messages accepted by these
437 statements are most likely to be submissions.
439 PH/02 Several tidies to the handling of ${prvs and ${prvscheck:
441 (1) Generate an error if the third argument for the ${prvs expansion is
444 (2) Treat a missing third argument of ${prvscheck as if it were an empty
447 (3) Reset the variables that are obtained from the first argument of
448 ${prvscheck and used in the second argument before leaving the code,
449 because their memory is reclaimed, so using them afterwards may do
452 (4) Tidy up the code for expanding the arguments of ${prvscheck one by
453 one (it's much easier than Tom thought :-).
455 (5) Because of (4), we can now allow for the use of $prvscheck_result
456 inside the third argument.
458 PH/03 For some reason, the default setting of PATH when running a command from
459 a pipe transport was just "/usr/bin". I have changed it to
462 PH/04 SUPPORT_TRANSLATE_IP_ADDRESS and MOVE_FROZEN_MESSAGES did not cause
463 anything to be listed in the output from -bV.
465 PH/05 When a filter generated an autoreply, the entire To: header line was
466 quoted in the delivery log line, like this:
468 => >A.N.Other <ano@some.domain> <original@ddress> ...
470 This has been changed so that it extracts the operative address. There
471 may be more than one such address. If so, they are comma-separated, like
474 => >ano@some.domain,ona@other.domain <original@ddress> ...
476 PH/06 When a client host used a correct literal IP address in a HELO or EHLO
477 command, (for example, EHLO [1.2.3.4]) and the client's IP address was
478 not being looked up in the rDNS to get a host name, Exim was showing the
479 IP address twice in Received: lines, even though the IP addresses were
480 identical. For example:
482 Received: from [1.2.3.4] (helo=[1.2.3.4])
484 However, if the real host name was known, it was omitting the HELO data
485 if it matched the actual IP address. This has been tidied up so that it
486 doesn't show the same IP address twice.
488 PH/07 When both +timestamp and +memory debugging was on, the value given by
489 $tod_xxx expansions could be wrong, because the tod_stamp() function was
490 called by the debug printing, thereby overwriting the timestamp buffer.
491 Debugging no longer uses the tod_stamp() function when +timestamp is set.
493 PH/08 When the original message was included in an autoreply transport, it
494 always said "this is a copy of the message, including all the headers",
495 even if body_only or headers_only was set. It now gives an appropriate
498 PH/09 Applied a patch from the Sieve maintainer which:
500 o fixes some comments
501 o adds the (disabled) notify extension core
502 o adds some debug output for the result of if/elsif tests
503 o points to the current vacation draft in the documentation
504 and documents the missing references header update
508 o fixes a bug in processing the envelope test (when testing
509 multiple envelope elements, the last element determinted the
512 PH/10 Exim was violating RFC 3834 ("Recommendations for Automatic Responses to
513 Electronic Mail") by including:
515 Auto-submitted: auto-generated
517 in the messages that it generates (bounce messages and others, such as
518 warnings). In the case of bounce messages for non-SMTP mesages, there was
519 also a typo: it was using "Auto_submitted" (underscore instead of
520 hyphen). Since every message generated by Exim is necessarily in response
521 to another message, thes have all been changed to:
523 Auto-Submitted: auto-replied
525 in accordance with these statements in the RFC:
527 The auto-replied keyword:
529 - SHOULD be used on messages sent in direct response to another
530 message by an automatic process,
532 - MUST NOT be used on manually-generated messages,
534 - MAY be used on Delivery Status Notifications (DSNs) and Message
535 Disposition Notifications (MDNs),
537 - MUST NOT be used on messages generated by automatic or periodic
538 processes, except for messages which are automatic responses to
541 PH/11 Added "${if def:sender_address {(envelope-from <$sender_address>)\n\t}}"
542 to the default Received: header definition.
544 PH/12 Added log selector acl_warn_skipped (default on).
546 PH/13 After a successful wildlsearch lookup, discard the values of numeric
547 variables because (a) they are in the wrong storage pool and (b) even if
548 they were copied, it wouldn't work properly because of the caching.
550 PH/14 Add check_rfc2047_length to disable enforcement of RFC 2047 length
551 checking when decoding. Apparently there are clients that generate
552 overlong encoded strings. Why am I not surprised?
554 PH/15 If the first argument of "${if match_address" was not empty, but did not
555 contain an "@" character, Exim crashed. Now it writes a panic log message
556 and treats the condition as false.
558 PH/16 In autoreply, treat an empty string for "once" the same as unset.
560 PH/17 A further patch from the Sieve maintainer: "Introduce the new Sieve
561 extension "envelope-auth". The code is finished and in agreement with
562 other implementations, but there is no documentation so far and in fact,
563 nobody wrote the draft yet. This extension is currently #undef'ed, thus
564 not changing the active code.
566 Print executed "if" and "elsif" statements when debugging is used. This
567 helps a great deal to understand what a filter does.
569 Document more things not specified clearly in RFC3028. I had all this
570 sorted out, when out of a sudden new issues came to my mind. Oops."
572 PH/18 Exim was not recognizing the "net-" search type prefix in match_ip lists
575 PH/19 Exim expands the IPv6 address given to -bh to its full non-abbreviated
576 canonical form (as documented). However, after a host name lookup from
577 the IP address, check_host() was doing a simple string comparison with
578 addresses acquired from the DNS when checking that the found name did
579 have the original IP as one of its addresses. Since any found IPv6
580 addresses are likely to be in abbreviated form, the comparison could
581 fail. Luckily, there already exists a function for doing the comparison
582 by converting both addresses to binary, so now that is used instead of
585 PH/20 There was another similar case to PH/19, when a complete host name was
586 given in a host list; looking up its IP address could give an abbreviated
587 form, whereas the current host's name might or might not be abbreviated.
588 The same fix has been applied.
594 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
595 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
598 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
599 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
601 PH/03 Typo: missing ".o" in src/pcre/Makefile.
603 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
604 header line, restrict the check to what is listed in RFCs 2369 and 2929.
605 Also, for "Auto-Submitted", treat anything other than "no" as
606 non-personal, in accordance with RFC 3834. (Previously it treated
607 anything starting "auto-" as non-personal.)
609 TF/01 The control=submission/name=... option had a problem with syntax
610 errors if the name included a slash character. The /name= option
611 now slurps the rest of the string, so it can include any characters
612 but it must come last in the list of options (after /sender_retain
615 PH/05 Some modifications to the interface to the fake nameserver for the new
623 TK/01 Added the "success_on_redirect" address verification option. See
624 NewStuff for rationale and an example.
626 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
628 PH/02 Patch to exigrep to allow it to work on syslog lines.
630 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
631 fread() to skip over the body file's header line, because in Cygwin the
632 header line is locked and is inaccessible.
634 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
635 co-exist for some time) to make it clear that it is the Exim ID that is
636 referenced, not the Message-ID: header line.
638 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
639 string_format() function, because snprintf() does not exist on all
642 PH/06 The use of forbid_filter_existstest now also locks out the use of the
643 ${stat: expansion item.
645 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
646 protocol synchronization error", to keep the pedants happy.
648 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
649 well as for IRIX systems, when gcc is being used. See the host.c source
652 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
654 PH/10 Named domain lists were not working if used in a queue_smtp_domains
657 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
658 transport and to the smtp transport in LMTP mode.
660 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
662 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
663 run a filter in a subprocess. This could lead to confusion in subsequent
664 lookups in the parent process. There should also be a search_tidyup() at
665 the end of the subprocess.
667 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
668 only if the host matched helo_try_verify_hosts, which caused the
669 verification to occur when the EHLO/HELO command was issued. The ACL just
670 tested the remembered result. Now, if a previous verification attempt has
671 not happened, "verify = helo" does it there and then.
673 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
675 TK/03 Fix log output including CR from clamd.
677 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
678 exist provoked a memory error which could cause a segfault.
680 PH/15 Installed PCRE 6.2
682 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
684 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
685 of the problem. Specifically, suggested +O2 rather than +O1 for the
688 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
690 PH/20 If a delivery was routed to a non-standard port by means of an SRV
691 record, the port was not correctly logged when the outgoing_port log
692 selector was set (it logged the transort's default port).
694 PH/21 Added support for host-specific ports to manualroute, queryprogram,
695 fallback_hosts, and "hosts" in the smtp transport.
697 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
698 host errors such as "Connection refused".
700 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
701 authentication with radiusclient 0.4.9:
703 - Error returned from rc_read_config was caught wrongly
704 - Username/password not passed on to radius server due to wrong length.
706 The presumption is that some radiusclient API changes for 4.51/PH/17
707 were not taken care of correctly. The code is still untested by me (my
708 Linux distribution still has 0.3.2 of radiusclient), but it was
709 contributed by a Radius user.
711 PH/24 When doing a callout, the value of $domain wasn't set correctly when
712 expanding the "port" option of the smtp transport.
714 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
715 while reading a MIME header. Thanks to Tom Hughes for a patch.
717 PH/24 Include config.h inside local_scan.h so that configuration settings are
720 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
721 that in an ACL for RCPT (for example), you can examine exactly what was
724 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
725 commands, but it was not correctly comparing the address with the actual
726 client host address. Thus, it would show the EHLO address in Received:
727 header lines when this was not necessary.
729 PH/27 Added the % operator to ${eval:}.
731 PH/28 Exim tries to create and chdir to its spool directory when it starts;
732 it should be ignoring failures (because with -C, for example, it has lost
733 privilege). It wasn't ignoring creation failures other than "already
736 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
739 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
740 because an input line was too long, either on its own, or by virtue of
741 too many continuations, the temporary file was not being removed, and the
742 return code was incorrect.
744 PH/31 Missing "BOOL" in function definition in filtertest.c.
746 PH/32 Applied Sieve patches from the maintainer.
748 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
750 PH/33 Added "verify = not_blind".
752 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
753 Local/Makefile (with some defaults set). These are used in built scripts
754 such as exicyclog, but they have never been used in the exim_install
755 script (though there are many overriding facilities there). I have
756 arranged that the exim_install script now takes note of these two
759 PH/35 Installed configuration files for Dragonfly.
761 PH/36 When a locally submitted message by a trusted user did not contain a
762 From: header, and the sender address was obtained from -f or from an SMTP
763 MAIL command, and the trusted user did not use -F to supply a sender
764 name, $originator_name was incorrectly used when constructing a From:
765 header. Furthermore, $originator_name was used for submission mode
766 messages from external hosts without From: headers in a similar way,
767 which is clearly wrong.
769 PH/37 Added control=suppress_local_fixups.
771 PH/38 When log_selector = +received_sender was set, and the addition of the
772 sender made the log line's construction buffer exactly full, or one byte
773 less than full, an overflow happened when the terminating "\n" was
776 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
777 when the result of a list match is failure because a DNS lookup failed.
779 PH/40 RM_COMMAND is now used in the building process.
781 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
782 the "build-* directories that it finds.
784 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
785 address in a domain literal was a prefix of an interface address.
787 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
788 when verifying a sender address, unless rewrite_headers is false.
790 PH/44 Wrote a long comment about why errors_to addresses are verified as
791 recipients, not senders.
793 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
794 the ratelimit ACL was added.
796 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
798 PH/46 Added extra information about PostgreSQL errors to the error string.
800 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
801 avoiding the need to install special zones in a real server. This is
802 backwards compatible; if it can't find the fake resolver, it drops back.
803 Thus, both old and new test suites can be run.
805 TF/02 Added util/ratelimit.pl
807 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
808 clock is set back in time.
810 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
811 Candler <B.Candler@pobox.com>.
813 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
814 OK for addresses that are the result of redirections.
816 PH/48 A number of further additions for the benefit of the new test suite,
817 including a fake gethostbyname() that interfaces to the fake DNS resolver
820 TF/06 The fix for widen_domains has also been applied to qualify_single and
821 search_parents which are the other dnslookup options that can cause
824 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
827 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
829 TF/07 Exim produced the error message "an SRV record indicated no SMTP
830 service" if it encountered an MX record with an empty target hostname.
831 The message is now "an MX or SRV record indicated no SMTP service".
833 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
834 if the DNS of the sending site is misconfigured. This is quite a
835 common situation. This change restores the behaviour of treating a
836 helo verification defer as a failure.
838 PH/51 If self=fail was set on a router, the bounce message did not include the
839 actual error message.
845 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
847 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
848 command itself ended in error, the underlying message about the transport
849 filter timeout was being overwritten with the pipe command error. Now the
850 underlying error message should be appended to the second error message.
852 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
854 PH/02 Reduce the amount of output that "make" produces by default. Full output
855 can still be requested.
857 PH/03 The warning log line about a condition test deferring for a "warn" verb
858 was being output only once per connection, rather than after each
859 occurrence (because it was using the same function as for successful
860 "warn" verbs). This seems wrong, so I have changed it.
862 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
863 it should not have, which might have caused a crash in the right
864 circumstances, but probably never did.
866 PH/04 Installed a modified version of Tony Finch's patch to make submission
867 mode fix the return path as well as the Sender: header line, and to
868 add a /name= option so that you can make the user's friendly name appear
871 TF/03 Added the control = fakedefer ACL modifier.
873 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
874 Mark Lowes for thorough testing.
876 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
878 TK/03 Merged latest SRS patch from Miles Wilton.
880 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
881 with the definition in sysexits.h (which is #included earlier).
882 Fortunately, Exim does not actually use EX_OK. The code used to try to
883 preserve the sysexits.h value, by assumimg that macro definitions were
884 scanned for macro replacements. I have been disabused of this notion,
885 so now the code just undefines EX_OK before #including unistd.h.
887 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
888 in the smtp transport. When a block could not be written in a single
889 write() function, the timeout was being re-applied to each part-write.
890 This seems wrong - if the receiver was accepting one byte at a time it
891 would take for ever. The timeout is now adjusted when this happens. It
892 doesn't have to be particularly precise.
894 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
895 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
897 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
898 without a domain if the check to <postmaster@domain> fails.
900 SC/01 Eximstats: added -xls and the ability to specify output files
901 (patch written by Frank Heydlauf).
903 SC/02 Eximstats: use FileHandles for outputing results.
905 SC/03 Eximstats: allow any combination of xls, txt, and html output.
907 SC/04 Eximstats: fixed display of large numbers with -nvr option
909 SC/05 Eximstats: fixed merging of reports with empty tables.
911 SC/06 Eximstats: added the -include_original_destination flag
913 SC/07 Eximstats: removed tabs and trailing whitespace.
915 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
917 TK/06 MBOX spool code: Add real "From " MBOX separator line
918 so the .eml file is really in mbox format (even though
919 most programs do not really care). Patch from Alex Miller.
921 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
922 The latter is generated from $received_to and is only set if the
923 message has one envelope recipient. SA can use these headers,
924 obviously out-of-the-box. Patch from Alex Miller.
926 PH/08 The ${def test on a variable was returning false if the variable's
927 value was "0", contrary to what the specification has always said!
928 The result should be true unless the variable is empty.
930 PH/09 The syntax error of a character other than { following "${if
931 def:variable_name" (after optional whitespace) was not being diagnosed.
932 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
933 accidental colon was present, for example, could give incorrect results.
935 PH/10 Tidied the code in a number of places where the st_size field of a stat()
936 result is used (not including appendfile, where other changes are about
939 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
940 This involved changing a lot of size variables from int to off_t. It
941 should work with maildirs and everything.
943 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
944 spamd dying while we are connected to it.
946 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
949 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
950 function for BDB 4.3.
952 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
954 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
955 the sources for PCRE have been re-arranged and more files are now
958 PH/15 The code I had for printing potentially long long variables in PH/11
959 above was not the best (it lost precision). The length of off_t variables
960 is now inspected at build time, and an appropriate printing format (%ld
961 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
962 to be "long long int" or "long int". This is needed for the internal
963 formatting function string_vformat().
965 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
966 the configuration file to be ":syslog", then the script "guesses" where
967 the logs files are, rather than using the compiled in default. In our
968 case the guess is not the same as the compiled default, so the script
969 suddenly stopped working when I started to use syslog. The patch checks
970 to see if log_file_path is "". If so, it attempts to read it from exim
971 with no configuration file to get the compiled in version, before it
972 falls back to the previous guessing code."
974 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
975 implementing BATV in an Exim configuration. See NewStuff for the gory
978 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
979 Makefile that are specific to HP-UX.
981 PH/18 If the "use_postmaster" option was set for a recipient callout together
982 with the "random" option, the postmaster address was used as the MAIL
983 FROM address for the random test, but not for the subsequent recipient
984 test. It is now used for both.
986 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
987 patch removes a few documentation additions to RFC 3028, because the
988 latest draft now contains them. It adds the new en;ascii-case comparator
989 and a new error check for 8bit text in MIME parts. Comparator and
990 require names are now matched exactly. I enabled the subaddress
991 extension, but it is not well tested yet (read: it works for me)."
993 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
994 rework some of the code of TK/09 above to avoid the hardwired use of
995 "%lld" and "long long". Replaced the call to snprintf() with a call to
998 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
999 records point to non-existent hosts", "retry timeout exceeded", and
1000 "retry time not reached for any host after a long failure period".
1002 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
1003 experimental DomainKeys support:
1005 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
1006 (2) On an error such as an illegally used "control", the wrong name for
1007 the control was given.
1009 These problems did NOT occur unless DomainKeys support was compiled.
1011 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
1013 PH/24 Added ${if match_ip condition.
1015 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
1016 databases so that it will be absolutely obvious if a crash occurs in the
1017 DB library. This is a regular occurrence (often caused by mis-matched
1020 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
1021 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
1022 picked up on a user's system that detects such things. There doesn't seem
1023 to be a gcc warning option for this - only an attribute that has to be
1024 put on the function's prototype. It seems that in Fedora Core 4 they have
1025 set this on a number of new functions. No doubt there will be more in due
1028 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
1029 specify a transport. However, if an address that was verified by such a
1030 router was the subject of a callout, Exim crashed because it tried to
1031 read the rcpt_include_affixes from the non-existent transport. Now it
1032 just assumes that the setting of that option is false. This bug was
1033 introduced by 4.51/PH/31.
1035 PH/28 Changed -d+all to exclude +memory, because that information is very
1036 rarely of interest, but it makes the output a lot bigger. People tend to
1037 do -d+all out of habit.
1039 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
1040 code in os-type was giving problems when libc.so lives in lib64, like on
1043 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
1044 aren't the modern standard, and it seems that some systems' include files
1045 don't always have them. Exim was already checking for some of the newer
1046 ones like T_AAAA, and defining it itself. I've added checks for all the
1047 record types that Exim uses.
1049 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
1050 not automatically generating a new one, as it is supposed to. This
1051 prevented TLS from working. If the file did exist, but contained invalid
1052 data, a new version was generated, as expected. It was only the case of a
1053 non-existent file that was broken.
1055 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
1056 with a change in libdomainkeys > 0.64.
1058 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
1059 from DNS. If the selector record carries the flag, it now has
1060 precedence over the domain-wide flag.
1062 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
1064 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
1065 the use of a port name that isn't defined in /etc/services) occurred, the
1066 message was deferred as in a normal delivery, and thus remained on the
1067 spool, instead of being failed because of the mua_wrapper setting. This
1068 is now fixed, and I tidied up some of the mua_wrapper messages at the
1071 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
1072 the messages in a hash of arrays rather than using individual hashes.
1073 This is a bit cleaner and results in dramatic memory savings, albeit
1074 at a slight CPU cost.
1076 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
1077 as requested by Marc Sherman.
1079 SC/10 Eximstats: added histograms for user specified patterns as requested
1082 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
1084 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
1085 fopen() in the content-scanning modules that did not already have it.
1091 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
1092 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
1094 TK/02 Fix ACL "control" statement not being available in MIME ACL.
1096 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
1098 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
1099 to test Sieve filters that use "vacation".
1101 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
1102 that changes the way the GnuTLS parameters are stored in the cache file.
1103 The new format can be generated externally. For backward compatibility,
1104 if the data in the cache doesn't make sense, Exim assumes it has read an
1105 old-format file, and it generates new data and writes a new file. This
1106 means that you can't go back to an older release without removing the
1109 PH/03 A redirect router that has both "unseen" and "one_time" set does not
1110 work if there are any delivery delays because "one_time" forces the
1111 parent to be marked "delivered", so its unseen clone is never tried
1112 again. For this reason, Exim now forbids the simultaneous setting of
1115 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
1116 redirected to themselves ("homonym" addresses). Read the long ChangeLog
1117 entry if you want to know the details. The fix, however, neglected to
1118 consider the case when local delivery batching is involved. The test for
1119 "previously delivered" was not happening when checking to see if an
1120 address could be batched with a previous (undelivered) one; under
1121 certain circumstances this could lead to multiple deliveries to the same
1124 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
1125 in its include files, and this causes problems building Exim.
1127 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
1128 header_syntax) but Exim was just ignoring anything given after a slash.
1129 In particular, this caused confusion with an attempt to use "verify =
1130 reverse_host_lookup/defer_ok". An error is now given when options are
1131 supplied for verify items that do not have them. (Maybe reverse_host_
1132 lookup should have a defer_ok option, but that's a different point.)
1134 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
1135 defined by RFC 821) to 2048, because there were problems with some AUTH
1136 commands, and RFC 1869 says the size should be increased for extended
1137 SMTP commands that take arguments.
1139 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
1142 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
1143 "unknown" error; now it says that the functionality isn't in the binary.
1145 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
1146 an address' error message when a string expansion fails (syntax or
1147 whatever). Otherwise the password may appear in the log. Following change
1148 PH/42 below, there is no longer a chance of it appearing in a bounce
1151 PH/11 Installed exipick version 20050225.0 from John Jetmore.
1153 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
1154 of its addresses was ever tried. (Bugzilla bug #2.)
1156 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
1157 the result incorrectly in the debug output. (It correctly added a newline
1158 to what was transported.)
1160 TF/01 Added $received_time.
1162 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
1163 commented out examples of how to interface to a virus scanner and to
1164 SpamAssassin. Also added commented examples of av_scanner and
1165 spamd_address settings.
1167 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
1168 and controls are allowed in which ACLs. There were a couple of minor
1169 errors. Some of the entries in the conditions table (which is a table of
1170 where they are NOT allowed) were getting very unwieldy; rewrote them as a
1171 negation of where the condition IS allowed.
1173 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
1175 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
1176 header file does not have a version number, so I've had to invent a new
1177 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
1178 API. The code is untested by me (my Linux distribution still has 0.3.2 of
1179 radiusclient), but it was contributed by a Radius user.
1181 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
1182 files or directories, for OpenSSL.
1184 PH/19 When an Exim process that is running as root has to create an Exim log
1185 file, it does so in a subprocess that runs as exim:exim so as to get the
1186 ownership right at creation (otherwise, other Exim processes might see
1187 the file with the wrong ownership). There was no test for failure of this
1188 fork() call, which would lead to the process getting stuck as it waited
1189 for a non-existent subprocess. Forks do occasionally fail when resources
1190 run out. I reviewed all the other calls to fork(); they all seem to check
1193 PH/20 When checking for unexpected SMTP input at connect time (before writing
1194 the banner), Exim was not dealing correctly with a non-positive return
1195 from the read() function. If the client had disconnected by this time,
1196 the result was a log entry for a synchronization error with an empty
1197 string after "input=" when read() returned zero. If read() returned -1
1198 (an event I could not check), uninitialized data bytes were printed.
1199 There were reports of junk text (parts of files, etc) appearing after
1202 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
1204 PH/22 Added support for macro redefinition, and (re)definition in between
1205 driver and ACL definitions.
1207 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
1208 forgetting to use the resulting value; it was using the unexpanded value.
1210 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
1211 hadn't been configured. The fix is from Juergen Kreileder, who
1212 understands it better than I do:
1214 "Here's what I see happening with three configured cyrus_sasl
1215 authenticators configured (plain, login, cram-md5):
1217 On startup auth_cyrus_sasl_init() gets called for each of these.
1218 This means three calls to sasl_listmech() without a specified mech_list.
1219 => SASL tests which mechs of all available mechs actually work
1220 => three warnings about OTP not working
1221 => the returned list contains: plain, login, cram-md5, digest-md5, ...
1223 With the patch, sasl_listmech() also gets called three times. But now
1224 SASL's mech_list option is set to the server_mech specified in the the
1225 authenticator. Or in other words, the answer from sasl_listmech()
1226 gets limited to just the mech you're testing for (which is different
1228 => the return list contains just 'plain' or 'login', 'cram-md5' or
1229 nothing depending on the value of ob->server_mech.
1231 I've just tested the patch: Authentication still works fine,
1232 unavailable mechs specified in the exim configuration are still
1233 caught, and the auth.log warnings about OTP are gone."
1235 PH/25 When debugging is enabled, the contents of the command line are added
1236 to the debugging output, even when log_selector=+arguments is not
1239 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
1240 answer is "GNU", and only if the return is "GNU/something" is the answer
1243 PH/27 $acl_verify_message is now set immediately after the failure of a
1244 verification in an ACL, and so is available in subsequent modifiers. In
1245 particular, the message can be preserved by coding like this:
1247 warn !verify = sender
1248 set acl_m0 = $acl_verify_message
1250 Previously, $acl_verify_message was set only while expanding "message"
1251 and "log_message" when a very denied access.
1253 PH/28 Modified OS/os.c-Linux with
1255 -#ifndef OS_LOAD_AVERAGE
1256 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
1258 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
1259 nomenclature these days.)
1261 PH/29 Installed patch from the Sieve maintainer that adds the options
1262 sieve_useraddress and sieve_subaddress to the redirect router.
1264 PH/30 In these circumstances:
1265 . Two addresses routed to the same list of hosts;
1266 . First host does not offer TLS;
1267 . First host accepts first address;
1268 . First host gives temporary error to second address;
1269 . Second host offers TLS and a TLS session is established;
1270 . Second host accepts second address.
1271 Exim incorrectly logged both deliveries with the TLS parameters (cipher
1272 and peerdn, if requested) that were in fact used only for the second
1275 PH/31 When doing a callout as part of verifying an address, Exim was not paying
1276 attention to any local part prefix or suffix that was matched by the
1277 router that accepted the address. It now behaves in the same way as it
1278 does for delivery: the affixes are removed from the local part unless
1279 rcpt_include_affixes is set on the transport.
1281 PH/32 Add the sender address, as F=<...>, to the log line when logging a
1282 timeout during the DATA phase of an incoming message.
1284 PH/33 Sieve envelope tests were broken for match types other than :is. I have
1285 applied a patch sanctioned by the Sieve maintainer.
1287 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
1288 the uid or gid is negative. A case of a negative gid caused this to be
1289 noticed. The fix allows for either to be negative.
1291 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
1292 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
1295 PH/36 The change PH/12 above was broken. Fixed it.
1297 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
1298 the grounds that routing the same address twice would always produce the
1299 same answer. This might have been true once, but it is certainly no
1300 longer true now. Routing a child address may depend on the previous
1301 routing that produced that child. Some complicated redirection strategies
1302 went wrong when messages had multiple recipients, and made Exim's
1303 behaviour dependent on the order in which the addresses were given.
1305 I have moved the duplicate checking until after the routing is complete.
1306 Exim scans the addresses that are assigned to local and remote
1307 transports, and removes any duplicates. This means that more work will be
1308 done, as duplicates will always all be routed, but duplicates are
1309 presumably rare, so I don't expect this is of any significance.
1311 For deliveries to pipes, files, and autoreplies, the duplicate checking
1312 still happens during the routing process, since they are not going to be
1315 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
1316 It corrects a timeout issue with spamd. This is Ian's comment: "The
1317 background is that sometimes spamd either never reads data from a
1318 connection it has accepted, or it never writes response data. The exiscan
1319 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
1320 blindly assumes that writes won't block so it may never time out."
1322 PH/39 Allow G after quota size as well as K and M.
1324 PH/40 The value set for $authenticated_id in an authenticator may not contain
1325 binary zeroes or newlines because the value is written to log lines and
1326 to spool files. There was no check on this. Now the value is run through
1327 the string_printing() function so that such characters are converted to
1328 printable escape sequences.
1330 PH/41 $message_linecount is a new variable that contains the total number of
1331 lines in the message. Compare $body_linecount, which is the count for the
1334 PH/42 Exim no longer gives details of delivery errors for specific addresses in
1335 bounce and delay warning messages, except in certain special cases, which
1338 (a) An SMTP error message from a remote host;
1339 (b) A message specified in a :fail: redirection;
1340 (c) A message specified in a "fail" command in a system filter;
1341 (d) A message specified in a FAIL return from the queryprogram router;
1342 (e) A message specified by the cannot_route_message router option.
1344 In these cases only, Exim does include the error details in bounce and
1345 warning messages. There are also a few cases where bland messages such
1346 as "unrouteable address" or "local delivery error" are given.
1348 PH/43 $value is now also set for the "else" part of a ${run expansion.
1350 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
1351 being worked on, but at least Exim now implements the latest version to
1354 PH/45 In a pipe transport, although a timeout while waiting for the pipe
1355 process to complete was treated as a delivery failure, a timeout while
1356 writing the message to the pipe was logged, but erroneously treated as a
1357 successful delivery. Such timeouts include transport filter timeouts. For
1358 consistency with the overall process timeout, these timeouts are now
1359 treated as errors, giving rise to delivery failures by default. However,
1360 there is now a new Boolean option for the pipe transport called
1361 timeout_defer, which, if set TRUE, converts the failures into defers for
1362 both kinds of timeout. A transport filter timeout is now identified in
1365 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
1366 systems where "make" and "gmake" are different, calling "gmake" at top
1367 level broke things. I've arranged for the value of $(MAKE) to be passed
1368 from the Makefile to this script so that it can call the same version of
1372 A note about Exim versions 4.44 and 4.50
1373 ----------------------------------------
1375 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
1376 changes of various kinds. As a consequence, a big documentation update was
1377 needed. This delayed the release for rather longer than seemed good, especially
1378 in the light of a couple of (minor) security issues. Therefore, the changes
1379 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
1380 release. So 4.44 and 4.50 are in effect two different branches that both start
1383 I have left the 4.50 change log unchanged; it contains all the changes since
1384 4.43. The change log for 4.44 is below; many of its items are identical to
1385 those for 4.50. This seems to be the most sensible way to preserve the
1386 historical information.
1392 1. Minor wording change to the doc/README.SIEVE file.
1394 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
1395 computation of the current number of files was incorrect.
1397 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1398 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1401 4. Give more explanation in the error message when the command for a transport
1402 filter fails to execute.
1404 5. There are several places where Exim runs a non-Exim command in a
1405 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1406 was being done only for the command run by the queryprogram router. It is
1407 now done for all such subprocesses. The other cases are: ${run, transport
1408 filters, and the commands run by the lmtp and pipe transports.
1410 6. Added CONFIGURE_GROUP build-time option.
1412 7. Some older OS have a limit of 256 on the maximum number of file
1413 descriptors. Exim was using setrlimit() to set 1000 as a large value
1414 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1415 systems. I've change it so that if it can't get 1000, it tries for 256.
1417 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1418 was an oversight, and furthermore, ever since the addition of extra
1419 controls (e.g. 4.43/32), the checks on when to allow different forms of
1420 "control" were broken. There should now be diagnostics for all cases when a
1421 control that does not make sense is encountered.
1423 9. Added the /retain_sender option to "control=submission".
1425 10. $recipients is now available in the predata ACL (oversight).
1427 11. Tidy the search cache before the fork to do a delivery from a message
1428 received from the command line. Otherwise the child will trigger a lookup
1429 failure and thereby defer the delivery if it tries to use (for example) a
1430 cached ldap connection that the parent has called unbind on.
1432 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1433 of $address_data from the recipient verification was clobbered by the
1434 sender verification.
1436 13. The value of address_data from a sender verification is now available in
1437 $sender_address_data in subsequent conditions in the ACL statement.
1439 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
1441 15. Added a new option "connect=<time>" to callout options, to set a different
1444 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1445 was its contents. (It was OK if the option was not defined at all.)
1447 17. A "Completed" log line is now written for messages that are removed from
1448 the spool by the -Mrm option.
1450 18. New variables $sender_verify_failure and $recipient_verify_failure contain
1451 information about exactly what failed.
1453 19. Added -dd to debug only the daemon process.
1455 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
1456 handles timeouts, both on the server side and network timeouts. Renamed the
1457 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
1459 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
1460 It is now set to "smtps".
1462 22. $host_address is now set to the target address during the checking of
1463 ignore_target_hosts.
1465 23. When checking ignore_target_hosts for an ipliteral router, no host name was
1466 being passed; this would have caused $sender_host_name to have been used if
1467 matching the list had actually called for a host name (not very likely,
1468 since this list is usually IP addresses). A host name is now passed as
1471 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1472 code that specifies a non-restarting handler (typically sigaction() in
1473 modern systems) in an attempt to fix a rare and obscure crash bug.
1475 25. Narrowed the window for a race in the daemon that could cause it to ignore
1476 SIGCHLD signals. This is not a major problem, because they are used only to
1477 wake it up if nothing else does.
1479 26. A malformed maildirsize file could cause Exim to calculate negative values
1480 for the mailbox size or file count. Odd effects could occur as a result.
1481 The maildirsize information is now recalculated if the size or filecount
1484 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1485 support for a long time. Removed HAVE_SYS_VFS_H.
1487 28. Installed the latest version of exipick from John Jetmore.
1489 29. In an address list, if the pattern was not a regular expression, an empty
1490 subject address (from a bounce message) matched only if the pattern was an
1491 empty string. Non-empty patterns were not even tested. This was the wrong
1492 because it is perfectly reasonable to use an empty address as part of a
1493 database query. An empty address is now tested by patterns that are
1494 lookups. However, all the other forms of pattern expect the subject to
1495 contain a local part and a domain, and therefore, for them, an empty
1496 address still always fails if the pattern is not itself empty.
1498 30. Exim went into a mad DNS loop when attempting to do a callout where the
1499 host was specified on an smtp transport, and looking it up yielded more
1500 than one IP address.
1502 31. Re-factored the code for checking spool and log partition space into a
1503 function that finds that data and another that does the check. The former
1504 is then used to implement four new variables: $spool_space, $log_space,
1505 $spool_inodes, and $log_inodes.
1507 32. The RFC2047 encoding function was originally intended for short strings
1508 such as real names; it was not keeping to the 75-character limit for
1509 encoded words that the RFC imposes. It now respects the limit, and
1510 generates multiple encoded words if necessary. To be on the safe side, I
1511 have increased the buffer size for the ${rfc2047: expansion operator from
1514 33. It is now permitted to omit both strings after an "if" condition; if the
1515 condition is true, the result is "true". As before, when the second string
1516 is omitted, a false condition yields an empty string. This makes it less
1517 cumbersome to write custom ACL and router conditions.
1519 34. Failure to deliver a bounce message always caused it to be frozen, even if
1520 there was an errors_to setting on the router. The errors_to setting is now
1523 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1524 canonical form (fully expanded) before being placed in
1525 $sender_host_address.
1527 36. The table in the code that translates DNS record types into text (T_A to
1528 "A" for instance) was missing entries for NS and CNAME. It is just possible
1529 that this could have caused confusion if both these types were looked up
1530 for the same domain, because the text type is used as part of Exim's
1531 per-process caching. But the chance of anyone hitting this buglet seems
1534 37. The dnsdb lookup has been extended in a number of ways.
1536 (1) There is a new type, "zns", which walks up the domain tree until it
1537 finds some nameserver records. It should be used with care.
1539 (2) There is a new type, "mxh", which is like "mx" except that it returns
1540 just the host names, not the priorities.
1542 (3) It is now possible to give a list of domains (or IP addresses) to be
1543 looked up. The behaviour when one of the lookups defers can be
1544 controlled by a keyword.
1546 (4) It is now possible to specify the separator character for use when
1547 multiple records are returned.
1549 38. The dnslists ACL condition has been extended: it is now possible to supply
1550 a list of IP addresses and/or domains to be looked up in a particular DNS
1553 39. Added log_selector=+queue_time_overall.
1555 40. When running the queue in the test harness, wait just a tad after forking a
1556 delivery process, to get repeatability of debugging output.
1558 41. Include certificate and key file names in error message when GnuTLS fails
1559 to set them up, because the GnuTLS error message doesn't include the name
1560 of the failing file when there is a problem reading it.
1562 42. Allow both -bf and -bF in the same test run.
1564 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
1566 44. The "Exiscan patch" is now merged into the mainline Exim source.
1568 45. Sometimes the final signoff response after QUIT could fail to get
1569 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1570 < 0 before doing a fflush(). This bug looks as though it goes back to the
1571 introduction of TLS in release 3.20, but "sometimes" must have been rare
1572 because the tests only now provoked it.
1574 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
1575 (this can affect the format of dates).
1577 47. exim_tidydb, when checking for the continued existence of a message for
1578 which it has found a message-specific retry record, was not finding
1579 messages that were in split spool directories. Consequently, it was
1580 deleting retry records that should have stayed in existence.
1582 48. Steve fixed some bugs in eximstats.
1584 49. The SPA authentication driver was not abandoning authentication and moving
1585 on to the next authenticator when an expansion was forced to fail,
1586 contradicting the general specification for all authenticators. Instead it
1587 was generating a temporary error. It now behaves as specified.
1589 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
1590 (the order specifies the preference for clients). The order is now AES256,
1591 AES128, 3DES, ARCFOUR128.
1593 51. Small patch to Sieve code - explicitly set From: when generating an
1596 52. Exim crashed if a remote delivery caused a very long error message to be
1597 recorded - for instance if somebody sent an entire SpamAssassin report back
1598 as a large number of 550 error lines. This bug was coincidentally fixed by
1599 increasing the size of one of Exim's internal buffers (big_buffer) that
1600 happened as part of the Exiscan merge. However, to be on the safe side, I
1601 have made the code more robust (and fixed the comments that describe what
1604 53. Now that there can be additional text after "Completed" in log lines (if
1605 the queue_time_overall log selector is set), a one-byte patch to exigrep
1606 was needed to allow it to recognize "Completed" as not the last thing in
1609 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
1610 patch that reportedly fixes this has been added. I am not expert enough to
1611 create a test for it. This is what the patch creator wrote:
1613 "I found a little strange behaviour of ldap code when working with
1614 Windows 2003 AD Domain, where users was placed in more than one
1615 Organization Units. When I tried to give exim partial DN, the exit code
1616 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
1617 But simultaneously result of request was absolutely normal ldap result,
1618 so I produce this patch..."
1620 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
1621 REFERENCE, so I have modified the code to exclude the patch when that macro
1624 55. Some experimental protocols are using DNS PTR records for new purposes. The
1625 keys for these records are domain names, not reversed IP addresses. The
1626 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1627 leaves it alone. Component reversal etc. now happens only for IP addresses.
1630 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
1632 57. Double the size of the debug message buffer (to 2048) so that more of very
1633 long debug lines gets shown.
1635 58. The exicyclog utility now does better if the number of log files to keep
1636 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
1638 59. Two changes related to the smtp_active_hostname option:
1640 (1) $smtp_active_hostname is now available as a variable.
1641 (2) The default for smtp_banner uses $smtp_active_hostname instead
1642 of $primary_hostname.
1644 60. The host_aton() function is supposed to be passed a string that is known
1645 to be a valid IP address. However, in the case of IPv6 addresses, it was
1646 not checking this. This is a hostage to fortune. Exim now panics and dies
1647 if the condition is not met. A case was found where this could be provoked
1648 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1649 components; fortuitously, this particular loophole had already been fixed
1650 by change 4.50/55 above.
1652 If there are any other similar loopholes, the new check in host_aton()
1653 itself should stop them being exploited. The report I received stated that
1654 data on the command line could provoke the exploit when Exim was running as
1655 exim, but did not say which command line option was involved. All I could
1656 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1657 running as the user.
1660 61. There was a buffer overflow vulnerability in the SPA authentication code
1661 (which came originally from the Samba project). I have added a test to the
1662 spa_base64_to_bits() function which I hope fixes it.
1665 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
1666 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
1668 63. The daemon start-up calls getloadavg() while still root for those OS that
1669 need the first call to be done as root, but it missed one case: when
1670 deliver_queue_load_max is set with deliver_drop_privilege. This is
1671 necessary for the benefit of the queue runner, because there is no re-exec
1672 when deliver_drop_privilege is set.
1674 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
1675 This has been fixed.
1677 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1678 was in use, was not putting the data itself into the right store pool;
1679 consequently, it could be overwritten for a subsequent message in the same
1680 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1683 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
1685 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
1686 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
1687 the function were treating the return as a boolean value, which happened to
1688 work because 0=false and not-0=true, but is not correct code.
1690 68. The host_aton() function was not handling scoped IPv6 addresses (those
1691 with, for example, "%eth0" on the end) correctly.
1693 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
1694 negated items (that is, ~something) in unsigned ints. Some compilers
1695 apparently mutter when there is no cast.
1697 70. If an address verification called from an ACL failed, and did not produce a
1698 user-specific message (i.e. there was only a "system" message), nothing was
1699 put in $acl_verify_message. In this situation, it now puts the system
1702 71. Change 4.23/11 added synchronization checking at the start of an SMTP
1703 session; change 4.31/43 added the unwanted input to the log line - except
1704 that it did not do this in the start of session case. It now does.
1706 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
1707 This is wrong and can cause the other end to generate a synchronization
1708 error if it is another Exim or anything else that does the synchronization
1709 check. A QUIT command is no longer sent after a timeout.
1711 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
1712 during host lookups.
1714 74. The defer_ok option of callout verification was not working if it was used
1715 when verifying addresses in header lines, that is, for this case:
1717 verify = header_sender/callout=defer_ok
1719 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
1720 those file descriptors could be used for SMTP connections. If anything
1721 wrote to stderr (the example that came up was "warn" in embedded Perl), it
1722 could be sent to the SMTP client, causing chaos. The daemon now opens
1723 stdin, stdout, and stderr to /dev/null when it puts itself into the
1726 76. Arrange for output from Perl's "warn" command to be written to Exim's main
1727 log by default. The user can override this with suitable Perl magic.
1729 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
1730 the log message when discard triggers, was not working for the DATA ACL or
1731 for the non-SMTP ACL.
1733 78. Error message wording change in sieve.c.
1735 79. If smtp_accept_max_per_host was set, the number of connections could be
1736 restricted to fewer than expected, because the daemon was trying to set up
1737 a new connection before checking whether the processes handling previous
1738 connections had finished. The check for completed processes is now done
1739 earlier. On busy systems, this bug wouldn't be noticed because something
1740 else would have woken the daemon, and it would have reaped the completed
1743 80. If a message was submitted locally by a user whose login name contained one
1744 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
1745 It caused a spool format error. I have fixed the spool reading code. A
1746 related problem was that the "from" clause in the Received: line became
1747 illegal because of the space(s). It is now covered by ${quote_local_part.
1749 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
1752 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
1754 83. Patch from Sieve maintainer to fix unterminated string problem in
1755 "vacation" handling.
1757 84. Some minor changes to the Linux configuration files to help with other
1758 OS variants using glibc.
1760 85. One more patch for Sieve to update vacation handling to latest spec.
1763 ----------------------------------------------------
1764 See the note above about the 4.44 and 4.50 releases.
1765 ----------------------------------------------------
1771 1. Change 4.43/35 introduced a bug that caused file counts to be
1772 incorrectly computed when quota_filecount was set in an appendfile
1775 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1776 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1779 3. Give more explanation in the error message when the command for a transport
1780 filter fails to execute.
1782 4. There are several places where Exim runs a non-Exim command in a
1783 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1784 was being done only for the command run by the queryprogram router. It is
1785 now done for all such subprocesses. The other cases are: ${run, transport
1786 filters, and the commands run by the lmtp and pipe transports.
1788 5. Some older OS have a limit of 256 on the maximum number of file
1789 descriptors. Exim was using setrlimit() to set 1000 as a large value
1790 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1791 systems. I've change it so that if it can't get 1000, it tries for 256.
1793 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1794 was an oversight, and furthermore, ever since the addition of extra
1795 controls (e.g. 4.43/32), the checks on when to allow different forms of
1796 "control" were broken. There should now be diagnostics for all cases when a
1797 control that does not make sense is encountered.
1799 7. $recipients is now available in the predata ACL (oversight).
1801 8. Tidy the search cache before the fork to do a delivery from a message
1802 received from the command line. Otherwise the child will trigger a lookup
1803 failure and thereby defer the delivery if it tries to use (for example) a
1804 cached ldap connection that the parent has called unbind on.
1806 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1807 of $address_data from the recipient verification was clobbered by the
1808 sender verification.
1810 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
1811 was its contents. (It was OK if the option was not defined at all.)
1813 11. A "Completed" log line is now written for messages that are removed from
1814 the spool by the -Mrm option.
1816 12. $host_address is now set to the target address during the checking of
1817 ignore_target_hosts.
1819 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1820 being passed; this would have caused $sender_host_name to have been used if
1821 matching the list had actually called for a host name (not very likely,
1822 since this list is usually IP addresses). A host name is now passed as
1825 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1826 code that specifies a non-restarting handler (typically sigaction() in
1827 modern systems) in an attempt to fix a rare and obscure crash bug.
1829 15. Narrowed the window for a race in the daemon that could cause it to ignore
1830 SIGCHLD signals. This is not a major problem, because they are used only to
1831 wake it up if nothing else does.
1833 16. A malformed maildirsize file could cause Exim to calculate negative values
1834 for the mailbox size or file count. Odd effects could occur as a result.
1835 The maildirsize information is now recalculated if the size or filecount
1838 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1839 support for a long time. Removed HAVE_SYS_VFS_H.
1841 18. Updated exipick to current release from John Jetmore.
1843 19. Allow an empty sender to be matched against a lookup in an address list.
1844 Previously the only cases considered were a regular expression, or an
1847 20. Exim went into a mad DNS lookup loop when doing a callout where the
1848 host was specified on the transport, if the DNS lookup yielded more than
1851 21. The RFC2047 encoding function was originally intended for short strings
1852 such as real names; it was not keeping to the 75-character limit for
1853 encoded words that the RFC imposes. It now respects the limit, and
1854 generates multiple encoded words if necessary. To be on the safe side, I
1855 have increased the buffer size for the ${rfc2047: expansion operator from
1858 22. Failure to deliver a bounce message always caused it to be frozen, even if
1859 there was an errors_to setting on the router. The errors_to setting is now
1862 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1863 canonical form (fully expanded) before being placed in
1864 $sender_host_address.
1866 24. Updated eximstats to version 1.33
1868 25. Include certificate and key file names in error message when GnuTLS fails
1869 to set them up, because the GnuTLS error message doesn't include the name
1870 of the failing file when there is a problem reading it.
1872 26. Expand error message when OpenSSL has problems setting up cert/key files.
1875 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1876 (this can affect the format of dates).
1878 28. exim_tidydb, when checking for the continued existence of a message for
1879 which it has found a message-specific retry record, was not finding
1880 messages that were in split spool directories. Consequently, it was
1881 deleting retry records that should have stayed in existence.
1883 29. eximstats updated to version 1.35
1884 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1885 1.35 - bugfix such that pie charts by volume are generated correctly
1887 30. The SPA authentication driver was not abandoning authentication and moving
1888 on to the next authenticator when an expansion was forced to fail,
1889 contradicting the general specification for all authenticators. Instead it
1890 was generating a temporary error. It now behaves as specified.
1892 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1893 (the order specifies the preference for clients). The order is now AES256,
1894 AES128, 3DES, ARCFOUR128.
1896 31. Small patch to Sieve code - explicitly set From: when generating an
1899 32. Exim crashed if a remote delivery caused a very long error message to be
1900 recorded - for instance if somebody sent an entire SpamAssassin report back
1901 as a large number of 550 error lines. This bug was coincidentally fixed by
1902 increasing the size of one of Exim's internal buffers (big_buffer) that
1903 happened as part of the Exiscan merge. However, to be on the safe side, I
1904 have made the code more robust (and fixed the comments that describe what
1907 33. Some experimental protocols are using DNS PTR records for new purposes. The
1908 keys for these records are domain names, not reversed IP addresses. The
1909 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1910 leaves it alone. Component reversal etc. now happens only for IP addresses.
1913 34. The host_aton() function is supposed to be passed a string that is known
1914 to be a valid IP address. However, in the case of IPv6 addresses, it was
1915 not checking this. This is a hostage to fortune. Exim now panics and dies
1916 if the condition is not met. A case was found where this could be provoked
1917 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1918 components; fortuitously, this particular loophole had already been fixed
1919 by change 4.50/55 or 4.44/33 above.
1921 If there are any other similar loopholes, the new check in host_aton()
1922 itself should stop them being exploited. The report I received stated that
1923 data on the command line could provoke the exploit when Exim was running as
1924 exim, but did not say which command line option was involved. All I could
1925 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1926 running as the user.
1929 35. There was a buffer overflow vulnerability in the SPA authentication code
1930 (which came originally from the Samba project). I have added a test to the
1931 spa_base64_to_bits() function which I hope fixes it.
1934 36. The daemon start-up calls getloadavg() while still root for those OS that
1935 need the first call to be done as root, but it missed one case: when
1936 deliver_queue_load_max is set with deliver_drop_privilege. This is
1937 necessary for the benefit of the queue runner, because there is no re-exec
1938 when deliver_drop_privilege is set.
1940 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1941 was in use, was not putting the data itself into the right store pool;
1942 consequently, it could be overwritten for a subsequent message in the same
1943 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1946 38. Sometimes the final signoff response after QUIT could fail to get
1947 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1948 < 0 before doing a fflush(). This bug looks as though it goes back to the
1949 introduction of TLS in release 3.20, but "sometimes" must have been rare
1950 because the tests only now provoked it.
1956 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1957 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1958 Now it returns an integer. A number of calls were still expecting a T/F
1959 return. Fortuitously, in all cases, the tests worked in OK situations,
1960 which is the norm. However, things would have gone wrong on any write
1961 failures on the smtp file descriptor. This function is used when sending
1962 messages over SMTP and also when doing verify callouts.
1964 2. When Exim is called to do synchronous delivery of a locally submitted
1965 message (the -odf or -odi options), it no longer closes stderr before doing
1968 3. Implemented the mua_wrapper option.
1970 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1972 5. Implemented the functions header_remove(), header_testname(),
1973 header_add_at_position(), and receive_remove_recipient(), and exported them
1976 6. If an ACL "warn" statement specified the addition of headers, Exim already
1977 inserted X-ACL-Warn: at the start if there was no header name. However, it
1978 was not making this test for the second and subsequent header lines if
1979 there were newlines in the string. This meant that an invalid header could
1980 be inserted if Exim was badly configured.
1982 7. Allow an ACL "warn" statement to add header lines at the start or after all
1983 the Received: headers, as well as at the end.
1985 8. Added the rcpt_4xx retry error code.
1987 9. Added postmaster_mailfrom=xxx to callout verification option.
1989 10. Added mailfrom=xxxx to the callout verification option, for verify=
1992 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1993 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1995 12. Inserted some casts to stop certain compilers warning when using pointer
1996 differences as field lengths or precisions in printf-type calls (mostly
1997 affecting debugging statements).
1999 13. Added optional readline() support for -be (dynamically loaded).
2001 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
2002 same clock tick as a message's arrival, so that its received time was the
2003 same as the "first fail" time on the retry record, and that message
2004 remained on the queue past the ultimate address timeout, every queue runner
2005 would try a delivery (because it was past the ultimate address timeout) but
2006 after another failure, the ultimate address timeout, which should have then
2007 bounced the address, did not kick in. This was a "< instead of <=" error;
2008 in most cases the first failure would have been in the next clock tick
2009 after the received time, and all would be well.
2011 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
2012 being recognized when the domain list was tested by the match_domain
2013 condition in an expansion string.
2015 16. Added the ${str2b64: operator.
2017 17. Exim was always calling setrlimit() to set a large limit for the number of
2018 processes, without checking whether the existing limit was already
2019 adequate. (It did check for the limit on file descriptors.) Furthermore,
2020 errors from getrlimit() and setrlimit() were being ignored. Now they are
2021 logged to the main and panic logs, but Exim does carry on, to try to do its
2022 job under whatever limits there are.
2024 18. Imported PCRE 5.0.
2026 19. Trivial typo in log message " temporarily refused connection" (the leading
2029 20. If the log selector return_path_on_delivery was set and an address was
2030 redirected to /dev/null, the delivery process crashed because it assumed
2031 that a return path would always be set for a "successful" delivery. In this
2032 case, the whole delivery is bypassed as an optimization, and therefore no
2035 21. Internal re-arrangement: the function for sending a challenge and reading
2036 a response while authentication was assuming a zero-terminated challenge
2037 string. It's now changed to take a pointer and a length, to allow for
2038 binary data in such strings.
2040 22. Added the cyrus_sasl authenticator (code supplied by MBM).
2042 23. Exim was not respecting finduser_retries when seeking the login of the
2043 uid under which it was called; it was always trying 10 times. (The default
2044 setting of finduser_retries is zero.) Also, it was sleeping after the final
2045 failure, which is pointless.
2047 24. Implemented tls_on_connect_ports.
2049 25. Implemented acl_smtp_predata.
2051 26. If the domain in control=submission is set empty, Exim assumes that the
2052 authenticated id is a complete email address when it generates From: or
2053 Sender: header lines.
2055 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
2056 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
2057 chown and chgrp in /bin and hostname in /usr/bin.
2059 28. Exim was keeping the "process log" file open after each use, just as it
2060 does for the main log. This opens the possibility of it remaining open for
2061 long periods when the USR1 signal hits a daemon. Occasional processlog
2062 errors were reported, that could have been caused by this. Anyway, it seems
2063 much more sensible not to leave this file open at all, so that is what now
2066 29. The long-running daemon process does not normally write to the log once it
2067 has entered its main loop, and it closes the log before doing so. This is
2068 so that log files can straightforwardly be renamed and moved. However,
2069 there are a couple of unusual error situations where the daemon does write
2070 log entries, and I had neglected to close the log afterwards.
2072 30. The text of an SMTP error response that was received during a remote
2073 delivery was being truncated at 512 bytes. This is too short for some of
2074 the long messages that one sometimes sees. I've increased the limit to
2077 31. It is now possible to make retry rules that apply only when a message has a
2078 specific sender, in particular, an empty sender.
2080 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
2081 it possible to be selective about when SMTP synchronization is enforced.
2083 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
2085 32. Implemented hosts_connection_nolog.
2087 33. Added an ACL for QUIT.
2089 34. Setting "delay_warning=" to disable warnings was not working; it gave a
2092 35. Added mailbox_size and mailbox_filecount to appendfile.
2094 36. Added control = no_multiline_responses to ACLs.
2096 37. There was a bug in the logic of the code that waits for the clock to tick
2097 in the case where the clock went backwards by a substantial amount such
2098 that the microsecond fraction of "now" was more than the microsecond
2099 fraction of "then" (but the whole seconds number was less).
2101 38. Added support for the libradius Radius client library this is found on
2102 FreeBSD (previously only the radiusclient library was supported).
2108 1. When certain lookups returned multiple values in the form name=value, the
2109 quoting of the values was not always being done properly. Specifically:
2110 (a) If the value started with a double quote, but contained no whitespace,
2112 (b) If the value contained whitespace other than a space character (i.e.
2113 tabs or newlines or carriage returns) it was not quoted.
2114 This fix has been applied to the mysql and pgsql lookups by writing a
2115 separate quoting function and calling it from the lookup code. The fix
2116 should probably also be applied to nisplus, ibase and oracle lookups, but
2117 since I cannot test any of those, I have not disturbed their existing code.
2119 2. A hit in the callout cache for a specific address caused a log line with no
2120 reason for rejecting RCPT. Now it says "Previous (cached) callout
2121 verification failure".
2123 3. There was an off-by-one bug in the queryprogram router. An over-long
2124 return line was truncated at 256 instead of 255 characters, thereby
2125 overflowing its buffer with the terminating zero. As well as fixing this, I
2126 have increased the buffer size to 1024 (and made a note to document this).
2128 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
2129 when Exim is waiting for an SMTP response from a remote server, Exim
2130 restarts its select() call on the socket, thereby resetting its timeout.
2131 This is not a problem when such interrupts are rare. Somebody set up a cron
2132 job to run exiwhat every 2 minutes, which is less than the normal select()
2133 timeout (5 or 10 minutes). This meant that the select() timeout never
2134 kicked in because it was always reset. I have fixed this by comparing the
2135 time when an interrupt arrives with the time at the start of the first call
2136 to select(). If more time than the timeout has elapsed, the interrupt is
2137 treated as a timeout.
2139 5. Some internal re-factoring in preparation for the addition of Sieve
2140 extensions (by MH). In particular, the "personal" test is moved to a
2141 separate function, and given an option for scanning Cc: and Bcc: (which is
2142 not set for Exim filters).
2144 6. When Exim created an email address using the login of the caller as the
2145 local part (e.g. when creating a From: or Sender: header line), it was not
2146 quoting the local part when it contained special characters such as @.
2148 7. Installed new OpenBSD configuration files.
2150 8. Reworded some messages for syntax errors in "and" and "or" conditions to
2151 try to make them clearer.
2153 9. Callout options, other than the timeout value, were being ignored when
2154 verifying sender addresses in header lines. For example, when using
2156 verify = header_sender/callout=no_cache
2158 the cache was (incorrectly) being used.
2160 10. Added a missing instance of ${EXE} to the exim_install script; this affects
2161 only the Cygwin environment.
2163 11. When return_path_on_delivery was set as a log selector, if different remote
2164 addresses in the same message used different return paths and parallel
2165 remote delivery occurred, the wrong values would sometimes be logged.
2166 (Whenever a remote delivery process finished, the return path value from
2167 the most recently started remote delivery process was logged.)
2169 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
2170 lines when AUTH and/or TLS are in use. This is the "received protocol"
2171 field. Exim used to use "asmtp" for authenticated SMTP, without any
2172 indication (in the protocol name) for TLS use. Now it follows the RFC and
2173 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
2174 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
2175 names appear in log lines as well as in Received: header lines.
2177 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
2178 extensions, and comparison tests, and to fix some bugs.
2180 14. Changes to the "personal" filter test:
2182 (1) The test was buggy in that it was just doing the equivalent of
2183 "contains" tests on header lines. For example, if a user's address was
2184 anne@some.where, the "personal" test would incorrectly be true for
2186 To: susanne@some.where
2188 This test is now done by extracting each address from the header in turn,
2189 and checking the entire address. Other tests that are part of "personal"
2190 are now done using regular expressions (for example, to check local parts
2191 of addresses in From: header lines).
2193 (2) The list of non-personal local parts in From: addresses has been
2194 extended to include "listserv", "majordomo", "*-request", and "owner-*",
2195 taken from the Sieve specification recommendations.
2197 (3) If the message contains any header line starting with "List-" it is
2198 treated as non-personal.
2200 (4) The test for "circular" in the Subject: header line has been removed
2201 because it now seems ill-conceived.
2203 15. Minor typos in src/EDITME comments corrected.
2205 16. Installed latest exipick from John Jetmore.
2207 17. If headers_add on a router specified a text string that was too long for
2208 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
2209 of string_sprintf() is now avoided.
2211 18. $message_body_size was not set (it was always zero) when running the DATA
2212 ACL and the local_scan() function.
2214 19. For the "mail" command in an Exim filter, no default was being set for
2215 the once_repeat time, causing a random time value to be used if "once" was
2216 specified. (If the value happened to be <= 0, no repeat happened.) The
2217 default is now 0s, meaning "never repeat". The "vacation" command was OK
2218 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
2219 (I found it when inspecting the code).
2221 20. There is now an overall timeout for performing a callout verification. It
2222 defaults to 4 times the callout timeout, which applies to individual SMTP
2223 commands during the callout. The overall timeout applies when there is more
2224 than one host that can be tried. The timeout is checked before trying the
2225 next host. This prevents very long delays if there are a large number of
2226 hosts and all are timing out (e.g. when the network connections are timing
2227 out). The value of the overall timeout can be changed by specifying an
2228 additional sub-option for "callout", called "maxwait". For example:
2230 verify = sender/callout=5s,maxwait=20s
2232 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
2233 to the end before writing, but this should make it even safer).
2235 22. Exim was forgetting that it had advertised PIPELINING for the second and
2236 subsequent messages on an SMTP connection. It was also not resetting its
2237 memory on STARTTLS and an internal HELO.
2239 23. When Exim logs an SMTP synchronization error within a session, it now
2240 records whether PIPELINING has been advertised or not.
2242 24. Added 3 instances of "(long int)" casts to time_t variables that were being
2243 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
2244 rather than long int.
2246 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
2248 26. Added the never_mail option to autoreply.
2254 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
2255 crash if the getsockname() call failed; this can happen if a connection is
2256 closed very soon after it is established. The problem was simply in the
2257 order in which certain operations were done, causing Exim to try to write
2258 to the SMTP stream before it had set up the file descriptor. The bug has
2259 been fixed by making things happen in the correct order.
2265 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
2266 before the connection was closed, thus losing the rejection response.
2268 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
2269 some early Solaris releases, but causes trouble in current releases where
2270 socklen_t is defined.
2272 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
2275 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
2278 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
2279 that releasing the top of it at the end releases what was used for sub-
2280 expansions (unless the block got too big). However, discard this block if
2281 the first thing is a variable or header, so that we can use its block when
2282 it is dynamic (useful for very large $message_headers, for example).
2284 6. Lookups now cache *every* query, not just the most recent. A new, separate
2285 store pool is used for this. It can be recovered when all lookup caches are
2286 flushed. Lookups now release memory at the end of their result strings.
2287 This has involved some general refactoring of the lookup sources.
2289 7. Some code has been added to the store_xxx() functions to reduce the amount
2290 of flapping under certain conditions.
2292 8. log_incoming_interface used to affect only the <= reception log lines. Now
2293 it causes the local interface and port to be added to several more SMTP log
2294 lines, for example "SMTP connection from", and rejection lines.
2296 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
2298 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
2300 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
2301 could be overwritten at the end of the current message (or the start of a
2302 new message if it was set in a HELO ACL). The value is now preserved for
2303 the duration of the SMTP connection.
2305 12. If a transport had a headers_rewrite setting, and a matching header line
2306 contained an unqualified address, that address was qualified, even if it
2307 did not match any rewriting rules. The underlying bug was that the values
2308 of the flags that permit the existence of unqualified sender and recipient
2309 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
2310 non-local messages, and by -bnq for local messages) were not being
2311 preserved with the message after it was received.
2313 13. When Exim was logging an SMTP synchronization error, it could sometimes log
2314 "next input=" as part of the text comprising the host identity instead of
2315 the correct text. The code was using the same buffer for two different
2316 strings. However, depending on which order the printing function evaluated
2317 its arguments, the bug did not always show up. Under Linux, for example, my
2318 test suite worked just fine.
2320 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
2321 doesn't work with some older versions of Perl. It has been changed to "my",
2322 which in any case is probably the better facility to use.
2324 15. A really picky compiler found some instances of statements for creating
2325 error messages that either had too many or two few arguments for the format
2328 16. The size of the buffer for calls to the DNS resolver has been increased
2329 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
2330 for addresses that have a lot of PTR records. This alleviates a problem; it
2331 does not fully solve it.
2333 17. A dnsdb lookup for PTR records that receives more data than will fit in the
2334 buffer now truncates the list and logs the incident, which is the same
2335 action as happens when Exim is looking up a host name and its aliases.
2336 Previously in this situation something unpredictable would happen;
2337 sometimes it was "internal error: store_reset failed".
2339 18. If a server dropped the connection unexpectedly when an Exim client was
2340 using GnuTLS and trying to read a response, the client delivery process
2341 crashed while trying to generate an error log message.
2343 19. If a "warn" verb in an ACL added multiple headers to a message in a single
2344 string, for example:
2346 warn message = H1: something\nH2: something
2348 the text was added as a single header line from Exim's point of view
2349 though it ended up OK in the delivered message. However, searching for the
2350 second and subsequent header lines using $h_h2: did not work. This has been
2351 fixed. Similarly, if a system filter added multiple headers in this way,
2352 the routers could not see them.
2354 20. Expanded the error message when iplsearch is called with an invalid key to
2355 suggest using net-iplsearch in a host list.
2357 21. When running tests using -bh, any delays imposed by "delay" modifiers in
2358 ACLs are no longer actually imposed (and a message to that effect is
2361 22. If a "gecos" field in a passwd entry contained escaped characters, in
2362 particular, if it contained a \" sequence, Exim got it wrong when building
2363 a From: or a Sender: header from that name. A second bug also caused
2364 incorrect handling when an unquoted " was present following a character
2365 that needed quoting.
2367 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
2368 was not being matched caselessly.
2370 24. Arranged for all hyphens in the exim.8 source to be escaped with
2373 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
2374 properly. Recipient callout cache records were still being keyed to include
2375 the sender, even when use_sender was set false. This led to far more
2376 callouts that were necessary. The sender is no longer included in the key
2377 when use_sender is false.
2379 26. Added "control = submission" modifier to ACLs.
2381 27. Added the ${base62d: operator to decode base 62 numbers.
2383 28. dnsdb lookups can now access SRV records.
2385 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
2386 the configuration file.
2388 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
2389 (-v) mode. This makes the output for a verbose queue run more intelligible.
2391 31. Added a use_postmaster feature to recipient callouts.
2393 32. Added the $body_zerocount variable, containing the number of binary zero
2394 bytes in the message body.
2396 33. The time of last modification of the "new" subdirectory is now used as the
2397 "mailbox time last read" when there is a quota error for a maildir
2400 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
2402 35. Added +ignore_unknown as a special item in host lists.
2404 36. Code for decoding IPv6 addresses in host lists is now included, even if
2405 IPv6 support is not being compiled. This fixes a bug in which an IPv6
2406 address was recognized as an IP address, but was then not correctly decoded
2407 into binary, causing unexpected and incorrect effects when compared with
2414 1. Very minor rewording of debugging text in manualroute to say "list of
2415 hosts" instead of "hostlist".
2417 2. If verify=header_syntax was set, and a header line with an unqualified
2418 address (no domain) and a large number of spaces between the end of the
2419 name and the colon was received, the reception process suffered a buffer
2420 overflow, and (when I tested it) crashed. This was caused by some obsolete
2421 code that should have been removed. The fix is to remove it!
2423 3. When running in the test harness, delay a bit after writing a bounce
2424 message to get a bit more predictability in the log output.
2426 4. Added a call to search_tidyup() just before forking a reception process. In
2427 theory, someone could use a lookup in the expansion of smtp_accept_max_
2428 per_host which, without the tidyup, could leave open a database connection.
2430 5. Added the variables $recipient_data and $sender_data which get set from a
2431 lookup success in an ACL "recipients" or "senders" condition, or a router
2432 "senders" option, similar to $domain_data and $local_part_data.
2434 6. Moved the writing of debug_print from before to after the "senders" test
2437 7. Change 4.31/66 (moving the time when the Received: is generated) caused
2438 problems for message scanning, either using a data ACL, or using
2439 local_scan() because the Received: header was not generated till after they
2440 were called (in order to set the time as the time of reception completion).
2441 I have revised the way this works. The header is now generated after the
2442 body is received, but before the ACL or local_scan() are called. After they
2443 are run, the timestamp in the header is updated.
2449 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
2450 before starting a queue runner without re-exec. This happened only when
2451 deliver_drop_privilege was set or when the Exim user was set to root. The
2452 effect of the bug was that timeouts during subsequent deliveries caused
2453 crashes instead of being properly handled. The handler is now left at its
2454 default (and expected) setting.
2456 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
2457 message, again when deliver_drop_privilege is set or Exim is run as root.
2458 The bug described in (1) was not present in this case, but the tidying up
2459 of the other signals was missing. I have made the two cases consistent.
2461 3. The ignore_target_hosts setting on a manualroute router was being ignored
2462 for hosts that were looked up using the /MX notation.
2464 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
2467 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
2468 operated on the sender address. After changing the $sender_address to <>
2469 for the sender address verify, Exim was re-instated it as the original
2470 (before rewriting) address, but remembering that it had rewritten it, so it
2471 wasn't rewriting it again. This bug also had the effect of breaking the
2472 sender address verification caching when the sender address was rewritten.
2474 6. The ignore_target_hosts option was being ignored by the ipliteral router.
2475 This has been changed so that if the ip literal address matches
2476 ignore_target_hosts, the router declines.
2478 7. Added expansion conditions match_domain, match_address, and match_local_
2479 part (NOT match_host).
2481 8. The placeholder for the Received: header didn't have a length field set.
2483 9. Added code to Exim itself and to exim_lock to test for a specific race
2484 condition that could lead to file corruption when using MBX delivery. The
2485 issue is with the lockfile that is created in /tmp. If this file is removed
2486 after a process has opened it but before that process has acquired a lock,
2487 there is the potential for a second process to recreate the file and also
2488 acquire a lock. This could lead to two Exim processes writing to the file
2489 at the same time. The added code performs the same test as UW imapd; it
2490 checks after acquiring the lock that its file descriptor still refers to
2491 the same named file.
2493 10. The buffer for building added header lines was of fixed size, 8192 bytes.
2494 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
2497 11. Added the smtp_active_hostname option. If used, this will typically be made
2498 to depend on the incoming interface address. Because $interface_address is
2499 not set up until the daemon has forked a reception process, error responses
2500 that can happen earlier (such as "too many connections") no longer contain
2503 12. If an expansion in a condition on a "warn" statement fails because a lookup
2504 defers, the "warn" statement is abandoned, and the next ACL statement is
2505 processed. Previously this caused the whole ACL to be aborted.
2507 13. Added the iplsearch lookup type.
2509 14. Added ident_timeout as a log selector.
2511 15. Added tls_certificate_verified as a log selector.
2513 16. Added a global option tls_require_ciphers (compare the smtp transport
2514 option of the same name). This controls incoming TLS connections.
2516 17. I finally figured out how to make tls_require_ciphers do a similar thing
2517 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
2518 before starting the TLS session.
2520 18. Tabs are now shown as \t in -bP output.
2522 19. If the log selector return_path_on_delivery was set, Exim crashed when
2523 bouncing a message because it had too many Received: header lines.
2525 20. If two routers both had headers_remove settings, and the first one included
2526 a superfluous trailing colon, the final name in the first list and the
2527 first name in the second list were incorrectly joined into one item (with a
2528 colon in the middle).
2534 1. Added -C and -D options to the exinext utility, mainly to make it easier
2535 to include in the automated testing, but these could be helpful when
2536 multiple configurations are in use.
2538 2. The exinext utility was not formatting the output nicely when there was
2539 an alternate port involved in the retry record key, nor when there was a
2540 message id as well (for retries that were specific to a specific message
2541 and a specific host). It was also confused by IPv6 addresses, because of
2542 the additional colons they contain. I have fixed the IPv4 problem, and
2543 patched it up to do a reasonable job for IPv6.
2545 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
2546 delivery, the log line now contains "pipelined" if PIPELINING was used.
2548 4. An SMTP transport process used to panic and die if the bind() call to set
2549 an explicit outgoing interface failed. This has been changed; it is now
2550 treated in the same way as a connect() failure.
2552 5. A reference to $sender_host_name in the part of a conditional expansion
2553 that was being skipped was still causing a DNS lookup. This no longer
2556 6. The def: expansion condition was not recognizing references to header lines
2557 that used bh_ and bheader_.
2559 7. Added the _cache feature to named lists.
2561 8. The code for checking quota_filecount in the appendfile transport was
2562 allowing one more file than it should have been.
2564 9. For compatibility with Sendmail, the command line option
2572 and sets the incoming protocol and host name (for trusted callers). The
2573 host name and its colon can be omitted when only the protocol is to be set.
2574 Note the Exim already has two private options, -pd and -ps, that refer to
2575 embedded Perl. It is therefore impossible to set a protocol value of "d" or
2576 "s", but I don't think that's a major issue.
2578 10. A number of refactoring changes to the code, none of which should affect
2581 (a) The number of logging options was getting close to filling up the
2582 32-bit word that was used as a bit map. I have split them into two classes:
2583 those that are passed in the argument to log_write(), and those that are
2584 only ever tested independently outside of that function. These are now in
2585 separate 32-bit words, so there is plenty of room for expansion again.
2586 There is no change in the user interface or the logging behaviour.
2588 (b) When building, for example, log lines, the code previously used a
2589 macro that called string_cat() twice, in order to add two strings. This is
2590 not really sufficiently general. Furthermore, there was one instance where
2591 it was actually wrong because one of the argument was used twice, and in
2592 one call a function was used. (As it happened, calling the function twice
2593 did not affect the overall behaviour.) The macro has been replaced by a
2594 function that can join an arbitrary number of extra strings onto a growing
2597 (c) The code for expansion conditions now uses a table and a binary chop
2598 instead of a serial search (which was left over from when there were very
2599 few conditions). Also, it now recognizes conditions like "pam" even when
2600 the relevant support is not compiled in: a suitably worded error message is
2601 given if an attempt is made to use such a condition.
2603 11. Added ${time_interval:xxxxx}.
2605 12. A bug was causing one of the ddress fields not to be passed back correctly
2606 from remote delivery subprocesses. The field in question was not being
2607 subsequently used, so this caused to problems in practice.
2609 13. Added new log selectors queue_time and deliver_time.
2611 14. Might have fixed a bug in maildirsizefile handling that threw up
2612 "unexpected character" debug warnings, and recalculated the data
2613 unnecessarily. In any case, I expanded the warning message to give more
2616 15. Added the message "Restricted characters in address" to the statements in
2617 the default ACL that block characters like @ and % in local parts.
2619 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
2620 Three changes have been made:
2622 (a) There was a serious bug; a negative response to MAIL caused the whole
2623 recipient domain to be cached as invalid, thereby blocking all messages
2624 to all local parts at the same domain, from all senders. This bug has
2625 been fixed. The domain is no longer cached after a negative response to
2626 MAIL if the sender used is not empty.
2628 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
2631 (c) A new callout option, "use_sender" has been added for people who want
2632 the modified behaviour.
2638 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
2641 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
2642 indeed breaks things for older releases.
2644 3. Added additional logging to the case where there is a problem reading data
2645 from a filter that is running in a subprocess using a pipe, in order to
2646 try to track down a specific problem.
2648 4. Testing facility fudge: when running in the test harness and attempting
2649 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
2650 getting "No route to host". Convert this to a timeout.
2652 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
2655 6. Some OS don't have socklen_t but use size_t instead. This affects the
2656 fifth argument of getsockopt() amongst other things. This is now
2657 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
2658 can be set for individual OS. I have set it for SunOS5, OSF1, and
2659 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
2660 some earlier ones do not.
2662 7. Change 4.30/15 was not doing the test caselessly.
2664 8. The standard form for an IPv6 address literal was being rejected by address
2665 parsing in, for example, MAIL and RCPT commands. An example of this kind of
2666 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2667 this, as well as the form without the "IPv6" on the front (but only when
2668 address literals are enabled, of course).
2670 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
2672 10. Exim crashed if a message with an empty sender address specified by -f
2673 encountered a router with an errors_to setting. This could be provoked only
2674 by a command such as
2678 where an empty string was supplied; "<>" did not hit this bug.
2680 11. Installed PCRE release 4.5.
2682 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
2683 remained set. It is now erased.
2685 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
2686 times from message ids (which are base 36 rather than the normal 62).
2688 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
2689 were being counted as actual protocol errors, and logged if the log
2690 selector +smtp_protocol_error was set. One cannot be perfect in this test,
2691 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
2692 and DATA following a set of rejected RCPTs do not count as protocol errors.
2693 In other words, Exim assumes they were pipelined, though this may not
2694 actually be the case. Of course, in all cases the client gets an
2695 appropriate error code.
2697 15. If a lookup fails in an ACL condition, a message about the failure may
2698 be available; it is used if testing the ACL cannot continue, because most
2699 such messages specify what the cause of the deferral is. However, some
2700 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
2701 that caused an old message to be retained and used if a later statement
2702 caused a defer, replacing the real cause of the deferral.
2704 16. If an IP address had so many PTR records that the DNS lookup buffer
2705 was not large enough to hold them, Exim could crash while trying to process
2706 the truncated data. It now detects and logs this case.
2708 17. Further to 4.21/58, another change has been made: if (and only if) the
2709 first line of a message (the first header line) ends with CRLF, a bare LF
2710 in a subsequent header line has a space inserted after it, so as not to
2711 terminate the header.
2713 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
2714 unnecessarily, used atoi() instead of strtol(), and didn't check the
2715 termination when getting file sizes from file names by regex.
2717 19. Completely re-implemented the support for maildirsize files, in the light
2718 of a number of problems with the previous contributed implementation
2719 (4.30/29). In particular:
2721 . If the quota is zero, the maildirsize file is maintained, but no quota is
2724 . If the maildir directory does not exist, it is created before any attempt
2725 to write a maildirsize file.
2727 . The quota value in the file is just a cache; if the quota is changed in
2728 the transport, the new value overrides.
2730 . A regular expression is available for excluding directories from the
2733 20. The autoreply transport checks the characters in options that define the
2734 message's headers; it allows continued headers, but it was checking with
2735 isspace() after an embedded newline instead of explicitly looking for a
2738 21. If all the "regular" hosts to which an address was routed had passed their
2739 expiry times, and had not reached their retry times, the address was
2740 bounced, even if fallback hosts were defined. Now Exim should go on to try
2743 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
2744 equivalent code in the SMTP transport. Some hosts send humungous responses
2745 to HELO/EHLO, more than 1024 it seems.
2747 23. Refactoring: code in filter.c used (void *) for "any old type" but this
2748 gives compiler warnings in some environments. I've now done it "properly",
2751 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
2752 (because of problems with the built-in one) was declared to return uschar *
2753 instead of char *, causing compiler failure.
2755 25. Fixed a file descriptor leak when processing alias/forward files.
2757 26. Fixed a minor format string issue in dbfn.c.
2759 27. Typo in exim.c: ("dmbnz" for "dbmnz").
2761 28. If a filter file refered to $h_xxx or $message_headers, and the headers
2762 contained RFC 2047 "words", Exim's memory could, under certain conditions,
2765 29. When a sender address is verified, it is cached, to save repeating the test
2766 when there is more than one recipient in a message. However, when the
2767 verification involves a callout, it is possible for different callout
2768 options to be set for different recipients. It is too complicated to keep
2769 track of this in the cache, so now Exim always runs a verification when a
2770 callout is required, relying on the callout cache for the optimization.
2771 The overhead is duplication of the address routing, but this should not be
2774 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
2775 to be verified with callout=postmaster, and the main callout worked but the
2776 postmaster check failed, the verification correctly failed. However, if a
2777 subsequent RCPT command asked for sender verification *without* the
2778 postmaster check, incorrect caching caused this verification also to fail,
2781 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
2782 it was not caching the DNS options (qualify_single, search_parents) that
2783 were used when the lookup failed. A subsequent lookup with different
2784 options therefore always gave the same answer, though there were cases
2785 where it should not have. (Example: a "domains = !$mx_any" option on a
2786 dnslookup router: the "domains" option is always processed without any
2787 widening, but the router might have qualify_single set.) Now Exim uses the
2788 cached value only when the same options are set.
2790 32. Added John Jetmore's "exipick" utility to the distribution.
2792 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
2793 than a timeout (e.g. a certificate is required, and is not provided), an
2794 Exim server now closes the connection immediately. Previously it waited for
2795 the client to close - but if the client is SSL, it seems that they each
2796 wait for each other, leading to a delay before one of them times out.
2798 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
2799 maintained 0.8.x compatibility because I don't think many are using it, and
2800 it is clearly obsolete.
2802 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
2805 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
2806 client certificate was expired. A simple patch fixes this, though I don't
2807 understand the full logic of why the verify callback is called multiple
2810 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
2811 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
2812 which causes problems with some clients (such as the Certicom SSL Plus
2813 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2814 disables the coutermeasure allowing Eudora to connect."
2816 38. Exim was not checking that a write() to a log file succeeded. This could
2817 lead to Bad Things if a log got too big, in particular if it hit a file
2818 size limit. Exim now panics and dies if it cannot write to a log file, just
2819 as it does if it cannot open a log file.
2821 39. Modified OS/Makefile-Linux so that it now contains
2823 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2825 The two -D definitions ensure that Exim is compiled with large file
2826 support, which makes it possible to handle log files that are bigger than
2829 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2830 instance) a domain was checked against a named list that involved a lookup,
2831 causing $domain_data to be set, then another domain was checked against the
2832 same list, then the first domain was re-checked, the value of $domain_data
2833 after the final check could be wrong. In particular, if the second check
2834 failed, it could be set empty. This bug probably also applied to
2837 41. The strip_trailing_dot option was not being applied to the address given
2838 with the -f command-line option.
2840 42. The code for reading a message's header from the spool was incrementing
2841 $received_count, but never initializing it. This meant that the value was
2842 incorrect (doubled) while delivering a message in the same process in which
2843 it was received. In the most common configuration of Exim, this never
2844 happens - a fresh exec is done - but it can happen when
2845 deliver_drop_privilege is set.
2847 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2848 it now includes up to 150 characters of the unexpected data in the log
2851 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2852 and building data strings. The size of both of these buffers was 10 000
2853 bytes - far larger than anybody would *ever* want, thought I. Needless to
2854 say, somebody hit the limit. I have increased the maximum line length to
2855 20 000 and the maximum data length of concatenated lines to 100 000. I have
2856 also fixed two bugs, because there was no checking on these buffers. Tsk,
2857 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2858 buffer is too small.
2860 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2861 lsearch lookups. Now it does.
2863 46. When parsing a route_list item in a manualroute router, a fixed-length
2864 buffer was used for the list of hosts. I made this 1024 bytes long,
2865 thinking that nobody would ever have a list of hosts that long. Wrong.
2866 Somebody had a whole pile of complicated expansion conditions, and the
2867 string was silently truncated, leading to an expansion error. It turns out
2868 that it is easier to change to an unlimited length (owing to other changes
2869 that have happened since this code was originally written) than to build
2870 structure for giving a limitation error. The length of the item that
2871 expands into the list of hosts is now unlimited.
2873 47. The lsearch lookup could not handle data where the length of text line was
2874 more than 4095 characters. Such lines were truncated, leading to shortened
2875 data being returned. It should now handle lines of any length.
2877 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2878 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2881 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2882 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2883 no longer happens in comment lines. A list of replacements is now placed
2884 at the head of all of the source files, except those whose only change is
2885 to replace PERL_COMMAND in the very first #! line.
2887 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2888 messages on the queue, with a bottom-up merge sort, using code contributed
2889 by Michael Haardt. This should make operations like -bp somewhat faster on
2890 large queues. It won't affect queue runners, except when queue_run_in_order
2893 51. Installed eximstats 1.31 in the distribution.
2895 52. Added support for SRV lookups to the dnslookup router.
2897 53. If an ACL referred to $message_body or $message_body_end, the value was not
2898 reset for any messages that followed in the same SMTP session.
2900 54. The store-handling optimization for building very long strings was not
2901 differentiating between the different store pools. I don't think this
2902 actually made any difference in practice, but I've tidied it.
2904 55. While running the routers to verify a sender address, $sender_address
2905 was still set to the sender address. This is wrong, because when routing to
2906 send a bounce to the sender, it would be empty. Therefore, I have changed
2907 it so that, while verifying a sender address, $sender_address is set to <>.
2908 (There is no change to what happens when verifying a recipient address.)
2910 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2911 target A or AAAA records (if not already returned) without resetting the
2912 qualify_single or search_parents options of the DNS resolver. These are
2913 inappropriate in this case because the targets of MX and SRV records must
2914 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2915 target that, when qualified, matched something in the local domain. These
2916 two options are now turned off when doing these lookups.
2918 57. It seems that at least some releases of Reiserfs (which does not have the
2919 concept of a fixed number of inodes) returns zero and not -1 for the
2920 number of available inodes. This interacted badly with check_spool_inodes,
2921 which assumed that -1 was the "no such thing" setting. What I have done is
2922 to check that the total number of inodes is greater than zero before doing
2923 the test of how many are available.
2925 58. When a "warn" ACL statement has a log_message modifier, the message is
2926 remembered, and not repeated. This is to avoid a lot of repetition when a
2927 message has many recipients that cause the same warning to be written.
2928 Howewer, Exim was preserving the list of already written lines for an
2929 entire SMTP session, which doesn't seem right. The memory is now reset if a
2930 new message is started.
2932 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2933 debugging output unless log_rewrite was also set.
2935 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2936 of (int)(handle) when we know that handle contains (void *)(-1).
2938 61. The Exim daemon panic-logs an error return when it closes the incoming
2939 connection. However "connection reset by peer" seems to be common, and
2940 isn't really an error worthy of noting specially, so that particular error
2943 62. When Exim is trying to find all the local interfaces, it used to panic and
2944 die if the ioctl to get the interface flags failed. However, it seems that
2945 on at least one OS (Solaris 9) it is possible to have an interface that is
2946 included in the list of interfaces, but for which you get a failure error
2947 for this call. This happens when the interface is not "plumbed" into a
2948 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2949 failure of the "get flags" call assumes that the interface is down.
2951 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2952 makes life easier for people who are doing arithmetic on fields extracted
2953 from dates, where you often get leading zeros that should not be
2954 interpreted as octal.
2956 64. Added qualify_domain to the redirect router, to override the global
2959 65. If a pathologically long header line contained very many addresses (the
2960 report of this problem mentioned 10 000) and each of them was rewritten,
2961 Exim could use up a very large amount of memory. (It kept on making new
2962 copies of the header line as it rewrote, and never released the old ones.)
2963 At the expense of a bit more processing, the header rewriting function has
2964 been changed so that it no longer eats memory in this way.
2966 66. The generation of the Received: header has been moved from the time that a
2967 message starts to be received, to the time that it finishes. The timestamp
2968 in the Received: header should now be very close to that of the <= log
2969 line. There are two side-effects of this change:
2971 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2972 logged header lines no longer include the local Received: line, because
2973 it has not yet been created. The same applies to a copy of the message
2974 that is returned to a non-SMTP sender when a message is rejected.
2976 (b) When a filter file is tested using -bf, no additional Received: header
2977 is added to the test message. After some thought, I decided that this
2980 This change does not affect the value of $received_for. It is still set
2981 after address rewriting, but before local_scan() is called.
2983 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2985 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2986 gave an unhelpful panic error message, and a defer error. I have managed to
2987 change this behaviour so that it now rejects any supplied certificate,
2988 which seems right, as the list of acceptable certificates is empty.
2990 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2991 gave an unhelpful defer error. I have not managed to make this reject any
2992 supplied certificates, but the error message it gives is "no certificate
2993 supplied", which is not helpful.
2995 70. exigrep's output now also includes lines that are not associated with any
2996 message, but which match the given pattern. Implemented by a patch from
2997 Martin Sluka, which also tidied up the Perl a bit.
2999 71. Recipient callout verification, like sender verification, was using <> in
3000 the MAIL FROM command. This isn't really the right thing, since the actual
3001 sender may affect whether the remote host accepts the recipient or not. I
3002 have changed it to use the actual sender in the callout; this means that
3003 the cache record is now keyed on a recipient/sender pair, not just the
3004 recipient address. There doesn't seem to be a real danger of callout loops,
3005 since a callout by the remote host to check the sender would use <>.
3006 [SEE ABOVE: changed after hitting problems.]
3008 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
3009 temporary errors. However, in the case of such a code being given after
3010 the end of a data transmission (i.e. after ".") Exim was failing to write
3011 a retry record for the message. (Yes, there was some broken host that was
3012 actually sending 8xx at this point.)
3014 73. An unknown lookup type in a host list could cause Exim to panic-die when
3015 the list was checked. (An example that provoked this was putting <; in the
3016 middle of a list instead of at the start.) If this happened during a DATA
3017 ACL check, a -D file could be left lying around. This kind of configuration
3018 error no longer causes Exim to die; instead it causes a defer errror. The
3019 incident is still logged to the main and panic logs.
3021 74. Buglet left over from Exim 3 conversion. The message "too many messages
3022 in one connection" was written to the rejectlog but not the mainlog, except
3023 when address rewriting (yes!) was being logged.
3025 75. Added write_rejectlog option.
3027 76. When a system filter was run not as root (that is, when system_filter_user
3028 was set), the values of the $n variables were not being returned to the
3029 main process; thus, they were not subsequently available in the $sn
3032 77. Added +return_path_on_delivery log selector.
3034 78. A connection timeout was being treated differently from recipients deferred
3035 when testing hosts_max_try with a message that was older than the host's
3036 retry timeout. (The host should not be counted, thus allowing all hosts to
3037 be tried at least once before bouncing.) This may have been the cause of an
3038 occasionally reported bug whereby a message would remain on the queue
3039 longer than the retry timeout, but would be bounced if a delivery was
3040 forced. I say "may" because I never totally pinned down the problem;
3041 setting up timeout/retry tests is difficult. See also the next item.
3043 79. The ultimate address timeout was not being applied to errors that involved
3044 a combination of host plus message (for example, a timeout on a MAIL
3045 command). When an address resolved to a number of possible hosts, and they
3046 were not all tried for each delivery (e.g. because of hosts_max_try), a
3047 message could remain on the queue longer than the retry timeout.
3049 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
3052 81. Fixed an obscure SMTP outgoing bug which required at least the following
3053 conditions: (a) there was another message waiting for the same server;
3054 (b) the server returned 5xx to all RCPT commands in the first message so
3055 that the message was not completed; (c) the server dropped the connection
3056 or gave a negative response to the RSET that Exim sends to abort the
3057 transaction. The observed case was a dropped connection after DATA that had
3058 been sent in pipelining mode. That is, the server had advertised PIPELINING
3059 but was not implementing it correctly. The effect of the bug was incorrect
3060 behaviour, such as trying another host, and this could lead to a crash.
3066 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
3067 and daemon.c were passed as pointers to ints; they should have been
3068 pointers to socklen_t variables (which are typically unsigned ints).
3070 2. Some signed/unsigned type warnings in the os.c file for Linux have been
3073 3. Fixed a really odd bug that affected only the testing scheme; patching a
3074 certain fixed string in the binary changed the value of another string that
3075 happened to be identical to the end of the original first string.
3077 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
3078 name", it returns that address as the IP address. On some operating
3079 systems (e.g. Solaris), it also passes back the IP address string as the
3080 "host name". However, on others (e.g. Linux), it passes back an empty
3081 string. Exim wasn't checking for this, and was changing the host name to an
3082 empty string, assuming it had been canonicized.
3084 5. Although rare, it is permitted to have more than one PTR record for a given
3085 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
3086 all the names associated with an address, because they do in Solaris.
3087 However, it seems that they do not in Linux for data that comes from the
3088 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
3089 I found this out when I moved to a new Linux workstation and tried to run
3090 the Exim test suite.
3092 To get round this problem I have changed the code so that it now does its
3093 own call to the DNS to look up PTR records when searching for a host name.
3094 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
3095 addresses that are only in /etc/hosts are still found.
3097 This behaviour is, however, controlled by an option called host_lookup_
3098 order, which defaults to "bydns:byaddr". If people want to use the other
3099 order, or indeed, just use one or the other means of lookup, they can
3100 specify it in this variable.
3102 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
3103 some operating systems, this comes back from gethostbyaddr() as an empty
3104 string, and this is what Exim used to test for. However, it seems that in
3105 other systems, "." is yielded. Exim now tests for this case too.
3107 7. The values of check_spool_space and check_log_space are now held internally
3108 as a number of kilobytes instead of an absolute number of bytes. If a
3109 numbers is specified without 'K' or 'M', it is rounded up to the nearest
3110 kilobyte. This means that much larger values can be stored.
3112 8. Exim monitor: an attempt to get the action menu when not actually pointing
3113 at a message produces an empty menu entitled "No message selected". This
3114 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
3115 no entries in it ("Shell widget menu has zero width and/or height"). So I
3116 have added a single, blank menu entry in this case.
3118 9. Added ${quote_local_part.
3120 10. MIME decoding is now applied to the contents of Subject: header lines when
3123 11. Now that a reference to $sender_host_address automatically causes a reverse
3124 lookup to occur if necessary (4.13/18), there is no need to arrange for a
3125 host lookup before query-style lookups in lists that might use this
3126 variable. This has therefore been abolished, and the "net-" prefix is no
3127 longer necessary for query-style lookups.
3129 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
3130 have been a typo for LFLAGS, so it has been changed.
3132 13. The install script calls Exim with "-C /dev/null" in order to find the
3133 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
3134 to be output. Howeve, since Exim outputs its version number before the
3135 error, it didn't break the script. It just looked ugly. I fixed this by
3136 always allowing "-C /dev/null" if the caller is root.
3138 14. Ignore overlarge ACL variable number when reading spool file - insurance
3139 against a later release with more variables having written the file.
3141 15. The standard form for an IPv6 address literal was being rejected by EHLO.
3142 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
3143 this, as well as the form without the "IPv6" on the front.
3145 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
3146 OS/Makefile-Darwin file.
3148 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
3149 only to LDAP libraries that do not have LDAP_OPT_DEREF.
3151 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
3152 variable. However, some OS use ints for inodes. I've added cast to long int
3153 to get rid of the compiler warning.
3155 19. I had forgotten to lock out "/../" in configuration file names when
3156 ALT_CONFIG_PREFIX was set.
3158 20. Routers used for verification do not need to specify transports. However,
3159 if such a router generated a host list, and callout was configured, Exim
3160 crashed, because it could not find a port number from the (non-existent)
3161 transport. It now assumes port 25 in this circumstance.
3163 21. Added the -t option to exigrep.
3165 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
3166 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
3167 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
3168 them is set, LOOKUP_LSEARCH is forced.
3170 23. "exim -bV" now outputs a list of lookups that are included in the binary.
3172 24. Added sender and host information to the "rejected by local_scan()" log
3173 line; previously there was no indication of these.
3175 25. Added .include_if_exists.
3177 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
3178 Linux. It turns out that not all file systems support this. Apparently some
3179 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
3180 do it.) To cope with this, the error EINVAL, which means that sync-ing is
3181 not supported on the file descriptor, is now ignored when Exim is trying to
3182 sync a directory. This applies only to Linux.
3184 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
3186 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
3187 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
3190 29. Added support for maildirsize files from supplied patch (modified a bit).
3192 30. The use of :fail: followed by an empty string could lead Exim to respond to
3193 sender verification failures with (e.g.):
3195 550 Verification failed for <xxx>
3196 550 Sender verify failed
3198 where the first response line was missing the '-' that indicates it is not
3199 the final line of the response.
3201 31. The loop for finding the name of the user that called Exim had a hardwired
3202 limit of 10; it now uses the value of finduser_retries, which is used for
3203 all other user lookups.
3205 32. Added $received_count variable, available in data and not_smtp ACLs, and at
3208 33. Exim was neglecting to zero errno before one call of strtol() when
3209 expanding a string and expecting an integer value. On some systems this
3210 resulted in spurious "integer overflow" errors. Also, it was casting the
3211 result into an int without checking.
3213 34. Testing for a connection timeout using "timeout_connect" in the retry rules
3214 did not work. The code looks as if it has *never* worked, though it appears
3215 to have been documented since at least releast 1.62. I have made it work.
3217 35. The "timeout_DNS" error in retry rules, also documented since at least
3218 1.62, also never worked. As it isn't clear exactly what this means, and
3219 clearly it isn't a major issue, I have abolished the feature by treating it
3220 as "timeout", and writing a warning to the main and panic logs.
3222 36. The display of retry rules for -brt wasn't always showing the error code
3225 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
3226 timeout_connect_A, timeout_connect_MX.
3228 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
3229 to the empty sender.
3231 39. The daemon was not analysing the content of -oX till after it had closed
3232 stderr and disconnected from the controlling terminal. This meant that any
3233 syntax errors were only noted on the panic log, and the return code from
3234 the command was 0. By re-arranging the code a little, I've made the
3235 decoding happen first, so such errors now appear on stderr, and the return
3236 code is 1. However, the actual setting up of the sockets still happens in
3237 the disconnected process, so errors there are still only recorded on the
3240 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
3241 connections (as happens on some IP stacks) was logged at start up time as
3242 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
3243 it from "IPv6 and IPv4", which means that two separate sockets are being
3246 41. The debug output for gethostbyname2() or getipnodebyname() failures now
3247 says whether AF_INET or AF_INET6 was passed as an argument.
3249 42. Exiwhat output was messed up when time zones were included in log
3252 43. Exiwhat now gives more information about the daemon's listening ports,
3253 and whether -tls-on-connect was used.
3255 44. The "port" option of the smtp transport is now expanded.
3257 45. A "message" modifier in a "warn" statement in a non-message ACL was being
3258 silently ignored. Now an error message is written to the main and panic
3261 46. There's a new ACL modifier called "logwrite" which writes to a log file
3262 as soon as it is encountered.
3264 47. Added $local_user_uid and $local_user_gid at routing time.
3266 48. Exim crashed when trying to verify a sender address that was being
3269 49. Exim was recognizing only a space character after ".include". It now also
3270 recognizes a tab character.
3272 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
3273 extracting the relevant information from the specification. The man page no
3274 longer contains scrambled data for the -d option, and I've added a section
3275 at the front about calling Exim under different names.
3277 51. Added "extra_headers" argument to the "mail" command in filter files.
3279 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
3282 53. Installed eximstats 1.29.
3284 54. Added transport_filter_timeout as a generic transport option.
3286 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
3287 Cc: header lines. This was required by RFC 822, but it not required by RFC
3290 56. Exim used to add From:, Date:, and Message-Id: header lines to any
3291 incoming messages that did not have them. Now it does so only if the
3292 message originates locally, that is, if there is no associated remote host
3293 address. When Resent- header lines are present, this applies to the Resent-
3294 lines rather than the non-Resent- lines.
3296 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
3297 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
3299 58. Messages for configuration errors now include the name of the main
3300 configuration file - useful now that there may be more than one file in a
3301 list (.included file names were always shown).
3303 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
3304 for those rare installations that do not start the daemon as root or run it
3305 setuid root. I've cut out the call to initgroups() if the daemon is not
3308 60. The Exim user and group can now be bound into the binary as text strings
3309 that are looked up at the start of Exim's processing.
3311 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
3313 62. Added $mailstore_basename variable.
3315 63. Installed patch to sieve.c from Michael Haardt.
3317 64. When Exim failed to open the panic log after failing to open the main log,
3318 the original message it was trying to log was written to stderr and debug
3319 output, but if they were not available (the usual case in production), it
3320 was lost. Now it is written to syslog before the two lines that record the
3321 failures to open the logs.
3323 65. Users' Exim filters run in subprocesses under the user's uid. It is
3324 possible for a "deliver" command or an alias in a "personal" command to
3325 provoke an address rewrite. If logging of address rewriting is configured,
3326 this fails because the process is not running as root or exim. There may be
3327 a better way of dealing with this, but for the moment (because 4.30 needs
3328 to be released), I have disabled address rewrite logging when running a
3329 filter in a non-root, non-exim process.
3335 1. The buildconfig auxiliary program wasn't quoting the value set for
3336 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
3337 not defined. This bug was masked in 4.22 by the effect that was fixed in
3340 2. Some messages that were rejected after a message id was allocated were
3341 shown as "incomplete" by exigrep. It no longer does this for messages that
3342 are rejected by local_scan() or the DATA or non-SMTP ACLs.
3344 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
3345 have allow_domain_literals set, the ID did not get logged in the <= line.
3346 Domain literals are now always recognized in Message-ID: header lines.
3348 4. The first argument for a ${extract expansion item is the key name or field
3349 number. Leading and trailing spaces in this item were not being ignored,
3350 causing some misleading effects.
3352 5. When deliver_drop_privilege was set, single queue runner processes started
3353 manually (i.e. by the command "exim -q") or by the daemon (which uses the
3354 same command in the process it spins off) were not dropping privilege.
3356 6. When the daemon running as "exim" started a queue runner, it always
3357 re-executed Exim in the spun-off process. This is a waste of effort when
3358 deliver_drop_privilege is set. The new process now just calls the
3359 queue-runner function directly.
3365 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
3368 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
3369 ignored. Though the use of -oP was forcing the writing of a pid file, it
3370 was always written to the default place.
3372 3. If the message "no IP address found for host xxxx" is generated during
3373 incoming verification, it is now followed by identification of the incoming
3374 connection (so you can more easily find what provoked it).
3376 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
3378 5. Added some features to "harden" Exim a bit more against certain attacks:
3380 (a) There is now a build-time option called FIXED_NEVER_USERS that can
3381 be put in Local/Makefile. This is like the never_users runtime option,
3382 but it cannot be overridden. The default setting is "root".
3384 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
3385 prefix string with which any file named in a -C command line option
3388 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
3389 is retained for -C and -D only if the caller of Exim is root. Without
3390 it, the exim user may also use -C and -D and retain privilege.
3392 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
3393 command line option is disabled.
3395 6. Macro names set by the -D option must start with an upper case letter, just
3396 like macro names defined in the configuration file.
3398 7. Added "dereference=" facility to LDAP.
3400 8. Two instances of the typo "uknown" in the source files are fixed.
3402 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
3403 the Configure-Makefile script screwed up while processing it.
3405 10. Incorporated PCRE 4.4.
3407 11. The SMTP synchronization check was not operating right at the start of an
3408 SMTP session. For example, it could not catch a HELO sent before the client
3409 waited for the greeting. There is now a check for outstanding input at the
3410 point when the greeting is written. Because of the duplex, asynchronous
3411 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
3412 way, but not yet received, when the check is performed.
3414 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
3415 on TCP/IP sockets, because this apparently causes some broken clients to
3418 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
3419 unchanged) from the Cygwin maintainer.
3421 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
3422 was supported instead of testing for mbx. Effectively a typo.
3424 15. The spa authenticator server code was not checking that the input it
3425 received was valid base64.
3427 16. The debug output line for the "set" modifier in ACLs was not showing the
3428 name of the variable that was being set.
3430 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
3432 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
3433 lookup on its own. Something else was needed to trigger the lookup. For
3434 example, a match in host_lookup or the need for a host name in a host list.
3435 Now, if $sender_host_name is referenced and the host name has not yet been
3436 looked up, a lookup is performed. If the lookup fails, the variable remains
3437 empty, and $host_lookup_failed is set to "1".
3439 19. Added "eqi" as a case-independent comparison operator.
3441 20. The saslauthd authentication condition could segfault if neither service
3442 nor realm was specified.
3444 21. If an overflowing value such as "2048M" was set for message_size_limit, the
3445 error message that was logged was misleading, and incoming SMTP
3446 connections were dropped. The message is now more accurate, and temporary
3447 errors are given to SMTP connections.
3449 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
3450 (except RSET) with a 421 error, until QUIT is received. However, it was
3451 failing to send a response to QUIT.
3453 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
3454 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
3455 ACL is now run after the helo_try_verify_hosts code.
3457 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
3458 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
3459 case-independent, so other case variants are also recognized). Apparently
3460 some systems use these upper case variants.
3462 25. If more than two messages were waiting for the same host, and a transport
3463 filter was specified for the transport, Exim sent two messages over the
3464 same TCP/IP connection, and then failed with "socket operation on non-
3465 socket" when it tried to send the third.
3467 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
3469 27. The extern definition of crypt16() in expand.c was not being excluded when
3470 the OS had its own crypt16() function.
3472 28. Added bounce_return_body as a new option, and bounce_return_size_limit
3473 as a preferred synonym for return_size_limit, both as an option and as an
3476 29. Added LIBS=-liconv to OS/Makefile-OSF1.
3478 30. Changed the default configuration ACL to relax the local part checking rule
3479 for addresses that are not in any local domains. For these addresses,
3480 slashes and pipe symbols are allowed within local parts, but the sequence
3481 /../ is explicitly forbidden.
3483 31. SPA server authentication was not clearing the challenge buffer before
3486 32. log_message in a "warn" ACL statement was writing to the reject log as
3487 well as to the main log, which contradicts the documentation and doesn't
3488 seem right (because no rejection is happening). So I have stopped it.
3490 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
3491 However, I am unable to do any testing of this.
3493 34. Fixed an infelicity in the appendfile transport. When checking directories
3494 for a mailbox, to see if any needed to be created, it was accidentally
3495 using path names with one or more superfluous leading slashes; tracing
3496 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
3498 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
3499 contents are added to the log line that is written for every discarded
3500 recipient. (Previously a log_message setting was ignored.)
3502 36. The ${quote: operator now quotes the string if it is empty.
3504 37. The install script runs exim in order to find its version number. If for
3505 some reason other than non-existence or emptiness, which it checks, it
3506 could not run './exim', it was installing it with an empty version number,
3507 i.e. as "exim-". This error state is now caught, and the installation is
3510 38. An argument was missing from the function that creates an error message
3511 when Exim fails to connect to the socket for saslauthd authentication.
3512 This could cause Exim to crash, or give a corrupted message.
3514 39. Added isip, isip4, and isip6 to ${if conditions.
3516 40. The ACL variables $acl_xx are now saved with the message, and can be
3517 accessed later in routers, transports, and filters.
3519 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
3520 strings in the file are not string-expanded.
3522 42. If a MAIL command specified a SIZE value that was too large to fit into an
3523 int variable, the check against message_size_limit failed. Such values are
3524 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
3525 day this will have to be increased, but I don't think I want to be around
3526 when emails are that large.
3533 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
3534 iconv() is not standard in FreeBSD.
3536 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
3537 IPv6 enabled. The observed symptom was a segmentation fault on return from
3538 the function os_common_find_running_interfaces() in src/os.c.
3540 3. In the check_special_case() function in daemon.c I had used "errno" as an
3541 argument name, which causes warnings on some systems. This was basically a
3542 typo, since it was named "eno" in the comments!
3544 4. The code that waits for the clock to tick (at a resolution of some fraction
3545 of a second) so as to ensure message-id uniqueness was always waiting for
3546 at least one whole tick, when it could have waited for less. [This is
3547 almost certainly not relevant at current processor speeds, where it is
3548 unlikely to ever wait at all. But we try to future-proof.]
3550 5. The function that sleeps for a time interval that includes fractions of a
3551 second contained a race. It did not block SIGALRM between setting the
3552 timer, and suspending (a couple of lines later). If the interval was short
3553 and the sigsuspend() was delayed until after it had expired, the suspension
3554 never ended. On busy systems this could lead to processes getting stuck for
3557 6. Some uncommon configurations may cause a lookup to happen in a queue runner
3558 process, before it forks any delivery processes. The open lookup caching
3559 mechanism meant that the open file or database connection was passed into
3560 the delivery process. The problem was that delivery processes always tidy
3561 up cached lookup data. This could cause a problem for the next delivery
3562 process started by the queue runner, because the external queue runner
3563 process does not know about the closure. So the next delivery process
3564 still has data in the lookup cache. In the case of a file lookup, there was
3565 no problem because closing a file descriptor in a subprocess doesn't affect
3566 the parent. However, if the lookup was caching a connection to a database,
3567 the connection was closed, and the second delivery process was likely to
3568 see errors such as "PGSQL: query failed: server closed the connection
3569 unexpectedly". The problem has been fixed by closing all cached lookups
3570 in a queue runner before running a delivery process.
3572 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
3573 seem to have the "const" qualifier which it has on other OS. I've
3576 8. Change 4.21/2 was too strict. It is only if there are two authenticators
3577 *of the same type* (client or server) with the same public name that an
3578 error should be diagnosed.
3580 9. When Exim looked up a host name for an IP address, but failed to find the
3581 original IP address when looking up the host name (a safety check), it
3582 output the message "<ip address> does not match any IP for NULL", which was
3583 confusing, to say the least. The bug was that the host name should have
3584 appeared instead of "NULL".
3586 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
3587 user that is built into the binary, and the -C or -D options is used, root
3588 privilege is dropped before the configuration file is read. In addition,
3589 logging is switched to stderr instead of the normal log files. If the
3590 configuration then re-defines the Exim user, the unprivileged environment
3591 is probably not what is expected, so Exim logs a panic warning message (but
3594 However, if deliver_drop_privilege is set, the unprivileged state may well
3595 be exactly what is intended, so the warning has been cut out in that case,
3596 and Exim is allowed to try to write to its normal log files.
3602 1. smtp_return_error_details was not giving details for temporary sender
3603 or receiver verification errors.
3605 2. Diagnose a configuration error if two authenticators have the same public
3608 3. Exim used not to create the message log file for a message until the first
3609 delivery attempt. This could be confusing when incoming messages were held
3610 for policy or load reasons. The message log file is now created at the time
3611 the message is received, and an initial "Received" line is written to it.
3613 4. The automatically generated man page for command line options had a minor
3614 bug that caused no ill effects; however, a more serious problem was that
3615 the procedure for building the man page automatically didn't always
3616 operate. Consequently, release 4.20 contains an out-of-date version. This
3617 shouldn't happen again.
3619 5. When building Exim with embedded Perl support, the script that builds the
3620 Makefile was calling 'perl' to find its compile-time parameters, ignoring
3621 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
3623 6. The freeze_tell option was not being used for messages that were frozen on
3624 arrival, either by an ACL or by local_scan().
3626 7. Added the smtp_incomplete_transaction log selector.
3628 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
3629 was accepting AUTH without a new EHLO.
3631 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
3632 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
3635 10. Logging of TCP/IP connections (when configured) now happens in the main
3636 daemon process instead of the child process, so that the TCP/IP connection
3637 count is more accurate (but it can never be perfect).
3639 11. The use of "drop" in a nested ACL was not being handled correctly in the
3640 outer ACL. Now, if condition failure induced by the nested "drop" causes
3641 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
3642 or "require"), the connection is dropped.
3644 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
3645 that yield "discard" can now be used with an "accept" or a "discard" verb,
3646 but an error is generated for any others (because I can't see a useful way
3647 to define what should happen).
3649 13. When an ACL is read dynamically from a file (or anywhere else), the lines
3650 are now processed in the same way as lines in the Exim configuration file.
3651 In particular, continuation lines are supported.
3653 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
3655 15. Added -ti meaning -t -i.
3657 16. Check for letters, digits, hyphens, and dots in the names of dnslist
3658 domains, and warn by logging if others are found.
3660 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
3661 returned from GIFCONF when Exim is trying to find the list of interfaces on
3662 a host. The code in os.c has been modified to copy each ifreq to an aligned
3663 structure in all cases.
3665 Also, in some cases, the returned ifreq's were being copied to a 'struct
3666 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
3667 means the last couple of bytes of an IPv6 address could be chopped if the
3668 ifreq contained only a normal sockaddr (14 bytes storage).
3670 18. Named domain lists were not supported in the hosts_treat_as_local option.
3671 An entry such as +xxxx was not recognized, and was treated as a literal
3674 19. Ensure that header lines added by a DATA ACL are included in the reject log
3675 if the ACL subsequently rejects the message.
3677 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
3678 MD5 (which is deprecated).
3680 21. When testing a filter file using -bf, Exim was writing a message when it
3681 took the sender from a "From " line in the message, but it was not doing so
3682 when it took $return_path from a Return-Path: header line. It now does.
3684 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
3685 with a valid header line field name (a series of printing characters
3686 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
3688 23. Changed "disc" in the source to "disk" to conform to the documentation and
3689 the book and for uniformity.
3691 24. Ignore Sendmail's -Ooption=value command line item.
3693 25. When execve() failed while trying to run a command in a pipe transport,
3694 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
3695 could be confused with a return value of 69 from the command itself. This
3696 has been changed to 127, the value the shell returns if it is asked to run
3697 a non-existent command. The wording for the related log line suggests a
3698 non-existent command as the problem.
3700 26. If received_header_text expands to an empty string, do not add a Received:
3701 header line to the message. (Well, it adds a token one on the spool, but
3702 marks it "old" so that it doesn't get used or transmitted.)
3704 27. Installed eximstats 1.28 (addition of -nt option).
3706 28. There was no check for failure on the call to getsockname() in the daemon
3707 code. This can fail if there is a shortage of resources on the system, with
3708 ENOMEM, for example. A temporary error is now given on failure.
3710 29. Contrary to the C standard, it seems that in some environments, the
3711 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
3712 program. Exim now does this explicitly; it affects the formatting of
3713 timestamps using strftime().
3715 30. If exiqsumm was given junk data, it threw up some uninitialized variable
3716 complaints. I've now initialized all the variables, to avoid this.
3718 32. Header lines added by a system filter were not being "seen" during
3719 transport-time rewrites.
3721 33. The info_callback() function passed to OpenSSL is set up with type void
3722 (*)(SSL *, int, int), as described somewhere. However, when calling the
3723 function (actually a macro) that sets it up, the type void(*)() is
3724 expected. I've put in a cast to prevent warnings from picky compilers.
3726 34. If a DNS black list lookup found a CNAME record, but there were no A
3727 records associated with the domain it pointed at, Exim crashed.
3729 35. If a DNS black list lookup returned more than one A record, Exim ignored
3730 all but the first. It now scans all returned addresses if a particular IP
3731 value is being sought. In this situation, the contents of the
3732 $dnslist_value variable are a list of all the addresses, separated by a
3735 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
3736 to accept a host name and all its aliases if the forward lookup for any of
3737 them yielded the IP address of the incoming connection. Now it accepts only
3738 those names whose forward lookup yields the correct IP address. Any other
3739 names are discarded. This closes a loophole whereby a rogue DNS
3740 administrator could create reverse DNS records to break through a
3741 wildcarded host restriction in an ACL.
3743 37. If a user filter or a system filter that ran in a subprocess used any of
3744 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
3745 the wrong values were passed to the pipe command ($thisaddress had the
3746 value of $0, $0 had the value of $1, etc). This bug was introduced by
3747 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
3749 38. Improved the line breaking for long SMTP error messages from ACLs.
3750 Previously, if there was no break point between 40 and 75 characters, Exim
3751 left the rest of the message alone. Two changes have been made: (a) I've
3752 reduced the minimum length to 35 characters; (b) if it can't find a break
3753 point between 35 and 75 characters, it looks ahead and uses the first one
3754 that it finds. This may give the occasional overlong line, but at least the
3755 remaining text gets split now.
3757 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
3758 file descriptors might be low, and that setting 1000 would always raise it.
3759 It turns out that in some environments, the limit is already over 1000 and
3760 that lowering it causes trouble. So now Exim takes care not to decrease it.
3762 40. When delivering a message, the value of $return_path is set to $sender_
3763 address at the start of routing (routers may change the value). By an
3764 oversight, this default was not being set up when an address was tested by
3765 -bt or -bv, which affected the outcome if any router or filter referred to
3768 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
3769 log only when "message" or "log_message" are set. However, if one of the
3770 conditions was an address verification, or a call to a nested ACL, the
3771 messages generated by the underlying test were being passed through. This
3772 no longer happens. The underlying message is available in $acl_verify_
3773 message for both "message" and "log_message" expansions, so it can be
3774 passed through if needed.
3776 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
3777 new expansion $bh_ to give the encoded byte string without charset
3778 translation. Translation happens only if iconv() is available; HAVE_ICONV
3779 indicates this at build time. HEADERS_CHARSET gives the charset to
3780 translate to; headers_charset can change it in the configuration, and
3781 "headers charset" can change it in an individual filter file.
3783 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
3784 that creates RFC 2047 encoded "words" labels them as that charset instead
3785 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
3786 expansion operator; (ii) when Exim creates a From: line for a local
3787 message; (iii) when a header line is rewritten to include a "phrase" part.
3789 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
3790 buggy, causing it to skip the first lines of messages whose message ID
3791 ended in 'D'. This would not have bitten before Exim release 4.14, because
3792 message IDs were unlikely to end in 'D' before then. The effect was to have
3793 incorrect size information for certain domains.
3795 45. #include "config.h" was missing at the start of the crypt16.c module. This
3796 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
3799 46. If there was a timeout during a "random" callout check, Exim treated it as
3800 a failure of the random address, and carried on sending RSET and the real
3801 address. If the delay was just some slowness somewhere, the response to the
3802 original RCPT would be taken as a response to RSET and so on, causing
3803 mayhem of various kinds.
3805 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
3806 when I implemented it. It didn't allow for the fact that some option values
3807 may legitimatetly be negative (e.g. size_addition), and it didn't even do
3808 the right test for positive values.
3810 48. Domain names in DNS records are case-independent. Exim always looks them up
3811 in lower case. Some resolvers return domain names in exactly the case they
3812 appear in the zone file, that is, they may contain uppercase letters. Not
3813 all resolvers do this - some return always lower case. Exim was treating a
3814 change of case by a resolver as a change of domain, similar to a widening
3815 of a domain abbreviation. This triggered its re-routing code and so it was
3816 trying to route what was effectively the same domain again. This normally
3817 caused routing to fail (because the router wouldn't handle the domain
3818 twice). Now Exim checks for this case specially, and just changes the
3819 casing of the domain that it ultimately uses when it transmits the message
3822 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3825 50. If a filter generated a file delivery with a non-absolute name (possible if
3826 no home directory exists for the router), the forbid_file option was not
3829 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3830 the existing equality matching.
3832 52. Exim was using ints instead of ino_t variables in some places where it was
3833 dealing with inode numbers.
3835 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3836 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3837 called TMPDIR, and if it finds it is different, it changes its value.
3839 54. The smtp_printf() function is now made available to local_scan() so
3840 additional output lines can be written before returning. There is also an
3841 smtp_fflush() function to enable the detection of a dropped connection.
3842 The variables smtp_input and smtp_batched_input are exported to
3845 55. Changed the default runtime configuration: the message "Unknown user"
3846 has been removed from the ACL, and instead placed on the localuser router,
3847 using the cannot_route_message feature. This means that any verification
3848 failures that generate their own messages won't get overridden. Similarly,
3849 the "Unrouteable address" message that was in the ACL for unverifiable
3850 relay addresses has also been removed.
3852 56. Added hosts_avoid_esmtp to the smtp transport.
3854 57. The exicyclog script was not checking for the esoteric option
3855 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3856 will work only if exicyclog is run under the appropriate euid.
3858 58. Following a discussion on the list, the rules by which Exim recognises line
3859 endings on incoming messages have been changed. The -dropcr and drop_cr
3860 options are now no-ops, retained only for backwards compatibility. The
3861 following line terminators are recognized: LF CRLF CR. However, special
3862 processing applies to CR:
3864 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3865 nor a local message in the state where . is a terminator.
3867 (ii) If a bare CR is encountered in a header line, an extra space is added
3868 after the line terminator so as not to end the header. The reasoning
3869 behind this is that bare CRs in header lines are most likely either
3870 to be mistakes, or people trying to play silly games.
3872 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3873 was being incorrectly given as 18 bytes larger than it should have been.
3874 This is a VOB (very old bug).
3876 60. This may never have affected anything current, but just in case it has:
3877 When the local host is found other than at the start of a list of hosts,
3878 the local host, those with the same MX, and any that follow, are discarded.
3879 When the list in question was part of a longer list of hosts, the following
3880 hosts (not currently being processed) were also being discarded. This no
3881 longer happens. I'm not sure if this situation could ever has previously
3884 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3887 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3888 header. This is something that is recommended in a new Internet Draft, and
3889 is something that is documented as being done by Sendmail. There are two
3890 possible values. For messages generated by the autoreply transport, Exim
3893 Auto-Submitted: auto-replied
3895 whereas for all other generated messages (e.g. bounces) it adds
3897 Auto-Submitted: auto-generated
3899 63. The "personal" condition in filters now includes a test for the
3900 Auto-Submitted: header. If it contains the string "auto-" the message it
3901 not considered personal.
3903 64. Added rcpt_include_affixes as a generic transport option.
3905 65. Added queue_only_override (default true).
3907 66. Added the syslog_duplication option.
3909 67. If what should have been the first header line of a message consisted of
3910 a space followed by a colon, Exim was mis-interpreting it as a header line.
3911 It isn't of course - it is syntactically invalid and should therefore be
3912 treated as the start of the message body. The misbehaviour could have
3913 caused a number of strange effects, including loss of data in subsequent
3914 header lines, and spool format errors.
3916 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3917 client host had authenticated. This control can now be exercised by an ACL
3918 for more flexibility.
3920 69. By default, callouts do not happen when testing with -bh. There is now a
3921 variant, -bhc, which does actually run the callout code, including
3922 consulting and updating the callout cache.
3924 70. Added support for saslauthd authentication, courtesy of Alexander
3927 71. If statvfs() failed on the spool or log directories while checking their
3928 size for availability, Exim confusingly gave the error "space shortage".
3929 Furthermore, in debugging mode it crashed with a floating point exception.
3930 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3931 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3932 serious problem, Exim now writes to the main and panic logs when this
3933 happens, with details of the failure. It then refuses to accept the
3934 incoming message, giving the message "spool directory problem" or "log
3935 directory problem" with a 421 code for SMTP messages.
3937 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3938 0, 1, and 2 exist, because some OS complain for execs without them (see
3939 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3940 descriptors. However, the code omitted to check that the open succeeded,
3941 causing mysterious errors if for some reason the permissions on /dev/null
3942 got screwed. Now Exim writes a message to the main and panic logs, and
3943 bombs out if it can't open /dev/null.
3945 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3946 interact so that it is all more flexible. It is supposed to remain
3947 backwards compatible. Also added extra_local_interfaces.
3949 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3950 to bomb out with an assertion failure - to the client this appears as a
3951 connection drop. This problem occurs in the part of the code that was taken
3952 from the Samba project. Fortunately, the assertion is in a very simple
3953 function, so I have fixed this by reproducing the function inline in the
3954 one place where it is called, and arranging for authentication to fail
3955 instead of killing the process with assert().
3957 75. The SPA client code was not working when the server requested OEM rather
3958 than Unicode encoding.
3960 76. Added code to make require_files with a specific uid setting more usable in
3961 the case where statting the file as root fails - usually a non-root-mounted
3962 NFS file system. When this happens and the failure is EACCES, Exim now
3963 forks a subprocess and does the per-uid checking as the relevant uid.
3965 77. Added process_log_path.
3967 78. If log_file_path was not explicitly set, a setting of check_log_space or
3968 check_log_inodes was ignored.
3970 79. If a space check for the spool or log partitions fails, the incident is now
3971 logged. Of course, in the latter case the data may get lost...
3973 80. Added the %p formatting code to string_format() so that it can be used to
3974 print addresses in debug_print(). Adjusted all the address printing in the
3975 debugging in store.c to use %p rather than %d.
3977 81. There was a concern that a line of code in smtp_in.c could overflow a
3978 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3979 initially expressed, the concern was not well-founded, because trailing
3980 spaces are removed early. However, if the trailing spaces were followed by
3981 a NULL, they did not get removed, so the overflow was possible. Two fixes
3984 (a) I re-wrote the offending code in a cleaner fashion.
3985 (b) If an incoming SMTP command contains a NULL character, it is rejected
3988 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3989 runs initgroups(), so that if the Exim user is in any additional groups,
3990 they will be used during message reception.
3996 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / heiko / exim.git / blob