1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.186 2005/08/01 14:00:34 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.
28 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
30 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
31 command itself ended in error, the underlying message about the transport
32 filter timeout was being overwritten with the pipe command error. Now the
33 underlying error message should be appended to the second error message.
35 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
37 PH/02 Reduce the amount of output that "make" produces by default. Full output
38 can still be requested.
40 PH/03 The warning log line about a condition test deferring for a "warn" verb
41 was being output only once per connection, rather than after each
42 occurrence (because it was using the same function as for successful
43 "warn" verbs). This seems wrong, so I have changed it.
45 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
46 it should not have, which might have caused a crash in the right
47 circumstances, but probably never did.
49 PH/04 Installed a modified version of Tony Finch's patch to make submission
50 mode fix the return path as well as the Sender: header line, and to
51 add a /name= option so that you can make the user's friendly name appear
54 TF/03 Added the control = fakedefer ACL modifier.
56 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
57 Mark Lowes for thorough testing.
59 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
61 TK/03 Merged latest SRS patch from Miles Wilton.
63 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
64 with the definition in sysexits.h (which is #included earlier).
65 Fortunately, Exim does not actually use EX_OK. The code used to try to
66 preserve the sysexits.h value, by assumimg that macro definitions were
67 scanned for macro replacements. I have been disabused of this notion,
68 so now the code just undefines EX_OK before #including unistd.h.
70 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
71 in the smtp transport. When a block could not be written in a single
72 write() function, the timeout was being re-applied to each part-write.
73 This seems wrong - if the receiver was accepting one byte at a time it
74 would take for ever. The timeout is now adjusted when this happens. It
75 doesn't have to be particularly precise.
77 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
78 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
80 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
81 without a domain if the check to <postmaster@domain> fails.
83 SC/01 Eximstats: added -xls and the ability to specify output files
84 (patch written by Frank Heydlauf).
86 SC/02 Eximstats: use FileHandles for outputing results.
88 SC/03 Eximstats: allow any combination of xls, txt, and html output.
90 SC/04 Eximstats: fixed display of large numbers with -nvr option
92 SC/05 Eximstats: fixed merging of reports with empty tables.
94 SC/06 Eximstats: added the -include_original_destination flag
96 SC/07 Eximstats: removed tabs and trailing whitespace.
98 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
100 TK/06 MBOX spool code: Add real "From " MBOX separator line
101 so the .eml file is really in mbox format (even though
102 most programs do not really care). Patch from Alex Miller.
104 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
105 The latter is generated from $received_to and is only set if the
106 message has one envelope recipient. SA can use these headers,
107 obviously out-of-the-box. Patch from Alex Miller.
109 PH/08 The ${def test on a variable was returning false if the variable's
110 value was "0", contrary to what the specification has always said!
111 The result should be true unless the variable is empty.
113 PH/09 The syntax error of a character other than { following "${if
114 def:variable_name" (after optional whitespace) was not being diagnosed.
115 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
116 accidental colon was present, for example, could give incorrect results.
118 PH/10 Tidied the code in a number of places where the st_size field of a stat()
119 result is used (not including appendfile, where other changes are about
122 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
123 This involved changing a lot of size variables from int to off_t. It
124 should work with maildirs and everything.
126 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
127 spamd dying while we are connected to it.
129 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
132 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
133 function for BDB 4.3.
135 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
137 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
138 the sources for PCRE have been re-arranged and more files are now
141 PH/15 The code I had for printing potentially long long variables in PH/11
142 above was not the best (it lost precision). The length of off_t variables
143 is now inspected at build time, and an appropriate printing format (%ld
144 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
145 to be "long long int" or "long int". This is needed for the internal
146 formatting function string_vformat().
148 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
149 the configuration file to be ":syslog", then the script "guesses" where
150 the logs files are, rather than using the compiled in default. In our
151 case the guess is not the same as the compiled default, so the script
152 suddenly stopped working when I started to use syslog. The patch checks
153 to see if log_file_path is "". If so, it attempts to read it from exim
154 with no configuration file to get the compiled in version, before it
155 falls back to the previous guessing code."
157 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
158 implementing BATV in an Exim configuration. See NewStuff for the gory
161 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
162 Makefile that are specific to HP-UX.
164 PH/18 If the "use_postmaster" option was set for a recipient callout together
165 with the "random" option, the postmaster address was used as the MAIL
166 FROM address for the random test, but not for the subsequent recipient
167 test. It is now used for both.
169 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
170 patch removes a few documentation additions to RFC 3028, because the
171 latest draft now contains them. It adds the new en;ascii-case comparator
172 and a new error check for 8bit text in MIME parts. Comparator and
173 require names are now matched exactly. I enabled the subaddress
174 extension, but it is not well tested yet (read: it works for me)."
176 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
177 rework some of the code of TK/09 above to avoid the hardwired use of
178 "%lld" and "long long". Replaced the call to snprintf() with a call to
181 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
182 records point to non-existent hosts", "retry timeout exceeded", and
183 "retry time not reached for any host after a long failure period".
185 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
186 experimental DomainKeys support:
188 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
189 (2) On an error such as an illegally used "control", the wrong name for
190 the control was given.
192 These problems did NOT occur unless DomainKeys support was compiled.
194 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
196 PH/24 Added ${if match_ip condition.
198 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
199 databases so that it will be absolutely obvious if a crash occurs in the
200 DB library. This is a regular occurrence (often caused by mis-matched
203 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
204 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
205 picked up on a user's system that detects such things. There doesn't seem
206 to be a gcc warning option for this - only an attribute that has to be
207 put on the function's prototype. It seems that in Fedora Core 4 they have
208 set this on a number of new functions. No doubt there will be more in due
211 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
212 specify a transport. However, if an address that was verified by such a
213 router was the subject of a callout, Exim crashed because it tried to
214 read the rcpt_include_affixes from the non-existent transport. Now it
215 just assumes that the setting of that option is false. This bug was
216 introduced by 4.51/PH/31.
218 PH/28 Changed -d+all to exclude +memory, because that information is very
219 rarely of interest, but it makes the output a lot bigger. People tend to
220 do -d+all out of habit.
222 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
223 code in os-type was giving problems when libc.so lives in lib64, like on
226 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
227 aren't the modern standard, and it seems that some systems' include files
228 don't always have them. Exim was already checking for some of the newer
229 ones like T_AAAA, and defining it itself. I've added checks for all the
230 record types that Exim uses.
232 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
233 not automatically generating a new one, as it is supposed to. This
234 prevented TLS from working. If the file did exist, but contained invalid
235 data, a new version was generated, as expected. It was only the case of a
236 non-existent file that was broken.
238 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
239 with a change in libdomainkeys > 0.64.
241 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
242 from DNS. If the selector record carries the flag, it now has
243 precedence over the domain-wide flag.
245 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
247 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
248 the use of a port name that isn't defined in /etc/services) occurred, the
249 message was deferred as in a normal delivery, and thus remained on the
250 spool, instead of being failed because of the mua_wrapper setting. This
251 is now fixed, and I tidied up some of the mua_wrapper messages at the
254 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
255 the messages in a hash of arrays rather than using individual hashes.
256 This is a bit cleaner and results in dramatic memory savings, albeit
257 at a slight CPU cost.
259 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
260 as requested by Marc Sherman.
262 SC/10 Eximstats: added histograms for user specified patterns as requested
265 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
267 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
268 fopen() in the content-scanning modules that did not already have it.
274 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
275 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
277 TK/02 Fix ACL "control" statement not being available in MIME ACL.
279 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
281 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
282 to test Sieve filters that use "vacation".
284 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
285 that changes the way the GnuTLS parameters are stored in the cache file.
286 The new format can be generated externally. For backward compatibility,
287 if the data in the cache doesn't make sense, Exim assumes it has read an
288 old-format file, and it generates new data and writes a new file. This
289 means that you can't go back to an older release without removing the
292 PH/03 A redirect router that has both "unseen" and "one_time" set does not
293 work if there are any delivery delays because "one_time" forces the
294 parent to be marked "delivered", so its unseen clone is never tried
295 again. For this reason, Exim now forbids the simultaneous setting of
298 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
299 redirected to themselves ("homonym" addresses). Read the long ChangeLog
300 entry if you want to know the details. The fix, however, neglected to
301 consider the case when local delivery batching is involved. The test for
302 "previously delivered" was not happening when checking to see if an
303 address could be batched with a previous (undelivered) one; under
304 certain circumstances this could lead to multiple deliveries to the same
307 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
308 in its include files, and this causes problems building Exim.
310 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
311 header_syntax) but Exim was just ignoring anything given after a slash.
312 In particular, this caused confusion with an attempt to use "verify =
313 reverse_host_lookup/defer_ok". An error is now given when options are
314 supplied for verify items that do not have them. (Maybe reverse_host_
315 lookup should have a defer_ok option, but that's a different point.)
317 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
318 defined by RFC 821) to 2048, because there were problems with some AUTH
319 commands, and RFC 1869 says the size should be increased for extended
320 SMTP commands that take arguments.
322 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
325 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
326 "unknown" error; now it says that the functionality isn't in the binary.
328 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
329 an address' error message when a string expansion fails (syntax or
330 whatever). Otherwise the password may appear in the log. Following change
331 PH/42 below, there is no longer a chance of it appearing in a bounce
334 PH/11 Installed exipick version 20050225.0 from John Jetmore.
336 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
337 of its addresses was ever tried. (Bugzilla bug #2.)
339 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
340 the result incorrectly in the debug output. (It correctly added a newline
341 to what was transported.)
343 TF/01 Added $received_time.
345 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
346 commented out examples of how to interface to a virus scanner and to
347 SpamAssassin. Also added commented examples of av_scanner and
348 spamd_address settings.
350 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
351 and controls are allowed in which ACLs. There were a couple of minor
352 errors. Some of the entries in the conditions table (which is a table of
353 where they are NOT allowed) were getting very unwieldy; rewrote them as a
354 negation of where the condition IS allowed.
356 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
358 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
359 header file does not have a version number, so I've had to invent a new
360 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
361 API. The code is untested by me (my Linux distribution still has 0.3.2 of
362 radiusclient), but it was contributed by a Radius user.
364 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
365 files or directories, for OpenSSL.
367 PH/19 When an Exim process that is running as root has to create an Exim log
368 file, it does so in a subprocess that runs as exim:exim so as to get the
369 ownership right at creation (otherwise, other Exim processes might see
370 the file with the wrong ownership). There was no test for failure of this
371 fork() call, which would lead to the process getting stuck as it waited
372 for a non-existent subprocess. Forks do occasionally fail when resources
373 run out. I reviewed all the other calls to fork(); they all seem to check
376 PH/20 When checking for unexpected SMTP input at connect time (before writing
377 the banner), Exim was not dealing correctly with a non-positive return
378 from the read() function. If the client had disconnected by this time,
379 the result was a log entry for a synchronization error with an empty
380 string after "input=" when read() returned zero. If read() returned -1
381 (an event I could not check), uninitialized data bytes were printed.
382 There were reports of junk text (parts of files, etc) appearing after
385 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
387 PH/22 Added support for macro redefinition, and (re)definition in between
388 driver and ACL definitions.
390 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
391 forgetting to use the resulting value; it was using the unexpanded value.
393 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
394 hadn't been configured. The fix is from Juergen Kreileder, who
395 understands it better than I do:
397 "Here's what I see happening with three configured cyrus_sasl
398 authenticators configured (plain, login, cram-md5):
400 On startup auth_cyrus_sasl_init() gets called for each of these.
401 This means three calls to sasl_listmech() without a specified mech_list.
402 => SASL tests which mechs of all available mechs actually work
403 => three warnings about OTP not working
404 => the returned list contains: plain, login, cram-md5, digest-md5, ...
406 With the patch, sasl_listmech() also gets called three times. But now
407 SASL's mech_list option is set to the server_mech specified in the the
408 authenticator. Or in other words, the answer from sasl_listmech()
409 gets limited to just the mech you're testing for (which is different
411 => the return list contains just 'plain' or 'login', 'cram-md5' or
412 nothing depending on the value of ob->server_mech.
414 I've just tested the patch: Authentication still works fine,
415 unavailable mechs specified in the exim configuration are still
416 caught, and the auth.log warnings about OTP are gone."
418 PH/25 When debugging is enabled, the contents of the command line are added
419 to the debugging output, even when log_selector=+arguments is not
422 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
423 answer is "GNU", and only if the return is "GNU/something" is the answer
426 PH/27 $acl_verify_message is now set immediately after the failure of a
427 verification in an ACL, and so is available in subsequent modifiers. In
428 particular, the message can be preserved by coding like this:
430 warn !verify = sender
431 set acl_m0 = $acl_verify_message
433 Previously, $acl_verify_message was set only while expanding "message"
434 and "log_message" when a very denied access.
436 PH/28 Modified OS/os.c-Linux with
438 -#ifndef OS_LOAD_AVERAGE
439 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
441 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
442 nomenclature these days.)
444 PH/29 Installed patch from the Sieve maintainer that adds the options
445 sieve_useraddress and sieve_subaddress to the redirect router.
447 PH/30 In these circumstances:
448 . Two addresses routed to the same list of hosts;
449 . First host does not offer TLS;
450 . First host accepts first address;
451 . First host gives temporary error to second address;
452 . Second host offers TLS and a TLS session is established;
453 . Second host accepts second address.
454 Exim incorrectly logged both deliveries with the TLS parameters (cipher
455 and peerdn, if requested) that were in fact used only for the second
458 PH/31 When doing a callout as part of verifying an address, Exim was not paying
459 attention to any local part prefix or suffix that was matched by the
460 router that accepted the address. It now behaves in the same way as it
461 does for delivery: the affixes are removed from the local part unless
462 rcpt_include_affixes is set on the transport.
464 PH/32 Add the sender address, as F=<...>, to the log line when logging a
465 timeout during the DATA phase of an incoming message.
467 PH/33 Sieve envelope tests were broken for match types other than :is. I have
468 applied a patch sanctioned by the Sieve maintainer.
470 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
471 the uid or gid is negative. A case of a negative gid caused this to be
472 noticed. The fix allows for either to be negative.
474 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
475 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
478 PH/36 The change PH/12 above was broken. Fixed it.
480 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
481 the grounds that routing the same address twice would always produce the
482 same answer. This might have been true once, but it is certainly no
483 longer true now. Routing a child address may depend on the previous
484 routing that produced that child. Some complicated redirection strategies
485 went wrong when messages had multiple recipients, and made Exim's
486 behaviour dependent on the order in which the addresses were given.
488 I have moved the duplicate checking until after the routing is complete.
489 Exim scans the addresses that are assigned to local and remote
490 transports, and removes any duplicates. This means that more work will be
491 done, as duplicates will always all be routed, but duplicates are
492 presumably rare, so I don't expect this is of any significance.
494 For deliveries to pipes, files, and autoreplies, the duplicate checking
495 still happens during the routing process, since they are not going to be
498 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
499 It corrects a timeout issue with spamd. This is Ian's comment: "The
500 background is that sometimes spamd either never reads data from a
501 connection it has accepted, or it never writes response data. The exiscan
502 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
503 blindly assumes that writes won't block so it may never time out."
505 PH/39 Allow G after quota size as well as K and M.
507 PH/40 The value set for $authenticated_id in an authenticator may not contain
508 binary zeroes or newlines because the value is written to log lines and
509 to spool files. There was no check on this. Now the value is run through
510 the string_printing() function so that such characters are converted to
511 printable escape sequences.
513 PH/41 $message_linecount is a new variable that contains the total number of
514 lines in the message. Compare $body_linecount, which is the count for the
517 PH/42 Exim no longer gives details of delivery errors for specific addresses in
518 bounce and delay warning messages, except in certain special cases, which
521 (a) An SMTP error message from a remote host;
522 (b) A message specified in a :fail: redirection;
523 (c) A message specified in a "fail" command in a system filter;
524 (d) A message specified in a FAIL return from the queryprogram router;
525 (e) A message specified by the cannot_route_message router option.
527 In these cases only, Exim does include the error details in bounce and
528 warning messages. There are also a few cases where bland messages such
529 as "unrouteable address" or "local delivery error" are given.
531 PH/43 $value is now also set for the "else" part of a ${run expansion.
533 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
534 being worked on, but at least Exim now implements the latest version to
537 PH/45 In a pipe transport, although a timeout while waiting for the pipe
538 process to complete was treated as a delivery failure, a timeout while
539 writing the message to the pipe was logged, but erroneously treated as a
540 successful delivery. Such timeouts include transport filter timeouts. For
541 consistency with the overall process timeout, these timeouts are now
542 treated as errors, giving rise to delivery failures by default. However,
543 there is now a new Boolean option for the pipe transport called
544 timeout_defer, which, if set TRUE, converts the failures into defers for
545 both kinds of timeout. A transport filter timeout is now identified in
548 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
549 systems where "make" and "gmake" are different, calling "gmake" at top
550 level broke things. I've arranged for the value of $(MAKE) to be passed
551 from the Makefile to this script so that it can call the same version of
555 A note about Exim versions 4.44 and 4.50
556 ----------------------------------------
558 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
559 changes of various kinds. As a consequence, a big documentation update was
560 needed. This delayed the release for rather longer than seemed good, especially
561 in the light of a couple of (minor) security issues. Therefore, the changes
562 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
563 release. So 4.44 and 4.50 are in effect two different branches that both start
566 I have left the 4.50 change log unchanged; it contains all the changes since
567 4.43. The change log for 4.44 is below; many of its items are identical to
568 those for 4.50. This seems to be the most sensible way to preserve the
569 historical information.
575 1. Minor wording change to the doc/README.SIEVE file.
577 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
578 computation of the current number of files was incorrect.
580 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
581 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
584 4. Give more explanation in the error message when the command for a transport
585 filter fails to execute.
587 5. There are several places where Exim runs a non-Exim command in a
588 subprocess. The SIGUSR1 signal should be disabled for these processes. This
589 was being done only for the command run by the queryprogram router. It is
590 now done for all such subprocesses. The other cases are: ${run, transport
591 filters, and the commands run by the lmtp and pipe transports.
593 6. Added CONFIGURE_GROUP build-time option.
595 7. Some older OS have a limit of 256 on the maximum number of file
596 descriptors. Exim was using setrlimit() to set 1000 as a large value
597 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
598 systems. I've change it so that if it can't get 1000, it tries for 256.
600 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
601 was an oversight, and furthermore, ever since the addition of extra
602 controls (e.g. 4.43/32), the checks on when to allow different forms of
603 "control" were broken. There should now be diagnostics for all cases when a
604 control that does not make sense is encountered.
606 9. Added the /retain_sender option to "control=submission".
608 10. $recipients is now available in the predata ACL (oversight).
610 11. Tidy the search cache before the fork to do a delivery from a message
611 received from the command line. Otherwise the child will trigger a lookup
612 failure and thereby defer the delivery if it tries to use (for example) a
613 cached ldap connection that the parent has called unbind on.
615 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
616 of $address_data from the recipient verification was clobbered by the
619 13. The value of address_data from a sender verification is now available in
620 $sender_address_data in subsequent conditions in the ACL statement.
622 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
624 15. Added a new option "connect=<time>" to callout options, to set a different
627 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
628 was its contents. (It was OK if the option was not defined at all.)
630 17. A "Completed" log line is now written for messages that are removed from
631 the spool by the -Mrm option.
633 18. New variables $sender_verify_failure and $recipient_verify_failure contain
634 information about exactly what failed.
636 19. Added -dd to debug only the daemon process.
638 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
639 handles timeouts, both on the server side and network timeouts. Renamed the
640 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
642 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
643 It is now set to "smtps".
645 22. $host_address is now set to the target address during the checking of
648 23. When checking ignore_target_hosts for an ipliteral router, no host name was
649 being passed; this would have caused $sender_host_name to have been used if
650 matching the list had actually called for a host name (not very likely,
651 since this list is usually IP addresses). A host name is now passed as
654 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
655 code that specifies a non-restarting handler (typically sigaction() in
656 modern systems) in an attempt to fix a rare and obscure crash bug.
658 25. Narrowed the window for a race in the daemon that could cause it to ignore
659 SIGCHLD signals. This is not a major problem, because they are used only to
660 wake it up if nothing else does.
662 26. A malformed maildirsize file could cause Exim to calculate negative values
663 for the mailbox size or file count. Odd effects could occur as a result.
664 The maildirsize information is now recalculated if the size or filecount
667 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
668 support for a long time. Removed HAVE_SYS_VFS_H.
670 28. Installed the latest version of exipick from John Jetmore.
672 29. In an address list, if the pattern was not a regular expression, an empty
673 subject address (from a bounce message) matched only if the pattern was an
674 empty string. Non-empty patterns were not even tested. This was the wrong
675 because it is perfectly reasonable to use an empty address as part of a
676 database query. An empty address is now tested by patterns that are
677 lookups. However, all the other forms of pattern expect the subject to
678 contain a local part and a domain, and therefore, for them, an empty
679 address still always fails if the pattern is not itself empty.
681 30. Exim went into a mad DNS loop when attempting to do a callout where the
682 host was specified on an smtp transport, and looking it up yielded more
685 31. Re-factored the code for checking spool and log partition space into a
686 function that finds that data and another that does the check. The former
687 is then used to implement four new variables: $spool_space, $log_space,
688 $spool_inodes, and $log_inodes.
690 32. The RFC2047 encoding function was originally intended for short strings
691 such as real names; it was not keeping to the 75-character limit for
692 encoded words that the RFC imposes. It now respects the limit, and
693 generates multiple encoded words if necessary. To be on the safe side, I
694 have increased the buffer size for the ${rfc2047: expansion operator from
697 33. It is now permitted to omit both strings after an "if" condition; if the
698 condition is true, the result is "true". As before, when the second string
699 is omitted, a false condition yields an empty string. This makes it less
700 cumbersome to write custom ACL and router conditions.
702 34. Failure to deliver a bounce message always caused it to be frozen, even if
703 there was an errors_to setting on the router. The errors_to setting is now
706 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
707 canonical form (fully expanded) before being placed in
708 $sender_host_address.
710 36. The table in the code that translates DNS record types into text (T_A to
711 "A" for instance) was missing entries for NS and CNAME. It is just possible
712 that this could have caused confusion if both these types were looked up
713 for the same domain, because the text type is used as part of Exim's
714 per-process caching. But the chance of anyone hitting this buglet seems
717 37. The dnsdb lookup has been extended in a number of ways.
719 (1) There is a new type, "zns", which walks up the domain tree until it
720 finds some nameserver records. It should be used with care.
722 (2) There is a new type, "mxh", which is like "mx" except that it returns
723 just the host names, not the priorities.
725 (3) It is now possible to give a list of domains (or IP addresses) to be
726 looked up. The behaviour when one of the lookups defers can be
727 controlled by a keyword.
729 (4) It is now possible to specify the separator character for use when
730 multiple records are returned.
732 38. The dnslists ACL condition has been extended: it is now possible to supply
733 a list of IP addresses and/or domains to be looked up in a particular DNS
736 39. Added log_selector=+queue_time_overall.
738 40. When running the queue in the test harness, wait just a tad after forking a
739 delivery process, to get repeatability of debugging output.
741 41. Include certificate and key file names in error message when GnuTLS fails
742 to set them up, because the GnuTLS error message doesn't include the name
743 of the failing file when there is a problem reading it.
745 42. Allow both -bf and -bF in the same test run.
747 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
749 44. The "Exiscan patch" is now merged into the mainline Exim source.
751 45. Sometimes the final signoff response after QUIT could fail to get
752 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
753 < 0 before doing a fflush(). This bug looks as though it goes back to the
754 introduction of TLS in release 3.20, but "sometimes" must have been rare
755 because the tests only now provoked it.
757 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
758 (this can affect the format of dates).
760 47. exim_tidydb, when checking for the continued existence of a message for
761 which it has found a message-specific retry record, was not finding
762 messages that were in split spool directories. Consequently, it was
763 deleting retry records that should have stayed in existence.
765 48. Steve fixed some bugs in eximstats.
767 49. The SPA authentication driver was not abandoning authentication and moving
768 on to the next authenticator when an expansion was forced to fail,
769 contradicting the general specification for all authenticators. Instead it
770 was generating a temporary error. It now behaves as specified.
772 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
773 (the order specifies the preference for clients). The order is now AES256,
774 AES128, 3DES, ARCFOUR128.
776 51. Small patch to Sieve code - explicitly set From: when generating an
779 52. Exim crashed if a remote delivery caused a very long error message to be
780 recorded - for instance if somebody sent an entire SpamAssassin report back
781 as a large number of 550 error lines. This bug was coincidentally fixed by
782 increasing the size of one of Exim's internal buffers (big_buffer) that
783 happened as part of the Exiscan merge. However, to be on the safe side, I
784 have made the code more robust (and fixed the comments that describe what
787 53. Now that there can be additional text after "Completed" in log lines (if
788 the queue_time_overall log selector is set), a one-byte patch to exigrep
789 was needed to allow it to recognize "Completed" as not the last thing in
792 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
793 patch that reportedly fixes this has been added. I am not expert enough to
794 create a test for it. This is what the patch creator wrote:
796 "I found a little strange behaviour of ldap code when working with
797 Windows 2003 AD Domain, where users was placed in more than one
798 Organization Units. When I tried to give exim partial DN, the exit code
799 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
800 But simultaneously result of request was absolutely normal ldap result,
801 so I produce this patch..."
803 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
804 REFERENCE, so I have modified the code to exclude the patch when that macro
807 55. Some experimental protocols are using DNS PTR records for new purposes. The
808 keys for these records are domain names, not reversed IP addresses. The
809 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
810 leaves it alone. Component reversal etc. now happens only for IP addresses.
813 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
815 57. Double the size of the debug message buffer (to 2048) so that more of very
816 long debug lines gets shown.
818 58. The exicyclog utility now does better if the number of log files to keep
819 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
821 59. Two changes related to the smtp_active_hostname option:
823 (1) $smtp_active_hostname is now available as a variable.
824 (2) The default for smtp_banner uses $smtp_active_hostname instead
825 of $primary_hostname.
827 60. The host_aton() function is supposed to be passed a string that is known
828 to be a valid IP address. However, in the case of IPv6 addresses, it was
829 not checking this. This is a hostage to fortune. Exim now panics and dies
830 if the condition is not met. A case was found where this could be provoked
831 from a dnsdb PTR lookup with an IPv6 address that had more than 8
832 components; fortuitously, this particular loophole had already been fixed
833 by change 4.50/55 above.
835 If there are any other similar loopholes, the new check in host_aton()
836 itself should stop them being exploited. The report I received stated that
837 data on the command line could provoke the exploit when Exim was running as
838 exim, but did not say which command line option was involved. All I could
839 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
843 61. There was a buffer overflow vulnerability in the SPA authentication code
844 (which came originally from the Samba project). I have added a test to the
845 spa_base64_to_bits() function which I hope fixes it.
848 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
849 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
851 63. The daemon start-up calls getloadavg() while still root for those OS that
852 need the first call to be done as root, but it missed one case: when
853 deliver_queue_load_max is set with deliver_drop_privilege. This is
854 necessary for the benefit of the queue runner, because there is no re-exec
855 when deliver_drop_privilege is set.
857 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
860 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
861 was in use, was not putting the data itself into the right store pool;
862 consequently, it could be overwritten for a subsequent message in the same
863 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
866 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
868 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
869 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
870 the function were treating the return as a boolean value, which happened to
871 work because 0=false and not-0=true, but is not correct code.
873 68. The host_aton() function was not handling scoped IPv6 addresses (those
874 with, for example, "%eth0" on the end) correctly.
876 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
877 negated items (that is, ~something) in unsigned ints. Some compilers
878 apparently mutter when there is no cast.
880 70. If an address verification called from an ACL failed, and did not produce a
881 user-specific message (i.e. there was only a "system" message), nothing was
882 put in $acl_verify_message. In this situation, it now puts the system
885 71. Change 4.23/11 added synchronization checking at the start of an SMTP
886 session; change 4.31/43 added the unwanted input to the log line - except
887 that it did not do this in the start of session case. It now does.
889 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
890 This is wrong and can cause the other end to generate a synchronization
891 error if it is another Exim or anything else that does the synchronization
892 check. A QUIT command is no longer sent after a timeout.
894 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
897 74. The defer_ok option of callout verification was not working if it was used
898 when verifying addresses in header lines, that is, for this case:
900 verify = header_sender/callout=defer_ok
902 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
903 those file descriptors could be used for SMTP connections. If anything
904 wrote to stderr (the example that came up was "warn" in embedded Perl), it
905 could be sent to the SMTP client, causing chaos. The daemon now opens
906 stdin, stdout, and stderr to /dev/null when it puts itself into the
909 76. Arrange for output from Perl's "warn" command to be written to Exim's main
910 log by default. The user can override this with suitable Perl magic.
912 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
913 the log message when discard triggers, was not working for the DATA ACL or
914 for the non-SMTP ACL.
916 78. Error message wording change in sieve.c.
918 79. If smtp_accept_max_per_host was set, the number of connections could be
919 restricted to fewer than expected, because the daemon was trying to set up
920 a new connection before checking whether the processes handling previous
921 connections had finished. The check for completed processes is now done
922 earlier. On busy systems, this bug wouldn't be noticed because something
923 else would have woken the daemon, and it would have reaped the completed
926 80. If a message was submitted locally by a user whose login name contained one
927 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
928 It caused a spool format error. I have fixed the spool reading code. A
929 related problem was that the "from" clause in the Received: line became
930 illegal because of the space(s). It is now covered by ${quote_local_part.
932 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
935 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
937 83. Patch from Sieve maintainer to fix unterminated string problem in
940 84. Some minor changes to the Linux configuration files to help with other
941 OS variants using glibc.
943 85. One more patch for Sieve to update vacation handling to latest spec.
946 ----------------------------------------------------
947 See the note above about the 4.44 and 4.50 releases.
948 ----------------------------------------------------
954 1. Change 4.43/35 introduced a bug that caused file counts to be
955 incorrectly computed when quota_filecount was set in an appendfile
958 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
959 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
962 3. Give more explanation in the error message when the command for a transport
963 filter fails to execute.
965 4. There are several places where Exim runs a non-Exim command in a
966 subprocess. The SIGUSR1 signal should be disabled for these processes. This
967 was being done only for the command run by the queryprogram router. It is
968 now done for all such subprocesses. The other cases are: ${run, transport
969 filters, and the commands run by the lmtp and pipe transports.
971 5. Some older OS have a limit of 256 on the maximum number of file
972 descriptors. Exim was using setrlimit() to set 1000 as a large value
973 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
974 systems. I've change it so that if it can't get 1000, it tries for 256.
976 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
977 was an oversight, and furthermore, ever since the addition of extra
978 controls (e.g. 4.43/32), the checks on when to allow different forms of
979 "control" were broken. There should now be diagnostics for all cases when a
980 control that does not make sense is encountered.
982 7. $recipients is now available in the predata ACL (oversight).
984 8. Tidy the search cache before the fork to do a delivery from a message
985 received from the command line. Otherwise the child will trigger a lookup
986 failure and thereby defer the delivery if it tries to use (for example) a
987 cached ldap connection that the parent has called unbind on.
989 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
990 of $address_data from the recipient verification was clobbered by the
993 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
994 was its contents. (It was OK if the option was not defined at all.)
996 11. A "Completed" log line is now written for messages that are removed from
997 the spool by the -Mrm option.
999 12. $host_address is now set to the target address during the checking of
1000 ignore_target_hosts.
1002 13. When checking ignore_target_hosts for an ipliteral router, no host name was
1003 being passed; this would have caused $sender_host_name to have been used if
1004 matching the list had actually called for a host name (not very likely,
1005 since this list is usually IP addresses). A host name is now passed as
1008 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
1009 code that specifies a non-restarting handler (typically sigaction() in
1010 modern systems) in an attempt to fix a rare and obscure crash bug.
1012 15. Narrowed the window for a race in the daemon that could cause it to ignore
1013 SIGCHLD signals. This is not a major problem, because they are used only to
1014 wake it up if nothing else does.
1016 16. A malformed maildirsize file could cause Exim to calculate negative values
1017 for the mailbox size or file count. Odd effects could occur as a result.
1018 The maildirsize information is now recalculated if the size or filecount
1021 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1022 support for a long time. Removed HAVE_SYS_VFS_H.
1024 18. Updated exipick to current release from John Jetmore.
1026 19. Allow an empty sender to be matched against a lookup in an address list.
1027 Previously the only cases considered were a regular expression, or an
1030 20. Exim went into a mad DNS lookup loop when doing a callout where the
1031 host was specified on the transport, if the DNS lookup yielded more than
1034 21. The RFC2047 encoding function was originally intended for short strings
1035 such as real names; it was not keeping to the 75-character limit for
1036 encoded words that the RFC imposes. It now respects the limit, and
1037 generates multiple encoded words if necessary. To be on the safe side, I
1038 have increased the buffer size for the ${rfc2047: expansion operator from
1041 22. Failure to deliver a bounce message always caused it to be frozen, even if
1042 there was an errors_to setting on the router. The errors_to setting is now
1045 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1046 canonical form (fully expanded) before being placed in
1047 $sender_host_address.
1049 24. Updated eximstats to version 1.33
1051 25. Include certificate and key file names in error message when GnuTLS fails
1052 to set them up, because the GnuTLS error message doesn't include the name
1053 of the failing file when there is a problem reading it.
1055 26. Expand error message when OpenSSL has problems setting up cert/key files.
1058 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1059 (this can affect the format of dates).
1061 28. exim_tidydb, when checking for the continued existence of a message for
1062 which it has found a message-specific retry record, was not finding
1063 messages that were in split spool directories. Consequently, it was
1064 deleting retry records that should have stayed in existence.
1066 29. eximstats updated to version 1.35
1067 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1068 1.35 - bugfix such that pie charts by volume are generated correctly
1070 30. The SPA authentication driver was not abandoning authentication and moving
1071 on to the next authenticator when an expansion was forced to fail,
1072 contradicting the general specification for all authenticators. Instead it
1073 was generating a temporary error. It now behaves as specified.
1075 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1076 (the order specifies the preference for clients). The order is now AES256,
1077 AES128, 3DES, ARCFOUR128.
1079 31. Small patch to Sieve code - explicitly set From: when generating an
1082 32. Exim crashed if a remote delivery caused a very long error message to be
1083 recorded - for instance if somebody sent an entire SpamAssassin report back
1084 as a large number of 550 error lines. This bug was coincidentally fixed by
1085 increasing the size of one of Exim's internal buffers (big_buffer) that
1086 happened as part of the Exiscan merge. However, to be on the safe side, I
1087 have made the code more robust (and fixed the comments that describe what
1090 33. Some experimental protocols are using DNS PTR records for new purposes. The
1091 keys for these records are domain names, not reversed IP addresses. The
1092 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1093 leaves it alone. Component reversal etc. now happens only for IP addresses.
1096 34. The host_aton() function is supposed to be passed a string that is known
1097 to be a valid IP address. However, in the case of IPv6 addresses, it was
1098 not checking this. This is a hostage to fortune. Exim now panics and dies
1099 if the condition is not met. A case was found where this could be provoked
1100 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1101 components; fortuitously, this particular loophole had already been fixed
1102 by change 4.50/55 or 4.44/33 above.
1104 If there are any other similar loopholes, the new check in host_aton()
1105 itself should stop them being exploited. The report I received stated that
1106 data on the command line could provoke the exploit when Exim was running as
1107 exim, but did not say which command line option was involved. All I could
1108 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1109 running as the user.
1112 35. There was a buffer overflow vulnerability in the SPA authentication code
1113 (which came originally from the Samba project). I have added a test to the
1114 spa_base64_to_bits() function which I hope fixes it.
1117 36. The daemon start-up calls getloadavg() while still root for those OS that
1118 need the first call to be done as root, but it missed one case: when
1119 deliver_queue_load_max is set with deliver_drop_privilege. This is
1120 necessary for the benefit of the queue runner, because there is no re-exec
1121 when deliver_drop_privilege is set.
1123 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1124 was in use, was not putting the data itself into the right store pool;
1125 consequently, it could be overwritten for a subsequent message in the same
1126 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1129 38. Sometimes the final signoff response after QUIT could fail to get
1130 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1131 < 0 before doing a fflush(). This bug looks as though it goes back to the
1132 introduction of TLS in release 3.20, but "sometimes" must have been rare
1133 because the tests only now provoked it.
1139 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1140 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1141 Now it returns an integer. A number of calls were still expecting a T/F
1142 return. Fortuitously, in all cases, the tests worked in OK situations,
1143 which is the norm. However, things would have gone wrong on any write
1144 failures on the smtp file descriptor. This function is used when sending
1145 messages over SMTP and also when doing verify callouts.
1147 2. When Exim is called to do synchronous delivery of a locally submitted
1148 message (the -odf or -odi options), it no longer closes stderr before doing
1151 3. Implemented the mua_wrapper option.
1153 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1155 5. Implemented the functions header_remove(), header_testname(),
1156 header_add_at_position(), and receive_remove_recipient(), and exported them
1159 6. If an ACL "warn" statement specified the addition of headers, Exim already
1160 inserted X-ACL-Warn: at the start if there was no header name. However, it
1161 was not making this test for the second and subsequent header lines if
1162 there were newlines in the string. This meant that an invalid header could
1163 be inserted if Exim was badly configured.
1165 7. Allow an ACL "warn" statement to add header lines at the start or after all
1166 the Received: headers, as well as at the end.
1168 8. Added the rcpt_4xx retry error code.
1170 9. Added postmaster_mailfrom=xxx to callout verification option.
1172 10. Added mailfrom=xxxx to the callout verification option, for verify=
1175 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1176 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1178 12. Inserted some casts to stop certain compilers warning when using pointer
1179 differences as field lengths or precisions in printf-type calls (mostly
1180 affecting debugging statements).
1182 13. Added optional readline() support for -be (dynamically loaded).
1184 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1185 same clock tick as a message's arrival, so that its received time was the
1186 same as the "first fail" time on the retry record, and that message
1187 remained on the queue past the ultimate address timeout, every queue runner
1188 would try a delivery (because it was past the ultimate address timeout) but
1189 after another failure, the ultimate address timeout, which should have then
1190 bounced the address, did not kick in. This was a "< instead of <=" error;
1191 in most cases the first failure would have been in the next clock tick
1192 after the received time, and all would be well.
1194 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1195 being recognized when the domain list was tested by the match_domain
1196 condition in an expansion string.
1198 16. Added the ${str2b64: operator.
1200 17. Exim was always calling setrlimit() to set a large limit for the number of
1201 processes, without checking whether the existing limit was already
1202 adequate. (It did check for the limit on file descriptors.) Furthermore,
1203 errors from getrlimit() and setrlimit() were being ignored. Now they are
1204 logged to the main and panic logs, but Exim does carry on, to try to do its
1205 job under whatever limits there are.
1207 18. Imported PCRE 5.0.
1209 19. Trivial typo in log message " temporarily refused connection" (the leading
1212 20. If the log selector return_path_on_delivery was set and an address was
1213 redirected to /dev/null, the delivery process crashed because it assumed
1214 that a return path would always be set for a "successful" delivery. In this
1215 case, the whole delivery is bypassed as an optimization, and therefore no
1218 21. Internal re-arrangement: the function for sending a challenge and reading
1219 a response while authentication was assuming a zero-terminated challenge
1220 string. It's now changed to take a pointer and a length, to allow for
1221 binary data in such strings.
1223 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1225 23. Exim was not respecting finduser_retries when seeking the login of the
1226 uid under which it was called; it was always trying 10 times. (The default
1227 setting of finduser_retries is zero.) Also, it was sleeping after the final
1228 failure, which is pointless.
1230 24. Implemented tls_on_connect_ports.
1232 25. Implemented acl_smtp_predata.
1234 26. If the domain in control=submission is set empty, Exim assumes that the
1235 authenticated id is a complete email address when it generates From: or
1236 Sender: header lines.
1238 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1239 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1240 chown and chgrp in /bin and hostname in /usr/bin.
1242 28. Exim was keeping the "process log" file open after each use, just as it
1243 does for the main log. This opens the possibility of it remaining open for
1244 long periods when the USR1 signal hits a daemon. Occasional processlog
1245 errors were reported, that could have been caused by this. Anyway, it seems
1246 much more sensible not to leave this file open at all, so that is what now
1249 29. The long-running daemon process does not normally write to the log once it
1250 has entered its main loop, and it closes the log before doing so. This is
1251 so that log files can straightforwardly be renamed and moved. However,
1252 there are a couple of unusual error situations where the daemon does write
1253 log entries, and I had neglected to close the log afterwards.
1255 30. The text of an SMTP error response that was received during a remote
1256 delivery was being truncated at 512 bytes. This is too short for some of
1257 the long messages that one sometimes sees. I've increased the limit to
1260 31. It is now possible to make retry rules that apply only when a message has a
1261 specific sender, in particular, an empty sender.
1263 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1264 it possible to be selective about when SMTP synchronization is enforced.
1266 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1268 32. Implemented hosts_connection_nolog.
1270 33. Added an ACL for QUIT.
1272 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1275 35. Added mailbox_size and mailbox_filecount to appendfile.
1277 36. Added control = no_multiline_responses to ACLs.
1279 37. There was a bug in the logic of the code that waits for the clock to tick
1280 in the case where the clock went backwards by a substantial amount such
1281 that the microsecond fraction of "now" was more than the microsecond
1282 fraction of "then" (but the whole seconds number was less).
1284 38. Added support for the libradius Radius client library this is found on
1285 FreeBSD (previously only the radiusclient library was supported).
1291 1. When certain lookups returned multiple values in the form name=value, the
1292 quoting of the values was not always being done properly. Specifically:
1293 (a) If the value started with a double quote, but contained no whitespace,
1295 (b) If the value contained whitespace other than a space character (i.e.
1296 tabs or newlines or carriage returns) it was not quoted.
1297 This fix has been applied to the mysql and pgsql lookups by writing a
1298 separate quoting function and calling it from the lookup code. The fix
1299 should probably also be applied to nisplus, ibase and oracle lookups, but
1300 since I cannot test any of those, I have not disturbed their existing code.
1302 2. A hit in the callout cache for a specific address caused a log line with no
1303 reason for rejecting RCPT. Now it says "Previous (cached) callout
1304 verification failure".
1306 3. There was an off-by-one bug in the queryprogram router. An over-long
1307 return line was truncated at 256 instead of 255 characters, thereby
1308 overflowing its buffer with the terminating zero. As well as fixing this, I
1309 have increased the buffer size to 1024 (and made a note to document this).
1311 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1312 when Exim is waiting for an SMTP response from a remote server, Exim
1313 restarts its select() call on the socket, thereby resetting its timeout.
1314 This is not a problem when such interrupts are rare. Somebody set up a cron
1315 job to run exiwhat every 2 minutes, which is less than the normal select()
1316 timeout (5 or 10 minutes). This meant that the select() timeout never
1317 kicked in because it was always reset. I have fixed this by comparing the
1318 time when an interrupt arrives with the time at the start of the first call
1319 to select(). If more time than the timeout has elapsed, the interrupt is
1320 treated as a timeout.
1322 5. Some internal re-factoring in preparation for the addition of Sieve
1323 extensions (by MH). In particular, the "personal" test is moved to a
1324 separate function, and given an option for scanning Cc: and Bcc: (which is
1325 not set for Exim filters).
1327 6. When Exim created an email address using the login of the caller as the
1328 local part (e.g. when creating a From: or Sender: header line), it was not
1329 quoting the local part when it contained special characters such as @.
1331 7. Installed new OpenBSD configuration files.
1333 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1334 try to make them clearer.
1336 9. Callout options, other than the timeout value, were being ignored when
1337 verifying sender addresses in header lines. For example, when using
1339 verify = header_sender/callout=no_cache
1341 the cache was (incorrectly) being used.
1343 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1344 only the Cygwin environment.
1346 11. When return_path_on_delivery was set as a log selector, if different remote
1347 addresses in the same message used different return paths and parallel
1348 remote delivery occurred, the wrong values would sometimes be logged.
1349 (Whenever a remote delivery process finished, the return path value from
1350 the most recently started remote delivery process was logged.)
1352 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1353 lines when AUTH and/or TLS are in use. This is the "received protocol"
1354 field. Exim used to use "asmtp" for authenticated SMTP, without any
1355 indication (in the protocol name) for TLS use. Now it follows the RFC and
1356 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1357 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1358 names appear in log lines as well as in Received: header lines.
1360 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1361 extensions, and comparison tests, and to fix some bugs.
1363 14. Changes to the "personal" filter test:
1365 (1) The test was buggy in that it was just doing the equivalent of
1366 "contains" tests on header lines. For example, if a user's address was
1367 anne@some.where, the "personal" test would incorrectly be true for
1369 To: susanne@some.where
1371 This test is now done by extracting each address from the header in turn,
1372 and checking the entire address. Other tests that are part of "personal"
1373 are now done using regular expressions (for example, to check local parts
1374 of addresses in From: header lines).
1376 (2) The list of non-personal local parts in From: addresses has been
1377 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1378 taken from the Sieve specification recommendations.
1380 (3) If the message contains any header line starting with "List-" it is
1381 treated as non-personal.
1383 (4) The test for "circular" in the Subject: header line has been removed
1384 because it now seems ill-conceived.
1386 15. Minor typos in src/EDITME comments corrected.
1388 16. Installed latest exipick from John Jetmore.
1390 17. If headers_add on a router specified a text string that was too long for
1391 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1392 of string_sprintf() is now avoided.
1394 18. $message_body_size was not set (it was always zero) when running the DATA
1395 ACL and the local_scan() function.
1397 19. For the "mail" command in an Exim filter, no default was being set for
1398 the once_repeat time, causing a random time value to be used if "once" was
1399 specified. (If the value happened to be <= 0, no repeat happened.) The
1400 default is now 0s, meaning "never repeat". The "vacation" command was OK
1401 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1402 (I found it when inspecting the code).
1404 20. There is now an overall timeout for performing a callout verification. It
1405 defaults to 4 times the callout timeout, which applies to individual SMTP
1406 commands during the callout. The overall timeout applies when there is more
1407 than one host that can be tried. The timeout is checked before trying the
1408 next host. This prevents very long delays if there are a large number of
1409 hosts and all are timing out (e.g. when the network connections are timing
1410 out). The value of the overall timeout can be changed by specifying an
1411 additional sub-option for "callout", called "maxwait". For example:
1413 verify = sender/callout=5s,maxwait=20s
1415 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1416 to the end before writing, but this should make it even safer).
1418 22. Exim was forgetting that it had advertised PIPELINING for the second and
1419 subsequent messages on an SMTP connection. It was also not resetting its
1420 memory on STARTTLS and an internal HELO.
1422 23. When Exim logs an SMTP synchronization error within a session, it now
1423 records whether PIPELINING has been advertised or not.
1425 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1426 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1427 rather than long int.
1429 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1431 26. Added the never_mail option to autoreply.
1437 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1438 crash if the getsockname() call failed; this can happen if a connection is
1439 closed very soon after it is established. The problem was simply in the
1440 order in which certain operations were done, causing Exim to try to write
1441 to the SMTP stream before it had set up the file descriptor. The bug has
1442 been fixed by making things happen in the correct order.
1448 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1449 before the connection was closed, thus losing the rejection response.
1451 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1452 some early Solaris releases, but causes trouble in current releases where
1453 socklen_t is defined.
1455 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1458 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1461 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1462 that releasing the top of it at the end releases what was used for sub-
1463 expansions (unless the block got too big). However, discard this block if
1464 the first thing is a variable or header, so that we can use its block when
1465 it is dynamic (useful for very large $message_headers, for example).
1467 6. Lookups now cache *every* query, not just the most recent. A new, separate
1468 store pool is used for this. It can be recovered when all lookup caches are
1469 flushed. Lookups now release memory at the end of their result strings.
1470 This has involved some general refactoring of the lookup sources.
1472 7. Some code has been added to the store_xxx() functions to reduce the amount
1473 of flapping under certain conditions.
1475 8. log_incoming_interface used to affect only the <= reception log lines. Now
1476 it causes the local interface and port to be added to several more SMTP log
1477 lines, for example "SMTP connection from", and rejection lines.
1479 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1481 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1483 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1484 could be overwritten at the end of the current message (or the start of a
1485 new message if it was set in a HELO ACL). The value is now preserved for
1486 the duration of the SMTP connection.
1488 12. If a transport had a headers_rewrite setting, and a matching header line
1489 contained an unqualified address, that address was qualified, even if it
1490 did not match any rewriting rules. The underlying bug was that the values
1491 of the flags that permit the existence of unqualified sender and recipient
1492 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1493 non-local messages, and by -bnq for local messages) were not being
1494 preserved with the message after it was received.
1496 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1497 "next input=" as part of the text comprising the host identity instead of
1498 the correct text. The code was using the same buffer for two different
1499 strings. However, depending on which order the printing function evaluated
1500 its arguments, the bug did not always show up. Under Linux, for example, my
1501 test suite worked just fine.
1503 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1504 doesn't work with some older versions of Perl. It has been changed to "my",
1505 which in any case is probably the better facility to use.
1507 15. A really picky compiler found some instances of statements for creating
1508 error messages that either had too many or two few arguments for the format
1511 16. The size of the buffer for calls to the DNS resolver has been increased
1512 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1513 for addresses that have a lot of PTR records. This alleviates a problem; it
1514 does not fully solve it.
1516 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1517 buffer now truncates the list and logs the incident, which is the same
1518 action as happens when Exim is looking up a host name and its aliases.
1519 Previously in this situation something unpredictable would happen;
1520 sometimes it was "internal error: store_reset failed".
1522 18. If a server dropped the connection unexpectedly when an Exim client was
1523 using GnuTLS and trying to read a response, the client delivery process
1524 crashed while trying to generate an error log message.
1526 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1527 string, for example:
1529 warn message = H1: something\nH2: something
1531 the text was added as a single header line from Exim's point of view
1532 though it ended up OK in the delivered message. However, searching for the
1533 second and subsequent header lines using $h_h2: did not work. This has been
1534 fixed. Similarly, if a system filter added multiple headers in this way,
1535 the routers could not see them.
1537 20. Expanded the error message when iplsearch is called with an invalid key to
1538 suggest using net-iplsearch in a host list.
1540 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1541 ACLs are no longer actually imposed (and a message to that effect is
1544 22. If a "gecos" field in a passwd entry contained escaped characters, in
1545 particular, if it contained a \" sequence, Exim got it wrong when building
1546 a From: or a Sender: header from that name. A second bug also caused
1547 incorrect handling when an unquoted " was present following a character
1548 that needed quoting.
1550 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1551 was not being matched caselessly.
1553 24. Arranged for all hyphens in the exim.8 source to be escaped with
1556 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1557 properly. Recipient callout cache records were still being keyed to include
1558 the sender, even when use_sender was set false. This led to far more
1559 callouts that were necessary. The sender is no longer included in the key
1560 when use_sender is false.
1562 26. Added "control = submission" modifier to ACLs.
1564 27. Added the ${base62d: operator to decode base 62 numbers.
1566 28. dnsdb lookups can now access SRV records.
1568 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1569 the configuration file.
1571 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1572 (-v) mode. This makes the output for a verbose queue run more intelligible.
1574 31. Added a use_postmaster feature to recipient callouts.
1576 32. Added the $body_zerocount variable, containing the number of binary zero
1577 bytes in the message body.
1579 33. The time of last modification of the "new" subdirectory is now used as the
1580 "mailbox time last read" when there is a quota error for a maildir
1583 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1585 35. Added +ignore_unknown as a special item in host lists.
1587 36. Code for decoding IPv6 addresses in host lists is now included, even if
1588 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1589 address was recognized as an IP address, but was then not correctly decoded
1590 into binary, causing unexpected and incorrect effects when compared with
1597 1. Very minor rewording of debugging text in manualroute to say "list of
1598 hosts" instead of "hostlist".
1600 2. If verify=header_syntax was set, and a header line with an unqualified
1601 address (no domain) and a large number of spaces between the end of the
1602 name and the colon was received, the reception process suffered a buffer
1603 overflow, and (when I tested it) crashed. This was caused by some obsolete
1604 code that should have been removed. The fix is to remove it!
1606 3. When running in the test harness, delay a bit after writing a bounce
1607 message to get a bit more predictability in the log output.
1609 4. Added a call to search_tidyup() just before forking a reception process. In
1610 theory, someone could use a lookup in the expansion of smtp_accept_max_
1611 per_host which, without the tidyup, could leave open a database connection.
1613 5. Added the variables $recipient_data and $sender_data which get set from a
1614 lookup success in an ACL "recipients" or "senders" condition, or a router
1615 "senders" option, similar to $domain_data and $local_part_data.
1617 6. Moved the writing of debug_print from before to after the "senders" test
1620 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1621 problems for message scanning, either using a data ACL, or using
1622 local_scan() because the Received: header was not generated till after they
1623 were called (in order to set the time as the time of reception completion).
1624 I have revised the way this works. The header is now generated after the
1625 body is received, but before the ACL or local_scan() are called. After they
1626 are run, the timestamp in the header is updated.
1632 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1633 before starting a queue runner without re-exec. This happened only when
1634 deliver_drop_privilege was set or when the Exim user was set to root. The
1635 effect of the bug was that timeouts during subsequent deliveries caused
1636 crashes instead of being properly handled. The handler is now left at its
1637 default (and expected) setting.
1639 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1640 message, again when deliver_drop_privilege is set or Exim is run as root.
1641 The bug described in (1) was not present in this case, but the tidying up
1642 of the other signals was missing. I have made the two cases consistent.
1644 3. The ignore_target_hosts setting on a manualroute router was being ignored
1645 for hosts that were looked up using the /MX notation.
1647 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1650 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1651 operated on the sender address. After changing the $sender_address to <>
1652 for the sender address verify, Exim was re-instated it as the original
1653 (before rewriting) address, but remembering that it had rewritten it, so it
1654 wasn't rewriting it again. This bug also had the effect of breaking the
1655 sender address verification caching when the sender address was rewritten.
1657 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1658 This has been changed so that if the ip literal address matches
1659 ignore_target_hosts, the router declines.
1661 7. Added expansion conditions match_domain, match_address, and match_local_
1662 part (NOT match_host).
1664 8. The placeholder for the Received: header didn't have a length field set.
1666 9. Added code to Exim itself and to exim_lock to test for a specific race
1667 condition that could lead to file corruption when using MBX delivery. The
1668 issue is with the lockfile that is created in /tmp. If this file is removed
1669 after a process has opened it but before that process has acquired a lock,
1670 there is the potential for a second process to recreate the file and also
1671 acquire a lock. This could lead to two Exim processes writing to the file
1672 at the same time. The added code performs the same test as UW imapd; it
1673 checks after acquiring the lock that its file descriptor still refers to
1674 the same named file.
1676 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1677 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1680 11. Added the smtp_active_hostname option. If used, this will typically be made
1681 to depend on the incoming interface address. Because $interface_address is
1682 not set up until the daemon has forked a reception process, error responses
1683 that can happen earlier (such as "too many connections") no longer contain
1686 12. If an expansion in a condition on a "warn" statement fails because a lookup
1687 defers, the "warn" statement is abandoned, and the next ACL statement is
1688 processed. Previously this caused the whole ACL to be aborted.
1690 13. Added the iplsearch lookup type.
1692 14. Added ident_timeout as a log selector.
1694 15. Added tls_certificate_verified as a log selector.
1696 16. Added a global option tls_require_ciphers (compare the smtp transport
1697 option of the same name). This controls incoming TLS connections.
1699 17. I finally figured out how to make tls_require_ciphers do a similar thing
1700 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1701 before starting the TLS session.
1703 18. Tabs are now shown as \t in -bP output.
1705 19. If the log selector return_path_on_delivery was set, Exim crashed when
1706 bouncing a message because it had too many Received: header lines.
1708 20. If two routers both had headers_remove settings, and the first one included
1709 a superfluous trailing colon, the final name in the first list and the
1710 first name in the second list were incorrectly joined into one item (with a
1711 colon in the middle).
1717 1. Added -C and -D options to the exinext utility, mainly to make it easier
1718 to include in the automated testing, but these could be helpful when
1719 multiple configurations are in use.
1721 2. The exinext utility was not formatting the output nicely when there was
1722 an alternate port involved in the retry record key, nor when there was a
1723 message id as well (for retries that were specific to a specific message
1724 and a specific host). It was also confused by IPv6 addresses, because of
1725 the additional colons they contain. I have fixed the IPv4 problem, and
1726 patched it up to do a reasonable job for IPv6.
1728 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1729 delivery, the log line now contains "pipelined" if PIPELINING was used.
1731 4. An SMTP transport process used to panic and die if the bind() call to set
1732 an explicit outgoing interface failed. This has been changed; it is now
1733 treated in the same way as a connect() failure.
1735 5. A reference to $sender_host_name in the part of a conditional expansion
1736 that was being skipped was still causing a DNS lookup. This no longer
1739 6. The def: expansion condition was not recognizing references to header lines
1740 that used bh_ and bheader_.
1742 7. Added the _cache feature to named lists.
1744 8. The code for checking quota_filecount in the appendfile transport was
1745 allowing one more file than it should have been.
1747 9. For compatibility with Sendmail, the command line option
1755 and sets the incoming protocol and host name (for trusted callers). The
1756 host name and its colon can be omitted when only the protocol is to be set.
1757 Note the Exim already has two private options, -pd and -ps, that refer to
1758 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1759 "s", but I don't think that's a major issue.
1761 10. A number of refactoring changes to the code, none of which should affect
1764 (a) The number of logging options was getting close to filling up the
1765 32-bit word that was used as a bit map. I have split them into two classes:
1766 those that are passed in the argument to log_write(), and those that are
1767 only ever tested independently outside of that function. These are now in
1768 separate 32-bit words, so there is plenty of room for expansion again.
1769 There is no change in the user interface or the logging behaviour.
1771 (b) When building, for example, log lines, the code previously used a
1772 macro that called string_cat() twice, in order to add two strings. This is
1773 not really sufficiently general. Furthermore, there was one instance where
1774 it was actually wrong because one of the argument was used twice, and in
1775 one call a function was used. (As it happened, calling the function twice
1776 did not affect the overall behaviour.) The macro has been replaced by a
1777 function that can join an arbitrary number of extra strings onto a growing
1780 (c) The code for expansion conditions now uses a table and a binary chop
1781 instead of a serial search (which was left over from when there were very
1782 few conditions). Also, it now recognizes conditions like "pam" even when
1783 the relevant support is not compiled in: a suitably worded error message is
1784 given if an attempt is made to use such a condition.
1786 11. Added ${time_interval:xxxxx}.
1788 12. A bug was causing one of the ddress fields not to be passed back correctly
1789 from remote delivery subprocesses. The field in question was not being
1790 subsequently used, so this caused to problems in practice.
1792 13. Added new log selectors queue_time and deliver_time.
1794 14. Might have fixed a bug in maildirsizefile handling that threw up
1795 "unexpected character" debug warnings, and recalculated the data
1796 unnecessarily. In any case, I expanded the warning message to give more
1799 15. Added the message "Restricted characters in address" to the statements in
1800 the default ACL that block characters like @ and % in local parts.
1802 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1803 Three changes have been made:
1805 (a) There was a serious bug; a negative response to MAIL caused the whole
1806 recipient domain to be cached as invalid, thereby blocking all messages
1807 to all local parts at the same domain, from all senders. This bug has
1808 been fixed. The domain is no longer cached after a negative response to
1809 MAIL if the sender used is not empty.
1811 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1814 (c) A new callout option, "use_sender" has been added for people who want
1815 the modified behaviour.
1821 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1824 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1825 indeed breaks things for older releases.
1827 3. Added additional logging to the case where there is a problem reading data
1828 from a filter that is running in a subprocess using a pipe, in order to
1829 try to track down a specific problem.
1831 4. Testing facility fudge: when running in the test harness and attempting
1832 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1833 getting "No route to host". Convert this to a timeout.
1835 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1838 6. Some OS don't have socklen_t but use size_t instead. This affects the
1839 fifth argument of getsockopt() amongst other things. This is now
1840 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1841 can be set for individual OS. I have set it for SunOS5, OSF1, and
1842 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1843 some earlier ones do not.
1845 7. Change 4.30/15 was not doing the test caselessly.
1847 8. The standard form for an IPv6 address literal was being rejected by address
1848 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1849 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1850 this, as well as the form without the "IPv6" on the front (but only when
1851 address literals are enabled, of course).
1853 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1855 10. Exim crashed if a message with an empty sender address specified by -f
1856 encountered a router with an errors_to setting. This could be provoked only
1857 by a command such as
1861 where an empty string was supplied; "<>" did not hit this bug.
1863 11. Installed PCRE release 4.5.
1865 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1866 remained set. It is now erased.
1868 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1869 times from message ids (which are base 36 rather than the normal 62).
1871 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1872 were being counted as actual protocol errors, and logged if the log
1873 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1874 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1875 and DATA following a set of rejected RCPTs do not count as protocol errors.
1876 In other words, Exim assumes they were pipelined, though this may not
1877 actually be the case. Of course, in all cases the client gets an
1878 appropriate error code.
1880 15. If a lookup fails in an ACL condition, a message about the failure may
1881 be available; it is used if testing the ACL cannot continue, because most
1882 such messages specify what the cause of the deferral is. However, some
1883 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1884 that caused an old message to be retained and used if a later statement
1885 caused a defer, replacing the real cause of the deferral.
1887 16. If an IP address had so many PTR records that the DNS lookup buffer
1888 was not large enough to hold them, Exim could crash while trying to process
1889 the truncated data. It now detects and logs this case.
1891 17. Further to 4.21/58, another change has been made: if (and only if) the
1892 first line of a message (the first header line) ends with CRLF, a bare LF
1893 in a subsequent header line has a space inserted after it, so as not to
1894 terminate the header.
1896 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1897 unnecessarily, used atoi() instead of strtol(), and didn't check the
1898 termination when getting file sizes from file names by regex.
1900 19. Completely re-implemented the support for maildirsize files, in the light
1901 of a number of problems with the previous contributed implementation
1902 (4.30/29). In particular:
1904 . If the quota is zero, the maildirsize file is maintained, but no quota is
1907 . If the maildir directory does not exist, it is created before any attempt
1908 to write a maildirsize file.
1910 . The quota value in the file is just a cache; if the quota is changed in
1911 the transport, the new value overrides.
1913 . A regular expression is available for excluding directories from the
1916 20. The autoreply transport checks the characters in options that define the
1917 message's headers; it allows continued headers, but it was checking with
1918 isspace() after an embedded newline instead of explicitly looking for a
1921 21. If all the "regular" hosts to which an address was routed had passed their
1922 expiry times, and had not reached their retry times, the address was
1923 bounced, even if fallback hosts were defined. Now Exim should go on to try
1926 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1927 equivalent code in the SMTP transport. Some hosts send humungous responses
1928 to HELO/EHLO, more than 1024 it seems.
1930 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1931 gives compiler warnings in some environments. I've now done it "properly",
1934 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1935 (because of problems with the built-in one) was declared to return uschar *
1936 instead of char *, causing compiler failure.
1938 25. Fixed a file descriptor leak when processing alias/forward files.
1940 26. Fixed a minor format string issue in dbfn.c.
1942 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1944 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1945 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1948 29. When a sender address is verified, it is cached, to save repeating the test
1949 when there is more than one recipient in a message. However, when the
1950 verification involves a callout, it is possible for different callout
1951 options to be set for different recipients. It is too complicated to keep
1952 track of this in the cache, so now Exim always runs a verification when a
1953 callout is required, relying on the callout cache for the optimization.
1954 The overhead is duplication of the address routing, but this should not be
1957 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1958 to be verified with callout=postmaster, and the main callout worked but the
1959 postmaster check failed, the verification correctly failed. However, if a
1960 subsequent RCPT command asked for sender verification *without* the
1961 postmaster check, incorrect caching caused this verification also to fail,
1964 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1965 it was not caching the DNS options (qualify_single, search_parents) that
1966 were used when the lookup failed. A subsequent lookup with different
1967 options therefore always gave the same answer, though there were cases
1968 where it should not have. (Example: a "domains = !$mx_any" option on a
1969 dnslookup router: the "domains" option is always processed without any
1970 widening, but the router might have qualify_single set.) Now Exim uses the
1971 cached value only when the same options are set.
1973 32. Added John Jetmore's "exipick" utility to the distribution.
1975 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1976 than a timeout (e.g. a certificate is required, and is not provided), an
1977 Exim server now closes the connection immediately. Previously it waited for
1978 the client to close - but if the client is SSL, it seems that they each
1979 wait for each other, leading to a delay before one of them times out.
1981 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1982 maintained 0.8.x compatibility because I don't think many are using it, and
1983 it is clearly obsolete.
1985 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1988 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1989 client certificate was expired. A simple patch fixes this, though I don't
1990 understand the full logic of why the verify callback is called multiple
1993 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1994 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1995 which causes problems with some clients (such as the Certicom SSL Plus
1996 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1997 disables the coutermeasure allowing Eudora to connect."
1999 38. Exim was not checking that a write() to a log file succeeded. This could
2000 lead to Bad Things if a log got too big, in particular if it hit a file
2001 size limit. Exim now panics and dies if it cannot write to a log file, just
2002 as it does if it cannot open a log file.
2004 39. Modified OS/Makefile-Linux so that it now contains
2006 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
2008 The two -D definitions ensure that Exim is compiled with large file
2009 support, which makes it possible to handle log files that are bigger than
2012 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2013 instance) a domain was checked against a named list that involved a lookup,
2014 causing $domain_data to be set, then another domain was checked against the
2015 same list, then the first domain was re-checked, the value of $domain_data
2016 after the final check could be wrong. In particular, if the second check
2017 failed, it could be set empty. This bug probably also applied to
2020 41. The strip_trailing_dot option was not being applied to the address given
2021 with the -f command-line option.
2023 42. The code for reading a message's header from the spool was incrementing
2024 $received_count, but never initializing it. This meant that the value was
2025 incorrect (doubled) while delivering a message in the same process in which
2026 it was received. In the most common configuration of Exim, this never
2027 happens - a fresh exec is done - but it can happen when
2028 deliver_drop_privilege is set.
2030 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2031 it now includes up to 150 characters of the unexpected data in the log
2034 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2035 and building data strings. The size of both of these buffers was 10 000
2036 bytes - far larger than anybody would *ever* want, thought I. Needless to
2037 say, somebody hit the limit. I have increased the maximum line length to
2038 20 000 and the maximum data length of concatenated lines to 100 000. I have
2039 also fixed two bugs, because there was no checking on these buffers. Tsk,
2040 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2041 buffer is too small.
2043 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2044 lsearch lookups. Now it does.
2046 46. When parsing a route_list item in a manualroute router, a fixed-length
2047 buffer was used for the list of hosts. I made this 1024 bytes long,
2048 thinking that nobody would ever have a list of hosts that long. Wrong.
2049 Somebody had a whole pile of complicated expansion conditions, and the
2050 string was silently truncated, leading to an expansion error. It turns out
2051 that it is easier to change to an unlimited length (owing to other changes
2052 that have happened since this code was originally written) than to build
2053 structure for giving a limitation error. The length of the item that
2054 expands into the list of hosts is now unlimited.
2056 47. The lsearch lookup could not handle data where the length of text line was
2057 more than 4095 characters. Such lines were truncated, leading to shortened
2058 data being returned. It should now handle lines of any length.
2060 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2061 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2064 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2065 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2066 no longer happens in comment lines. A list of replacements is now placed
2067 at the head of all of the source files, except those whose only change is
2068 to replace PERL_COMMAND in the very first #! line.
2070 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2071 messages on the queue, with a bottom-up merge sort, using code contributed
2072 by Michael Haardt. This should make operations like -bp somewhat faster on
2073 large queues. It won't affect queue runners, except when queue_run_in_order
2076 51. Installed eximstats 1.31 in the distribution.
2078 52. Added support for SRV lookups to the dnslookup router.
2080 53. If an ACL referred to $message_body or $message_body_end, the value was not
2081 reset for any messages that followed in the same SMTP session.
2083 54. The store-handling optimization for building very long strings was not
2084 differentiating between the different store pools. I don't think this
2085 actually made any difference in practice, but I've tidied it.
2087 55. While running the routers to verify a sender address, $sender_address
2088 was still set to the sender address. This is wrong, because when routing to
2089 send a bounce to the sender, it would be empty. Therefore, I have changed
2090 it so that, while verifying a sender address, $sender_address is set to <>.
2091 (There is no change to what happens when verifying a recipient address.)
2093 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2094 target A or AAAA records (if not already returned) without resetting the
2095 qualify_single or search_parents options of the DNS resolver. These are
2096 inappropriate in this case because the targets of MX and SRV records must
2097 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2098 target that, when qualified, matched something in the local domain. These
2099 two options are now turned off when doing these lookups.
2101 57. It seems that at least some releases of Reiserfs (which does not have the
2102 concept of a fixed number of inodes) returns zero and not -1 for the
2103 number of available inodes. This interacted badly with check_spool_inodes,
2104 which assumed that -1 was the "no such thing" setting. What I have done is
2105 to check that the total number of inodes is greater than zero before doing
2106 the test of how many are available.
2108 58. When a "warn" ACL statement has a log_message modifier, the message is
2109 remembered, and not repeated. This is to avoid a lot of repetition when a
2110 message has many recipients that cause the same warning to be written.
2111 Howewer, Exim was preserving the list of already written lines for an
2112 entire SMTP session, which doesn't seem right. The memory is now reset if a
2113 new message is started.
2115 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2116 debugging output unless log_rewrite was also set.
2118 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2119 of (int)(handle) when we know that handle contains (void *)(-1).
2121 61. The Exim daemon panic-logs an error return when it closes the incoming
2122 connection. However "connection reset by peer" seems to be common, and
2123 isn't really an error worthy of noting specially, so that particular error
2126 62. When Exim is trying to find all the local interfaces, it used to panic and
2127 die if the ioctl to get the interface flags failed. However, it seems that
2128 on at least one OS (Solaris 9) it is possible to have an interface that is
2129 included in the list of interfaces, but for which you get a failure error
2130 for this call. This happens when the interface is not "plumbed" into a
2131 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2132 failure of the "get flags" call assumes that the interface is down.
2134 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2135 makes life easier for people who are doing arithmetic on fields extracted
2136 from dates, where you often get leading zeros that should not be
2137 interpreted as octal.
2139 64. Added qualify_domain to the redirect router, to override the global
2142 65. If a pathologically long header line contained very many addresses (the
2143 report of this problem mentioned 10 000) and each of them was rewritten,
2144 Exim could use up a very large amount of memory. (It kept on making new
2145 copies of the header line as it rewrote, and never released the old ones.)
2146 At the expense of a bit more processing, the header rewriting function has
2147 been changed so that it no longer eats memory in this way.
2149 66. The generation of the Received: header has been moved from the time that a
2150 message starts to be received, to the time that it finishes. The timestamp
2151 in the Received: header should now be very close to that of the <= log
2152 line. There are two side-effects of this change:
2154 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2155 logged header lines no longer include the local Received: line, because
2156 it has not yet been created. The same applies to a copy of the message
2157 that is returned to a non-SMTP sender when a message is rejected.
2159 (b) When a filter file is tested using -bf, no additional Received: header
2160 is added to the test message. After some thought, I decided that this
2163 This change does not affect the value of $received_for. It is still set
2164 after address rewriting, but before local_scan() is called.
2166 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2168 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2169 gave an unhelpful panic error message, and a defer error. I have managed to
2170 change this behaviour so that it now rejects any supplied certificate,
2171 which seems right, as the list of acceptable certificates is empty.
2173 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2174 gave an unhelpful defer error. I have not managed to make this reject any
2175 supplied certificates, but the error message it gives is "no certificate
2176 supplied", which is not helpful.
2178 70. exigrep's output now also includes lines that are not associated with any
2179 message, but which match the given pattern. Implemented by a patch from
2180 Martin Sluka, which also tidied up the Perl a bit.
2182 71. Recipient callout verification, like sender verification, was using <> in
2183 the MAIL FROM command. This isn't really the right thing, since the actual
2184 sender may affect whether the remote host accepts the recipient or not. I
2185 have changed it to use the actual sender in the callout; this means that
2186 the cache record is now keyed on a recipient/sender pair, not just the
2187 recipient address. There doesn't seem to be a real danger of callout loops,
2188 since a callout by the remote host to check the sender would use <>.
2189 [SEE ABOVE: changed after hitting problems.]
2191 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2192 temporary errors. However, in the case of such a code being given after
2193 the end of a data transmission (i.e. after ".") Exim was failing to write
2194 a retry record for the message. (Yes, there was some broken host that was
2195 actually sending 8xx at this point.)
2197 73. An unknown lookup type in a host list could cause Exim to panic-die when
2198 the list was checked. (An example that provoked this was putting <; in the
2199 middle of a list instead of at the start.) If this happened during a DATA
2200 ACL check, a -D file could be left lying around. This kind of configuration
2201 error no longer causes Exim to die; instead it causes a defer errror. The
2202 incident is still logged to the main and panic logs.
2204 74. Buglet left over from Exim 3 conversion. The message "too many messages
2205 in one connection" was written to the rejectlog but not the mainlog, except
2206 when address rewriting (yes!) was being logged.
2208 75. Added write_rejectlog option.
2210 76. When a system filter was run not as root (that is, when system_filter_user
2211 was set), the values of the $n variables were not being returned to the
2212 main process; thus, they were not subsequently available in the $sn
2215 77. Added +return_path_on_delivery log selector.
2217 78. A connection timeout was being treated differently from recipients deferred
2218 when testing hosts_max_try with a message that was older than the host's
2219 retry timeout. (The host should not be counted, thus allowing all hosts to
2220 be tried at least once before bouncing.) This may have been the cause of an
2221 occasionally reported bug whereby a message would remain on the queue
2222 longer than the retry timeout, but would be bounced if a delivery was
2223 forced. I say "may" because I never totally pinned down the problem;
2224 setting up timeout/retry tests is difficult. See also the next item.
2226 79. The ultimate address timeout was not being applied to errors that involved
2227 a combination of host plus message (for example, a timeout on a MAIL
2228 command). When an address resolved to a number of possible hosts, and they
2229 were not all tried for each delivery (e.g. because of hosts_max_try), a
2230 message could remain on the queue longer than the retry timeout.
2232 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2235 81. Fixed an obscure SMTP outgoing bug which required at least the following
2236 conditions: (a) there was another message waiting for the same server;
2237 (b) the server returned 5xx to all RCPT commands in the first message so
2238 that the message was not completed; (c) the server dropped the connection
2239 or gave a negative response to the RSET that Exim sends to abort the
2240 transaction. The observed case was a dropped connection after DATA that had
2241 been sent in pipelining mode. That is, the server had advertised PIPELINING
2242 but was not implementing it correctly. The effect of the bug was incorrect
2243 behaviour, such as trying another host, and this could lead to a crash.
2249 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2250 and daemon.c were passed as pointers to ints; they should have been
2251 pointers to socklen_t variables (which are typically unsigned ints).
2253 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2256 3. Fixed a really odd bug that affected only the testing scheme; patching a
2257 certain fixed string in the binary changed the value of another string that
2258 happened to be identical to the end of the original first string.
2260 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2261 name", it returns that address as the IP address. On some operating
2262 systems (e.g. Solaris), it also passes back the IP address string as the
2263 "host name". However, on others (e.g. Linux), it passes back an empty
2264 string. Exim wasn't checking for this, and was changing the host name to an
2265 empty string, assuming it had been canonicized.
2267 5. Although rare, it is permitted to have more than one PTR record for a given
2268 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2269 all the names associated with an address, because they do in Solaris.
2270 However, it seems that they do not in Linux for data that comes from the
2271 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2272 I found this out when I moved to a new Linux workstation and tried to run
2273 the Exim test suite.
2275 To get round this problem I have changed the code so that it now does its
2276 own call to the DNS to look up PTR records when searching for a host name.
2277 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2278 addresses that are only in /etc/hosts are still found.
2280 This behaviour is, however, controlled by an option called host_lookup_
2281 order, which defaults to "bydns:byaddr". If people want to use the other
2282 order, or indeed, just use one or the other means of lookup, they can
2283 specify it in this variable.
2285 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2286 some operating systems, this comes back from gethostbyaddr() as an empty
2287 string, and this is what Exim used to test for. However, it seems that in
2288 other systems, "." is yielded. Exim now tests for this case too.
2290 7. The values of check_spool_space and check_log_space are now held internally
2291 as a number of kilobytes instead of an absolute number of bytes. If a
2292 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2293 kilobyte. This means that much larger values can be stored.
2295 8. Exim monitor: an attempt to get the action menu when not actually pointing
2296 at a message produces an empty menu entitled "No message selected". This
2297 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2298 no entries in it ("Shell widget menu has zero width and/or height"). So I
2299 have added a single, blank menu entry in this case.
2301 9. Added ${quote_local_part.
2303 10. MIME decoding is now applied to the contents of Subject: header lines when
2306 11. Now that a reference to $sender_host_address automatically causes a reverse
2307 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2308 host lookup before query-style lookups in lists that might use this
2309 variable. This has therefore been abolished, and the "net-" prefix is no
2310 longer necessary for query-style lookups.
2312 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2313 have been a typo for LFLAGS, so it has been changed.
2315 13. The install script calls Exim with "-C /dev/null" in order to find the
2316 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2317 to be output. Howeve, since Exim outputs its version number before the
2318 error, it didn't break the script. It just looked ugly. I fixed this by
2319 always allowing "-C /dev/null" if the caller is root.
2321 14. Ignore overlarge ACL variable number when reading spool file - insurance
2322 against a later release with more variables having written the file.
2324 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2325 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2326 this, as well as the form without the "IPv6" on the front.
2328 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2329 OS/Makefile-Darwin file.
2331 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2332 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2334 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2335 variable. However, some OS use ints for inodes. I've added cast to long int
2336 to get rid of the compiler warning.
2338 19. I had forgotten to lock out "/../" in configuration file names when
2339 ALT_CONFIG_PREFIX was set.
2341 20. Routers used for verification do not need to specify transports. However,
2342 if such a router generated a host list, and callout was configured, Exim
2343 crashed, because it could not find a port number from the (non-existent)
2344 transport. It now assumes port 25 in this circumstance.
2346 21. Added the -t option to exigrep.
2348 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2349 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2350 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2351 them is set, LOOKUP_LSEARCH is forced.
2353 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2355 24. Added sender and host information to the "rejected by local_scan()" log
2356 line; previously there was no indication of these.
2358 25. Added .include_if_exists.
2360 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2361 Linux. It turns out that not all file systems support this. Apparently some
2362 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2363 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2364 not supported on the file descriptor, is now ignored when Exim is trying to
2365 sync a directory. This applies only to Linux.
2367 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2369 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2370 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2373 29. Added support for maildirsize files from supplied patch (modified a bit).
2375 30. The use of :fail: followed by an empty string could lead Exim to respond to
2376 sender verification failures with (e.g.):
2378 550 Verification failed for <xxx>
2379 550 Sender verify failed
2381 where the first response line was missing the '-' that indicates it is not
2382 the final line of the response.
2384 31. The loop for finding the name of the user that called Exim had a hardwired
2385 limit of 10; it now uses the value of finduser_retries, which is used for
2386 all other user lookups.
2388 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2391 33. Exim was neglecting to zero errno before one call of strtol() when
2392 expanding a string and expecting an integer value. On some systems this
2393 resulted in spurious "integer overflow" errors. Also, it was casting the
2394 result into an int without checking.
2396 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2397 did not work. The code looks as if it has *never* worked, though it appears
2398 to have been documented since at least releast 1.62. I have made it work.
2400 35. The "timeout_DNS" error in retry rules, also documented since at least
2401 1.62, also never worked. As it isn't clear exactly what this means, and
2402 clearly it isn't a major issue, I have abolished the feature by treating it
2403 as "timeout", and writing a warning to the main and panic logs.
2405 36. The display of retry rules for -brt wasn't always showing the error code
2408 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2409 timeout_connect_A, timeout_connect_MX.
2411 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2412 to the empty sender.
2414 39. The daemon was not analysing the content of -oX till after it had closed
2415 stderr and disconnected from the controlling terminal. This meant that any
2416 syntax errors were only noted on the panic log, and the return code from
2417 the command was 0. By re-arranging the code a little, I've made the
2418 decoding happen first, so such errors now appear on stderr, and the return
2419 code is 1. However, the actual setting up of the sockets still happens in
2420 the disconnected process, so errors there are still only recorded on the
2423 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2424 connections (as happens on some IP stacks) was logged at start up time as
2425 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2426 it from "IPv6 and IPv4", which means that two separate sockets are being
2429 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2430 says whether AF_INET or AF_INET6 was passed as an argument.
2432 42. Exiwhat output was messed up when time zones were included in log
2435 43. Exiwhat now gives more information about the daemon's listening ports,
2436 and whether -tls-on-connect was used.
2438 44. The "port" option of the smtp transport is now expanded.
2440 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2441 silently ignored. Now an error message is written to the main and panic
2444 46. There's a new ACL modifier called "logwrite" which writes to a log file
2445 as soon as it is encountered.
2447 47. Added $local_user_uid and $local_user_gid at routing time.
2449 48. Exim crashed when trying to verify a sender address that was being
2452 49. Exim was recognizing only a space character after ".include". It now also
2453 recognizes a tab character.
2455 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2456 extracting the relevant information from the specification. The man page no
2457 longer contains scrambled data for the -d option, and I've added a section
2458 at the front about calling Exim under different names.
2460 51. Added "extra_headers" argument to the "mail" command in filter files.
2462 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2465 53. Installed eximstats 1.29.
2467 54. Added transport_filter_timeout as a generic transport option.
2469 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2470 Cc: header lines. This was required by RFC 822, but it not required by RFC
2473 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2474 incoming messages that did not have them. Now it does so only if the
2475 message originates locally, that is, if there is no associated remote host
2476 address. When Resent- header lines are present, this applies to the Resent-
2477 lines rather than the non-Resent- lines.
2479 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2480 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2482 58. Messages for configuration errors now include the name of the main
2483 configuration file - useful now that there may be more than one file in a
2484 list (.included file names were always shown).
2486 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2487 for those rare installations that do not start the daemon as root or run it
2488 setuid root. I've cut out the call to initgroups() if the daemon is not
2491 60. The Exim user and group can now be bound into the binary as text strings
2492 that are looked up at the start of Exim's processing.
2494 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2496 62. Added $mailstore_basename variable.
2498 63. Installed patch to sieve.c from Michael Haardt.
2500 64. When Exim failed to open the panic log after failing to open the main log,
2501 the original message it was trying to log was written to stderr and debug
2502 output, but if they were not available (the usual case in production), it
2503 was lost. Now it is written to syslog before the two lines that record the
2504 failures to open the logs.
2506 65. Users' Exim filters run in subprocesses under the user's uid. It is
2507 possible for a "deliver" command or an alias in a "personal" command to
2508 provoke an address rewrite. If logging of address rewriting is configured,
2509 this fails because the process is not running as root or exim. There may be
2510 a better way of dealing with this, but for the moment (because 4.30 needs
2511 to be released), I have disabled address rewrite logging when running a
2512 filter in a non-root, non-exim process.
2518 1. The buildconfig auxiliary program wasn't quoting the value set for
2519 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2520 not defined. This bug was masked in 4.22 by the effect that was fixed in
2523 2. Some messages that were rejected after a message id was allocated were
2524 shown as "incomplete" by exigrep. It no longer does this for messages that
2525 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2527 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2528 have allow_domain_literals set, the ID did not get logged in the <= line.
2529 Domain literals are now always recognized in Message-ID: header lines.
2531 4. The first argument for a ${extract expansion item is the key name or field
2532 number. Leading and trailing spaces in this item were not being ignored,
2533 causing some misleading effects.
2535 5. When deliver_drop_privilege was set, single queue runner processes started
2536 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2537 same command in the process it spins off) were not dropping privilege.
2539 6. When the daemon running as "exim" started a queue runner, it always
2540 re-executed Exim in the spun-off process. This is a waste of effort when
2541 deliver_drop_privilege is set. The new process now just calls the
2542 queue-runner function directly.
2548 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2551 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2552 ignored. Though the use of -oP was forcing the writing of a pid file, it
2553 was always written to the default place.
2555 3. If the message "no IP address found for host xxxx" is generated during
2556 incoming verification, it is now followed by identification of the incoming
2557 connection (so you can more easily find what provoked it).
2559 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2561 5. Added some features to "harden" Exim a bit more against certain attacks:
2563 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2564 be put in Local/Makefile. This is like the never_users runtime option,
2565 but it cannot be overridden. The default setting is "root".
2567 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2568 prefix string with which any file named in a -C command line option
2571 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2572 is retained for -C and -D only if the caller of Exim is root. Without
2573 it, the exim user may also use -C and -D and retain privilege.
2575 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2576 command line option is disabled.
2578 6. Macro names set by the -D option must start with an upper case letter, just
2579 like macro names defined in the configuration file.
2581 7. Added "dereference=" facility to LDAP.
2583 8. Two instances of the typo "uknown" in the source files are fixed.
2585 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2586 the Configure-Makefile script screwed up while processing it.
2588 10. Incorporated PCRE 4.4.
2590 11. The SMTP synchronization check was not operating right at the start of an
2591 SMTP session. For example, it could not catch a HELO sent before the client
2592 waited for the greeting. There is now a check for outstanding input at the
2593 point when the greeting is written. Because of the duplex, asynchronous
2594 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2595 way, but not yet received, when the check is performed.
2597 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2598 on TCP/IP sockets, because this apparently causes some broken clients to
2601 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2602 unchanged) from the Cygwin maintainer.
2604 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2605 was supported instead of testing for mbx. Effectively a typo.
2607 15. The spa authenticator server code was not checking that the input it
2608 received was valid base64.
2610 16. The debug output line for the "set" modifier in ACLs was not showing the
2611 name of the variable that was being set.
2613 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2615 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2616 lookup on its own. Something else was needed to trigger the lookup. For
2617 example, a match in host_lookup or the need for a host name in a host list.
2618 Now, if $sender_host_name is referenced and the host name has not yet been
2619 looked up, a lookup is performed. If the lookup fails, the variable remains
2620 empty, and $host_lookup_failed is set to "1".
2622 19. Added "eqi" as a case-independent comparison operator.
2624 20. The saslauthd authentication condition could segfault if neither service
2625 nor realm was specified.
2627 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2628 error message that was logged was misleading, and incoming SMTP
2629 connections were dropped. The message is now more accurate, and temporary
2630 errors are given to SMTP connections.
2632 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2633 (except RSET) with a 421 error, until QUIT is received. However, it was
2634 failing to send a response to QUIT.
2636 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2637 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2638 ACL is now run after the helo_try_verify_hosts code.
2640 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2641 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2642 case-independent, so other case variants are also recognized). Apparently
2643 some systems use these upper case variants.
2645 25. If more than two messages were waiting for the same host, and a transport
2646 filter was specified for the transport, Exim sent two messages over the
2647 same TCP/IP connection, and then failed with "socket operation on non-
2648 socket" when it tried to send the third.
2650 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2652 27. The extern definition of crypt16() in expand.c was not being excluded when
2653 the OS had its own crypt16() function.
2655 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2656 as a preferred synonym for return_size_limit, both as an option and as an
2659 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2661 30. Changed the default configuration ACL to relax the local part checking rule
2662 for addresses that are not in any local domains. For these addresses,
2663 slashes and pipe symbols are allowed within local parts, but the sequence
2664 /../ is explicitly forbidden.
2666 31. SPA server authentication was not clearing the challenge buffer before
2669 32. log_message in a "warn" ACL statement was writing to the reject log as
2670 well as to the main log, which contradicts the documentation and doesn't
2671 seem right (because no rejection is happening). So I have stopped it.
2673 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2674 However, I am unable to do any testing of this.
2676 34. Fixed an infelicity in the appendfile transport. When checking directories
2677 for a mailbox, to see if any needed to be created, it was accidentally
2678 using path names with one or more superfluous leading slashes; tracing
2679 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2681 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2682 contents are added to the log line that is written for every discarded
2683 recipient. (Previously a log_message setting was ignored.)
2685 36. The ${quote: operator now quotes the string if it is empty.
2687 37. The install script runs exim in order to find its version number. If for
2688 some reason other than non-existence or emptiness, which it checks, it
2689 could not run './exim', it was installing it with an empty version number,
2690 i.e. as "exim-". This error state is now caught, and the installation is
2693 38. An argument was missing from the function that creates an error message
2694 when Exim fails to connect to the socket for saslauthd authentication.
2695 This could cause Exim to crash, or give a corrupted message.
2697 39. Added isip, isip4, and isip6 to ${if conditions.
2699 40. The ACL variables $acl_xx are now saved with the message, and can be
2700 accessed later in routers, transports, and filters.
2702 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2703 strings in the file are not string-expanded.
2705 42. If a MAIL command specified a SIZE value that was too large to fit into an
2706 int variable, the check against message_size_limit failed. Such values are
2707 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2708 day this will have to be increased, but I don't think I want to be around
2709 when emails are that large.
2716 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2717 iconv() is not standard in FreeBSD.
2719 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2720 IPv6 enabled. The observed symptom was a segmentation fault on return from
2721 the function os_common_find_running_interfaces() in src/os.c.
2723 3. In the check_special_case() function in daemon.c I had used "errno" as an
2724 argument name, which causes warnings on some systems. This was basically a
2725 typo, since it was named "eno" in the comments!
2727 4. The code that waits for the clock to tick (at a resolution of some fraction
2728 of a second) so as to ensure message-id uniqueness was always waiting for
2729 at least one whole tick, when it could have waited for less. [This is
2730 almost certainly not relevant at current processor speeds, where it is
2731 unlikely to ever wait at all. But we try to future-proof.]
2733 5. The function that sleeps for a time interval that includes fractions of a
2734 second contained a race. It did not block SIGALRM between setting the
2735 timer, and suspending (a couple of lines later). If the interval was short
2736 and the sigsuspend() was delayed until after it had expired, the suspension
2737 never ended. On busy systems this could lead to processes getting stuck for
2740 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2741 process, before it forks any delivery processes. The open lookup caching
2742 mechanism meant that the open file or database connection was passed into
2743 the delivery process. The problem was that delivery processes always tidy
2744 up cached lookup data. This could cause a problem for the next delivery
2745 process started by the queue runner, because the external queue runner
2746 process does not know about the closure. So the next delivery process
2747 still has data in the lookup cache. In the case of a file lookup, there was
2748 no problem because closing a file descriptor in a subprocess doesn't affect
2749 the parent. However, if the lookup was caching a connection to a database,
2750 the connection was closed, and the second delivery process was likely to
2751 see errors such as "PGSQL: query failed: server closed the connection
2752 unexpectedly". The problem has been fixed by closing all cached lookups
2753 in a queue runner before running a delivery process.
2755 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2756 seem to have the "const" qualifier which it has on other OS. I've
2759 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2760 *of the same type* (client or server) with the same public name that an
2761 error should be diagnosed.
2763 9. When Exim looked up a host name for an IP address, but failed to find the
2764 original IP address when looking up the host name (a safety check), it
2765 output the message "<ip address> does not match any IP for NULL", which was
2766 confusing, to say the least. The bug was that the host name should have
2767 appeared instead of "NULL".
2769 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2770 user that is built into the binary, and the -C or -D options is used, root
2771 privilege is dropped before the configuration file is read. In addition,
2772 logging is switched to stderr instead of the normal log files. If the
2773 configuration then re-defines the Exim user, the unprivileged environment
2774 is probably not what is expected, so Exim logs a panic warning message (but
2777 However, if deliver_drop_privilege is set, the unprivileged state may well
2778 be exactly what is intended, so the warning has been cut out in that case,
2779 and Exim is allowed to try to write to its normal log files.
2785 1. smtp_return_error_details was not giving details for temporary sender
2786 or receiver verification errors.
2788 2. Diagnose a configuration error if two authenticators have the same public
2791 3. Exim used not to create the message log file for a message until the first
2792 delivery attempt. This could be confusing when incoming messages were held
2793 for policy or load reasons. The message log file is now created at the time
2794 the message is received, and an initial "Received" line is written to it.
2796 4. The automatically generated man page for command line options had a minor
2797 bug that caused no ill effects; however, a more serious problem was that
2798 the procedure for building the man page automatically didn't always
2799 operate. Consequently, release 4.20 contains an out-of-date version. This
2800 shouldn't happen again.
2802 5. When building Exim with embedded Perl support, the script that builds the
2803 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2804 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2806 6. The freeze_tell option was not being used for messages that were frozen on
2807 arrival, either by an ACL or by local_scan().
2809 7. Added the smtp_incomplete_transaction log selector.
2811 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2812 was accepting AUTH without a new EHLO.
2814 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2815 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2818 10. Logging of TCP/IP connections (when configured) now happens in the main
2819 daemon process instead of the child process, so that the TCP/IP connection
2820 count is more accurate (but it can never be perfect).
2822 11. The use of "drop" in a nested ACL was not being handled correctly in the
2823 outer ACL. Now, if condition failure induced by the nested "drop" causes
2824 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2825 or "require"), the connection is dropped.
2827 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2828 that yield "discard" can now be used with an "accept" or a "discard" verb,
2829 but an error is generated for any others (because I can't see a useful way
2830 to define what should happen).
2832 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2833 are now processed in the same way as lines in the Exim configuration file.
2834 In particular, continuation lines are supported.
2836 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2838 15. Added -ti meaning -t -i.
2840 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2841 domains, and warn by logging if others are found.
2843 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2844 returned from GIFCONF when Exim is trying to find the list of interfaces on
2845 a host. The code in os.c has been modified to copy each ifreq to an aligned
2846 structure in all cases.
2848 Also, in some cases, the returned ifreq's were being copied to a 'struct
2849 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2850 means the last couple of bytes of an IPv6 address could be chopped if the
2851 ifreq contained only a normal sockaddr (14 bytes storage).
2853 18. Named domain lists were not supported in the hosts_treat_as_local option.
2854 An entry such as +xxxx was not recognized, and was treated as a literal
2857 19. Ensure that header lines added by a DATA ACL are included in the reject log
2858 if the ACL subsequently rejects the message.
2860 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2861 MD5 (which is deprecated).
2863 21. When testing a filter file using -bf, Exim was writing a message when it
2864 took the sender from a "From " line in the message, but it was not doing so
2865 when it took $return_path from a Return-Path: header line. It now does.
2867 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2868 with a valid header line field name (a series of printing characters
2869 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2871 23. Changed "disc" in the source to "disk" to conform to the documentation and
2872 the book and for uniformity.
2874 24. Ignore Sendmail's -Ooption=value command line item.
2876 25. When execve() failed while trying to run a command in a pipe transport,
2877 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2878 could be confused with a return value of 69 from the command itself. This
2879 has been changed to 127, the value the shell returns if it is asked to run
2880 a non-existent command. The wording for the related log line suggests a
2881 non-existent command as the problem.
2883 26. If received_header_text expands to an empty string, do not add a Received:
2884 header line to the message. (Well, it adds a token one on the spool, but
2885 marks it "old" so that it doesn't get used or transmitted.)
2887 27. Installed eximstats 1.28 (addition of -nt option).
2889 28. There was no check for failure on the call to getsockname() in the daemon
2890 code. This can fail if there is a shortage of resources on the system, with
2891 ENOMEM, for example. A temporary error is now given on failure.
2893 29. Contrary to the C standard, it seems that in some environments, the
2894 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2895 program. Exim now does this explicitly; it affects the formatting of
2896 timestamps using strftime().
2898 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2899 complaints. I've now initialized all the variables, to avoid this.
2901 32. Header lines added by a system filter were not being "seen" during
2902 transport-time rewrites.
2904 33. The info_callback() function passed to OpenSSL is set up with type void
2905 (*)(SSL *, int, int), as described somewhere. However, when calling the
2906 function (actually a macro) that sets it up, the type void(*)() is
2907 expected. I've put in a cast to prevent warnings from picky compilers.
2909 34. If a DNS black list lookup found a CNAME record, but there were no A
2910 records associated with the domain it pointed at, Exim crashed.
2912 35. If a DNS black list lookup returned more than one A record, Exim ignored
2913 all but the first. It now scans all returned addresses if a particular IP
2914 value is being sought. In this situation, the contents of the
2915 $dnslist_value variable are a list of all the addresses, separated by a
2918 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2919 to accept a host name and all its aliases if the forward lookup for any of
2920 them yielded the IP address of the incoming connection. Now it accepts only
2921 those names whose forward lookup yields the correct IP address. Any other
2922 names are discarded. This closes a loophole whereby a rogue DNS
2923 administrator could create reverse DNS records to break through a
2924 wildcarded host restriction in an ACL.
2926 37. If a user filter or a system filter that ran in a subprocess used any of
2927 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2928 the wrong values were passed to the pipe command ($thisaddress had the
2929 value of $0, $0 had the value of $1, etc). This bug was introduced by
2930 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2932 38. Improved the line breaking for long SMTP error messages from ACLs.
2933 Previously, if there was no break point between 40 and 75 characters, Exim
2934 left the rest of the message alone. Two changes have been made: (a) I've
2935 reduced the minimum length to 35 characters; (b) if it can't find a break
2936 point between 35 and 75 characters, it looks ahead and uses the first one
2937 that it finds. This may give the occasional overlong line, but at least the
2938 remaining text gets split now.
2940 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2941 file descriptors might be low, and that setting 1000 would always raise it.
2942 It turns out that in some environments, the limit is already over 1000 and
2943 that lowering it causes trouble. So now Exim takes care not to decrease it.
2945 40. When delivering a message, the value of $return_path is set to $sender_
2946 address at the start of routing (routers may change the value). By an
2947 oversight, this default was not being set up when an address was tested by
2948 -bt or -bv, which affected the outcome if any router or filter referred to
2951 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2952 log only when "message" or "log_message" are set. However, if one of the
2953 conditions was an address verification, or a call to a nested ACL, the
2954 messages generated by the underlying test were being passed through. This
2955 no longer happens. The underlying message is available in $acl_verify_
2956 message for both "message" and "log_message" expansions, so it can be
2957 passed through if needed.
2959 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2960 new expansion $bh_ to give the encoded byte string without charset
2961 translation. Translation happens only if iconv() is available; HAVE_ICONV
2962 indicates this at build time. HEADERS_CHARSET gives the charset to
2963 translate to; headers_charset can change it in the configuration, and
2964 "headers charset" can change it in an individual filter file.
2966 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2967 that creates RFC 2047 encoded "words" labels them as that charset instead
2968 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2969 expansion operator; (ii) when Exim creates a From: line for a local
2970 message; (iii) when a header line is rewritten to include a "phrase" part.
2972 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2973 buggy, causing it to skip the first lines of messages whose message ID
2974 ended in 'D'. This would not have bitten before Exim release 4.14, because
2975 message IDs were unlikely to end in 'D' before then. The effect was to have
2976 incorrect size information for certain domains.
2978 45. #include "config.h" was missing at the start of the crypt16.c module. This
2979 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2982 46. If there was a timeout during a "random" callout check, Exim treated it as
2983 a failure of the random address, and carried on sending RSET and the real
2984 address. If the delay was just some slowness somewhere, the response to the
2985 original RCPT would be taken as a response to RSET and so on, causing
2986 mayhem of various kinds.
2988 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2989 when I implemented it. It didn't allow for the fact that some option values
2990 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2991 the right test for positive values.
2993 48. Domain names in DNS records are case-independent. Exim always looks them up
2994 in lower case. Some resolvers return domain names in exactly the case they
2995 appear in the zone file, that is, they may contain uppercase letters. Not
2996 all resolvers do this - some return always lower case. Exim was treating a
2997 change of case by a resolver as a change of domain, similar to a widening
2998 of a domain abbreviation. This triggered its re-routing code and so it was
2999 trying to route what was effectively the same domain again. This normally
3000 caused routing to fail (because the router wouldn't handle the domain
3001 twice). Now Exim checks for this case specially, and just changes the
3002 casing of the domain that it ultimately uses when it transmits the message
3005 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
3008 50. If a filter generated a file delivery with a non-absolute name (possible if
3009 no home directory exists for the router), the forbid_file option was not
3012 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3013 the existing equality matching.
3015 52. Exim was using ints instead of ino_t variables in some places where it was
3016 dealing with inode numbers.
3018 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3019 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3020 called TMPDIR, and if it finds it is different, it changes its value.
3022 54. The smtp_printf() function is now made available to local_scan() so
3023 additional output lines can be written before returning. There is also an
3024 smtp_fflush() function to enable the detection of a dropped connection.
3025 The variables smtp_input and smtp_batched_input are exported to
3028 55. Changed the default runtime configuration: the message "Unknown user"
3029 has been removed from the ACL, and instead placed on the localuser router,
3030 using the cannot_route_message feature. This means that any verification
3031 failures that generate their own messages won't get overridden. Similarly,
3032 the "Unrouteable address" message that was in the ACL for unverifiable
3033 relay addresses has also been removed.
3035 56. Added hosts_avoid_esmtp to the smtp transport.
3037 57. The exicyclog script was not checking for the esoteric option
3038 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3039 will work only if exicyclog is run under the appropriate euid.
3041 58. Following a discussion on the list, the rules by which Exim recognises line
3042 endings on incoming messages have been changed. The -dropcr and drop_cr
3043 options are now no-ops, retained only for backwards compatibility. The
3044 following line terminators are recognized: LF CRLF CR. However, special
3045 processing applies to CR:
3047 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3048 nor a local message in the state where . is a terminator.
3050 (ii) If a bare CR is encountered in a header line, an extra space is added
3051 after the line terminator so as not to end the header. The reasoning
3052 behind this is that bare CRs in header lines are most likely either
3053 to be mistakes, or people trying to play silly games.
3055 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3056 was being incorrectly given as 18 bytes larger than it should have been.
3057 This is a VOB (very old bug).
3059 60. This may never have affected anything current, but just in case it has:
3060 When the local host is found other than at the start of a list of hosts,
3061 the local host, those with the same MX, and any that follow, are discarded.
3062 When the list in question was part of a longer list of hosts, the following
3063 hosts (not currently being processed) were also being discarded. This no
3064 longer happens. I'm not sure if this situation could ever has previously
3067 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3070 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3071 header. This is something that is recommended in a new Internet Draft, and
3072 is something that is documented as being done by Sendmail. There are two
3073 possible values. For messages generated by the autoreply transport, Exim
3076 Auto-Submitted: auto-replied
3078 whereas for all other generated messages (e.g. bounces) it adds
3080 Auto-Submitted: auto-generated
3082 63. The "personal" condition in filters now includes a test for the
3083 Auto-Submitted: header. If it contains the string "auto-" the message it
3084 not considered personal.
3086 64. Added rcpt_include_affixes as a generic transport option.
3088 65. Added queue_only_override (default true).
3090 66. Added the syslog_duplication option.
3092 67. If what should have been the first header line of a message consisted of
3093 a space followed by a colon, Exim was mis-interpreting it as a header line.
3094 It isn't of course - it is syntactically invalid and should therefore be
3095 treated as the start of the message body. The misbehaviour could have
3096 caused a number of strange effects, including loss of data in subsequent
3097 header lines, and spool format errors.
3099 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3100 client host had authenticated. This control can now be exercised by an ACL
3101 for more flexibility.
3103 69. By default, callouts do not happen when testing with -bh. There is now a
3104 variant, -bhc, which does actually run the callout code, including
3105 consulting and updating the callout cache.
3107 70. Added support for saslauthd authentication, courtesy of Alexander
3110 71. If statvfs() failed on the spool or log directories while checking their
3111 size for availability, Exim confusingly gave the error "space shortage".
3112 Furthermore, in debugging mode it crashed with a floating point exception.
3113 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3114 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3115 serious problem, Exim now writes to the main and panic logs when this
3116 happens, with details of the failure. It then refuses to accept the
3117 incoming message, giving the message "spool directory problem" or "log
3118 directory problem" with a 421 code for SMTP messages.
3120 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3121 0, 1, and 2 exist, because some OS complain for execs without them (see
3122 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3123 descriptors. However, the code omitted to check that the open succeeded,
3124 causing mysterious errors if for some reason the permissions on /dev/null
3125 got screwed. Now Exim writes a message to the main and panic logs, and
3126 bombs out if it can't open /dev/null.
3128 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3129 interact so that it is all more flexible. It is supposed to remain
3130 backwards compatible. Also added extra_local_interfaces.
3132 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3133 to bomb out with an assertion failure - to the client this appears as a
3134 connection drop. This problem occurs in the part of the code that was taken
3135 from the Samba project. Fortunately, the assertion is in a very simple
3136 function, so I have fixed this by reproducing the function inline in the
3137 one place where it is called, and arranging for authentication to fail
3138 instead of killing the process with assert().
3140 75. The SPA client code was not working when the server requested OEM rather
3141 than Unicode encoding.
3143 76. Added code to make require_files with a specific uid setting more usable in
3144 the case where statting the file as root fails - usually a non-root-mounted
3145 NFS file system. When this happens and the failure is EACCES, Exim now
3146 forks a subprocess and does the per-uid checking as the relevant uid.
3148 77. Added process_log_path.
3150 78. If log_file_path was not explicitly set, a setting of check_log_space or
3151 check_log_inodes was ignored.
3153 79. If a space check for the spool or log partitions fails, the incident is now
3154 logged. Of course, in the latter case the data may get lost...
3156 80. Added the %p formatting code to string_format() so that it can be used to
3157 print addresses in debug_print(). Adjusted all the address printing in the
3158 debugging in store.c to use %p rather than %d.
3160 81. There was a concern that a line of code in smtp_in.c could overflow a
3161 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3162 initially expressed, the concern was not well-founded, because trailing
3163 spaces are removed early. However, if the trailing spaces were followed by
3164 a NULL, they did not get removed, so the overflow was possible. Two fixes
3167 (a) I re-wrote the offending code in a cleaner fashion.
3168 (b) If an incoming SMTP command contains a NULL character, it is rejected
3171 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3172 runs initgroups(), so that if the Exim user is in any additional groups,
3173 they will be used during message reception.
3179 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / heiko / exim.git / blob