1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.187 2005/08/01 14:41:25 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 TK/01 Added the "success_on_redirect" address verification option. See
10 NewStuff for rationale and an example.
12 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
14 PH/02 Patch to exigrep to allow it to work on syslog lines.
16 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
17 fread() to skip over the body file's header line, because in Cygwin the
18 header line is locked and is inaccessible.
20 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
21 co-exist for some time) to make it clear that it is the Exim ID that is
22 referenced, not the Message-ID: header line.
24 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
25 string_format() function, because snprintf() does not exist on all
32 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
34 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
35 command itself ended in error, the underlying message about the transport
36 filter timeout was being overwritten with the pipe command error. Now the
37 underlying error message should be appended to the second error message.
39 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
41 PH/02 Reduce the amount of output that "make" produces by default. Full output
42 can still be requested.
44 PH/03 The warning log line about a condition test deferring for a "warn" verb
45 was being output only once per connection, rather than after each
46 occurrence (because it was using the same function as for successful
47 "warn" verbs). This seems wrong, so I have changed it.
49 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
50 it should not have, which might have caused a crash in the right
51 circumstances, but probably never did.
53 PH/04 Installed a modified version of Tony Finch's patch to make submission
54 mode fix the return path as well as the Sender: header line, and to
55 add a /name= option so that you can make the user's friendly name appear
58 TF/03 Added the control = fakedefer ACL modifier.
60 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
61 Mark Lowes for thorough testing.
63 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
65 TK/03 Merged latest SRS patch from Miles Wilton.
67 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
68 with the definition in sysexits.h (which is #included earlier).
69 Fortunately, Exim does not actually use EX_OK. The code used to try to
70 preserve the sysexits.h value, by assumimg that macro definitions were
71 scanned for macro replacements. I have been disabused of this notion,
72 so now the code just undefines EX_OK before #including unistd.h.
74 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
75 in the smtp transport. When a block could not be written in a single
76 write() function, the timeout was being re-applied to each part-write.
77 This seems wrong - if the receiver was accepting one byte at a time it
78 would take for ever. The timeout is now adjusted when this happens. It
79 doesn't have to be particularly precise.
81 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
82 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
84 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
85 without a domain if the check to <postmaster@domain> fails.
87 SC/01 Eximstats: added -xls and the ability to specify output files
88 (patch written by Frank Heydlauf).
90 SC/02 Eximstats: use FileHandles for outputing results.
92 SC/03 Eximstats: allow any combination of xls, txt, and html output.
94 SC/04 Eximstats: fixed display of large numbers with -nvr option
96 SC/05 Eximstats: fixed merging of reports with empty tables.
98 SC/06 Eximstats: added the -include_original_destination flag
100 SC/07 Eximstats: removed tabs and trailing whitespace.
102 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
104 TK/06 MBOX spool code: Add real "From " MBOX separator line
105 so the .eml file is really in mbox format (even though
106 most programs do not really care). Patch from Alex Miller.
108 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
109 The latter is generated from $received_to and is only set if the
110 message has one envelope recipient. SA can use these headers,
111 obviously out-of-the-box. Patch from Alex Miller.
113 PH/08 The ${def test on a variable was returning false if the variable's
114 value was "0", contrary to what the specification has always said!
115 The result should be true unless the variable is empty.
117 PH/09 The syntax error of a character other than { following "${if
118 def:variable_name" (after optional whitespace) was not being diagnosed.
119 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
120 accidental colon was present, for example, could give incorrect results.
122 PH/10 Tidied the code in a number of places where the st_size field of a stat()
123 result is used (not including appendfile, where other changes are about
126 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
127 This involved changing a lot of size variables from int to off_t. It
128 should work with maildirs and everything.
130 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
131 spamd dying while we are connected to it.
133 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
136 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
137 function for BDB 4.3.
139 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
141 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
142 the sources for PCRE have been re-arranged and more files are now
145 PH/15 The code I had for printing potentially long long variables in PH/11
146 above was not the best (it lost precision). The length of off_t variables
147 is now inspected at build time, and an appropriate printing format (%ld
148 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
149 to be "long long int" or "long int". This is needed for the internal
150 formatting function string_vformat().
152 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
153 the configuration file to be ":syslog", then the script "guesses" where
154 the logs files are, rather than using the compiled in default. In our
155 case the guess is not the same as the compiled default, so the script
156 suddenly stopped working when I started to use syslog. The patch checks
157 to see if log_file_path is "". If so, it attempts to read it from exim
158 with no configuration file to get the compiled in version, before it
159 falls back to the previous guessing code."
161 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
162 implementing BATV in an Exim configuration. See NewStuff for the gory
165 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
166 Makefile that are specific to HP-UX.
168 PH/18 If the "use_postmaster" option was set for a recipient callout together
169 with the "random" option, the postmaster address was used as the MAIL
170 FROM address for the random test, but not for the subsequent recipient
171 test. It is now used for both.
173 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
174 patch removes a few documentation additions to RFC 3028, because the
175 latest draft now contains them. It adds the new en;ascii-case comparator
176 and a new error check for 8bit text in MIME parts. Comparator and
177 require names are now matched exactly. I enabled the subaddress
178 extension, but it is not well tested yet (read: it works for me)."
180 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
181 rework some of the code of TK/09 above to avoid the hardwired use of
182 "%lld" and "long long". Replaced the call to snprintf() with a call to
185 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
186 records point to non-existent hosts", "retry timeout exceeded", and
187 "retry time not reached for any host after a long failure period".
189 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
190 experimental DomainKeys support:
192 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
193 (2) On an error such as an illegally used "control", the wrong name for
194 the control was given.
196 These problems did NOT occur unless DomainKeys support was compiled.
198 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
200 PH/24 Added ${if match_ip condition.
202 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
203 databases so that it will be absolutely obvious if a crash occurs in the
204 DB library. This is a regular occurrence (often caused by mis-matched
207 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
208 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
209 picked up on a user's system that detects such things. There doesn't seem
210 to be a gcc warning option for this - only an attribute that has to be
211 put on the function's prototype. It seems that in Fedora Core 4 they have
212 set this on a number of new functions. No doubt there will be more in due
215 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
216 specify a transport. However, if an address that was verified by such a
217 router was the subject of a callout, Exim crashed because it tried to
218 read the rcpt_include_affixes from the non-existent transport. Now it
219 just assumes that the setting of that option is false. This bug was
220 introduced by 4.51/PH/31.
222 PH/28 Changed -d+all to exclude +memory, because that information is very
223 rarely of interest, but it makes the output a lot bigger. People tend to
224 do -d+all out of habit.
226 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
227 code in os-type was giving problems when libc.so lives in lib64, like on
230 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
231 aren't the modern standard, and it seems that some systems' include files
232 don't always have them. Exim was already checking for some of the newer
233 ones like T_AAAA, and defining it itself. I've added checks for all the
234 record types that Exim uses.
236 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
237 not automatically generating a new one, as it is supposed to. This
238 prevented TLS from working. If the file did exist, but contained invalid
239 data, a new version was generated, as expected. It was only the case of a
240 non-existent file that was broken.
242 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
243 with a change in libdomainkeys > 0.64.
245 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
246 from DNS. If the selector record carries the flag, it now has
247 precedence over the domain-wide flag.
249 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
251 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
252 the use of a port name that isn't defined in /etc/services) occurred, the
253 message was deferred as in a normal delivery, and thus remained on the
254 spool, instead of being failed because of the mua_wrapper setting. This
255 is now fixed, and I tidied up some of the mua_wrapper messages at the
258 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
259 the messages in a hash of arrays rather than using individual hashes.
260 This is a bit cleaner and results in dramatic memory savings, albeit
261 at a slight CPU cost.
263 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
264 as requested by Marc Sherman.
266 SC/10 Eximstats: added histograms for user specified patterns as requested
269 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
271 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
272 fopen() in the content-scanning modules that did not already have it.
278 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
279 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
281 TK/02 Fix ACL "control" statement not being available in MIME ACL.
283 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
285 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
286 to test Sieve filters that use "vacation".
288 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
289 that changes the way the GnuTLS parameters are stored in the cache file.
290 The new format can be generated externally. For backward compatibility,
291 if the data in the cache doesn't make sense, Exim assumes it has read an
292 old-format file, and it generates new data and writes a new file. This
293 means that you can't go back to an older release without removing the
296 PH/03 A redirect router that has both "unseen" and "one_time" set does not
297 work if there are any delivery delays because "one_time" forces the
298 parent to be marked "delivered", so its unseen clone is never tried
299 again. For this reason, Exim now forbids the simultaneous setting of
302 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
303 redirected to themselves ("homonym" addresses). Read the long ChangeLog
304 entry if you want to know the details. The fix, however, neglected to
305 consider the case when local delivery batching is involved. The test for
306 "previously delivered" was not happening when checking to see if an
307 address could be batched with a previous (undelivered) one; under
308 certain circumstances this could lead to multiple deliveries to the same
311 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
312 in its include files, and this causes problems building Exim.
314 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
315 header_syntax) but Exim was just ignoring anything given after a slash.
316 In particular, this caused confusion with an attempt to use "verify =
317 reverse_host_lookup/defer_ok". An error is now given when options are
318 supplied for verify items that do not have them. (Maybe reverse_host_
319 lookup should have a defer_ok option, but that's a different point.)
321 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
322 defined by RFC 821) to 2048, because there were problems with some AUTH
323 commands, and RFC 1869 says the size should be increased for extended
324 SMTP commands that take arguments.
326 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
329 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
330 "unknown" error; now it says that the functionality isn't in the binary.
332 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
333 an address' error message when a string expansion fails (syntax or
334 whatever). Otherwise the password may appear in the log. Following change
335 PH/42 below, there is no longer a chance of it appearing in a bounce
338 PH/11 Installed exipick version 20050225.0 from John Jetmore.
340 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
341 of its addresses was ever tried. (Bugzilla bug #2.)
343 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
344 the result incorrectly in the debug output. (It correctly added a newline
345 to what was transported.)
347 TF/01 Added $received_time.
349 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
350 commented out examples of how to interface to a virus scanner and to
351 SpamAssassin. Also added commented examples of av_scanner and
352 spamd_address settings.
354 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
355 and controls are allowed in which ACLs. There were a couple of minor
356 errors. Some of the entries in the conditions table (which is a table of
357 where they are NOT allowed) were getting very unwieldy; rewrote them as a
358 negation of where the condition IS allowed.
360 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
362 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
363 header file does not have a version number, so I've had to invent a new
364 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
365 API. The code is untested by me (my Linux distribution still has 0.3.2 of
366 radiusclient), but it was contributed by a Radius user.
368 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
369 files or directories, for OpenSSL.
371 PH/19 When an Exim process that is running as root has to create an Exim log
372 file, it does so in a subprocess that runs as exim:exim so as to get the
373 ownership right at creation (otherwise, other Exim processes might see
374 the file with the wrong ownership). There was no test for failure of this
375 fork() call, which would lead to the process getting stuck as it waited
376 for a non-existent subprocess. Forks do occasionally fail when resources
377 run out. I reviewed all the other calls to fork(); they all seem to check
380 PH/20 When checking for unexpected SMTP input at connect time (before writing
381 the banner), Exim was not dealing correctly with a non-positive return
382 from the read() function. If the client had disconnected by this time,
383 the result was a log entry for a synchronization error with an empty
384 string after "input=" when read() returned zero. If read() returned -1
385 (an event I could not check), uninitialized data bytes were printed.
386 There were reports of junk text (parts of files, etc) appearing after
389 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
391 PH/22 Added support for macro redefinition, and (re)definition in between
392 driver and ACL definitions.
394 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
395 forgetting to use the resulting value; it was using the unexpanded value.
397 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
398 hadn't been configured. The fix is from Juergen Kreileder, who
399 understands it better than I do:
401 "Here's what I see happening with three configured cyrus_sasl
402 authenticators configured (plain, login, cram-md5):
404 On startup auth_cyrus_sasl_init() gets called for each of these.
405 This means three calls to sasl_listmech() without a specified mech_list.
406 => SASL tests which mechs of all available mechs actually work
407 => three warnings about OTP not working
408 => the returned list contains: plain, login, cram-md5, digest-md5, ...
410 With the patch, sasl_listmech() also gets called three times. But now
411 SASL's mech_list option is set to the server_mech specified in the the
412 authenticator. Or in other words, the answer from sasl_listmech()
413 gets limited to just the mech you're testing for (which is different
415 => the return list contains just 'plain' or 'login', 'cram-md5' or
416 nothing depending on the value of ob->server_mech.
418 I've just tested the patch: Authentication still works fine,
419 unavailable mechs specified in the exim configuration are still
420 caught, and the auth.log warnings about OTP are gone."
422 PH/25 When debugging is enabled, the contents of the command line are added
423 to the debugging output, even when log_selector=+arguments is not
426 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
427 answer is "GNU", and only if the return is "GNU/something" is the answer
430 PH/27 $acl_verify_message is now set immediately after the failure of a
431 verification in an ACL, and so is available in subsequent modifiers. In
432 particular, the message can be preserved by coding like this:
434 warn !verify = sender
435 set acl_m0 = $acl_verify_message
437 Previously, $acl_verify_message was set only while expanding "message"
438 and "log_message" when a very denied access.
440 PH/28 Modified OS/os.c-Linux with
442 -#ifndef OS_LOAD_AVERAGE
443 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
445 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
446 nomenclature these days.)
448 PH/29 Installed patch from the Sieve maintainer that adds the options
449 sieve_useraddress and sieve_subaddress to the redirect router.
451 PH/30 In these circumstances:
452 . Two addresses routed to the same list of hosts;
453 . First host does not offer TLS;
454 . First host accepts first address;
455 . First host gives temporary error to second address;
456 . Second host offers TLS and a TLS session is established;
457 . Second host accepts second address.
458 Exim incorrectly logged both deliveries with the TLS parameters (cipher
459 and peerdn, if requested) that were in fact used only for the second
462 PH/31 When doing a callout as part of verifying an address, Exim was not paying
463 attention to any local part prefix or suffix that was matched by the
464 router that accepted the address. It now behaves in the same way as it
465 does for delivery: the affixes are removed from the local part unless
466 rcpt_include_affixes is set on the transport.
468 PH/32 Add the sender address, as F=<...>, to the log line when logging a
469 timeout during the DATA phase of an incoming message.
471 PH/33 Sieve envelope tests were broken for match types other than :is. I have
472 applied a patch sanctioned by the Sieve maintainer.
474 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
475 the uid or gid is negative. A case of a negative gid caused this to be
476 noticed. The fix allows for either to be negative.
478 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
479 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
482 PH/36 The change PH/12 above was broken. Fixed it.
484 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
485 the grounds that routing the same address twice would always produce the
486 same answer. This might have been true once, but it is certainly no
487 longer true now. Routing a child address may depend on the previous
488 routing that produced that child. Some complicated redirection strategies
489 went wrong when messages had multiple recipients, and made Exim's
490 behaviour dependent on the order in which the addresses were given.
492 I have moved the duplicate checking until after the routing is complete.
493 Exim scans the addresses that are assigned to local and remote
494 transports, and removes any duplicates. This means that more work will be
495 done, as duplicates will always all be routed, but duplicates are
496 presumably rare, so I don't expect this is of any significance.
498 For deliveries to pipes, files, and autoreplies, the duplicate checking
499 still happens during the routing process, since they are not going to be
502 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
503 It corrects a timeout issue with spamd. This is Ian's comment: "The
504 background is that sometimes spamd either never reads data from a
505 connection it has accepted, or it never writes response data. The exiscan
506 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
507 blindly assumes that writes won't block so it may never time out."
509 PH/39 Allow G after quota size as well as K and M.
511 PH/40 The value set for $authenticated_id in an authenticator may not contain
512 binary zeroes or newlines because the value is written to log lines and
513 to spool files. There was no check on this. Now the value is run through
514 the string_printing() function so that such characters are converted to
515 printable escape sequences.
517 PH/41 $message_linecount is a new variable that contains the total number of
518 lines in the message. Compare $body_linecount, which is the count for the
521 PH/42 Exim no longer gives details of delivery errors for specific addresses in
522 bounce and delay warning messages, except in certain special cases, which
525 (a) An SMTP error message from a remote host;
526 (b) A message specified in a :fail: redirection;
527 (c) A message specified in a "fail" command in a system filter;
528 (d) A message specified in a FAIL return from the queryprogram router;
529 (e) A message specified by the cannot_route_message router option.
531 In these cases only, Exim does include the error details in bounce and
532 warning messages. There are also a few cases where bland messages such
533 as "unrouteable address" or "local delivery error" are given.
535 PH/43 $value is now also set for the "else" part of a ${run expansion.
537 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
538 being worked on, but at least Exim now implements the latest version to
541 PH/45 In a pipe transport, although a timeout while waiting for the pipe
542 process to complete was treated as a delivery failure, a timeout while
543 writing the message to the pipe was logged, but erroneously treated as a
544 successful delivery. Such timeouts include transport filter timeouts. For
545 consistency with the overall process timeout, these timeouts are now
546 treated as errors, giving rise to delivery failures by default. However,
547 there is now a new Boolean option for the pipe transport called
548 timeout_defer, which, if set TRUE, converts the failures into defers for
549 both kinds of timeout. A transport filter timeout is now identified in
552 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
553 systems where "make" and "gmake" are different, calling "gmake" at top
554 level broke things. I've arranged for the value of $(MAKE) to be passed
555 from the Makefile to this script so that it can call the same version of
559 A note about Exim versions 4.44 and 4.50
560 ----------------------------------------
562 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
563 changes of various kinds. As a consequence, a big documentation update was
564 needed. This delayed the release for rather longer than seemed good, especially
565 in the light of a couple of (minor) security issues. Therefore, the changes
566 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
567 release. So 4.44 and 4.50 are in effect two different branches that both start
570 I have left the 4.50 change log unchanged; it contains all the changes since
571 4.43. The change log for 4.44 is below; many of its items are identical to
572 those for 4.50. This seems to be the most sensible way to preserve the
573 historical information.
579 1. Minor wording change to the doc/README.SIEVE file.
581 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
582 computation of the current number of files was incorrect.
584 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
585 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
588 4. Give more explanation in the error message when the command for a transport
589 filter fails to execute.
591 5. There are several places where Exim runs a non-Exim command in a
592 subprocess. The SIGUSR1 signal should be disabled for these processes. This
593 was being done only for the command run by the queryprogram router. It is
594 now done for all such subprocesses. The other cases are: ${run, transport
595 filters, and the commands run by the lmtp and pipe transports.
597 6. Added CONFIGURE_GROUP build-time option.
599 7. Some older OS have a limit of 256 on the maximum number of file
600 descriptors. Exim was using setrlimit() to set 1000 as a large value
601 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
602 systems. I've change it so that if it can't get 1000, it tries for 256.
604 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
605 was an oversight, and furthermore, ever since the addition of extra
606 controls (e.g. 4.43/32), the checks on when to allow different forms of
607 "control" were broken. There should now be diagnostics for all cases when a
608 control that does not make sense is encountered.
610 9. Added the /retain_sender option to "control=submission".
612 10. $recipients is now available in the predata ACL (oversight).
614 11. Tidy the search cache before the fork to do a delivery from a message
615 received from the command line. Otherwise the child will trigger a lookup
616 failure and thereby defer the delivery if it tries to use (for example) a
617 cached ldap connection that the parent has called unbind on.
619 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
620 of $address_data from the recipient verification was clobbered by the
623 13. The value of address_data from a sender verification is now available in
624 $sender_address_data in subsequent conditions in the ACL statement.
626 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
628 15. Added a new option "connect=<time>" to callout options, to set a different
631 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
632 was its contents. (It was OK if the option was not defined at all.)
634 17. A "Completed" log line is now written for messages that are removed from
635 the spool by the -Mrm option.
637 18. New variables $sender_verify_failure and $recipient_verify_failure contain
638 information about exactly what failed.
640 19. Added -dd to debug only the daemon process.
642 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
643 handles timeouts, both on the server side and network timeouts. Renamed the
644 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
646 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
647 It is now set to "smtps".
649 22. $host_address is now set to the target address during the checking of
652 23. When checking ignore_target_hosts for an ipliteral router, no host name was
653 being passed; this would have caused $sender_host_name to have been used if
654 matching the list had actually called for a host name (not very likely,
655 since this list is usually IP addresses). A host name is now passed as
658 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
659 code that specifies a non-restarting handler (typically sigaction() in
660 modern systems) in an attempt to fix a rare and obscure crash bug.
662 25. Narrowed the window for a race in the daemon that could cause it to ignore
663 SIGCHLD signals. This is not a major problem, because they are used only to
664 wake it up if nothing else does.
666 26. A malformed maildirsize file could cause Exim to calculate negative values
667 for the mailbox size or file count. Odd effects could occur as a result.
668 The maildirsize information is now recalculated if the size or filecount
671 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
672 support for a long time. Removed HAVE_SYS_VFS_H.
674 28. Installed the latest version of exipick from John Jetmore.
676 29. In an address list, if the pattern was not a regular expression, an empty
677 subject address (from a bounce message) matched only if the pattern was an
678 empty string. Non-empty patterns were not even tested. This was the wrong
679 because it is perfectly reasonable to use an empty address as part of a
680 database query. An empty address is now tested by patterns that are
681 lookups. However, all the other forms of pattern expect the subject to
682 contain a local part and a domain, and therefore, for them, an empty
683 address still always fails if the pattern is not itself empty.
685 30. Exim went into a mad DNS loop when attempting to do a callout where the
686 host was specified on an smtp transport, and looking it up yielded more
689 31. Re-factored the code for checking spool and log partition space into a
690 function that finds that data and another that does the check. The former
691 is then used to implement four new variables: $spool_space, $log_space,
692 $spool_inodes, and $log_inodes.
694 32. The RFC2047 encoding function was originally intended for short strings
695 such as real names; it was not keeping to the 75-character limit for
696 encoded words that the RFC imposes. It now respects the limit, and
697 generates multiple encoded words if necessary. To be on the safe side, I
698 have increased the buffer size for the ${rfc2047: expansion operator from
701 33. It is now permitted to omit both strings after an "if" condition; if the
702 condition is true, the result is "true". As before, when the second string
703 is omitted, a false condition yields an empty string. This makes it less
704 cumbersome to write custom ACL and router conditions.
706 34. Failure to deliver a bounce message always caused it to be frozen, even if
707 there was an errors_to setting on the router. The errors_to setting is now
710 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
711 canonical form (fully expanded) before being placed in
712 $sender_host_address.
714 36. The table in the code that translates DNS record types into text (T_A to
715 "A" for instance) was missing entries for NS and CNAME. It is just possible
716 that this could have caused confusion if both these types were looked up
717 for the same domain, because the text type is used as part of Exim's
718 per-process caching. But the chance of anyone hitting this buglet seems
721 37. The dnsdb lookup has been extended in a number of ways.
723 (1) There is a new type, "zns", which walks up the domain tree until it
724 finds some nameserver records. It should be used with care.
726 (2) There is a new type, "mxh", which is like "mx" except that it returns
727 just the host names, not the priorities.
729 (3) It is now possible to give a list of domains (or IP addresses) to be
730 looked up. The behaviour when one of the lookups defers can be
731 controlled by a keyword.
733 (4) It is now possible to specify the separator character for use when
734 multiple records are returned.
736 38. The dnslists ACL condition has been extended: it is now possible to supply
737 a list of IP addresses and/or domains to be looked up in a particular DNS
740 39. Added log_selector=+queue_time_overall.
742 40. When running the queue in the test harness, wait just a tad after forking a
743 delivery process, to get repeatability of debugging output.
745 41. Include certificate and key file names in error message when GnuTLS fails
746 to set them up, because the GnuTLS error message doesn't include the name
747 of the failing file when there is a problem reading it.
749 42. Allow both -bf and -bF in the same test run.
751 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
753 44. The "Exiscan patch" is now merged into the mainline Exim source.
755 45. Sometimes the final signoff response after QUIT could fail to get
756 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
757 < 0 before doing a fflush(). This bug looks as though it goes back to the
758 introduction of TLS in release 3.20, but "sometimes" must have been rare
759 because the tests only now provoked it.
761 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
762 (this can affect the format of dates).
764 47. exim_tidydb, when checking for the continued existence of a message for
765 which it has found a message-specific retry record, was not finding
766 messages that were in split spool directories. Consequently, it was
767 deleting retry records that should have stayed in existence.
769 48. Steve fixed some bugs in eximstats.
771 49. The SPA authentication driver was not abandoning authentication and moving
772 on to the next authenticator when an expansion was forced to fail,
773 contradicting the general specification for all authenticators. Instead it
774 was generating a temporary error. It now behaves as specified.
776 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
777 (the order specifies the preference for clients). The order is now AES256,
778 AES128, 3DES, ARCFOUR128.
780 51. Small patch to Sieve code - explicitly set From: when generating an
783 52. Exim crashed if a remote delivery caused a very long error message to be
784 recorded - for instance if somebody sent an entire SpamAssassin report back
785 as a large number of 550 error lines. This bug was coincidentally fixed by
786 increasing the size of one of Exim's internal buffers (big_buffer) that
787 happened as part of the Exiscan merge. However, to be on the safe side, I
788 have made the code more robust (and fixed the comments that describe what
791 53. Now that there can be additional text after "Completed" in log lines (if
792 the queue_time_overall log selector is set), a one-byte patch to exigrep
793 was needed to allow it to recognize "Completed" as not the last thing in
796 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
797 patch that reportedly fixes this has been added. I am not expert enough to
798 create a test for it. This is what the patch creator wrote:
800 "I found a little strange behaviour of ldap code when working with
801 Windows 2003 AD Domain, where users was placed in more than one
802 Organization Units. When I tried to give exim partial DN, the exit code
803 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
804 But simultaneously result of request was absolutely normal ldap result,
805 so I produce this patch..."
807 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
808 REFERENCE, so I have modified the code to exclude the patch when that macro
811 55. Some experimental protocols are using DNS PTR records for new purposes. The
812 keys for these records are domain names, not reversed IP addresses. The
813 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
814 leaves it alone. Component reversal etc. now happens only for IP addresses.
817 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
819 57. Double the size of the debug message buffer (to 2048) so that more of very
820 long debug lines gets shown.
822 58. The exicyclog utility now does better if the number of log files to keep
823 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
825 59. Two changes related to the smtp_active_hostname option:
827 (1) $smtp_active_hostname is now available as a variable.
828 (2) The default for smtp_banner uses $smtp_active_hostname instead
829 of $primary_hostname.
831 60. The host_aton() function is supposed to be passed a string that is known
832 to be a valid IP address. However, in the case of IPv6 addresses, it was
833 not checking this. This is a hostage to fortune. Exim now panics and dies
834 if the condition is not met. A case was found where this could be provoked
835 from a dnsdb PTR lookup with an IPv6 address that had more than 8
836 components; fortuitously, this particular loophole had already been fixed
837 by change 4.50/55 above.
839 If there are any other similar loopholes, the new check in host_aton()
840 itself should stop them being exploited. The report I received stated that
841 data on the command line could provoke the exploit when Exim was running as
842 exim, but did not say which command line option was involved. All I could
843 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
847 61. There was a buffer overflow vulnerability in the SPA authentication code
848 (which came originally from the Samba project). I have added a test to the
849 spa_base64_to_bits() function which I hope fixes it.
852 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
853 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
855 63. The daemon start-up calls getloadavg() while still root for those OS that
856 need the first call to be done as root, but it missed one case: when
857 deliver_queue_load_max is set with deliver_drop_privilege. This is
858 necessary for the benefit of the queue runner, because there is no re-exec
859 when deliver_drop_privilege is set.
861 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
864 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
865 was in use, was not putting the data itself into the right store pool;
866 consequently, it could be overwritten for a subsequent message in the same
867 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
870 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
872 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
873 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
874 the function were treating the return as a boolean value, which happened to
875 work because 0=false and not-0=true, but is not correct code.
877 68. The host_aton() function was not handling scoped IPv6 addresses (those
878 with, for example, "%eth0" on the end) correctly.
880 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
881 negated items (that is, ~something) in unsigned ints. Some compilers
882 apparently mutter when there is no cast.
884 70. If an address verification called from an ACL failed, and did not produce a
885 user-specific message (i.e. there was only a "system" message), nothing was
886 put in $acl_verify_message. In this situation, it now puts the system
889 71. Change 4.23/11 added synchronization checking at the start of an SMTP
890 session; change 4.31/43 added the unwanted input to the log line - except
891 that it did not do this in the start of session case. It now does.
893 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
894 This is wrong and can cause the other end to generate a synchronization
895 error if it is another Exim or anything else that does the synchronization
896 check. A QUIT command is no longer sent after a timeout.
898 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
901 74. The defer_ok option of callout verification was not working if it was used
902 when verifying addresses in header lines, that is, for this case:
904 verify = header_sender/callout=defer_ok
906 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
907 those file descriptors could be used for SMTP connections. If anything
908 wrote to stderr (the example that came up was "warn" in embedded Perl), it
909 could be sent to the SMTP client, causing chaos. The daemon now opens
910 stdin, stdout, and stderr to /dev/null when it puts itself into the
913 76. Arrange for output from Perl's "warn" command to be written to Exim's main
914 log by default. The user can override this with suitable Perl magic.
916 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
917 the log message when discard triggers, was not working for the DATA ACL or
918 for the non-SMTP ACL.
920 78. Error message wording change in sieve.c.
922 79. If smtp_accept_max_per_host was set, the number of connections could be
923 restricted to fewer than expected, because the daemon was trying to set up
924 a new connection before checking whether the processes handling previous
925 connections had finished. The check for completed processes is now done
926 earlier. On busy systems, this bug wouldn't be noticed because something
927 else would have woken the daemon, and it would have reaped the completed
930 80. If a message was submitted locally by a user whose login name contained one
931 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
932 It caused a spool format error. I have fixed the spool reading code. A
933 related problem was that the "from" clause in the Received: line became
934 illegal because of the space(s). It is now covered by ${quote_local_part.
936 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
939 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
941 83. Patch from Sieve maintainer to fix unterminated string problem in
944 84. Some minor changes to the Linux configuration files to help with other
945 OS variants using glibc.
947 85. One more patch for Sieve to update vacation handling to latest spec.
950 ----------------------------------------------------
951 See the note above about the 4.44 and 4.50 releases.
952 ----------------------------------------------------
958 1. Change 4.43/35 introduced a bug that caused file counts to be
959 incorrectly computed when quota_filecount was set in an appendfile
962 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
963 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
966 3. Give more explanation in the error message when the command for a transport
967 filter fails to execute.
969 4. There are several places where Exim runs a non-Exim command in a
970 subprocess. The SIGUSR1 signal should be disabled for these processes. This
971 was being done only for the command run by the queryprogram router. It is
972 now done for all such subprocesses. The other cases are: ${run, transport
973 filters, and the commands run by the lmtp and pipe transports.
975 5. Some older OS have a limit of 256 on the maximum number of file
976 descriptors. Exim was using setrlimit() to set 1000 as a large value
977 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
978 systems. I've change it so that if it can't get 1000, it tries for 256.
980 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
981 was an oversight, and furthermore, ever since the addition of extra
982 controls (e.g. 4.43/32), the checks on when to allow different forms of
983 "control" were broken. There should now be diagnostics for all cases when a
984 control that does not make sense is encountered.
986 7. $recipients is now available in the predata ACL (oversight).
988 8. Tidy the search cache before the fork to do a delivery from a message
989 received from the command line. Otherwise the child will trigger a lookup
990 failure and thereby defer the delivery if it tries to use (for example) a
991 cached ldap connection that the parent has called unbind on.
993 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
994 of $address_data from the recipient verification was clobbered by the
997 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
998 was its contents. (It was OK if the option was not defined at all.)
1000 11. A "Completed" log line is now written for messages that are removed from
1001 the spool by the -Mrm option.
1003 12. $host_address is now set to the target address during the checking of
1004 ignore_target_hosts.
1006 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1007 being passed; this would have caused $sender_host_name to have been used if
1008 matching the list had actually called for a host name (not very likely,
1009 since this list is usually IP addresses). A host name is now passed as
1012 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1013 code that specifies a non-restarting handler (typically sigaction() in
1014 modern systems) in an attempt to fix a rare and obscure crash bug.
1016 15. Narrowed the window for a race in the daemon that could cause it to ignore
1017 SIGCHLD signals. This is not a major problem, because they are used only to
1018 wake it up if nothing else does.
1020 16. A malformed maildirsize file could cause Exim to calculate negative values
1021 for the mailbox size or file count. Odd effects could occur as a result.
1022 The maildirsize information is now recalculated if the size or filecount
1025 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1026 support for a long time. Removed HAVE_SYS_VFS_H.
1028 18. Updated exipick to current release from John Jetmore.
1030 19. Allow an empty sender to be matched against a lookup in an address list.
1031 Previously the only cases considered were a regular expression, or an
1034 20. Exim went into a mad DNS lookup loop when doing a callout where the
1035 host was specified on the transport, if the DNS lookup yielded more than
1038 21. The RFC2047 encoding function was originally intended for short strings
1039 such as real names; it was not keeping to the 75-character limit for
1040 encoded words that the RFC imposes. It now respects the limit, and
1041 generates multiple encoded words if necessary. To be on the safe side, I
1042 have increased the buffer size for the ${rfc2047: expansion operator from
1045 22. Failure to deliver a bounce message always caused it to be frozen, even if
1046 there was an errors_to setting on the router. The errors_to setting is now
1049 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1050 canonical form (fully expanded) before being placed in
1051 $sender_host_address.
1053 24. Updated eximstats to version 1.33
1055 25. Include certificate and key file names in error message when GnuTLS fails
1056 to set them up, because the GnuTLS error message doesn't include the name
1057 of the failing file when there is a problem reading it.
1059 26. Expand error message when OpenSSL has problems setting up cert/key files.
1062 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1063 (this can affect the format of dates).
1065 28. exim_tidydb, when checking for the continued existence of a message for
1066 which it has found a message-specific retry record, was not finding
1067 messages that were in split spool directories. Consequently, it was
1068 deleting retry records that should have stayed in existence.
1070 29. eximstats updated to version 1.35
1071 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1072 1.35 - bugfix such that pie charts by volume are generated correctly
1074 30. The SPA authentication driver was not abandoning authentication and moving
1075 on to the next authenticator when an expansion was forced to fail,
1076 contradicting the general specification for all authenticators. Instead it
1077 was generating a temporary error. It now behaves as specified.
1079 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1080 (the order specifies the preference for clients). The order is now AES256,
1081 AES128, 3DES, ARCFOUR128.
1083 31. Small patch to Sieve code - explicitly set From: when generating an
1086 32. Exim crashed if a remote delivery caused a very long error message to be
1087 recorded - for instance if somebody sent an entire SpamAssassin report back
1088 as a large number of 550 error lines. This bug was coincidentally fixed by
1089 increasing the size of one of Exim's internal buffers (big_buffer) that
1090 happened as part of the Exiscan merge. However, to be on the safe side, I
1091 have made the code more robust (and fixed the comments that describe what
1094 33. Some experimental protocols are using DNS PTR records for new purposes. The
1095 keys for these records are domain names, not reversed IP addresses. The
1096 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1097 leaves it alone. Component reversal etc. now happens only for IP addresses.
1100 34. The host_aton() function is supposed to be passed a string that is known
1101 to be a valid IP address. However, in the case of IPv6 addresses, it was
1102 not checking this. This is a hostage to fortune. Exim now panics and dies
1103 if the condition is not met. A case was found where this could be provoked
1104 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1105 components; fortuitously, this particular loophole had already been fixed
1106 by change 4.50/55 or 4.44/33 above.
1108 If there are any other similar loopholes, the new check in host_aton()
1109 itself should stop them being exploited. The report I received stated that
1110 data on the command line could provoke the exploit when Exim was running as
1111 exim, but did not say which command line option was involved. All I could
1112 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1113 running as the user.
1116 35. There was a buffer overflow vulnerability in the SPA authentication code
1117 (which came originally from the Samba project). I have added a test to the
1118 spa_base64_to_bits() function which I hope fixes it.
1121 36. The daemon start-up calls getloadavg() while still root for those OS that
1122 need the first call to be done as root, but it missed one case: when
1123 deliver_queue_load_max is set with deliver_drop_privilege. This is
1124 necessary for the benefit of the queue runner, because there is no re-exec
1125 when deliver_drop_privilege is set.
1127 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1128 was in use, was not putting the data itself into the right store pool;
1129 consequently, it could be overwritten for a subsequent message in the same
1130 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1133 38. Sometimes the final signoff response after QUIT could fail to get
1134 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1135 < 0 before doing a fflush(). This bug looks as though it goes back to the
1136 introduction of TLS in release 3.20, but "sometimes" must have been rare
1137 because the tests only now provoked it.
1143 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1144 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1145 Now it returns an integer. A number of calls were still expecting a T/F
1146 return. Fortuitously, in all cases, the tests worked in OK situations,
1147 which is the norm. However, things would have gone wrong on any write
1148 failures on the smtp file descriptor. This function is used when sending
1149 messages over SMTP and also when doing verify callouts.
1151 2. When Exim is called to do synchronous delivery of a locally submitted
1152 message (the -odf or -odi options), it no longer closes stderr before doing
1155 3. Implemented the mua_wrapper option.
1157 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1159 5. Implemented the functions header_remove(), header_testname(),
1160 header_add_at_position(), and receive_remove_recipient(), and exported them
1163 6. If an ACL "warn" statement specified the addition of headers, Exim already
1164 inserted X-ACL-Warn: at the start if there was no header name. However, it
1165 was not making this test for the second and subsequent header lines if
1166 there were newlines in the string. This meant that an invalid header could
1167 be inserted if Exim was badly configured.
1169 7. Allow an ACL "warn" statement to add header lines at the start or after all
1170 the Received: headers, as well as at the end.
1172 8. Added the rcpt_4xx retry error code.
1174 9. Added postmaster_mailfrom=xxx to callout verification option.
1176 10. Added mailfrom=xxxx to the callout verification option, for verify=
1179 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1180 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1182 12. Inserted some casts to stop certain compilers warning when using pointer
1183 differences as field lengths or precisions in printf-type calls (mostly
1184 affecting debugging statements).
1186 13. Added optional readline() support for -be (dynamically loaded).
1188 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1189 same clock tick as a message's arrival, so that its received time was the
1190 same as the "first fail" time on the retry record, and that message
1191 remained on the queue past the ultimate address timeout, every queue runner
1192 would try a delivery (because it was past the ultimate address timeout) but
1193 after another failure, the ultimate address timeout, which should have then
1194 bounced the address, did not kick in. This was a "< instead of <=" error;
1195 in most cases the first failure would have been in the next clock tick
1196 after the received time, and all would be well.
1198 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1199 being recognized when the domain list was tested by the match_domain
1200 condition in an expansion string.
1202 16. Added the ${str2b64: operator.
1204 17. Exim was always calling setrlimit() to set a large limit for the number of
1205 processes, without checking whether the existing limit was already
1206 adequate. (It did check for the limit on file descriptors.) Furthermore,
1207 errors from getrlimit() and setrlimit() were being ignored. Now they are
1208 logged to the main and panic logs, but Exim does carry on, to try to do its
1209 job under whatever limits there are.
1211 18. Imported PCRE 5.0.
1213 19. Trivial typo in log message " temporarily refused connection" (the leading
1216 20. If the log selector return_path_on_delivery was set and an address was
1217 redirected to /dev/null, the delivery process crashed because it assumed
1218 that a return path would always be set for a "successful" delivery. In this
1219 case, the whole delivery is bypassed as an optimization, and therefore no
1222 21. Internal re-arrangement: the function for sending a challenge and reading
1223 a response while authentication was assuming a zero-terminated challenge
1224 string. It's now changed to take a pointer and a length, to allow for
1225 binary data in such strings.
1227 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1229 23. Exim was not respecting finduser_retries when seeking the login of the
1230 uid under which it was called; it was always trying 10 times. (The default
1231 setting of finduser_retries is zero.) Also, it was sleeping after the final
1232 failure, which is pointless.
1234 24. Implemented tls_on_connect_ports.
1236 25. Implemented acl_smtp_predata.
1238 26. If the domain in control=submission is set empty, Exim assumes that the
1239 authenticated id is a complete email address when it generates From: or
1240 Sender: header lines.
1242 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1243 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1244 chown and chgrp in /bin and hostname in /usr/bin.
1246 28. Exim was keeping the "process log" file open after each use, just as it
1247 does for the main log. This opens the possibility of it remaining open for
1248 long periods when the USR1 signal hits a daemon. Occasional processlog
1249 errors were reported, that could have been caused by this. Anyway, it seems
1250 much more sensible not to leave this file open at all, so that is what now
1253 29. The long-running daemon process does not normally write to the log once it
1254 has entered its main loop, and it closes the log before doing so. This is
1255 so that log files can straightforwardly be renamed and moved. However,
1256 there are a couple of unusual error situations where the daemon does write
1257 log entries, and I had neglected to close the log afterwards.
1259 30. The text of an SMTP error response that was received during a remote
1260 delivery was being truncated at 512 bytes. This is too short for some of
1261 the long messages that one sometimes sees. I've increased the limit to
1264 31. It is now possible to make retry rules that apply only when a message has a
1265 specific sender, in particular, an empty sender.
1267 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1268 it possible to be selective about when SMTP synchronization is enforced.
1270 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1272 32. Implemented hosts_connection_nolog.
1274 33. Added an ACL for QUIT.
1276 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1279 35. Added mailbox_size and mailbox_filecount to appendfile.
1281 36. Added control = no_multiline_responses to ACLs.
1283 37. There was a bug in the logic of the code that waits for the clock to tick
1284 in the case where the clock went backwards by a substantial amount such
1285 that the microsecond fraction of "now" was more than the microsecond
1286 fraction of "then" (but the whole seconds number was less).
1288 38. Added support for the libradius Radius client library this is found on
1289 FreeBSD (previously only the radiusclient library was supported).
1295 1. When certain lookups returned multiple values in the form name=value, the
1296 quoting of the values was not always being done properly. Specifically:
1297 (a) If the value started with a double quote, but contained no whitespace,
1299 (b) If the value contained whitespace other than a space character (i.e.
1300 tabs or newlines or carriage returns) it was not quoted.
1301 This fix has been applied to the mysql and pgsql lookups by writing a
1302 separate quoting function and calling it from the lookup code. The fix
1303 should probably also be applied to nisplus, ibase and oracle lookups, but
1304 since I cannot test any of those, I have not disturbed their existing code.
1306 2. A hit in the callout cache for a specific address caused a log line with no
1307 reason for rejecting RCPT. Now it says "Previous (cached) callout
1308 verification failure".
1310 3. There was an off-by-one bug in the queryprogram router. An over-long
1311 return line was truncated at 256 instead of 255 characters, thereby
1312 overflowing its buffer with the terminating zero. As well as fixing this, I
1313 have increased the buffer size to 1024 (and made a note to document this).
1315 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1316 when Exim is waiting for an SMTP response from a remote server, Exim
1317 restarts its select() call on the socket, thereby resetting its timeout.
1318 This is not a problem when such interrupts are rare. Somebody set up a cron
1319 job to run exiwhat every 2 minutes, which is less than the normal select()
1320 timeout (5 or 10 minutes). This meant that the select() timeout never
1321 kicked in because it was always reset. I have fixed this by comparing the
1322 time when an interrupt arrives with the time at the start of the first call
1323 to select(). If more time than the timeout has elapsed, the interrupt is
1324 treated as a timeout.
1326 5. Some internal re-factoring in preparation for the addition of Sieve
1327 extensions (by MH). In particular, the "personal" test is moved to a
1328 separate function, and given an option for scanning Cc: and Bcc: (which is
1329 not set for Exim filters).
1331 6. When Exim created an email address using the login of the caller as the
1332 local part (e.g. when creating a From: or Sender: header line), it was not
1333 quoting the local part when it contained special characters such as @.
1335 7. Installed new OpenBSD configuration files.
1337 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1338 try to make them clearer.
1340 9. Callout options, other than the timeout value, were being ignored when
1341 verifying sender addresses in header lines. For example, when using
1343 verify = header_sender/callout=no_cache
1345 the cache was (incorrectly) being used.
1347 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1348 only the Cygwin environment.
1350 11. When return_path_on_delivery was set as a log selector, if different remote
1351 addresses in the same message used different return paths and parallel
1352 remote delivery occurred, the wrong values would sometimes be logged.
1353 (Whenever a remote delivery process finished, the return path value from
1354 the most recently started remote delivery process was logged.)
1356 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1357 lines when AUTH and/or TLS are in use. This is the "received protocol"
1358 field. Exim used to use "asmtp" for authenticated SMTP, without any
1359 indication (in the protocol name) for TLS use. Now it follows the RFC and
1360 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1361 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1362 names appear in log lines as well as in Received: header lines.
1364 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1365 extensions, and comparison tests, and to fix some bugs.
1367 14. Changes to the "personal" filter test:
1369 (1) The test was buggy in that it was just doing the equivalent of
1370 "contains" tests on header lines. For example, if a user's address was
1371 anne@some.where, the "personal" test would incorrectly be true for
1373 To: susanne@some.where
1375 This test is now done by extracting each address from the header in turn,
1376 and checking the entire address. Other tests that are part of "personal"
1377 are now done using regular expressions (for example, to check local parts
1378 of addresses in From: header lines).
1380 (2) The list of non-personal local parts in From: addresses has been
1381 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1382 taken from the Sieve specification recommendations.
1384 (3) If the message contains any header line starting with "List-" it is
1385 treated as non-personal.
1387 (4) The test for "circular" in the Subject: header line has been removed
1388 because it now seems ill-conceived.
1390 15. Minor typos in src/EDITME comments corrected.
1392 16. Installed latest exipick from John Jetmore.
1394 17. If headers_add on a router specified a text string that was too long for
1395 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1396 of string_sprintf() is now avoided.
1398 18. $message_body_size was not set (it was always zero) when running the DATA
1399 ACL and the local_scan() function.
1401 19. For the "mail" command in an Exim filter, no default was being set for
1402 the once_repeat time, causing a random time value to be used if "once" was
1403 specified. (If the value happened to be <= 0, no repeat happened.) The
1404 default is now 0s, meaning "never repeat". The "vacation" command was OK
1405 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1406 (I found it when inspecting the code).
1408 20. There is now an overall timeout for performing a callout verification. It
1409 defaults to 4 times the callout timeout, which applies to individual SMTP
1410 commands during the callout. The overall timeout applies when there is more
1411 than one host that can be tried. The timeout is checked before trying the
1412 next host. This prevents very long delays if there are a large number of
1413 hosts and all are timing out (e.g. when the network connections are timing
1414 out). The value of the overall timeout can be changed by specifying an
1415 additional sub-option for "callout", called "maxwait". For example:
1417 verify = sender/callout=5s,maxwait=20s
1419 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1420 to the end before writing, but this should make it even safer).
1422 22. Exim was forgetting that it had advertised PIPELINING for the second and
1423 subsequent messages on an SMTP connection. It was also not resetting its
1424 memory on STARTTLS and an internal HELO.
1426 23. When Exim logs an SMTP synchronization error within a session, it now
1427 records whether PIPELINING has been advertised or not.
1429 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1430 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1431 rather than long int.
1433 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1435 26. Added the never_mail option to autoreply.
1441 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1442 crash if the getsockname() call failed; this can happen if a connection is
1443 closed very soon after it is established. The problem was simply in the
1444 order in which certain operations were done, causing Exim to try to write
1445 to the SMTP stream before it had set up the file descriptor. The bug has
1446 been fixed by making things happen in the correct order.
1452 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1453 before the connection was closed, thus losing the rejection response.
1455 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1456 some early Solaris releases, but causes trouble in current releases where
1457 socklen_t is defined.
1459 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1462 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1465 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1466 that releasing the top of it at the end releases what was used for sub-
1467 expansions (unless the block got too big). However, discard this block if
1468 the first thing is a variable or header, so that we can use its block when
1469 it is dynamic (useful for very large $message_headers, for example).
1471 6. Lookups now cache *every* query, not just the most recent. A new, separate
1472 store pool is used for this. It can be recovered when all lookup caches are
1473 flushed. Lookups now release memory at the end of their result strings.
1474 This has involved some general refactoring of the lookup sources.
1476 7. Some code has been added to the store_xxx() functions to reduce the amount
1477 of flapping under certain conditions.
1479 8. log_incoming_interface used to affect only the <= reception log lines. Now
1480 it causes the local interface and port to be added to several more SMTP log
1481 lines, for example "SMTP connection from", and rejection lines.
1483 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1485 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1487 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1488 could be overwritten at the end of the current message (or the start of a
1489 new message if it was set in a HELO ACL). The value is now preserved for
1490 the duration of the SMTP connection.
1492 12. If a transport had a headers_rewrite setting, and a matching header line
1493 contained an unqualified address, that address was qualified, even if it
1494 did not match any rewriting rules. The underlying bug was that the values
1495 of the flags that permit the existence of unqualified sender and recipient
1496 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1497 non-local messages, and by -bnq for local messages) were not being
1498 preserved with the message after it was received.
1500 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1501 "next input=" as part of the text comprising the host identity instead of
1502 the correct text. The code was using the same buffer for two different
1503 strings. However, depending on which order the printing function evaluated
1504 its arguments, the bug did not always show up. Under Linux, for example, my
1505 test suite worked just fine.
1507 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1508 doesn't work with some older versions of Perl. It has been changed to "my",
1509 which in any case is probably the better facility to use.
1511 15. A really picky compiler found some instances of statements for creating
1512 error messages that either had too many or two few arguments for the format
1515 16. The size of the buffer for calls to the DNS resolver has been increased
1516 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1517 for addresses that have a lot of PTR records. This alleviates a problem; it
1518 does not fully solve it.
1520 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1521 buffer now truncates the list and logs the incident, which is the same
1522 action as happens when Exim is looking up a host name and its aliases.
1523 Previously in this situation something unpredictable would happen;
1524 sometimes it was "internal error: store_reset failed".
1526 18. If a server dropped the connection unexpectedly when an Exim client was
1527 using GnuTLS and trying to read a response, the client delivery process
1528 crashed while trying to generate an error log message.
1530 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1531 string, for example:
1533 warn message = H1: something\nH2: something
1535 the text was added as a single header line from Exim's point of view
1536 though it ended up OK in the delivered message. However, searching for the
1537 second and subsequent header lines using $h_h2: did not work. This has been
1538 fixed. Similarly, if a system filter added multiple headers in this way,
1539 the routers could not see them.
1541 20. Expanded the error message when iplsearch is called with an invalid key to
1542 suggest using net-iplsearch in a host list.
1544 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1545 ACLs are no longer actually imposed (and a message to that effect is
1548 22. If a "gecos" field in a passwd entry contained escaped characters, in
1549 particular, if it contained a \" sequence, Exim got it wrong when building
1550 a From: or a Sender: header from that name. A second bug also caused
1551 incorrect handling when an unquoted " was present following a character
1552 that needed quoting.
1554 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1555 was not being matched caselessly.
1557 24. Arranged for all hyphens in the exim.8 source to be escaped with
1560 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1561 properly. Recipient callout cache records were still being keyed to include
1562 the sender, even when use_sender was set false. This led to far more
1563 callouts that were necessary. The sender is no longer included in the key
1564 when use_sender is false.
1566 26. Added "control = submission" modifier to ACLs.
1568 27. Added the ${base62d: operator to decode base 62 numbers.
1570 28. dnsdb lookups can now access SRV records.
1572 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1573 the configuration file.
1575 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1576 (-v) mode. This makes the output for a verbose queue run more intelligible.
1578 31. Added a use_postmaster feature to recipient callouts.
1580 32. Added the $body_zerocount variable, containing the number of binary zero
1581 bytes in the message body.
1583 33. The time of last modification of the "new" subdirectory is now used as the
1584 "mailbox time last read" when there is a quota error for a maildir
1587 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1589 35. Added +ignore_unknown as a special item in host lists.
1591 36. Code for decoding IPv6 addresses in host lists is now included, even if
1592 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1593 address was recognized as an IP address, but was then not correctly decoded
1594 into binary, causing unexpected and incorrect effects when compared with
1601 1. Very minor rewording of debugging text in manualroute to say "list of
1602 hosts" instead of "hostlist".
1604 2. If verify=header_syntax was set, and a header line with an unqualified
1605 address (no domain) and a large number of spaces between the end of the
1606 name and the colon was received, the reception process suffered a buffer
1607 overflow, and (when I tested it) crashed. This was caused by some obsolete
1608 code that should have been removed. The fix is to remove it!
1610 3. When running in the test harness, delay a bit after writing a bounce
1611 message to get a bit more predictability in the log output.
1613 4. Added a call to search_tidyup() just before forking a reception process. In
1614 theory, someone could use a lookup in the expansion of smtp_accept_max_
1615 per_host which, without the tidyup, could leave open a database connection.
1617 5. Added the variables $recipient_data and $sender_data which get set from a
1618 lookup success in an ACL "recipients" or "senders" condition, or a router
1619 "senders" option, similar to $domain_data and $local_part_data.
1621 6. Moved the writing of debug_print from before to after the "senders" test
1624 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1625 problems for message scanning, either using a data ACL, or using
1626 local_scan() because the Received: header was not generated till after they
1627 were called (in order to set the time as the time of reception completion).
1628 I have revised the way this works. The header is now generated after the
1629 body is received, but before the ACL or local_scan() are called. After they
1630 are run, the timestamp in the header is updated.
1636 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1637 before starting a queue runner without re-exec. This happened only when
1638 deliver_drop_privilege was set or when the Exim user was set to root. The
1639 effect of the bug was that timeouts during subsequent deliveries caused
1640 crashes instead of being properly handled. The handler is now left at its
1641 default (and expected) setting.
1643 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1644 message, again when deliver_drop_privilege is set or Exim is run as root.
1645 The bug described in (1) was not present in this case, but the tidying up
1646 of the other signals was missing. I have made the two cases consistent.
1648 3. The ignore_target_hosts setting on a manualroute router was being ignored
1649 for hosts that were looked up using the /MX notation.
1651 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1654 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1655 operated on the sender address. After changing the $sender_address to <>
1656 for the sender address verify, Exim was re-instated it as the original
1657 (before rewriting) address, but remembering that it had rewritten it, so it
1658 wasn't rewriting it again. This bug also had the effect of breaking the
1659 sender address verification caching when the sender address was rewritten.
1661 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1662 This has been changed so that if the ip literal address matches
1663 ignore_target_hosts, the router declines.
1665 7. Added expansion conditions match_domain, match_address, and match_local_
1666 part (NOT match_host).
1668 8. The placeholder for the Received: header didn't have a length field set.
1670 9. Added code to Exim itself and to exim_lock to test for a specific race
1671 condition that could lead to file corruption when using MBX delivery. The
1672 issue is with the lockfile that is created in /tmp. If this file is removed
1673 after a process has opened it but before that process has acquired a lock,
1674 there is the potential for a second process to recreate the file and also
1675 acquire a lock. This could lead to two Exim processes writing to the file
1676 at the same time. The added code performs the same test as UW imapd; it
1677 checks after acquiring the lock that its file descriptor still refers to
1678 the same named file.
1680 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1681 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1684 11. Added the smtp_active_hostname option. If used, this will typically be made
1685 to depend on the incoming interface address. Because $interface_address is
1686 not set up until the daemon has forked a reception process, error responses
1687 that can happen earlier (such as "too many connections") no longer contain
1690 12. If an expansion in a condition on a "warn" statement fails because a lookup
1691 defers, the "warn" statement is abandoned, and the next ACL statement is
1692 processed. Previously this caused the whole ACL to be aborted.
1694 13. Added the iplsearch lookup type.
1696 14. Added ident_timeout as a log selector.
1698 15. Added tls_certificate_verified as a log selector.
1700 16. Added a global option tls_require_ciphers (compare the smtp transport
1701 option of the same name). This controls incoming TLS connections.
1703 17. I finally figured out how to make tls_require_ciphers do a similar thing
1704 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1705 before starting the TLS session.
1707 18. Tabs are now shown as \t in -bP output.
1709 19. If the log selector return_path_on_delivery was set, Exim crashed when
1710 bouncing a message because it had too many Received: header lines.
1712 20. If two routers both had headers_remove settings, and the first one included
1713 a superfluous trailing colon, the final name in the first list and the
1714 first name in the second list were incorrectly joined into one item (with a
1715 colon in the middle).
1721 1. Added -C and -D options to the exinext utility, mainly to make it easier
1722 to include in the automated testing, but these could be helpful when
1723 multiple configurations are in use.
1725 2. The exinext utility was not formatting the output nicely when there was
1726 an alternate port involved in the retry record key, nor when there was a
1727 message id as well (for retries that were specific to a specific message
1728 and a specific host). It was also confused by IPv6 addresses, because of
1729 the additional colons they contain. I have fixed the IPv4 problem, and
1730 patched it up to do a reasonable job for IPv6.
1732 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1733 delivery, the log line now contains "pipelined" if PIPELINING was used.
1735 4. An SMTP transport process used to panic and die if the bind() call to set
1736 an explicit outgoing interface failed. This has been changed; it is now
1737 treated in the same way as a connect() failure.
1739 5. A reference to $sender_host_name in the part of a conditional expansion
1740 that was being skipped was still causing a DNS lookup. This no longer
1743 6. The def: expansion condition was not recognizing references to header lines
1744 that used bh_ and bheader_.
1746 7. Added the _cache feature to named lists.
1748 8. The code for checking quota_filecount in the appendfile transport was
1749 allowing one more file than it should have been.
1751 9. For compatibility with Sendmail, the command line option
1759 and sets the incoming protocol and host name (for trusted callers). The
1760 host name and its colon can be omitted when only the protocol is to be set.
1761 Note the Exim already has two private options, -pd and -ps, that refer to
1762 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1763 "s", but I don't think that's a major issue.
1765 10. A number of refactoring changes to the code, none of which should affect
1768 (a) The number of logging options was getting close to filling up the
1769 32-bit word that was used as a bit map. I have split them into two classes:
1770 those that are passed in the argument to log_write(), and those that are
1771 only ever tested independently outside of that function. These are now in
1772 separate 32-bit words, so there is plenty of room for expansion again.
1773 There is no change in the user interface or the logging behaviour.
1775 (b) When building, for example, log lines, the code previously used a
1776 macro that called string_cat() twice, in order to add two strings. This is
1777 not really sufficiently general. Furthermore, there was one instance where
1778 it was actually wrong because one of the argument was used twice, and in
1779 one call a function was used. (As it happened, calling the function twice
1780 did not affect the overall behaviour.) The macro has been replaced by a
1781 function that can join an arbitrary number of extra strings onto a growing
1784 (c) The code for expansion conditions now uses a table and a binary chop
1785 instead of a serial search (which was left over from when there were very
1786 few conditions). Also, it now recognizes conditions like "pam" even when
1787 the relevant support is not compiled in: a suitably worded error message is
1788 given if an attempt is made to use such a condition.
1790 11. Added ${time_interval:xxxxx}.
1792 12. A bug was causing one of the ddress fields not to be passed back correctly
1793 from remote delivery subprocesses. The field in question was not being
1794 subsequently used, so this caused to problems in practice.
1796 13. Added new log selectors queue_time and deliver_time.
1798 14. Might have fixed a bug in maildirsizefile handling that threw up
1799 "unexpected character" debug warnings, and recalculated the data
1800 unnecessarily. In any case, I expanded the warning message to give more
1803 15. Added the message "Restricted characters in address" to the statements in
1804 the default ACL that block characters like @ and % in local parts.
1806 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1807 Three changes have been made:
1809 (a) There was a serious bug; a negative response to MAIL caused the whole
1810 recipient domain to be cached as invalid, thereby blocking all messages
1811 to all local parts at the same domain, from all senders. This bug has
1812 been fixed. The domain is no longer cached after a negative response to
1813 MAIL if the sender used is not empty.
1815 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1818 (c) A new callout option, "use_sender" has been added for people who want
1819 the modified behaviour.
1825 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1828 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1829 indeed breaks things for older releases.
1831 3. Added additional logging to the case where there is a problem reading data
1832 from a filter that is running in a subprocess using a pipe, in order to
1833 try to track down a specific problem.
1835 4. Testing facility fudge: when running in the test harness and attempting
1836 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1837 getting "No route to host". Convert this to a timeout.
1839 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1842 6. Some OS don't have socklen_t but use size_t instead. This affects the
1843 fifth argument of getsockopt() amongst other things. This is now
1844 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1845 can be set for individual OS. I have set it for SunOS5, OSF1, and
1846 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1847 some earlier ones do not.
1849 7. Change 4.30/15 was not doing the test caselessly.
1851 8. The standard form for an IPv6 address literal was being rejected by address
1852 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1853 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1854 this, as well as the form without the "IPv6" on the front (but only when
1855 address literals are enabled, of course).
1857 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1859 10. Exim crashed if a message with an empty sender address specified by -f
1860 encountered a router with an errors_to setting. This could be provoked only
1861 by a command such as
1865 where an empty string was supplied; "<>" did not hit this bug.
1867 11. Installed PCRE release 4.5.
1869 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1870 remained set. It is now erased.
1872 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1873 times from message ids (which are base 36 rather than the normal 62).
1875 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1876 were being counted as actual protocol errors, and logged if the log
1877 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1878 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1879 and DATA following a set of rejected RCPTs do not count as protocol errors.
1880 In other words, Exim assumes they were pipelined, though this may not
1881 actually be the case. Of course, in all cases the client gets an
1882 appropriate error code.
1884 15. If a lookup fails in an ACL condition, a message about the failure may
1885 be available; it is used if testing the ACL cannot continue, because most
1886 such messages specify what the cause of the deferral is. However, some
1887 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1888 that caused an old message to be retained and used if a later statement
1889 caused a defer, replacing the real cause of the deferral.
1891 16. If an IP address had so many PTR records that the DNS lookup buffer
1892 was not large enough to hold them, Exim could crash while trying to process
1893 the truncated data. It now detects and logs this case.
1895 17. Further to 4.21/58, another change has been made: if (and only if) the
1896 first line of a message (the first header line) ends with CRLF, a bare LF
1897 in a subsequent header line has a space inserted after it, so as not to
1898 terminate the header.
1900 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1901 unnecessarily, used atoi() instead of strtol(), and didn't check the
1902 termination when getting file sizes from file names by regex.
1904 19. Completely re-implemented the support for maildirsize files, in the light
1905 of a number of problems with the previous contributed implementation
1906 (4.30/29). In particular:
1908 . If the quota is zero, the maildirsize file is maintained, but no quota is
1911 . If the maildir directory does not exist, it is created before any attempt
1912 to write a maildirsize file.
1914 . The quota value in the file is just a cache; if the quota is changed in
1915 the transport, the new value overrides.
1917 . A regular expression is available for excluding directories from the
1920 20. The autoreply transport checks the characters in options that define the
1921 message's headers; it allows continued headers, but it was checking with
1922 isspace() after an embedded newline instead of explicitly looking for a
1925 21. If all the "regular" hosts to which an address was routed had passed their
1926 expiry times, and had not reached their retry times, the address was
1927 bounced, even if fallback hosts were defined. Now Exim should go on to try
1930 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1931 equivalent code in the SMTP transport. Some hosts send humungous responses
1932 to HELO/EHLO, more than 1024 it seems.
1934 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1935 gives compiler warnings in some environments. I've now done it "properly",
1938 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1939 (because of problems with the built-in one) was declared to return uschar *
1940 instead of char *, causing compiler failure.
1942 25. Fixed a file descriptor leak when processing alias/forward files.
1944 26. Fixed a minor format string issue in dbfn.c.
1946 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1948 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1949 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1952 29. When a sender address is verified, it is cached, to save repeating the test
1953 when there is more than one recipient in a message. However, when the
1954 verification involves a callout, it is possible for different callout
1955 options to be set for different recipients. It is too complicated to keep
1956 track of this in the cache, so now Exim always runs a verification when a
1957 callout is required, relying on the callout cache for the optimization.
1958 The overhead is duplication of the address routing, but this should not be
1961 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1962 to be verified with callout=postmaster, and the main callout worked but the
1963 postmaster check failed, the verification correctly failed. However, if a
1964 subsequent RCPT command asked for sender verification *without* the
1965 postmaster check, incorrect caching caused this verification also to fail,
1968 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1969 it was not caching the DNS options (qualify_single, search_parents) that
1970 were used when the lookup failed. A subsequent lookup with different
1971 options therefore always gave the same answer, though there were cases
1972 where it should not have. (Example: a "domains = !$mx_any" option on a
1973 dnslookup router: the "domains" option is always processed without any
1974 widening, but the router might have qualify_single set.) Now Exim uses the
1975 cached value only when the same options are set.
1977 32. Added John Jetmore's "exipick" utility to the distribution.
1979 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1980 than a timeout (e.g. a certificate is required, and is not provided), an
1981 Exim server now closes the connection immediately. Previously it waited for
1982 the client to close - but if the client is SSL, it seems that they each
1983 wait for each other, leading to a delay before one of them times out.
1985 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1986 maintained 0.8.x compatibility because I don't think many are using it, and
1987 it is clearly obsolete.
1989 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1992 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1993 client certificate was expired. A simple patch fixes this, though I don't
1994 understand the full logic of why the verify callback is called multiple
1997 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1998 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1999 which causes problems with some clients (such as the Certicom SSL Plus
2000 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
2001 disables the coutermeasure allowing Eudora to connect."
2003 38. Exim was not checking that a write() to a log file succeeded. This could
2004 lead to Bad Things if a log got too big, in particular if it hit a file
2005 size limit. Exim now panics and dies if it cannot write to a log file, just
2006 as it does if it cannot open a log file.
2008 39. Modified OS/Makefile-Linux so that it now contains
2010 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2012 The two -D definitions ensure that Exim is compiled with large file
2013 support, which makes it possible to handle log files that are bigger than
2016 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2017 instance) a domain was checked against a named list that involved a lookup,
2018 causing $domain_data to be set, then another domain was checked against the
2019 same list, then the first domain was re-checked, the value of $domain_data
2020 after the final check could be wrong. In particular, if the second check
2021 failed, it could be set empty. This bug probably also applied to
2024 41. The strip_trailing_dot option was not being applied to the address given
2025 with the -f command-line option.
2027 42. The code for reading a message's header from the spool was incrementing
2028 $received_count, but never initializing it. This meant that the value was
2029 incorrect (doubled) while delivering a message in the same process in which
2030 it was received. In the most common configuration of Exim, this never
2031 happens - a fresh exec is done - but it can happen when
2032 deliver_drop_privilege is set.
2034 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2035 it now includes up to 150 characters of the unexpected data in the log
2038 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2039 and building data strings. The size of both of these buffers was 10 000
2040 bytes - far larger than anybody would *ever* want, thought I. Needless to
2041 say, somebody hit the limit. I have increased the maximum line length to
2042 20 000 and the maximum data length of concatenated lines to 100 000. I have
2043 also fixed two bugs, because there was no checking on these buffers. Tsk,
2044 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2045 buffer is too small.
2047 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2048 lsearch lookups. Now it does.
2050 46. When parsing a route_list item in a manualroute router, a fixed-length
2051 buffer was used for the list of hosts. I made this 1024 bytes long,
2052 thinking that nobody would ever have a list of hosts that long. Wrong.
2053 Somebody had a whole pile of complicated expansion conditions, and the
2054 string was silently truncated, leading to an expansion error. It turns out
2055 that it is easier to change to an unlimited length (owing to other changes
2056 that have happened since this code was originally written) than to build
2057 structure for giving a limitation error. The length of the item that
2058 expands into the list of hosts is now unlimited.
2060 47. The lsearch lookup could not handle data where the length of text line was
2061 more than 4095 characters. Such lines were truncated, leading to shortened
2062 data being returned. It should now handle lines of any length.
2064 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2065 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2068 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2069 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2070 no longer happens in comment lines. A list of replacements is now placed
2071 at the head of all of the source files, except those whose only change is
2072 to replace PERL_COMMAND in the very first #! line.
2074 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2075 messages on the queue, with a bottom-up merge sort, using code contributed
2076 by Michael Haardt. This should make operations like -bp somewhat faster on
2077 large queues. It won't affect queue runners, except when queue_run_in_order
2080 51. Installed eximstats 1.31 in the distribution.
2082 52. Added support for SRV lookups to the dnslookup router.
2084 53. If an ACL referred to $message_body or $message_body_end, the value was not
2085 reset for any messages that followed in the same SMTP session.
2087 54. The store-handling optimization for building very long strings was not
2088 differentiating between the different store pools. I don't think this
2089 actually made any difference in practice, but I've tidied it.
2091 55. While running the routers to verify a sender address, $sender_address
2092 was still set to the sender address. This is wrong, because when routing to
2093 send a bounce to the sender, it would be empty. Therefore, I have changed
2094 it so that, while verifying a sender address, $sender_address is set to <>.
2095 (There is no change to what happens when verifying a recipient address.)
2097 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2098 target A or AAAA records (if not already returned) without resetting the
2099 qualify_single or search_parents options of the DNS resolver. These are
2100 inappropriate in this case because the targets of MX and SRV records must
2101 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2102 target that, when qualified, matched something in the local domain. These
2103 two options are now turned off when doing these lookups.
2105 57. It seems that at least some releases of Reiserfs (which does not have the
2106 concept of a fixed number of inodes) returns zero and not -1 for the
2107 number of available inodes. This interacted badly with check_spool_inodes,
2108 which assumed that -1 was the "no such thing" setting. What I have done is
2109 to check that the total number of inodes is greater than zero before doing
2110 the test of how many are available.
2112 58. When a "warn" ACL statement has a log_message modifier, the message is
2113 remembered, and not repeated. This is to avoid a lot of repetition when a
2114 message has many recipients that cause the same warning to be written.
2115 Howewer, Exim was preserving the list of already written lines for an
2116 entire SMTP session, which doesn't seem right. The memory is now reset if a
2117 new message is started.
2119 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2120 debugging output unless log_rewrite was also set.
2122 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2123 of (int)(handle) when we know that handle contains (void *)(-1).
2125 61. The Exim daemon panic-logs an error return when it closes the incoming
2126 connection. However "connection reset by peer" seems to be common, and
2127 isn't really an error worthy of noting specially, so that particular error
2130 62. When Exim is trying to find all the local interfaces, it used to panic and
2131 die if the ioctl to get the interface flags failed. However, it seems that
2132 on at least one OS (Solaris 9) it is possible to have an interface that is
2133 included in the list of interfaces, but for which you get a failure error
2134 for this call. This happens when the interface is not "plumbed" into a
2135 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2136 failure of the "get flags" call assumes that the interface is down.
2138 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2139 makes life easier for people who are doing arithmetic on fields extracted
2140 from dates, where you often get leading zeros that should not be
2141 interpreted as octal.
2143 64. Added qualify_domain to the redirect router, to override the global
2146 65. If a pathologically long header line contained very many addresses (the
2147 report of this problem mentioned 10 000) and each of them was rewritten,
2148 Exim could use up a very large amount of memory. (It kept on making new
2149 copies of the header line as it rewrote, and never released the old ones.)
2150 At the expense of a bit more processing, the header rewriting function has
2151 been changed so that it no longer eats memory in this way.
2153 66. The generation of the Received: header has been moved from the time that a
2154 message starts to be received, to the time that it finishes. The timestamp
2155 in the Received: header should now be very close to that of the <= log
2156 line. There are two side-effects of this change:
2158 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2159 logged header lines no longer include the local Received: line, because
2160 it has not yet been created. The same applies to a copy of the message
2161 that is returned to a non-SMTP sender when a message is rejected.
2163 (b) When a filter file is tested using -bf, no additional Received: header
2164 is added to the test message. After some thought, I decided that this
2167 This change does not affect the value of $received_for. It is still set
2168 after address rewriting, but before local_scan() is called.
2170 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2172 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2173 gave an unhelpful panic error message, and a defer error. I have managed to
2174 change this behaviour so that it now rejects any supplied certificate,
2175 which seems right, as the list of acceptable certificates is empty.
2177 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2178 gave an unhelpful defer error. I have not managed to make this reject any
2179 supplied certificates, but the error message it gives is "no certificate
2180 supplied", which is not helpful.
2182 70. exigrep's output now also includes lines that are not associated with any
2183 message, but which match the given pattern. Implemented by a patch from
2184 Martin Sluka, which also tidied up the Perl a bit.
2186 71. Recipient callout verification, like sender verification, was using <> in
2187 the MAIL FROM command. This isn't really the right thing, since the actual
2188 sender may affect whether the remote host accepts the recipient or not. I
2189 have changed it to use the actual sender in the callout; this means that
2190 the cache record is now keyed on a recipient/sender pair, not just the
2191 recipient address. There doesn't seem to be a real danger of callout loops,
2192 since a callout by the remote host to check the sender would use <>.
2193 [SEE ABOVE: changed after hitting problems.]
2195 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2196 temporary errors. However, in the case of such a code being given after
2197 the end of a data transmission (i.e. after ".") Exim was failing to write
2198 a retry record for the message. (Yes, there was some broken host that was
2199 actually sending 8xx at this point.)
2201 73. An unknown lookup type in a host list could cause Exim to panic-die when
2202 the list was checked. (An example that provoked this was putting <; in the
2203 middle of a list instead of at the start.) If this happened during a DATA
2204 ACL check, a -D file could be left lying around. This kind of configuration
2205 error no longer causes Exim to die; instead it causes a defer errror. The
2206 incident is still logged to the main and panic logs.
2208 74. Buglet left over from Exim 3 conversion. The message "too many messages
2209 in one connection" was written to the rejectlog but not the mainlog, except
2210 when address rewriting (yes!) was being logged.
2212 75. Added write_rejectlog option.
2214 76. When a system filter was run not as root (that is, when system_filter_user
2215 was set), the values of the $n variables were not being returned to the
2216 main process; thus, they were not subsequently available in the $sn
2219 77. Added +return_path_on_delivery log selector.
2221 78. A connection timeout was being treated differently from recipients deferred
2222 when testing hosts_max_try with a message that was older than the host's
2223 retry timeout. (The host should not be counted, thus allowing all hosts to
2224 be tried at least once before bouncing.) This may have been the cause of an
2225 occasionally reported bug whereby a message would remain on the queue
2226 longer than the retry timeout, but would be bounced if a delivery was
2227 forced. I say "may" because I never totally pinned down the problem;
2228 setting up timeout/retry tests is difficult. See also the next item.
2230 79. The ultimate address timeout was not being applied to errors that involved
2231 a combination of host plus message (for example, a timeout on a MAIL
2232 command). When an address resolved to a number of possible hosts, and they
2233 were not all tried for each delivery (e.g. because of hosts_max_try), a
2234 message could remain on the queue longer than the retry timeout.
2236 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2239 81. Fixed an obscure SMTP outgoing bug which required at least the following
2240 conditions: (a) there was another message waiting for the same server;
2241 (b) the server returned 5xx to all RCPT commands in the first message so
2242 that the message was not completed; (c) the server dropped the connection
2243 or gave a negative response to the RSET that Exim sends to abort the
2244 transaction. The observed case was a dropped connection after DATA that had
2245 been sent in pipelining mode. That is, the server had advertised PIPELINING
2246 but was not implementing it correctly. The effect of the bug was incorrect
2247 behaviour, such as trying another host, and this could lead to a crash.
2253 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2254 and daemon.c were passed as pointers to ints; they should have been
2255 pointers to socklen_t variables (which are typically unsigned ints).
2257 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2260 3. Fixed a really odd bug that affected only the testing scheme; patching a
2261 certain fixed string in the binary changed the value of another string that
2262 happened to be identical to the end of the original first string.
2264 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2265 name", it returns that address as the IP address. On some operating
2266 systems (e.g. Solaris), it also passes back the IP address string as the
2267 "host name". However, on others (e.g. Linux), it passes back an empty
2268 string. Exim wasn't checking for this, and was changing the host name to an
2269 empty string, assuming it had been canonicized.
2271 5. Although rare, it is permitted to have more than one PTR record for a given
2272 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2273 all the names associated with an address, because they do in Solaris.
2274 However, it seems that they do not in Linux for data that comes from the
2275 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2276 I found this out when I moved to a new Linux workstation and tried to run
2277 the Exim test suite.
2279 To get round this problem I have changed the code so that it now does its
2280 own call to the DNS to look up PTR records when searching for a host name.
2281 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2282 addresses that are only in /etc/hosts are still found.
2284 This behaviour is, however, controlled by an option called host_lookup_
2285 order, which defaults to "bydns:byaddr". If people want to use the other
2286 order, or indeed, just use one or the other means of lookup, they can
2287 specify it in this variable.
2289 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2290 some operating systems, this comes back from gethostbyaddr() as an empty
2291 string, and this is what Exim used to test for. However, it seems that in
2292 other systems, "." is yielded. Exim now tests for this case too.
2294 7. The values of check_spool_space and check_log_space are now held internally
2295 as a number of kilobytes instead of an absolute number of bytes. If a
2296 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2297 kilobyte. This means that much larger values can be stored.
2299 8. Exim monitor: an attempt to get the action menu when not actually pointing
2300 at a message produces an empty menu entitled "No message selected". This
2301 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2302 no entries in it ("Shell widget menu has zero width and/or height"). So I
2303 have added a single, blank menu entry in this case.
2305 9. Added ${quote_local_part.
2307 10. MIME decoding is now applied to the contents of Subject: header lines when
2310 11. Now that a reference to $sender_host_address automatically causes a reverse
2311 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2312 host lookup before query-style lookups in lists that might use this
2313 variable. This has therefore been abolished, and the "net-" prefix is no
2314 longer necessary for query-style lookups.
2316 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2317 have been a typo for LFLAGS, so it has been changed.
2319 13. The install script calls Exim with "-C /dev/null" in order to find the
2320 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2321 to be output. Howeve, since Exim outputs its version number before the
2322 error, it didn't break the script. It just looked ugly. I fixed this by
2323 always allowing "-C /dev/null" if the caller is root.
2325 14. Ignore overlarge ACL variable number when reading spool file - insurance
2326 against a later release with more variables having written the file.
2328 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2329 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2330 this, as well as the form without the "IPv6" on the front.
2332 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2333 OS/Makefile-Darwin file.
2335 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2336 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2338 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2339 variable. However, some OS use ints for inodes. I've added cast to long int
2340 to get rid of the compiler warning.
2342 19. I had forgotten to lock out "/../" in configuration file names when
2343 ALT_CONFIG_PREFIX was set.
2345 20. Routers used for verification do not need to specify transports. However,
2346 if such a router generated a host list, and callout was configured, Exim
2347 crashed, because it could not find a port number from the (non-existent)
2348 transport. It now assumes port 25 in this circumstance.
2350 21. Added the -t option to exigrep.
2352 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2353 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2354 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2355 them is set, LOOKUP_LSEARCH is forced.
2357 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2359 24. Added sender and host information to the "rejected by local_scan()" log
2360 line; previously there was no indication of these.
2362 25. Added .include_if_exists.
2364 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2365 Linux. It turns out that not all file systems support this. Apparently some
2366 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2367 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2368 not supported on the file descriptor, is now ignored when Exim is trying to
2369 sync a directory. This applies only to Linux.
2371 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2373 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2374 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2377 29. Added support for maildirsize files from supplied patch (modified a bit).
2379 30. The use of :fail: followed by an empty string could lead Exim to respond to
2380 sender verification failures with (e.g.):
2382 550 Verification failed for <xxx>
2383 550 Sender verify failed
2385 where the first response line was missing the '-' that indicates it is not
2386 the final line of the response.
2388 31. The loop for finding the name of the user that called Exim had a hardwired
2389 limit of 10; it now uses the value of finduser_retries, which is used for
2390 all other user lookups.
2392 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2395 33. Exim was neglecting to zero errno before one call of strtol() when
2396 expanding a string and expecting an integer value. On some systems this
2397 resulted in spurious "integer overflow" errors. Also, it was casting the
2398 result into an int without checking.
2400 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2401 did not work. The code looks as if it has *never* worked, though it appears
2402 to have been documented since at least releast 1.62. I have made it work.
2404 35. The "timeout_DNS" error in retry rules, also documented since at least
2405 1.62, also never worked. As it isn't clear exactly what this means, and
2406 clearly it isn't a major issue, I have abolished the feature by treating it
2407 as "timeout", and writing a warning to the main and panic logs.
2409 36. The display of retry rules for -brt wasn't always showing the error code
2412 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2413 timeout_connect_A, timeout_connect_MX.
2415 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2416 to the empty sender.
2418 39. The daemon was not analysing the content of -oX till after it had closed
2419 stderr and disconnected from the controlling terminal. This meant that any
2420 syntax errors were only noted on the panic log, and the return code from
2421 the command was 0. By re-arranging the code a little, I've made the
2422 decoding happen first, so such errors now appear on stderr, and the return
2423 code is 1. However, the actual setting up of the sockets still happens in
2424 the disconnected process, so errors there are still only recorded on the
2427 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2428 connections (as happens on some IP stacks) was logged at start up time as
2429 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2430 it from "IPv6 and IPv4", which means that two separate sockets are being
2433 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2434 says whether AF_INET or AF_INET6 was passed as an argument.
2436 42. Exiwhat output was messed up when time zones were included in log
2439 43. Exiwhat now gives more information about the daemon's listening ports,
2440 and whether -tls-on-connect was used.
2442 44. The "port" option of the smtp transport is now expanded.
2444 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2445 silently ignored. Now an error message is written to the main and panic
2448 46. There's a new ACL modifier called "logwrite" which writes to a log file
2449 as soon as it is encountered.
2451 47. Added $local_user_uid and $local_user_gid at routing time.
2453 48. Exim crashed when trying to verify a sender address that was being
2456 49. Exim was recognizing only a space character after ".include". It now also
2457 recognizes a tab character.
2459 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2460 extracting the relevant information from the specification. The man page no
2461 longer contains scrambled data for the -d option, and I've added a section
2462 at the front about calling Exim under different names.
2464 51. Added "extra_headers" argument to the "mail" command in filter files.
2466 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2469 53. Installed eximstats 1.29.
2471 54. Added transport_filter_timeout as a generic transport option.
2473 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2474 Cc: header lines. This was required by RFC 822, but it not required by RFC
2477 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2478 incoming messages that did not have them. Now it does so only if the
2479 message originates locally, that is, if there is no associated remote host
2480 address. When Resent- header lines are present, this applies to the Resent-
2481 lines rather than the non-Resent- lines.
2483 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2484 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2486 58. Messages for configuration errors now include the name of the main
2487 configuration file - useful now that there may be more than one file in a
2488 list (.included file names were always shown).
2490 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2491 for those rare installations that do not start the daemon as root or run it
2492 setuid root. I've cut out the call to initgroups() if the daemon is not
2495 60. The Exim user and group can now be bound into the binary as text strings
2496 that are looked up at the start of Exim's processing.
2498 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2500 62. Added $mailstore_basename variable.
2502 63. Installed patch to sieve.c from Michael Haardt.
2504 64. When Exim failed to open the panic log after failing to open the main log,
2505 the original message it was trying to log was written to stderr and debug
2506 output, but if they were not available (the usual case in production), it
2507 was lost. Now it is written to syslog before the two lines that record the
2508 failures to open the logs.
2510 65. Users' Exim filters run in subprocesses under the user's uid. It is
2511 possible for a "deliver" command or an alias in a "personal" command to
2512 provoke an address rewrite. If logging of address rewriting is configured,
2513 this fails because the process is not running as root or exim. There may be
2514 a better way of dealing with this, but for the moment (because 4.30 needs
2515 to be released), I have disabled address rewrite logging when running a
2516 filter in a non-root, non-exim process.
2522 1. The buildconfig auxiliary program wasn't quoting the value set for
2523 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2524 not defined. This bug was masked in 4.22 by the effect that was fixed in
2527 2. Some messages that were rejected after a message id was allocated were
2528 shown as "incomplete" by exigrep. It no longer does this for messages that
2529 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2531 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2532 have allow_domain_literals set, the ID did not get logged in the <= line.
2533 Domain literals are now always recognized in Message-ID: header lines.
2535 4. The first argument for a ${extract expansion item is the key name or field
2536 number. Leading and trailing spaces in this item were not being ignored,
2537 causing some misleading effects.
2539 5. When deliver_drop_privilege was set, single queue runner processes started
2540 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2541 same command in the process it spins off) were not dropping privilege.
2543 6. When the daemon running as "exim" started a queue runner, it always
2544 re-executed Exim in the spun-off process. This is a waste of effort when
2545 deliver_drop_privilege is set. The new process now just calls the
2546 queue-runner function directly.
2552 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2555 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2556 ignored. Though the use of -oP was forcing the writing of a pid file, it
2557 was always written to the default place.
2559 3. If the message "no IP address found for host xxxx" is generated during
2560 incoming verification, it is now followed by identification of the incoming
2561 connection (so you can more easily find what provoked it).
2563 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2565 5. Added some features to "harden" Exim a bit more against certain attacks:
2567 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2568 be put in Local/Makefile. This is like the never_users runtime option,
2569 but it cannot be overridden. The default setting is "root".
2571 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2572 prefix string with which any file named in a -C command line option
2575 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2576 is retained for -C and -D only if the caller of Exim is root. Without
2577 it, the exim user may also use -C and -D and retain privilege.
2579 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2580 command line option is disabled.
2582 6. Macro names set by the -D option must start with an upper case letter, just
2583 like macro names defined in the configuration file.
2585 7. Added "dereference=" facility to LDAP.
2587 8. Two instances of the typo "uknown" in the source files are fixed.
2589 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2590 the Configure-Makefile script screwed up while processing it.
2592 10. Incorporated PCRE 4.4.
2594 11. The SMTP synchronization check was not operating right at the start of an
2595 SMTP session. For example, it could not catch a HELO sent before the client
2596 waited for the greeting. There is now a check for outstanding input at the
2597 point when the greeting is written. Because of the duplex, asynchronous
2598 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2599 way, but not yet received, when the check is performed.
2601 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2602 on TCP/IP sockets, because this apparently causes some broken clients to
2605 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2606 unchanged) from the Cygwin maintainer.
2608 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2609 was supported instead of testing for mbx. Effectively a typo.
2611 15. The spa authenticator server code was not checking that the input it
2612 received was valid base64.
2614 16. The debug output line for the "set" modifier in ACLs was not showing the
2615 name of the variable that was being set.
2617 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2619 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2620 lookup on its own. Something else was needed to trigger the lookup. For
2621 example, a match in host_lookup or the need for a host name in a host list.
2622 Now, if $sender_host_name is referenced and the host name has not yet been
2623 looked up, a lookup is performed. If the lookup fails, the variable remains
2624 empty, and $host_lookup_failed is set to "1".
2626 19. Added "eqi" as a case-independent comparison operator.
2628 20. The saslauthd authentication condition could segfault if neither service
2629 nor realm was specified.
2631 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2632 error message that was logged was misleading, and incoming SMTP
2633 connections were dropped. The message is now more accurate, and temporary
2634 errors are given to SMTP connections.
2636 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2637 (except RSET) with a 421 error, until QUIT is received. However, it was
2638 failing to send a response to QUIT.
2640 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2641 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2642 ACL is now run after the helo_try_verify_hosts code.
2644 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2645 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2646 case-independent, so other case variants are also recognized). Apparently
2647 some systems use these upper case variants.
2649 25. If more than two messages were waiting for the same host, and a transport
2650 filter was specified for the transport, Exim sent two messages over the
2651 same TCP/IP connection, and then failed with "socket operation on non-
2652 socket" when it tried to send the third.
2654 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2656 27. The extern definition of crypt16() in expand.c was not being excluded when
2657 the OS had its own crypt16() function.
2659 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2660 as a preferred synonym for return_size_limit, both as an option and as an
2663 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2665 30. Changed the default configuration ACL to relax the local part checking rule
2666 for addresses that are not in any local domains. For these addresses,
2667 slashes and pipe symbols are allowed within local parts, but the sequence
2668 /../ is explicitly forbidden.
2670 31. SPA server authentication was not clearing the challenge buffer before
2673 32. log_message in a "warn" ACL statement was writing to the reject log as
2674 well as to the main log, which contradicts the documentation and doesn't
2675 seem right (because no rejection is happening). So I have stopped it.
2677 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2678 However, I am unable to do any testing of this.
2680 34. Fixed an infelicity in the appendfile transport. When checking directories
2681 for a mailbox, to see if any needed to be created, it was accidentally
2682 using path names with one or more superfluous leading slashes; tracing
2683 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2685 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2686 contents are added to the log line that is written for every discarded
2687 recipient. (Previously a log_message setting was ignored.)
2689 36. The ${quote: operator now quotes the string if it is empty.
2691 37. The install script runs exim in order to find its version number. If for
2692 some reason other than non-existence or emptiness, which it checks, it
2693 could not run './exim', it was installing it with an empty version number,
2694 i.e. as "exim-". This error state is now caught, and the installation is
2697 38. An argument was missing from the function that creates an error message
2698 when Exim fails to connect to the socket for saslauthd authentication.
2699 This could cause Exim to crash, or give a corrupted message.
2701 39. Added isip, isip4, and isip6 to ${if conditions.
2703 40. The ACL variables $acl_xx are now saved with the message, and can be
2704 accessed later in routers, transports, and filters.
2706 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2707 strings in the file are not string-expanded.
2709 42. If a MAIL command specified a SIZE value that was too large to fit into an
2710 int variable, the check against message_size_limit failed. Such values are
2711 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2712 day this will have to be increased, but I don't think I want to be around
2713 when emails are that large.
2720 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2721 iconv() is not standard in FreeBSD.
2723 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2724 IPv6 enabled. The observed symptom was a segmentation fault on return from
2725 the function os_common_find_running_interfaces() in src/os.c.
2727 3. In the check_special_case() function in daemon.c I had used "errno" as an
2728 argument name, which causes warnings on some systems. This was basically a
2729 typo, since it was named "eno" in the comments!
2731 4. The code that waits for the clock to tick (at a resolution of some fraction
2732 of a second) so as to ensure message-id uniqueness was always waiting for
2733 at least one whole tick, when it could have waited for less. [This is
2734 almost certainly not relevant at current processor speeds, where it is
2735 unlikely to ever wait at all. But we try to future-proof.]
2737 5. The function that sleeps for a time interval that includes fractions of a
2738 second contained a race. It did not block SIGALRM between setting the
2739 timer, and suspending (a couple of lines later). If the interval was short
2740 and the sigsuspend() was delayed until after it had expired, the suspension
2741 never ended. On busy systems this could lead to processes getting stuck for
2744 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2745 process, before it forks any delivery processes. The open lookup caching
2746 mechanism meant that the open file or database connection was passed into
2747 the delivery process. The problem was that delivery processes always tidy
2748 up cached lookup data. This could cause a problem for the next delivery
2749 process started by the queue runner, because the external queue runner
2750 process does not know about the closure. So the next delivery process
2751 still has data in the lookup cache. In the case of a file lookup, there was
2752 no problem because closing a file descriptor in a subprocess doesn't affect
2753 the parent. However, if the lookup was caching a connection to a database,
2754 the connection was closed, and the second delivery process was likely to
2755 see errors such as "PGSQL: query failed: server closed the connection
2756 unexpectedly". The problem has been fixed by closing all cached lookups
2757 in a queue runner before running a delivery process.
2759 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2760 seem to have the "const" qualifier which it has on other OS. I've
2763 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2764 *of the same type* (client or server) with the same public name that an
2765 error should be diagnosed.
2767 9. When Exim looked up a host name for an IP address, but failed to find the
2768 original IP address when looking up the host name (a safety check), it
2769 output the message "<ip address> does not match any IP for NULL", which was
2770 confusing, to say the least. The bug was that the host name should have
2771 appeared instead of "NULL".
2773 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2774 user that is built into the binary, and the -C or -D options is used, root
2775 privilege is dropped before the configuration file is read. In addition,
2776 logging is switched to stderr instead of the normal log files. If the
2777 configuration then re-defines the Exim user, the unprivileged environment
2778 is probably not what is expected, so Exim logs a panic warning message (but
2781 However, if deliver_drop_privilege is set, the unprivileged state may well
2782 be exactly what is intended, so the warning has been cut out in that case,
2783 and Exim is allowed to try to write to its normal log files.
2789 1. smtp_return_error_details was not giving details for temporary sender
2790 or receiver verification errors.
2792 2. Diagnose a configuration error if two authenticators have the same public
2795 3. Exim used not to create the message log file for a message until the first
2796 delivery attempt. This could be confusing when incoming messages were held
2797 for policy or load reasons. The message log file is now created at the time
2798 the message is received, and an initial "Received" line is written to it.
2800 4. The automatically generated man page for command line options had a minor
2801 bug that caused no ill effects; however, a more serious problem was that
2802 the procedure for building the man page automatically didn't always
2803 operate. Consequently, release 4.20 contains an out-of-date version. This
2804 shouldn't happen again.
2806 5. When building Exim with embedded Perl support, the script that builds the
2807 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2808 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2810 6. The freeze_tell option was not being used for messages that were frozen on
2811 arrival, either by an ACL or by local_scan().
2813 7. Added the smtp_incomplete_transaction log selector.
2815 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2816 was accepting AUTH without a new EHLO.
2818 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2819 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2822 10. Logging of TCP/IP connections (when configured) now happens in the main
2823 daemon process instead of the child process, so that the TCP/IP connection
2824 count is more accurate (but it can never be perfect).
2826 11. The use of "drop" in a nested ACL was not being handled correctly in the
2827 outer ACL. Now, if condition failure induced by the nested "drop" causes
2828 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2829 or "require"), the connection is dropped.
2831 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2832 that yield "discard" can now be used with an "accept" or a "discard" verb,
2833 but an error is generated for any others (because I can't see a useful way
2834 to define what should happen).
2836 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2837 are now processed in the same way as lines in the Exim configuration file.
2838 In particular, continuation lines are supported.
2840 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2842 15. Added -ti meaning -t -i.
2844 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2845 domains, and warn by logging if others are found.
2847 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2848 returned from GIFCONF when Exim is trying to find the list of interfaces on
2849 a host. The code in os.c has been modified to copy each ifreq to an aligned
2850 structure in all cases.
2852 Also, in some cases, the returned ifreq's were being copied to a 'struct
2853 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2854 means the last couple of bytes of an IPv6 address could be chopped if the
2855 ifreq contained only a normal sockaddr (14 bytes storage).
2857 18. Named domain lists were not supported in the hosts_treat_as_local option.
2858 An entry such as +xxxx was not recognized, and was treated as a literal
2861 19. Ensure that header lines added by a DATA ACL are included in the reject log
2862 if the ACL subsequently rejects the message.
2864 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2865 MD5 (which is deprecated).
2867 21. When testing a filter file using -bf, Exim was writing a message when it
2868 took the sender from a "From " line in the message, but it was not doing so
2869 when it took $return_path from a Return-Path: header line. It now does.
2871 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2872 with a valid header line field name (a series of printing characters
2873 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2875 23. Changed "disc" in the source to "disk" to conform to the documentation and
2876 the book and for uniformity.
2878 24. Ignore Sendmail's -Ooption=value command line item.
2880 25. When execve() failed while trying to run a command in a pipe transport,
2881 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2882 could be confused with a return value of 69 from the command itself. This
2883 has been changed to 127, the value the shell returns if it is asked to run
2884 a non-existent command. The wording for the related log line suggests a
2885 non-existent command as the problem.
2887 26. If received_header_text expands to an empty string, do not add a Received:
2888 header line to the message. (Well, it adds a token one on the spool, but
2889 marks it "old" so that it doesn't get used or transmitted.)
2891 27. Installed eximstats 1.28 (addition of -nt option).
2893 28. There was no check for failure on the call to getsockname() in the daemon
2894 code. This can fail if there is a shortage of resources on the system, with
2895 ENOMEM, for example. A temporary error is now given on failure.
2897 29. Contrary to the C standard, it seems that in some environments, the
2898 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2899 program. Exim now does this explicitly; it affects the formatting of
2900 timestamps using strftime().
2902 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2903 complaints. I've now initialized all the variables, to avoid this.
2905 32. Header lines added by a system filter were not being "seen" during
2906 transport-time rewrites.
2908 33. The info_callback() function passed to OpenSSL is set up with type void
2909 (*)(SSL *, int, int), as described somewhere. However, when calling the
2910 function (actually a macro) that sets it up, the type void(*)() is
2911 expected. I've put in a cast to prevent warnings from picky compilers.
2913 34. If a DNS black list lookup found a CNAME record, but there were no A
2914 records associated with the domain it pointed at, Exim crashed.
2916 35. If a DNS black list lookup returned more than one A record, Exim ignored
2917 all but the first. It now scans all returned addresses if a particular IP
2918 value is being sought. In this situation, the contents of the
2919 $dnslist_value variable are a list of all the addresses, separated by a
2922 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2923 to accept a host name and all its aliases if the forward lookup for any of
2924 them yielded the IP address of the incoming connection. Now it accepts only
2925 those names whose forward lookup yields the correct IP address. Any other
2926 names are discarded. This closes a loophole whereby a rogue DNS
2927 administrator could create reverse DNS records to break through a
2928 wildcarded host restriction in an ACL.
2930 37. If a user filter or a system filter that ran in a subprocess used any of
2931 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2932 the wrong values were passed to the pipe command ($thisaddress had the
2933 value of $0, $0 had the value of $1, etc). This bug was introduced by
2934 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2936 38. Improved the line breaking for long SMTP error messages from ACLs.
2937 Previously, if there was no break point between 40 and 75 characters, Exim
2938 left the rest of the message alone. Two changes have been made: (a) I've
2939 reduced the minimum length to 35 characters; (b) if it can't find a break
2940 point between 35 and 75 characters, it looks ahead and uses the first one
2941 that it finds. This may give the occasional overlong line, but at least the
2942 remaining text gets split now.
2944 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2945 file descriptors might be low, and that setting 1000 would always raise it.
2946 It turns out that in some environments, the limit is already over 1000 and
2947 that lowering it causes trouble. So now Exim takes care not to decrease it.
2949 40. When delivering a message, the value of $return_path is set to $sender_
2950 address at the start of routing (routers may change the value). By an
2951 oversight, this default was not being set up when an address was tested by
2952 -bt or -bv, which affected the outcome if any router or filter referred to
2955 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2956 log only when "message" or "log_message" are set. However, if one of the
2957 conditions was an address verification, or a call to a nested ACL, the
2958 messages generated by the underlying test were being passed through. This
2959 no longer happens. The underlying message is available in $acl_verify_
2960 message for both "message" and "log_message" expansions, so it can be
2961 passed through if needed.
2963 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2964 new expansion $bh_ to give the encoded byte string without charset
2965 translation. Translation happens only if iconv() is available; HAVE_ICONV
2966 indicates this at build time. HEADERS_CHARSET gives the charset to
2967 translate to; headers_charset can change it in the configuration, and
2968 "headers charset" can change it in an individual filter file.
2970 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2971 that creates RFC 2047 encoded "words" labels them as that charset instead
2972 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2973 expansion operator; (ii) when Exim creates a From: line for a local
2974 message; (iii) when a header line is rewritten to include a "phrase" part.
2976 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2977 buggy, causing it to skip the first lines of messages whose message ID
2978 ended in 'D'. This would not have bitten before Exim release 4.14, because
2979 message IDs were unlikely to end in 'D' before then. The effect was to have
2980 incorrect size information for certain domains.
2982 45. #include "config.h" was missing at the start of the crypt16.c module. This
2983 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2986 46. If there was a timeout during a "random" callout check, Exim treated it as
2987 a failure of the random address, and carried on sending RSET and the real
2988 address. If the delay was just some slowness somewhere, the response to the
2989 original RCPT would be taken as a response to RSET and so on, causing
2990 mayhem of various kinds.
2992 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2993 when I implemented it. It didn't allow for the fact that some option values
2994 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2995 the right test for positive values.
2997 48. Domain names in DNS records are case-independent. Exim always looks them up
2998 in lower case. Some resolvers return domain names in exactly the case they
2999 appear in the zone file, that is, they may contain uppercase letters. Not
3000 all resolvers do this - some return always lower case. Exim was treating a
3001 change of case by a resolver as a change of domain, similar to a widening
3002 of a domain abbreviation. This triggered its re-routing code and so it was
3003 trying to route what was effectively the same domain again. This normally
3004 caused routing to fail (because the router wouldn't handle the domain
3005 twice). Now Exim checks for this case specially, and just changes the
3006 casing of the domain that it ultimately uses when it transmits the message
3009 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3012 50. If a filter generated a file delivery with a non-absolute name (possible if
3013 no home directory exists for the router), the forbid_file option was not
3016 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3017 the existing equality matching.
3019 52. Exim was using ints instead of ino_t variables in some places where it was
3020 dealing with inode numbers.
3022 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3023 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3024 called TMPDIR, and if it finds it is different, it changes its value.
3026 54. The smtp_printf() function is now made available to local_scan() so
3027 additional output lines can be written before returning. There is also an
3028 smtp_fflush() function to enable the detection of a dropped connection.
3029 The variables smtp_input and smtp_batched_input are exported to
3032 55. Changed the default runtime configuration: the message "Unknown user"
3033 has been removed from the ACL, and instead placed on the localuser router,
3034 using the cannot_route_message feature. This means that any verification
3035 failures that generate their own messages won't get overridden. Similarly,
3036 the "Unrouteable address" message that was in the ACL for unverifiable
3037 relay addresses has also been removed.
3039 56. Added hosts_avoid_esmtp to the smtp transport.
3041 57. The exicyclog script was not checking for the esoteric option
3042 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3043 will work only if exicyclog is run under the appropriate euid.
3045 58. Following a discussion on the list, the rules by which Exim recognises line
3046 endings on incoming messages have been changed. The -dropcr and drop_cr
3047 options are now no-ops, retained only for backwards compatibility. The
3048 following line terminators are recognized: LF CRLF CR. However, special
3049 processing applies to CR:
3051 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3052 nor a local message in the state where . is a terminator.
3054 (ii) If a bare CR is encountered in a header line, an extra space is added
3055 after the line terminator so as not to end the header. The reasoning
3056 behind this is that bare CRs in header lines are most likely either
3057 to be mistakes, or people trying to play silly games.
3059 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3060 was being incorrectly given as 18 bytes larger than it should have been.
3061 This is a VOB (very old bug).
3063 60. This may never have affected anything current, but just in case it has:
3064 When the local host is found other than at the start of a list of hosts,
3065 the local host, those with the same MX, and any that follow, are discarded.
3066 When the list in question was part of a longer list of hosts, the following
3067 hosts (not currently being processed) were also being discarded. This no
3068 longer happens. I'm not sure if this situation could ever has previously
3071 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3074 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3075 header. This is something that is recommended in a new Internet Draft, and
3076 is something that is documented as being done by Sendmail. There are two
3077 possible values. For messages generated by the autoreply transport, Exim
3080 Auto-Submitted: auto-replied
3082 whereas for all other generated messages (e.g. bounces) it adds
3084 Auto-Submitted: auto-generated
3086 63. The "personal" condition in filters now includes a test for the
3087 Auto-Submitted: header. If it contains the string "auto-" the message it
3088 not considered personal.
3090 64. Added rcpt_include_affixes as a generic transport option.
3092 65. Added queue_only_override (default true).
3094 66. Added the syslog_duplication option.
3096 67. If what should have been the first header line of a message consisted of
3097 a space followed by a colon, Exim was mis-interpreting it as a header line.
3098 It isn't of course - it is syntactically invalid and should therefore be
3099 treated as the start of the message body. The misbehaviour could have
3100 caused a number of strange effects, including loss of data in subsequent
3101 header lines, and spool format errors.
3103 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3104 client host had authenticated. This control can now be exercised by an ACL
3105 for more flexibility.
3107 69. By default, callouts do not happen when testing with -bh. There is now a
3108 variant, -bhc, which does actually run the callout code, including
3109 consulting and updating the callout cache.
3111 70. Added support for saslauthd authentication, courtesy of Alexander
3114 71. If statvfs() failed on the spool or log directories while checking their
3115 size for availability, Exim confusingly gave the error "space shortage".
3116 Furthermore, in debugging mode it crashed with a floating point exception.
3117 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3118 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3119 serious problem, Exim now writes to the main and panic logs when this
3120 happens, with details of the failure. It then refuses to accept the
3121 incoming message, giving the message "spool directory problem" or "log
3122 directory problem" with a 421 code for SMTP messages.
3124 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3125 0, 1, and 2 exist, because some OS complain for execs without them (see
3126 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3127 descriptors. However, the code omitted to check that the open succeeded,
3128 causing mysterious errors if for some reason the permissions on /dev/null
3129 got screwed. Now Exim writes a message to the main and panic logs, and
3130 bombs out if it can't open /dev/null.
3132 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3133 interact so that it is all more flexible. It is supposed to remain
3134 backwards compatible. Also added extra_local_interfaces.
3136 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3137 to bomb out with an assertion failure - to the client this appears as a
3138 connection drop. This problem occurs in the part of the code that was taken
3139 from the Samba project. Fortunately, the assertion is in a very simple
3140 function, so I have fixed this by reproducing the function inline in the
3141 one place where it is called, and arranging for authentication to fail
3142 instead of killing the process with assert().
3144 75. The SPA client code was not working when the server requested OEM rather
3145 than Unicode encoding.
3147 76. Added code to make require_files with a specific uid setting more usable in
3148 the case where statting the file as root fails - usually a non-root-mounted
3149 NFS file system. When this happens and the failure is EACCES, Exim now
3150 forks a subprocess and does the per-uid checking as the relevant uid.
3152 77. Added process_log_path.
3154 78. If log_file_path was not explicitly set, a setting of check_log_space or
3155 check_log_inodes was ignored.
3157 79. If a space check for the spool or log partitions fails, the incident is now
3158 logged. Of course, in the latter case the data may get lost...
3160 80. Added the %p formatting code to string_format() so that it can be used to
3161 print addresses in debug_print(). Adjusted all the address printing in the
3162 debugging in store.c to use %p rather than %d.
3164 81. There was a concern that a line of code in smtp_in.c could overflow a
3165 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3166 initially expressed, the concern was not well-founded, because trailing
3167 spaces are removed early. However, if the trailing spaces were followed by
3168 a NULL, they did not get removed, so the overflow was possible. Two fixes
3171 (a) I re-wrote the offending code in a cleaner fashion.
3172 (b) If an incoming SMTP command contains a NULL character, it is rejected
3175 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3176 runs initgroups(), so that if the Exim user is in any additional groups,
3177 they will be used during message reception.
3183 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / heiko / exim.git / blob