1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.183 2005/08/01 13:20:28 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.
18 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
20 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
21 command itself ended in error, the underlying message about the transport
22 filter timeout was being overwritten with the pipe command error. Now the
23 underlying error message should be appended to the second error message.
25 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
27 PH/02 Reduce the amount of output that "make" produces by default. Full output
28 can still be requested.
30 PH/03 The warning log line about a condition test deferring for a "warn" verb
31 was being output only once per connection, rather than after each
32 occurrence (because it was using the same function as for successful
33 "warn" verbs). This seems wrong, so I have changed it.
35 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
36 it should not have, which might have caused a crash in the right
37 circumstances, but probably never did.
39 PH/04 Installed a modified version of Tony Finch's patch to make submission
40 mode fix the return path as well as the Sender: header line, and to
41 add a /name= option so that you can make the user's friendly name appear
44 TF/03 Added the control = fakedefer ACL modifier.
46 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
47 Mark Lowes for thorough testing.
49 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
51 TK/03 Merged latest SRS patch from Miles Wilton.
53 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
54 with the definition in sysexits.h (which is #included earlier).
55 Fortunately, Exim does not actually use EX_OK. The code used to try to
56 preserve the sysexits.h value, by assumimg that macro definitions were
57 scanned for macro replacements. I have been disabused of this notion,
58 so now the code just undefines EX_OK before #including unistd.h.
60 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
61 in the smtp transport. When a block could not be written in a single
62 write() function, the timeout was being re-applied to each part-write.
63 This seems wrong - if the receiver was accepting one byte at a time it
64 would take for ever. The timeout is now adjusted when this happens. It
65 doesn't have to be particularly precise.
67 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
68 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
70 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
71 without a domain if the check to <postmaster@domain> fails.
73 SC/01 Eximstats: added -xls and the ability to specify output files
74 (patch written by Frank Heydlauf).
76 SC/02 Eximstats: use FileHandles for outputing results.
78 SC/03 Eximstats: allow any combination of xls, txt, and html output.
80 SC/04 Eximstats: fixed display of large numbers with -nvr option
82 SC/05 Eximstats: fixed merging of reports with empty tables.
84 SC/06 Eximstats: added the -include_original_destination flag
86 SC/07 Eximstats: removed tabs and trailing whitespace.
88 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
90 TK/06 MBOX spool code: Add real "From " MBOX separator line
91 so the .eml file is really in mbox format (even though
92 most programs do not really care). Patch from Alex Miller.
94 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
95 The latter is generated from $received_to and is only set if the
96 message has one envelope recipient. SA can use these headers,
97 obviously out-of-the-box. Patch from Alex Miller.
99 PH/08 The ${def test on a variable was returning false if the variable's
100 value was "0", contrary to what the specification has always said!
101 The result should be true unless the variable is empty.
103 PH/09 The syntax error of a character other than { following "${if
104 def:variable_name" (after optional whitespace) was not being diagnosed.
105 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
106 accidental colon was present, for example, could give incorrect results.
108 PH/10 Tidied the code in a number of places where the st_size field of a stat()
109 result is used (not including appendfile, where other changes are about
112 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
113 This involved changing a lot of size variables from int to off_t. It
114 should work with maildirs and everything.
116 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
117 spamd dying while we are connected to it.
119 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
122 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
123 function for BDB 4.3.
125 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
127 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
128 the sources for PCRE have been re-arranged and more files are now
131 PH/15 The code I had for printing potentially long long variables in PH/11
132 above was not the best (it lost precision). The length of off_t variables
133 is now inspected at build time, and an appropriate printing format (%ld
134 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
135 to be "long long int" or "long int". This is needed for the internal
136 formatting function string_vformat().
138 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
139 the configuration file to be ":syslog", then the script "guesses" where
140 the logs files are, rather than using the compiled in default. In our
141 case the guess is not the same as the compiled default, so the script
142 suddenly stopped working when I started to use syslog. The patch checks
143 to see if log_file_path is "". If so, it attempts to read it from exim
144 with no configuration file to get the compiled in version, before it
145 falls back to the previous guessing code."
147 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
148 implementing BATV in an Exim configuration. See NewStuff for the gory
151 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
152 Makefile that are specific to HP-UX.
154 PH/18 If the "use_postmaster" option was set for a recipient callout together
155 with the "random" option, the postmaster address was used as the MAIL
156 FROM address for the random test, but not for the subsequent recipient
157 test. It is now used for both.
159 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
160 patch removes a few documentation additions to RFC 3028, because the
161 latest draft now contains them. It adds the new en;ascii-case comparator
162 and a new error check for 8bit text in MIME parts. Comparator and
163 require names are now matched exactly. I enabled the subaddress
164 extension, but it is not well tested yet (read: it works for me)."
166 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
167 rework some of the code of TK/09 above to avoid the hardwired use of
168 "%lld" and "long long". Replaced the call to snprintf() with a call to
171 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
172 records point to non-existent hosts", "retry timeout exceeded", and
173 "retry time not reached for any host after a long failure period".
175 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
176 experimental DomainKeys support:
178 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
179 (2) On an error such as an illegally used "control", the wrong name for
180 the control was given.
182 These problems did NOT occur unless DomainKeys support was compiled.
184 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
186 PH/24 Added ${if match_ip condition.
188 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
189 databases so that it will be absolutely obvious if a crash occurs in the
190 DB library. This is a regular occurrence (often caused by mis-matched
193 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
194 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
195 picked up on a user's system that detects such things. There doesn't seem
196 to be a gcc warning option for this - only an attribute that has to be
197 put on the function's prototype. It seems that in Fedora Core 4 they have
198 set this on a number of new functions. No doubt there will be more in due
201 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
202 specify a transport. However, if an address that was verified by such a
203 router was the subject of a callout, Exim crashed because it tried to
204 read the rcpt_include_affixes from the non-existent transport. Now it
205 just assumes that the setting of that option is false. This bug was
206 introduced by 4.51/PH/31.
208 PH/28 Changed -d+all to exclude +memory, because that information is very
209 rarely of interest, but it makes the output a lot bigger. People tend to
210 do -d+all out of habit.
212 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
213 code in os-type was giving problems when libc.so lives in lib64, like on
216 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
217 aren't the modern standard, and it seems that some systems' include files
218 don't always have them. Exim was already checking for some of the newer
219 ones like T_AAAA, and defining it itself. I've added checks for all the
220 record types that Exim uses.
222 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
223 not automatically generating a new one, as it is supposed to. This
224 prevented TLS from working. If the file did exist, but contained invalid
225 data, a new version was generated, as expected. It was only the case of a
226 non-existent file that was broken.
228 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
229 with a change in libdomainkeys > 0.64.
231 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
232 from DNS. If the selector record carries the flag, it now has
233 precedence over the domain-wide flag.
235 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
237 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
238 the use of a port name that isn't defined in /etc/services) occurred, the
239 message was deferred as in a normal delivery, and thus remained on the
240 spool, instead of being failed because of the mua_wrapper setting. This
241 is now fixed, and I tidied up some of the mua_wrapper messages at the
244 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
245 the messages in a hash of arrays rather than using individual hashes.
246 This is a bit cleaner and results in dramatic memory savings, albeit
247 at a slight CPU cost.
249 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
250 as requested by Marc Sherman.
252 SC/10 Eximstats: added histograms for user specified patterns as requested
255 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
257 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
258 fopen() in the content-scanning modules that did not already have it.
264 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
265 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
267 TK/02 Fix ACL "control" statement not being available in MIME ACL.
269 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
271 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
272 to test Sieve filters that use "vacation".
274 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
275 that changes the way the GnuTLS parameters are stored in the cache file.
276 The new format can be generated externally. For backward compatibility,
277 if the data in the cache doesn't make sense, Exim assumes it has read an
278 old-format file, and it generates new data and writes a new file. This
279 means that you can't go back to an older release without removing the
282 PH/03 A redirect router that has both "unseen" and "one_time" set does not
283 work if there are any delivery delays because "one_time" forces the
284 parent to be marked "delivered", so its unseen clone is never tried
285 again. For this reason, Exim now forbids the simultaneous setting of
288 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
289 redirected to themselves ("homonym" addresses). Read the long ChangeLog
290 entry if you want to know the details. The fix, however, neglected to
291 consider the case when local delivery batching is involved. The test for
292 "previously delivered" was not happening when checking to see if an
293 address could be batched with a previous (undelivered) one; under
294 certain circumstances this could lead to multiple deliveries to the same
297 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
298 in its include files, and this causes problems building Exim.
300 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
301 header_syntax) but Exim was just ignoring anything given after a slash.
302 In particular, this caused confusion with an attempt to use "verify =
303 reverse_host_lookup/defer_ok". An error is now given when options are
304 supplied for verify items that do not have them. (Maybe reverse_host_
305 lookup should have a defer_ok option, but that's a different point.)
307 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
308 defined by RFC 821) to 2048, because there were problems with some AUTH
309 commands, and RFC 1869 says the size should be increased for extended
310 SMTP commands that take arguments.
312 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
315 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
316 "unknown" error; now it says that the functionality isn't in the binary.
318 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
319 an address' error message when a string expansion fails (syntax or
320 whatever). Otherwise the password may appear in the log. Following change
321 PH/42 below, there is no longer a chance of it appearing in a bounce
324 PH/11 Installed exipick version 20050225.0 from John Jetmore.
326 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
327 of its addresses was ever tried. (Bugzilla bug #2.)
329 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
330 the result incorrectly in the debug output. (It correctly added a newline
331 to what was transported.)
333 TF/01 Added $received_time.
335 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
336 commented out examples of how to interface to a virus scanner and to
337 SpamAssassin. Also added commented examples of av_scanner and
338 spamd_address settings.
340 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
341 and controls are allowed in which ACLs. There were a couple of minor
342 errors. Some of the entries in the conditions table (which is a table of
343 where they are NOT allowed) were getting very unwieldy; rewrote them as a
344 negation of where the condition IS allowed.
346 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
348 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
349 header file does not have a version number, so I've had to invent a new
350 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
351 API. The code is untested by me (my Linux distribution still has 0.3.2 of
352 radiusclient), but it was contributed by a Radius user.
354 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
355 files or directories, for OpenSSL.
357 PH/19 When an Exim process that is running as root has to create an Exim log
358 file, it does so in a subprocess that runs as exim:exim so as to get the
359 ownership right at creation (otherwise, other Exim processes might see
360 the file with the wrong ownership). There was no test for failure of this
361 fork() call, which would lead to the process getting stuck as it waited
362 for a non-existent subprocess. Forks do occasionally fail when resources
363 run out. I reviewed all the other calls to fork(); they all seem to check
366 PH/20 When checking for unexpected SMTP input at connect time (before writing
367 the banner), Exim was not dealing correctly with a non-positive return
368 from the read() function. If the client had disconnected by this time,
369 the result was a log entry for a synchronization error with an empty
370 string after "input=" when read() returned zero. If read() returned -1
371 (an event I could not check), uninitialized data bytes were printed.
372 There were reports of junk text (parts of files, etc) appearing after
375 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
377 PH/22 Added support for macro redefinition, and (re)definition in between
378 driver and ACL definitions.
380 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
381 forgetting to use the resulting value; it was using the unexpanded value.
383 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
384 hadn't been configured. The fix is from Juergen Kreileder, who
385 understands it better than I do:
387 "Here's what I see happening with three configured cyrus_sasl
388 authenticators configured (plain, login, cram-md5):
390 On startup auth_cyrus_sasl_init() gets called for each of these.
391 This means three calls to sasl_listmech() without a specified mech_list.
392 => SASL tests which mechs of all available mechs actually work
393 => three warnings about OTP not working
394 => the returned list contains: plain, login, cram-md5, digest-md5, ...
396 With the patch, sasl_listmech() also gets called three times. But now
397 SASL's mech_list option is set to the server_mech specified in the the
398 authenticator. Or in other words, the answer from sasl_listmech()
399 gets limited to just the mech you're testing for (which is different
401 => the return list contains just 'plain' or 'login', 'cram-md5' or
402 nothing depending on the value of ob->server_mech.
404 I've just tested the patch: Authentication still works fine,
405 unavailable mechs specified in the exim configuration are still
406 caught, and the auth.log warnings about OTP are gone."
408 PH/25 When debugging is enabled, the contents of the command line are added
409 to the debugging output, even when log_selector=+arguments is not
412 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
413 answer is "GNU", and only if the return is "GNU/something" is the answer
416 PH/27 $acl_verify_message is now set immediately after the failure of a
417 verification in an ACL, and so is available in subsequent modifiers. In
418 particular, the message can be preserved by coding like this:
420 warn !verify = sender
421 set acl_m0 = $acl_verify_message
423 Previously, $acl_verify_message was set only while expanding "message"
424 and "log_message" when a very denied access.
426 PH/28 Modified OS/os.c-Linux with
428 -#ifndef OS_LOAD_AVERAGE
429 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
431 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
432 nomenclature these days.)
434 PH/29 Installed patch from the Sieve maintainer that adds the options
435 sieve_useraddress and sieve_subaddress to the redirect router.
437 PH/30 In these circumstances:
438 . Two addresses routed to the same list of hosts;
439 . First host does not offer TLS;
440 . First host accepts first address;
441 . First host gives temporary error to second address;
442 . Second host offers TLS and a TLS session is established;
443 . Second host accepts second address.
444 Exim incorrectly logged both deliveries with the TLS parameters (cipher
445 and peerdn, if requested) that were in fact used only for the second
448 PH/31 When doing a callout as part of verifying an address, Exim was not paying
449 attention to any local part prefix or suffix that was matched by the
450 router that accepted the address. It now behaves in the same way as it
451 does for delivery: the affixes are removed from the local part unless
452 rcpt_include_affixes is set on the transport.
454 PH/32 Add the sender address, as F=<...>, to the log line when logging a
455 timeout during the DATA phase of an incoming message.
457 PH/33 Sieve envelope tests were broken for match types other than :is. I have
458 applied a patch sanctioned by the Sieve maintainer.
460 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
461 the uid or gid is negative. A case of a negative gid caused this to be
462 noticed. The fix allows for either to be negative.
464 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
465 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
468 PH/36 The change PH/12 above was broken. Fixed it.
470 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
471 the grounds that routing the same address twice would always produce the
472 same answer. This might have been true once, but it is certainly no
473 longer true now. Routing a child address may depend on the previous
474 routing that produced that child. Some complicated redirection strategies
475 went wrong when messages had multiple recipients, and made Exim's
476 behaviour dependent on the order in which the addresses were given.
478 I have moved the duplicate checking until after the routing is complete.
479 Exim scans the addresses that are assigned to local and remote
480 transports, and removes any duplicates. This means that more work will be
481 done, as duplicates will always all be routed, but duplicates are
482 presumably rare, so I don't expect this is of any significance.
484 For deliveries to pipes, files, and autoreplies, the duplicate checking
485 still happens during the routing process, since they are not going to be
488 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
489 It corrects a timeout issue with spamd. This is Ian's comment: "The
490 background is that sometimes spamd either never reads data from a
491 connection it has accepted, or it never writes response data. The exiscan
492 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
493 blindly assumes that writes won't block so it may never time out."
495 PH/39 Allow G after quota size as well as K and M.
497 PH/40 The value set for $authenticated_id in an authenticator may not contain
498 binary zeroes or newlines because the value is written to log lines and
499 to spool files. There was no check on this. Now the value is run through
500 the string_printing() function so that such characters are converted to
501 printable escape sequences.
503 PH/41 $message_linecount is a new variable that contains the total number of
504 lines in the message. Compare $body_linecount, which is the count for the
507 PH/42 Exim no longer gives details of delivery errors for specific addresses in
508 bounce and delay warning messages, except in certain special cases, which
511 (a) An SMTP error message from a remote host;
512 (b) A message specified in a :fail: redirection;
513 (c) A message specified in a "fail" command in a system filter;
514 (d) A message specified in a FAIL return from the queryprogram router;
515 (e) A message specified by the cannot_route_message router option.
517 In these cases only, Exim does include the error details in bounce and
518 warning messages. There are also a few cases where bland messages such
519 as "unrouteable address" or "local delivery error" are given.
521 PH/43 $value is now also set for the "else" part of a ${run expansion.
523 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
524 being worked on, but at least Exim now implements the latest version to
527 PH/45 In a pipe transport, although a timeout while waiting for the pipe
528 process to complete was treated as a delivery failure, a timeout while
529 writing the message to the pipe was logged, but erroneously treated as a
530 successful delivery. Such timeouts include transport filter timeouts. For
531 consistency with the overall process timeout, these timeouts are now
532 treated as errors, giving rise to delivery failures by default. However,
533 there is now a new Boolean option for the pipe transport called
534 timeout_defer, which, if set TRUE, converts the failures into defers for
535 both kinds of timeout. A transport filter timeout is now identified in
538 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
539 systems where "make" and "gmake" are different, calling "gmake" at top
540 level broke things. I've arranged for the value of $(MAKE) to be passed
541 from the Makefile to this script so that it can call the same version of
545 A note about Exim versions 4.44 and 4.50
546 ----------------------------------------
548 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
549 changes of various kinds. As a consequence, a big documentation update was
550 needed. This delayed the release for rather longer than seemed good, especially
551 in the light of a couple of (minor) security issues. Therefore, the changes
552 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
553 release. So 4.44 and 4.50 are in effect two different branches that both start
556 I have left the 4.50 change log unchanged; it contains all the changes since
557 4.43. The change log for 4.44 is below; many of its items are identical to
558 those for 4.50. This seems to be the most sensible way to preserve the
559 historical information.
565 1. Minor wording change to the doc/README.SIEVE file.
567 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
568 computation of the current number of files was incorrect.
570 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
571 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
574 4. Give more explanation in the error message when the command for a transport
575 filter fails to execute.
577 5. There are several places where Exim runs a non-Exim command in a
578 subprocess. The SIGUSR1 signal should be disabled for these processes. This
579 was being done only for the command run by the queryprogram router. It is
580 now done for all such subprocesses. The other cases are: ${run, transport
581 filters, and the commands run by the lmtp and pipe transports.
583 6. Added CONFIGURE_GROUP build-time option.
585 7. Some older OS have a limit of 256 on the maximum number of file
586 descriptors. Exim was using setrlimit() to set 1000 as a large value
587 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
588 systems. I've change it so that if it can't get 1000, it tries for 256.
590 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
591 was an oversight, and furthermore, ever since the addition of extra
592 controls (e.g. 4.43/32), the checks on when to allow different forms of
593 "control" were broken. There should now be diagnostics for all cases when a
594 control that does not make sense is encountered.
596 9. Added the /retain_sender option to "control=submission".
598 10. $recipients is now available in the predata ACL (oversight).
600 11. Tidy the search cache before the fork to do a delivery from a message
601 received from the command line. Otherwise the child will trigger a lookup
602 failure and thereby defer the delivery if it tries to use (for example) a
603 cached ldap connection that the parent has called unbind on.
605 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
606 of $address_data from the recipient verification was clobbered by the
609 13. The value of address_data from a sender verification is now available in
610 $sender_address_data in subsequent conditions in the ACL statement.
612 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
614 15. Added a new option "connect=<time>" to callout options, to set a different
617 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
618 was its contents. (It was OK if the option was not defined at all.)
620 17. A "Completed" log line is now written for messages that are removed from
621 the spool by the -Mrm option.
623 18. New variables $sender_verify_failure and $recipient_verify_failure contain
624 information about exactly what failed.
626 19. Added -dd to debug only the daemon process.
628 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
629 handles timeouts, both on the server side and network timeouts. Renamed the
630 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
632 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
633 It is now set to "smtps".
635 22. $host_address is now set to the target address during the checking of
638 23. When checking ignore_target_hosts for an ipliteral router, no host name was
639 being passed; this would have caused $sender_host_name to have been used if
640 matching the list had actually called for a host name (not very likely,
641 since this list is usually IP addresses). A host name is now passed as
644 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
645 code that specifies a non-restarting handler (typically sigaction() in
646 modern systems) in an attempt to fix a rare and obscure crash bug.
648 25. Narrowed the window for a race in the daemon that could cause it to ignore
649 SIGCHLD signals. This is not a major problem, because they are used only to
650 wake it up if nothing else does.
652 26. A malformed maildirsize file could cause Exim to calculate negative values
653 for the mailbox size or file count. Odd effects could occur as a result.
654 The maildirsize information is now recalculated if the size or filecount
657 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
658 support for a long time. Removed HAVE_SYS_VFS_H.
660 28. Installed the latest version of exipick from John Jetmore.
662 29. In an address list, if the pattern was not a regular expression, an empty
663 subject address (from a bounce message) matched only if the pattern was an
664 empty string. Non-empty patterns were not even tested. This was the wrong
665 because it is perfectly reasonable to use an empty address as part of a
666 database query. An empty address is now tested by patterns that are
667 lookups. However, all the other forms of pattern expect the subject to
668 contain a local part and a domain, and therefore, for them, an empty
669 address still always fails if the pattern is not itself empty.
671 30. Exim went into a mad DNS loop when attempting to do a callout where the
672 host was specified on an smtp transport, and looking it up yielded more
675 31. Re-factored the code for checking spool and log partition space into a
676 function that finds that data and another that does the check. The former
677 is then used to implement four new variables: $spool_space, $log_space,
678 $spool_inodes, and $log_inodes.
680 32. The RFC2047 encoding function was originally intended for short strings
681 such as real names; it was not keeping to the 75-character limit for
682 encoded words that the RFC imposes. It now respects the limit, and
683 generates multiple encoded words if necessary. To be on the safe side, I
684 have increased the buffer size for the ${rfc2047: expansion operator from
687 33. It is now permitted to omit both strings after an "if" condition; if the
688 condition is true, the result is "true". As before, when the second string
689 is omitted, a false condition yields an empty string. This makes it less
690 cumbersome to write custom ACL and router conditions.
692 34. Failure to deliver a bounce message always caused it to be frozen, even if
693 there was an errors_to setting on the router. The errors_to setting is now
696 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
697 canonical form (fully expanded) before being placed in
698 $sender_host_address.
700 36. The table in the code that translates DNS record types into text (T_A to
701 "A" for instance) was missing entries for NS and CNAME. It is just possible
702 that this could have caused confusion if both these types were looked up
703 for the same domain, because the text type is used as part of Exim's
704 per-process caching. But the chance of anyone hitting this buglet seems
707 37. The dnsdb lookup has been extended in a number of ways.
709 (1) There is a new type, "zns", which walks up the domain tree until it
710 finds some nameserver records. It should be used with care.
712 (2) There is a new type, "mxh", which is like "mx" except that it returns
713 just the host names, not the priorities.
715 (3) It is now possible to give a list of domains (or IP addresses) to be
716 looked up. The behaviour when one of the lookups defers can be
717 controlled by a keyword.
719 (4) It is now possible to specify the separator character for use when
720 multiple records are returned.
722 38. The dnslists ACL condition has been extended: it is now possible to supply
723 a list of IP addresses and/or domains to be looked up in a particular DNS
726 39. Added log_selector=+queue_time_overall.
728 40. When running the queue in the test harness, wait just a tad after forking a
729 delivery process, to get repeatability of debugging output.
731 41. Include certificate and key file names in error message when GnuTLS fails
732 to set them up, because the GnuTLS error message doesn't include the name
733 of the failing file when there is a problem reading it.
735 42. Allow both -bf and -bF in the same test run.
737 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
739 44. The "Exiscan patch" is now merged into the mainline Exim source.
741 45. Sometimes the final signoff response after QUIT could fail to get
742 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
743 < 0 before doing a fflush(). This bug looks as though it goes back to the
744 introduction of TLS in release 3.20, but "sometimes" must have been rare
745 because the tests only now provoked it.
747 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
748 (this can affect the format of dates).
750 47. exim_tidydb, when checking for the continued existence of a message for
751 which it has found a message-specific retry record, was not finding
752 messages that were in split spool directories. Consequently, it was
753 deleting retry records that should have stayed in existence.
755 48. Steve fixed some bugs in eximstats.
757 49. The SPA authentication driver was not abandoning authentication and moving
758 on to the next authenticator when an expansion was forced to fail,
759 contradicting the general specification for all authenticators. Instead it
760 was generating a temporary error. It now behaves as specified.
762 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
763 (the order specifies the preference for clients). The order is now AES256,
764 AES128, 3DES, ARCFOUR128.
766 51. Small patch to Sieve code - explicitly set From: when generating an
769 52. Exim crashed if a remote delivery caused a very long error message to be
770 recorded - for instance if somebody sent an entire SpamAssassin report back
771 as a large number of 550 error lines. This bug was coincidentally fixed by
772 increasing the size of one of Exim's internal buffers (big_buffer) that
773 happened as part of the Exiscan merge. However, to be on the safe side, I
774 have made the code more robust (and fixed the comments that describe what
777 53. Now that there can be additional text after "Completed" in log lines (if
778 the queue_time_overall log selector is set), a one-byte patch to exigrep
779 was needed to allow it to recognize "Completed" as not the last thing in
782 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
783 patch that reportedly fixes this has been added. I am not expert enough to
784 create a test for it. This is what the patch creator wrote:
786 "I found a little strange behaviour of ldap code when working with
787 Windows 2003 AD Domain, where users was placed in more than one
788 Organization Units. When I tried to give exim partial DN, the exit code
789 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
790 But simultaneously result of request was absolutely normal ldap result,
791 so I produce this patch..."
793 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
794 REFERENCE, so I have modified the code to exclude the patch when that macro
797 55. Some experimental protocols are using DNS PTR records for new purposes. The
798 keys for these records are domain names, not reversed IP addresses. The
799 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
800 leaves it alone. Component reversal etc. now happens only for IP addresses.
803 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
805 57. Double the size of the debug message buffer (to 2048) so that more of very
806 long debug lines gets shown.
808 58. The exicyclog utility now does better if the number of log files to keep
809 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
811 59. Two changes related to the smtp_active_hostname option:
813 (1) $smtp_active_hostname is now available as a variable.
814 (2) The default for smtp_banner uses $smtp_active_hostname instead
815 of $primary_hostname.
817 60. The host_aton() function is supposed to be passed a string that is known
818 to be a valid IP address. However, in the case of IPv6 addresses, it was
819 not checking this. This is a hostage to fortune. Exim now panics and dies
820 if the condition is not met. A case was found where this could be provoked
821 from a dnsdb PTR lookup with an IPv6 address that had more than 8
822 components; fortuitously, this particular loophole had already been fixed
823 by change 4.50/55 above.
825 If there are any other similar loopholes, the new check in host_aton()
826 itself should stop them being exploited. The report I received stated that
827 data on the command line could provoke the exploit when Exim was running as
828 exim, but did not say which command line option was involved. All I could
829 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
833 61. There was a buffer overflow vulnerability in the SPA authentication code
834 (which came originally from the Samba project). I have added a test to the
835 spa_base64_to_bits() function which I hope fixes it.
838 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
839 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
841 63. The daemon start-up calls getloadavg() while still root for those OS that
842 need the first call to be done as root, but it missed one case: when
843 deliver_queue_load_max is set with deliver_drop_privilege. This is
844 necessary for the benefit of the queue runner, because there is no re-exec
845 when deliver_drop_privilege is set.
847 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
850 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
851 was in use, was not putting the data itself into the right store pool;
852 consequently, it could be overwritten for a subsequent message in the same
853 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
856 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
858 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
859 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
860 the function were treating the return as a boolean value, which happened to
861 work because 0=false and not-0=true, but is not correct code.
863 68. The host_aton() function was not handling scoped IPv6 addresses (those
864 with, for example, "%eth0" on the end) correctly.
866 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
867 negated items (that is, ~something) in unsigned ints. Some compilers
868 apparently mutter when there is no cast.
870 70. If an address verification called from an ACL failed, and did not produce a
871 user-specific message (i.e. there was only a "system" message), nothing was
872 put in $acl_verify_message. In this situation, it now puts the system
875 71. Change 4.23/11 added synchronization checking at the start of an SMTP
876 session; change 4.31/43 added the unwanted input to the log line - except
877 that it did not do this in the start of session case. It now does.
879 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
880 This is wrong and can cause the other end to generate a synchronization
881 error if it is another Exim or anything else that does the synchronization
882 check. A QUIT command is no longer sent after a timeout.
884 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
887 74. The defer_ok option of callout verification was not working if it was used
888 when verifying addresses in header lines, that is, for this case:
890 verify = header_sender/callout=defer_ok
892 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
893 those file descriptors could be used for SMTP connections. If anything
894 wrote to stderr (the example that came up was "warn" in embedded Perl), it
895 could be sent to the SMTP client, causing chaos. The daemon now opens
896 stdin, stdout, and stderr to /dev/null when it puts itself into the
899 76. Arrange for output from Perl's "warn" command to be written to Exim's main
900 log by default. The user can override this with suitable Perl magic.
902 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
903 the log message when discard triggers, was not working for the DATA ACL or
904 for the non-SMTP ACL.
906 78. Error message wording change in sieve.c.
908 79. If smtp_accept_max_per_host was set, the number of connections could be
909 restricted to fewer than expected, because the daemon was trying to set up
910 a new connection before checking whether the processes handling previous
911 connections had finished. The check for completed processes is now done
912 earlier. On busy systems, this bug wouldn't be noticed because something
913 else would have woken the daemon, and it would have reaped the completed
916 80. If a message was submitted locally by a user whose login name contained one
917 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
918 It caused a spool format error. I have fixed the spool reading code. A
919 related problem was that the "from" clause in the Received: line became
920 illegal because of the space(s). It is now covered by ${quote_local_part.
922 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
925 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
927 83. Patch from Sieve maintainer to fix unterminated string problem in
930 84. Some minor changes to the Linux configuration files to help with other
931 OS variants using glibc.
933 85. One more patch for Sieve to update vacation handling to latest spec.
936 ----------------------------------------------------
937 See the note above about the 4.44 and 4.50 releases.
938 ----------------------------------------------------
944 1. Change 4.43/35 introduced a bug that caused file counts to be
945 incorrectly computed when quota_filecount was set in an appendfile
948 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
949 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
952 3. Give more explanation in the error message when the command for a transport
953 filter fails to execute.
955 4. There are several places where Exim runs a non-Exim command in a
956 subprocess. The SIGUSR1 signal should be disabled for these processes. This
957 was being done only for the command run by the queryprogram router. It is
958 now done for all such subprocesses. The other cases are: ${run, transport
959 filters, and the commands run by the lmtp and pipe transports.
961 5. Some older OS have a limit of 256 on the maximum number of file
962 descriptors. Exim was using setrlimit() to set 1000 as a large value
963 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
964 systems. I've change it so that if it can't get 1000, it tries for 256.
966 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
967 was an oversight, and furthermore, ever since the addition of extra
968 controls (e.g. 4.43/32), the checks on when to allow different forms of
969 "control" were broken. There should now be diagnostics for all cases when a
970 control that does not make sense is encountered.
972 7. $recipients is now available in the predata ACL (oversight).
974 8. Tidy the search cache before the fork to do a delivery from a message
975 received from the command line. Otherwise the child will trigger a lookup
976 failure and thereby defer the delivery if it tries to use (for example) a
977 cached ldap connection that the parent has called unbind on.
979 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
980 of $address_data from the recipient verification was clobbered by the
983 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
984 was its contents. (It was OK if the option was not defined at all.)
986 11. A "Completed" log line is now written for messages that are removed from
987 the spool by the -Mrm option.
989 12. $host_address is now set to the target address during the checking of
992 13. When checking ignore_target_hosts for an ipliteral router, no host name was
993 being passed; this would have caused $sender_host_name to have been used if
994 matching the list had actually called for a host name (not very likely,
995 since this list is usually IP addresses). A host name is now passed as
998 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
999 code that specifies a non-restarting handler (typically sigaction() in
1000 modern systems) in an attempt to fix a rare and obscure crash bug.
1002 15. Narrowed the window for a race in the daemon that could cause it to ignore
1003 SIGCHLD signals. This is not a major problem, because they are used only to
1004 wake it up if nothing else does.
1006 16. A malformed maildirsize file could cause Exim to calculate negative values
1007 for the mailbox size or file count. Odd effects could occur as a result.
1008 The maildirsize information is now recalculated if the size or filecount
1011 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
1012 support for a long time. Removed HAVE_SYS_VFS_H.
1014 18. Updated exipick to current release from John Jetmore.
1016 19. Allow an empty sender to be matched against a lookup in an address list.
1017 Previously the only cases considered were a regular expression, or an
1020 20. Exim went into a mad DNS lookup loop when doing a callout where the
1021 host was specified on the transport, if the DNS lookup yielded more than
1024 21. The RFC2047 encoding function was originally intended for short strings
1025 such as real names; it was not keeping to the 75-character limit for
1026 encoded words that the RFC imposes. It now respects the limit, and
1027 generates multiple encoded words if necessary. To be on the safe side, I
1028 have increased the buffer size for the ${rfc2047: expansion operator from
1031 22. Failure to deliver a bounce message always caused it to be frozen, even if
1032 there was an errors_to setting on the router. The errors_to setting is now
1035 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1036 canonical form (fully expanded) before being placed in
1037 $sender_host_address.
1039 24. Updated eximstats to version 1.33
1041 25. Include certificate and key file names in error message when GnuTLS fails
1042 to set them up, because the GnuTLS error message doesn't include the name
1043 of the failing file when there is a problem reading it.
1045 26. Expand error message when OpenSSL has problems setting up cert/key files.
1048 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1049 (this can affect the format of dates).
1051 28. exim_tidydb, when checking for the continued existence of a message for
1052 which it has found a message-specific retry record, was not finding
1053 messages that were in split spool directories. Consequently, it was
1054 deleting retry records that should have stayed in existence.
1056 29. eximstats updated to version 1.35
1057 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1058 1.35 - bugfix such that pie charts by volume are generated correctly
1060 30. The SPA authentication driver was not abandoning authentication and moving
1061 on to the next authenticator when an expansion was forced to fail,
1062 contradicting the general specification for all authenticators. Instead it
1063 was generating a temporary error. It now behaves as specified.
1065 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1066 (the order specifies the preference for clients). The order is now AES256,
1067 AES128, 3DES, ARCFOUR128.
1069 31. Small patch to Sieve code - explicitly set From: when generating an
1072 32. Exim crashed if a remote delivery caused a very long error message to be
1073 recorded - for instance if somebody sent an entire SpamAssassin report back
1074 as a large number of 550 error lines. This bug was coincidentally fixed by
1075 increasing the size of one of Exim's internal buffers (big_buffer) that
1076 happened as part of the Exiscan merge. However, to be on the safe side, I
1077 have made the code more robust (and fixed the comments that describe what
1080 33. Some experimental protocols are using DNS PTR records for new purposes. The
1081 keys for these records are domain names, not reversed IP addresses. The
1082 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1083 leaves it alone. Component reversal etc. now happens only for IP addresses.
1086 34. The host_aton() function is supposed to be passed a string that is known
1087 to be a valid IP address. However, in the case of IPv6 addresses, it was
1088 not checking this. This is a hostage to fortune. Exim now panics and dies
1089 if the condition is not met. A case was found where this could be provoked
1090 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1091 components; fortuitously, this particular loophole had already been fixed
1092 by change 4.50/55 or 4.44/33 above.
1094 If there are any other similar loopholes, the new check in host_aton()
1095 itself should stop them being exploited. The report I received stated that
1096 data on the command line could provoke the exploit when Exim was running as
1097 exim, but did not say which command line option was involved. All I could
1098 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1099 running as the user.
1102 35. There was a buffer overflow vulnerability in the SPA authentication code
1103 (which came originally from the Samba project). I have added a test to the
1104 spa_base64_to_bits() function which I hope fixes it.
1107 36. The daemon start-up calls getloadavg() while still root for those OS that
1108 need the first call to be done as root, but it missed one case: when
1109 deliver_queue_load_max is set with deliver_drop_privilege. This is
1110 necessary for the benefit of the queue runner, because there is no re-exec
1111 when deliver_drop_privilege is set.
1113 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1114 was in use, was not putting the data itself into the right store pool;
1115 consequently, it could be overwritten for a subsequent message in the same
1116 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1119 38. Sometimes the final signoff response after QUIT could fail to get
1120 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1121 < 0 before doing a fflush(). This bug looks as though it goes back to the
1122 introduction of TLS in release 3.20, but "sometimes" must have been rare
1123 because the tests only now provoked it.
1129 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1130 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1131 Now it returns an integer. A number of calls were still expecting a T/F
1132 return. Fortuitously, in all cases, the tests worked in OK situations,
1133 which is the norm. However, things would have gone wrong on any write
1134 failures on the smtp file descriptor. This function is used when sending
1135 messages over SMTP and also when doing verify callouts.
1137 2. When Exim is called to do synchronous delivery of a locally submitted
1138 message (the -odf or -odi options), it no longer closes stderr before doing
1141 3. Implemented the mua_wrapper option.
1143 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1145 5. Implemented the functions header_remove(), header_testname(),
1146 header_add_at_position(), and receive_remove_recipient(), and exported them
1149 6. If an ACL "warn" statement specified the addition of headers, Exim already
1150 inserted X-ACL-Warn: at the start if there was no header name. However, it
1151 was not making this test for the second and subsequent header lines if
1152 there were newlines in the string. This meant that an invalid header could
1153 be inserted if Exim was badly configured.
1155 7. Allow an ACL "warn" statement to add header lines at the start or after all
1156 the Received: headers, as well as at the end.
1158 8. Added the rcpt_4xx retry error code.
1160 9. Added postmaster_mailfrom=xxx to callout verification option.
1162 10. Added mailfrom=xxxx to the callout verification option, for verify=
1165 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1166 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1168 12. Inserted some casts to stop certain compilers warning when using pointer
1169 differences as field lengths or precisions in printf-type calls (mostly
1170 affecting debugging statements).
1172 13. Added optional readline() support for -be (dynamically loaded).
1174 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1175 same clock tick as a message's arrival, so that its received time was the
1176 same as the "first fail" time on the retry record, and that message
1177 remained on the queue past the ultimate address timeout, every queue runner
1178 would try a delivery (because it was past the ultimate address timeout) but
1179 after another failure, the ultimate address timeout, which should have then
1180 bounced the address, did not kick in. This was a "< instead of <=" error;
1181 in most cases the first failure would have been in the next clock tick
1182 after the received time, and all would be well.
1184 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1185 being recognized when the domain list was tested by the match_domain
1186 condition in an expansion string.
1188 16. Added the ${str2b64: operator.
1190 17. Exim was always calling setrlimit() to set a large limit for the number of
1191 processes, without checking whether the existing limit was already
1192 adequate. (It did check for the limit on file descriptors.) Furthermore,
1193 errors from getrlimit() and setrlimit() were being ignored. Now they are
1194 logged to the main and panic logs, but Exim does carry on, to try to do its
1195 job under whatever limits there are.
1197 18. Imported PCRE 5.0.
1199 19. Trivial typo in log message " temporarily refused connection" (the leading
1202 20. If the log selector return_path_on_delivery was set and an address was
1203 redirected to /dev/null, the delivery process crashed because it assumed
1204 that a return path would always be set for a "successful" delivery. In this
1205 case, the whole delivery is bypassed as an optimization, and therefore no
1208 21. Internal re-arrangement: the function for sending a challenge and reading
1209 a response while authentication was assuming a zero-terminated challenge
1210 string. It's now changed to take a pointer and a length, to allow for
1211 binary data in such strings.
1213 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1215 23. Exim was not respecting finduser_retries when seeking the login of the
1216 uid under which it was called; it was always trying 10 times. (The default
1217 setting of finduser_retries is zero.) Also, it was sleeping after the final
1218 failure, which is pointless.
1220 24. Implemented tls_on_connect_ports.
1222 25. Implemented acl_smtp_predata.
1224 26. If the domain in control=submission is set empty, Exim assumes that the
1225 authenticated id is a complete email address when it generates From: or
1226 Sender: header lines.
1228 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1229 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1230 chown and chgrp in /bin and hostname in /usr/bin.
1232 28. Exim was keeping the "process log" file open after each use, just as it
1233 does for the main log. This opens the possibility of it remaining open for
1234 long periods when the USR1 signal hits a daemon. Occasional processlog
1235 errors were reported, that could have been caused by this. Anyway, it seems
1236 much more sensible not to leave this file open at all, so that is what now
1239 29. The long-running daemon process does not normally write to the log once it
1240 has entered its main loop, and it closes the log before doing so. This is
1241 so that log files can straightforwardly be renamed and moved. However,
1242 there are a couple of unusual error situations where the daemon does write
1243 log entries, and I had neglected to close the log afterwards.
1245 30. The text of an SMTP error response that was received during a remote
1246 delivery was being truncated at 512 bytes. This is too short for some of
1247 the long messages that one sometimes sees. I've increased the limit to
1250 31. It is now possible to make retry rules that apply only when a message has a
1251 specific sender, in particular, an empty sender.
1253 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1254 it possible to be selective about when SMTP synchronization is enforced.
1256 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1258 32. Implemented hosts_connection_nolog.
1260 33. Added an ACL for QUIT.
1262 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1265 35. Added mailbox_size and mailbox_filecount to appendfile.
1267 36. Added control = no_multiline_responses to ACLs.
1269 37. There was a bug in the logic of the code that waits for the clock to tick
1270 in the case where the clock went backwards by a substantial amount such
1271 that the microsecond fraction of "now" was more than the microsecond
1272 fraction of "then" (but the whole seconds number was less).
1274 38. Added support for the libradius Radius client library this is found on
1275 FreeBSD (previously only the radiusclient library was supported).
1281 1. When certain lookups returned multiple values in the form name=value, the
1282 quoting of the values was not always being done properly. Specifically:
1283 (a) If the value started with a double quote, but contained no whitespace,
1285 (b) If the value contained whitespace other than a space character (i.e.
1286 tabs or newlines or carriage returns) it was not quoted.
1287 This fix has been applied to the mysql and pgsql lookups by writing a
1288 separate quoting function and calling it from the lookup code. The fix
1289 should probably also be applied to nisplus, ibase and oracle lookups, but
1290 since I cannot test any of those, I have not disturbed their existing code.
1292 2. A hit in the callout cache for a specific address caused a log line with no
1293 reason for rejecting RCPT. Now it says "Previous (cached) callout
1294 verification failure".
1296 3. There was an off-by-one bug in the queryprogram router. An over-long
1297 return line was truncated at 256 instead of 255 characters, thereby
1298 overflowing its buffer with the terminating zero. As well as fixing this, I
1299 have increased the buffer size to 1024 (and made a note to document this).
1301 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1302 when Exim is waiting for an SMTP response from a remote server, Exim
1303 restarts its select() call on the socket, thereby resetting its timeout.
1304 This is not a problem when such interrupts are rare. Somebody set up a cron
1305 job to run exiwhat every 2 minutes, which is less than the normal select()
1306 timeout (5 or 10 minutes). This meant that the select() timeout never
1307 kicked in because it was always reset. I have fixed this by comparing the
1308 time when an interrupt arrives with the time at the start of the first call
1309 to select(). If more time than the timeout has elapsed, the interrupt is
1310 treated as a timeout.
1312 5. Some internal re-factoring in preparation for the addition of Sieve
1313 extensions (by MH). In particular, the "personal" test is moved to a
1314 separate function, and given an option for scanning Cc: and Bcc: (which is
1315 not set for Exim filters).
1317 6. When Exim created an email address using the login of the caller as the
1318 local part (e.g. when creating a From: or Sender: header line), it was not
1319 quoting the local part when it contained special characters such as @.
1321 7. Installed new OpenBSD configuration files.
1323 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1324 try to make them clearer.
1326 9. Callout options, other than the timeout value, were being ignored when
1327 verifying sender addresses in header lines. For example, when using
1329 verify = header_sender/callout=no_cache
1331 the cache was (incorrectly) being used.
1333 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1334 only the Cygwin environment.
1336 11. When return_path_on_delivery was set as a log selector, if different remote
1337 addresses in the same message used different return paths and parallel
1338 remote delivery occurred, the wrong values would sometimes be logged.
1339 (Whenever a remote delivery process finished, the return path value from
1340 the most recently started remote delivery process was logged.)
1342 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1343 lines when AUTH and/or TLS are in use. This is the "received protocol"
1344 field. Exim used to use "asmtp" for authenticated SMTP, without any
1345 indication (in the protocol name) for TLS use. Now it follows the RFC and
1346 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1347 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1348 names appear in log lines as well as in Received: header lines.
1350 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1351 extensions, and comparison tests, and to fix some bugs.
1353 14. Changes to the "personal" filter test:
1355 (1) The test was buggy in that it was just doing the equivalent of
1356 "contains" tests on header lines. For example, if a user's address was
1357 anne@some.where, the "personal" test would incorrectly be true for
1359 To: susanne@some.where
1361 This test is now done by extracting each address from the header in turn,
1362 and checking the entire address. Other tests that are part of "personal"
1363 are now done using regular expressions (for example, to check local parts
1364 of addresses in From: header lines).
1366 (2) The list of non-personal local parts in From: addresses has been
1367 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1368 taken from the Sieve specification recommendations.
1370 (3) If the message contains any header line starting with "List-" it is
1371 treated as non-personal.
1373 (4) The test for "circular" in the Subject: header line has been removed
1374 because it now seems ill-conceived.
1376 15. Minor typos in src/EDITME comments corrected.
1378 16. Installed latest exipick from John Jetmore.
1380 17. If headers_add on a router specified a text string that was too long for
1381 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1382 of string_sprintf() is now avoided.
1384 18. $message_body_size was not set (it was always zero) when running the DATA
1385 ACL and the local_scan() function.
1387 19. For the "mail" command in an Exim filter, no default was being set for
1388 the once_repeat time, causing a random time value to be used if "once" was
1389 specified. (If the value happened to be <= 0, no repeat happened.) The
1390 default is now 0s, meaning "never repeat". The "vacation" command was OK
1391 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1392 (I found it when inspecting the code).
1394 20. There is now an overall timeout for performing a callout verification. It
1395 defaults to 4 times the callout timeout, which applies to individual SMTP
1396 commands during the callout. The overall timeout applies when there is more
1397 than one host that can be tried. The timeout is checked before trying the
1398 next host. This prevents very long delays if there are a large number of
1399 hosts and all are timing out (e.g. when the network connections are timing
1400 out). The value of the overall timeout can be changed by specifying an
1401 additional sub-option for "callout", called "maxwait". For example:
1403 verify = sender/callout=5s,maxwait=20s
1405 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1406 to the end before writing, but this should make it even safer).
1408 22. Exim was forgetting that it had advertised PIPELINING for the second and
1409 subsequent messages on an SMTP connection. It was also not resetting its
1410 memory on STARTTLS and an internal HELO.
1412 23. When Exim logs an SMTP synchronization error within a session, it now
1413 records whether PIPELINING has been advertised or not.
1415 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1416 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1417 rather than long int.
1419 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1421 26. Added the never_mail option to autoreply.
1427 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1428 crash if the getsockname() call failed; this can happen if a connection is
1429 closed very soon after it is established. The problem was simply in the
1430 order in which certain operations were done, causing Exim to try to write
1431 to the SMTP stream before it had set up the file descriptor. The bug has
1432 been fixed by making things happen in the correct order.
1438 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1439 before the connection was closed, thus losing the rejection response.
1441 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1442 some early Solaris releases, but causes trouble in current releases where
1443 socklen_t is defined.
1445 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1448 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1451 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1452 that releasing the top of it at the end releases what was used for sub-
1453 expansions (unless the block got too big). However, discard this block if
1454 the first thing is a variable or header, so that we can use its block when
1455 it is dynamic (useful for very large $message_headers, for example).
1457 6. Lookups now cache *every* query, not just the most recent. A new, separate
1458 store pool is used for this. It can be recovered when all lookup caches are
1459 flushed. Lookups now release memory at the end of their result strings.
1460 This has involved some general refactoring of the lookup sources.
1462 7. Some code has been added to the store_xxx() functions to reduce the amount
1463 of flapping under certain conditions.
1465 8. log_incoming_interface used to affect only the <= reception log lines. Now
1466 it causes the local interface and port to be added to several more SMTP log
1467 lines, for example "SMTP connection from", and rejection lines.
1469 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1471 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1473 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1474 could be overwritten at the end of the current message (or the start of a
1475 new message if it was set in a HELO ACL). The value is now preserved for
1476 the duration of the SMTP connection.
1478 12. If a transport had a headers_rewrite setting, and a matching header line
1479 contained an unqualified address, that address was qualified, even if it
1480 did not match any rewriting rules. The underlying bug was that the values
1481 of the flags that permit the existence of unqualified sender and recipient
1482 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1483 non-local messages, and by -bnq for local messages) were not being
1484 preserved with the message after it was received.
1486 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1487 "next input=" as part of the text comprising the host identity instead of
1488 the correct text. The code was using the same buffer for two different
1489 strings. However, depending on which order the printing function evaluated
1490 its arguments, the bug did not always show up. Under Linux, for example, my
1491 test suite worked just fine.
1493 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1494 doesn't work with some older versions of Perl. It has been changed to "my",
1495 which in any case is probably the better facility to use.
1497 15. A really picky compiler found some instances of statements for creating
1498 error messages that either had too many or two few arguments for the format
1501 16. The size of the buffer for calls to the DNS resolver has been increased
1502 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1503 for addresses that have a lot of PTR records. This alleviates a problem; it
1504 does not fully solve it.
1506 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1507 buffer now truncates the list and logs the incident, which is the same
1508 action as happens when Exim is looking up a host name and its aliases.
1509 Previously in this situation something unpredictable would happen;
1510 sometimes it was "internal error: store_reset failed".
1512 18. If a server dropped the connection unexpectedly when an Exim client was
1513 using GnuTLS and trying to read a response, the client delivery process
1514 crashed while trying to generate an error log message.
1516 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1517 string, for example:
1519 warn message = H1: something\nH2: something
1521 the text was added as a single header line from Exim's point of view
1522 though it ended up OK in the delivered message. However, searching for the
1523 second and subsequent header lines using $h_h2: did not work. This has been
1524 fixed. Similarly, if a system filter added multiple headers in this way,
1525 the routers could not see them.
1527 20. Expanded the error message when iplsearch is called with an invalid key to
1528 suggest using net-iplsearch in a host list.
1530 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1531 ACLs are no longer actually imposed (and a message to that effect is
1534 22. If a "gecos" field in a passwd entry contained escaped characters, in
1535 particular, if it contained a \" sequence, Exim got it wrong when building
1536 a From: or a Sender: header from that name. A second bug also caused
1537 incorrect handling when an unquoted " was present following a character
1538 that needed quoting.
1540 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1541 was not being matched caselessly.
1543 24. Arranged for all hyphens in the exim.8 source to be escaped with
1546 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1547 properly. Recipient callout cache records were still being keyed to include
1548 the sender, even when use_sender was set false. This led to far more
1549 callouts that were necessary. The sender is no longer included in the key
1550 when use_sender is false.
1552 26. Added "control = submission" modifier to ACLs.
1554 27. Added the ${base62d: operator to decode base 62 numbers.
1556 28. dnsdb lookups can now access SRV records.
1558 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1559 the configuration file.
1561 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1562 (-v) mode. This makes the output for a verbose queue run more intelligible.
1564 31. Added a use_postmaster feature to recipient callouts.
1566 32. Added the $body_zerocount variable, containing the number of binary zero
1567 bytes in the message body.
1569 33. The time of last modification of the "new" subdirectory is now used as the
1570 "mailbox time last read" when there is a quota error for a maildir
1573 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1575 35. Added +ignore_unknown as a special item in host lists.
1577 36. Code for decoding IPv6 addresses in host lists is now included, even if
1578 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1579 address was recognized as an IP address, but was then not correctly decoded
1580 into binary, causing unexpected and incorrect effects when compared with
1587 1. Very minor rewording of debugging text in manualroute to say "list of
1588 hosts" instead of "hostlist".
1590 2. If verify=header_syntax was set, and a header line with an unqualified
1591 address (no domain) and a large number of spaces between the end of the
1592 name and the colon was received, the reception process suffered a buffer
1593 overflow, and (when I tested it) crashed. This was caused by some obsolete
1594 code that should have been removed. The fix is to remove it!
1596 3. When running in the test harness, delay a bit after writing a bounce
1597 message to get a bit more predictability in the log output.
1599 4. Added a call to search_tidyup() just before forking a reception process. In
1600 theory, someone could use a lookup in the expansion of smtp_accept_max_
1601 per_host which, without the tidyup, could leave open a database connection.
1603 5. Added the variables $recipient_data and $sender_data which get set from a
1604 lookup success in an ACL "recipients" or "senders" condition, or a router
1605 "senders" option, similar to $domain_data and $local_part_data.
1607 6. Moved the writing of debug_print from before to after the "senders" test
1610 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1611 problems for message scanning, either using a data ACL, or using
1612 local_scan() because the Received: header was not generated till after they
1613 were called (in order to set the time as the time of reception completion).
1614 I have revised the way this works. The header is now generated after the
1615 body is received, but before the ACL or local_scan() are called. After they
1616 are run, the timestamp in the header is updated.
1622 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1623 before starting a queue runner without re-exec. This happened only when
1624 deliver_drop_privilege was set or when the Exim user was set to root. The
1625 effect of the bug was that timeouts during subsequent deliveries caused
1626 crashes instead of being properly handled. The handler is now left at its
1627 default (and expected) setting.
1629 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1630 message, again when deliver_drop_privilege is set or Exim is run as root.
1631 The bug described in (1) was not present in this case, but the tidying up
1632 of the other signals was missing. I have made the two cases consistent.
1634 3. The ignore_target_hosts setting on a manualroute router was being ignored
1635 for hosts that were looked up using the /MX notation.
1637 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1640 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1641 operated on the sender address. After changing the $sender_address to <>
1642 for the sender address verify, Exim was re-instated it as the original
1643 (before rewriting) address, but remembering that it had rewritten it, so it
1644 wasn't rewriting it again. This bug also had the effect of breaking the
1645 sender address verification caching when the sender address was rewritten.
1647 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1648 This has been changed so that if the ip literal address matches
1649 ignore_target_hosts, the router declines.
1651 7. Added expansion conditions match_domain, match_address, and match_local_
1652 part (NOT match_host).
1654 8. The placeholder for the Received: header didn't have a length field set.
1656 9. Added code to Exim itself and to exim_lock to test for a specific race
1657 condition that could lead to file corruption when using MBX delivery. The
1658 issue is with the lockfile that is created in /tmp. If this file is removed
1659 after a process has opened it but before that process has acquired a lock,
1660 there is the potential for a second process to recreate the file and also
1661 acquire a lock. This could lead to two Exim processes writing to the file
1662 at the same time. The added code performs the same test as UW imapd; it
1663 checks after acquiring the lock that its file descriptor still refers to
1664 the same named file.
1666 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1667 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1670 11. Added the smtp_active_hostname option. If used, this will typically be made
1671 to depend on the incoming interface address. Because $interface_address is
1672 not set up until the daemon has forked a reception process, error responses
1673 that can happen earlier (such as "too many connections") no longer contain
1676 12. If an expansion in a condition on a "warn" statement fails because a lookup
1677 defers, the "warn" statement is abandoned, and the next ACL statement is
1678 processed. Previously this caused the whole ACL to be aborted.
1680 13. Added the iplsearch lookup type.
1682 14. Added ident_timeout as a log selector.
1684 15. Added tls_certificate_verified as a log selector.
1686 16. Added a global option tls_require_ciphers (compare the smtp transport
1687 option of the same name). This controls incoming TLS connections.
1689 17. I finally figured out how to make tls_require_ciphers do a similar thing
1690 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1691 before starting the TLS session.
1693 18. Tabs are now shown as \t in -bP output.
1695 19. If the log selector return_path_on_delivery was set, Exim crashed when
1696 bouncing a message because it had too many Received: header lines.
1698 20. If two routers both had headers_remove settings, and the first one included
1699 a superfluous trailing colon, the final name in the first list and the
1700 first name in the second list were incorrectly joined into one item (with a
1701 colon in the middle).
1707 1. Added -C and -D options to the exinext utility, mainly to make it easier
1708 to include in the automated testing, but these could be helpful when
1709 multiple configurations are in use.
1711 2. The exinext utility was not formatting the output nicely when there was
1712 an alternate port involved in the retry record key, nor when there was a
1713 message id as well (for retries that were specific to a specific message
1714 and a specific host). It was also confused by IPv6 addresses, because of
1715 the additional colons they contain. I have fixed the IPv4 problem, and
1716 patched it up to do a reasonable job for IPv6.
1718 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1719 delivery, the log line now contains "pipelined" if PIPELINING was used.
1721 4. An SMTP transport process used to panic and die if the bind() call to set
1722 an explicit outgoing interface failed. This has been changed; it is now
1723 treated in the same way as a connect() failure.
1725 5. A reference to $sender_host_name in the part of a conditional expansion
1726 that was being skipped was still causing a DNS lookup. This no longer
1729 6. The def: expansion condition was not recognizing references to header lines
1730 that used bh_ and bheader_.
1732 7. Added the _cache feature to named lists.
1734 8. The code for checking quota_filecount in the appendfile transport was
1735 allowing one more file than it should have been.
1737 9. For compatibility with Sendmail, the command line option
1745 and sets the incoming protocol and host name (for trusted callers). The
1746 host name and its colon can be omitted when only the protocol is to be set.
1747 Note the Exim already has two private options, -pd and -ps, that refer to
1748 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1749 "s", but I don't think that's a major issue.
1751 10. A number of refactoring changes to the code, none of which should affect
1754 (a) The number of logging options was getting close to filling up the
1755 32-bit word that was used as a bit map. I have split them into two classes:
1756 those that are passed in the argument to log_write(), and those that are
1757 only ever tested independently outside of that function. These are now in
1758 separate 32-bit words, so there is plenty of room for expansion again.
1759 There is no change in the user interface or the logging behaviour.
1761 (b) When building, for example, log lines, the code previously used a
1762 macro that called string_cat() twice, in order to add two strings. This is
1763 not really sufficiently general. Furthermore, there was one instance where
1764 it was actually wrong because one of the argument was used twice, and in
1765 one call a function was used. (As it happened, calling the function twice
1766 did not affect the overall behaviour.) The macro has been replaced by a
1767 function that can join an arbitrary number of extra strings onto a growing
1770 (c) The code for expansion conditions now uses a table and a binary chop
1771 instead of a serial search (which was left over from when there were very
1772 few conditions). Also, it now recognizes conditions like "pam" even when
1773 the relevant support is not compiled in: a suitably worded error message is
1774 given if an attempt is made to use such a condition.
1776 11. Added ${time_interval:xxxxx}.
1778 12. A bug was causing one of the ddress fields not to be passed back correctly
1779 from remote delivery subprocesses. The field in question was not being
1780 subsequently used, so this caused to problems in practice.
1782 13. Added new log selectors queue_time and deliver_time.
1784 14. Might have fixed a bug in maildirsizefile handling that threw up
1785 "unexpected character" debug warnings, and recalculated the data
1786 unnecessarily. In any case, I expanded the warning message to give more
1789 15. Added the message "Restricted characters in address" to the statements in
1790 the default ACL that block characters like @ and % in local parts.
1792 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1793 Three changes have been made:
1795 (a) There was a serious bug; a negative response to MAIL caused the whole
1796 recipient domain to be cached as invalid, thereby blocking all messages
1797 to all local parts at the same domain, from all senders. This bug has
1798 been fixed. The domain is no longer cached after a negative response to
1799 MAIL if the sender used is not empty.
1801 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1804 (c) A new callout option, "use_sender" has been added for people who want
1805 the modified behaviour.
1811 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1814 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1815 indeed breaks things for older releases.
1817 3. Added additional logging to the case where there is a problem reading data
1818 from a filter that is running in a subprocess using a pipe, in order to
1819 try to track down a specific problem.
1821 4. Testing facility fudge: when running in the test harness and attempting
1822 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1823 getting "No route to host". Convert this to a timeout.
1825 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1828 6. Some OS don't have socklen_t but use size_t instead. This affects the
1829 fifth argument of getsockopt() amongst other things. This is now
1830 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1831 can be set for individual OS. I have set it for SunOS5, OSF1, and
1832 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1833 some earlier ones do not.
1835 7. Change 4.30/15 was not doing the test caselessly.
1837 8. The standard form for an IPv6 address literal was being rejected by address
1838 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1839 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1840 this, as well as the form without the "IPv6" on the front (but only when
1841 address literals are enabled, of course).
1843 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1845 10. Exim crashed if a message with an empty sender address specified by -f
1846 encountered a router with an errors_to setting. This could be provoked only
1847 by a command such as
1851 where an empty string was supplied; "<>" did not hit this bug.
1853 11. Installed PCRE release 4.5.
1855 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1856 remained set. It is now erased.
1858 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1859 times from message ids (which are base 36 rather than the normal 62).
1861 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1862 were being counted as actual protocol errors, and logged if the log
1863 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1864 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1865 and DATA following a set of rejected RCPTs do not count as protocol errors.
1866 In other words, Exim assumes they were pipelined, though this may not
1867 actually be the case. Of course, in all cases the client gets an
1868 appropriate error code.
1870 15. If a lookup fails in an ACL condition, a message about the failure may
1871 be available; it is used if testing the ACL cannot continue, because most
1872 such messages specify what the cause of the deferral is. However, some
1873 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1874 that caused an old message to be retained and used if a later statement
1875 caused a defer, replacing the real cause of the deferral.
1877 16. If an IP address had so many PTR records that the DNS lookup buffer
1878 was not large enough to hold them, Exim could crash while trying to process
1879 the truncated data. It now detects and logs this case.
1881 17. Further to 4.21/58, another change has been made: if (and only if) the
1882 first line of a message (the first header line) ends with CRLF, a bare LF
1883 in a subsequent header line has a space inserted after it, so as not to
1884 terminate the header.
1886 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1887 unnecessarily, used atoi() instead of strtol(), and didn't check the
1888 termination when getting file sizes from file names by regex.
1890 19. Completely re-implemented the support for maildirsize files, in the light
1891 of a number of problems with the previous contributed implementation
1892 (4.30/29). In particular:
1894 . If the quota is zero, the maildirsize file is maintained, but no quota is
1897 . If the maildir directory does not exist, it is created before any attempt
1898 to write a maildirsize file.
1900 . The quota value in the file is just a cache; if the quota is changed in
1901 the transport, the new value overrides.
1903 . A regular expression is available for excluding directories from the
1906 20. The autoreply transport checks the characters in options that define the
1907 message's headers; it allows continued headers, but it was checking with
1908 isspace() after an embedded newline instead of explicitly looking for a
1911 21. If all the "regular" hosts to which an address was routed had passed their
1912 expiry times, and had not reached their retry times, the address was
1913 bounced, even if fallback hosts were defined. Now Exim should go on to try
1916 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1917 equivalent code in the SMTP transport. Some hosts send humungous responses
1918 to HELO/EHLO, more than 1024 it seems.
1920 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1921 gives compiler warnings in some environments. I've now done it "properly",
1924 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1925 (because of problems with the built-in one) was declared to return uschar *
1926 instead of char *, causing compiler failure.
1928 25. Fixed a file descriptor leak when processing alias/forward files.
1930 26. Fixed a minor format string issue in dbfn.c.
1932 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1934 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1935 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1938 29. When a sender address is verified, it is cached, to save repeating the test
1939 when there is more than one recipient in a message. However, when the
1940 verification involves a callout, it is possible for different callout
1941 options to be set for different recipients. It is too complicated to keep
1942 track of this in the cache, so now Exim always runs a verification when a
1943 callout is required, relying on the callout cache for the optimization.
1944 The overhead is duplication of the address routing, but this should not be
1947 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1948 to be verified with callout=postmaster, and the main callout worked but the
1949 postmaster check failed, the verification correctly failed. However, if a
1950 subsequent RCPT command asked for sender verification *without* the
1951 postmaster check, incorrect caching caused this verification also to fail,
1954 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1955 it was not caching the DNS options (qualify_single, search_parents) that
1956 were used when the lookup failed. A subsequent lookup with different
1957 options therefore always gave the same answer, though there were cases
1958 where it should not have. (Example: a "domains = !$mx_any" option on a
1959 dnslookup router: the "domains" option is always processed without any
1960 widening, but the router might have qualify_single set.) Now Exim uses the
1961 cached value only when the same options are set.
1963 32. Added John Jetmore's "exipick" utility to the distribution.
1965 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1966 than a timeout (e.g. a certificate is required, and is not provided), an
1967 Exim server now closes the connection immediately. Previously it waited for
1968 the client to close - but if the client is SSL, it seems that they each
1969 wait for each other, leading to a delay before one of them times out.
1971 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1972 maintained 0.8.x compatibility because I don't think many are using it, and
1973 it is clearly obsolete.
1975 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1978 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1979 client certificate was expired. A simple patch fixes this, though I don't
1980 understand the full logic of why the verify callback is called multiple
1983 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1984 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1985 which causes problems with some clients (such as the Certicom SSL Plus
1986 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1987 disables the coutermeasure allowing Eudora to connect."
1989 38. Exim was not checking that a write() to a log file succeeded. This could
1990 lead to Bad Things if a log got too big, in particular if it hit a file
1991 size limit. Exim now panics and dies if it cannot write to a log file, just
1992 as it does if it cannot open a log file.
1994 39. Modified OS/Makefile-Linux so that it now contains
1996 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1998 The two -D definitions ensure that Exim is compiled with large file
1999 support, which makes it possible to handle log files that are bigger than
2002 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
2003 instance) a domain was checked against a named list that involved a lookup,
2004 causing $domain_data to be set, then another domain was checked against the
2005 same list, then the first domain was re-checked, the value of $domain_data
2006 after the final check could be wrong. In particular, if the second check
2007 failed, it could be set empty. This bug probably also applied to
2010 41. The strip_trailing_dot option was not being applied to the address given
2011 with the -f command-line option.
2013 42. The code for reading a message's header from the spool was incrementing
2014 $received_count, but never initializing it. This meant that the value was
2015 incorrect (doubled) while delivering a message in the same process in which
2016 it was received. In the most common configuration of Exim, this never
2017 happens - a fresh exec is done - but it can happen when
2018 deliver_drop_privilege is set.
2020 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2021 it now includes up to 150 characters of the unexpected data in the log
2024 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2025 and building data strings. The size of both of these buffers was 10 000
2026 bytes - far larger than anybody would *ever* want, thought I. Needless to
2027 say, somebody hit the limit. I have increased the maximum line length to
2028 20 000 and the maximum data length of concatenated lines to 100 000. I have
2029 also fixed two bugs, because there was no checking on these buffers. Tsk,
2030 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2031 buffer is too small.
2033 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2034 lsearch lookups. Now it does.
2036 46. When parsing a route_list item in a manualroute router, a fixed-length
2037 buffer was used for the list of hosts. I made this 1024 bytes long,
2038 thinking that nobody would ever have a list of hosts that long. Wrong.
2039 Somebody had a whole pile of complicated expansion conditions, and the
2040 string was silently truncated, leading to an expansion error. It turns out
2041 that it is easier to change to an unlimited length (owing to other changes
2042 that have happened since this code was originally written) than to build
2043 structure for giving a limitation error. The length of the item that
2044 expands into the list of hosts is now unlimited.
2046 47. The lsearch lookup could not handle data where the length of text line was
2047 more than 4095 characters. Such lines were truncated, leading to shortened
2048 data being returned. It should now handle lines of any length.
2050 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2051 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2054 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2055 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2056 no longer happens in comment lines. A list of replacements is now placed
2057 at the head of all of the source files, except those whose only change is
2058 to replace PERL_COMMAND in the very first #! line.
2060 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2061 messages on the queue, with a bottom-up merge sort, using code contributed
2062 by Michael Haardt. This should make operations like -bp somewhat faster on
2063 large queues. It won't affect queue runners, except when queue_run_in_order
2066 51. Installed eximstats 1.31 in the distribution.
2068 52. Added support for SRV lookups to the dnslookup router.
2070 53. If an ACL referred to $message_body or $message_body_end, the value was not
2071 reset for any messages that followed in the same SMTP session.
2073 54. The store-handling optimization for building very long strings was not
2074 differentiating between the different store pools. I don't think this
2075 actually made any difference in practice, but I've tidied it.
2077 55. While running the routers to verify a sender address, $sender_address
2078 was still set to the sender address. This is wrong, because when routing to
2079 send a bounce to the sender, it would be empty. Therefore, I have changed
2080 it so that, while verifying a sender address, $sender_address is set to <>.
2081 (There is no change to what happens when verifying a recipient address.)
2083 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2084 target A or AAAA records (if not already returned) without resetting the
2085 qualify_single or search_parents options of the DNS resolver. These are
2086 inappropriate in this case because the targets of MX and SRV records must
2087 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2088 target that, when qualified, matched something in the local domain. These
2089 two options are now turned off when doing these lookups.
2091 57. It seems that at least some releases of Reiserfs (which does not have the
2092 concept of a fixed number of inodes) returns zero and not -1 for the
2093 number of available inodes. This interacted badly with check_spool_inodes,
2094 which assumed that -1 was the "no such thing" setting. What I have done is
2095 to check that the total number of inodes is greater than zero before doing
2096 the test of how many are available.
2098 58. When a "warn" ACL statement has a log_message modifier, the message is
2099 remembered, and not repeated. This is to avoid a lot of repetition when a
2100 message has many recipients that cause the same warning to be written.
2101 Howewer, Exim was preserving the list of already written lines for an
2102 entire SMTP session, which doesn't seem right. The memory is now reset if a
2103 new message is started.
2105 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2106 debugging output unless log_rewrite was also set.
2108 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2109 of (int)(handle) when we know that handle contains (void *)(-1).
2111 61. The Exim daemon panic-logs an error return when it closes the incoming
2112 connection. However "connection reset by peer" seems to be common, and
2113 isn't really an error worthy of noting specially, so that particular error
2116 62. When Exim is trying to find all the local interfaces, it used to panic and
2117 die if the ioctl to get the interface flags failed. However, it seems that
2118 on at least one OS (Solaris 9) it is possible to have an interface that is
2119 included in the list of interfaces, but for which you get a failure error
2120 for this call. This happens when the interface is not "plumbed" into a
2121 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2122 failure of the "get flags" call assumes that the interface is down.
2124 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2125 makes life easier for people who are doing arithmetic on fields extracted
2126 from dates, where you often get leading zeros that should not be
2127 interpreted as octal.
2129 64. Added qualify_domain to the redirect router, to override the global
2132 65. If a pathologically long header line contained very many addresses (the
2133 report of this problem mentioned 10 000) and each of them was rewritten,
2134 Exim could use up a very large amount of memory. (It kept on making new
2135 copies of the header line as it rewrote, and never released the old ones.)
2136 At the expense of a bit more processing, the header rewriting function has
2137 been changed so that it no longer eats memory in this way.
2139 66. The generation of the Received: header has been moved from the time that a
2140 message starts to be received, to the time that it finishes. The timestamp
2141 in the Received: header should now be very close to that of the <= log
2142 line. There are two side-effects of this change:
2144 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2145 logged header lines no longer include the local Received: line, because
2146 it has not yet been created. The same applies to a copy of the message
2147 that is returned to a non-SMTP sender when a message is rejected.
2149 (b) When a filter file is tested using -bf, no additional Received: header
2150 is added to the test message. After some thought, I decided that this
2153 This change does not affect the value of $received_for. It is still set
2154 after address rewriting, but before local_scan() is called.
2156 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2158 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2159 gave an unhelpful panic error message, and a defer error. I have managed to
2160 change this behaviour so that it now rejects any supplied certificate,
2161 which seems right, as the list of acceptable certificates is empty.
2163 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2164 gave an unhelpful defer error. I have not managed to make this reject any
2165 supplied certificates, but the error message it gives is "no certificate
2166 supplied", which is not helpful.
2168 70. exigrep's output now also includes lines that are not associated with any
2169 message, but which match the given pattern. Implemented by a patch from
2170 Martin Sluka, which also tidied up the Perl a bit.
2172 71. Recipient callout verification, like sender verification, was using <> in
2173 the MAIL FROM command. This isn't really the right thing, since the actual
2174 sender may affect whether the remote host accepts the recipient or not. I
2175 have changed it to use the actual sender in the callout; this means that
2176 the cache record is now keyed on a recipient/sender pair, not just the
2177 recipient address. There doesn't seem to be a real danger of callout loops,
2178 since a callout by the remote host to check the sender would use <>.
2179 [SEE ABOVE: changed after hitting problems.]
2181 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2182 temporary errors. However, in the case of such a code being given after
2183 the end of a data transmission (i.e. after ".") Exim was failing to write
2184 a retry record for the message. (Yes, there was some broken host that was
2185 actually sending 8xx at this point.)
2187 73. An unknown lookup type in a host list could cause Exim to panic-die when
2188 the list was checked. (An example that provoked this was putting <; in the
2189 middle of a list instead of at the start.) If this happened during a DATA
2190 ACL check, a -D file could be left lying around. This kind of configuration
2191 error no longer causes Exim to die; instead it causes a defer errror. The
2192 incident is still logged to the main and panic logs.
2194 74. Buglet left over from Exim 3 conversion. The message "too many messages
2195 in one connection" was written to the rejectlog but not the mainlog, except
2196 when address rewriting (yes!) was being logged.
2198 75. Added write_rejectlog option.
2200 76. When a system filter was run not as root (that is, when system_filter_user
2201 was set), the values of the $n variables were not being returned to the
2202 main process; thus, they were not subsequently available in the $sn
2205 77. Added +return_path_on_delivery log selector.
2207 78. A connection timeout was being treated differently from recipients deferred
2208 when testing hosts_max_try with a message that was older than the host's
2209 retry timeout. (The host should not be counted, thus allowing all hosts to
2210 be tried at least once before bouncing.) This may have been the cause of an
2211 occasionally reported bug whereby a message would remain on the queue
2212 longer than the retry timeout, but would be bounced if a delivery was
2213 forced. I say "may" because I never totally pinned down the problem;
2214 setting up timeout/retry tests is difficult. See also the next item.
2216 79. The ultimate address timeout was not being applied to errors that involved
2217 a combination of host plus message (for example, a timeout on a MAIL
2218 command). When an address resolved to a number of possible hosts, and they
2219 were not all tried for each delivery (e.g. because of hosts_max_try), a
2220 message could remain on the queue longer than the retry timeout.
2222 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2225 81. Fixed an obscure SMTP outgoing bug which required at least the following
2226 conditions: (a) there was another message waiting for the same server;
2227 (b) the server returned 5xx to all RCPT commands in the first message so
2228 that the message was not completed; (c) the server dropped the connection
2229 or gave a negative response to the RSET that Exim sends to abort the
2230 transaction. The observed case was a dropped connection after DATA that had
2231 been sent in pipelining mode. That is, the server had advertised PIPELINING
2232 but was not implementing it correctly. The effect of the bug was incorrect
2233 behaviour, such as trying another host, and this could lead to a crash.
2239 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2240 and daemon.c were passed as pointers to ints; they should have been
2241 pointers to socklen_t variables (which are typically unsigned ints).
2243 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2246 3. Fixed a really odd bug that affected only the testing scheme; patching a
2247 certain fixed string in the binary changed the value of another string that
2248 happened to be identical to the end of the original first string.
2250 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2251 name", it returns that address as the IP address. On some operating
2252 systems (e.g. Solaris), it also passes back the IP address string as the
2253 "host name". However, on others (e.g. Linux), it passes back an empty
2254 string. Exim wasn't checking for this, and was changing the host name to an
2255 empty string, assuming it had been canonicized.
2257 5. Although rare, it is permitted to have more than one PTR record for a given
2258 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2259 all the names associated with an address, because they do in Solaris.
2260 However, it seems that they do not in Linux for data that comes from the
2261 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2262 I found this out when I moved to a new Linux workstation and tried to run
2263 the Exim test suite.
2265 To get round this problem I have changed the code so that it now does its
2266 own call to the DNS to look up PTR records when searching for a host name.
2267 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2268 addresses that are only in /etc/hosts are still found.
2270 This behaviour is, however, controlled by an option called host_lookup_
2271 order, which defaults to "bydns:byaddr". If people want to use the other
2272 order, or indeed, just use one or the other means of lookup, they can
2273 specify it in this variable.
2275 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2276 some operating systems, this comes back from gethostbyaddr() as an empty
2277 string, and this is what Exim used to test for. However, it seems that in
2278 other systems, "." is yielded. Exim now tests for this case too.
2280 7. The values of check_spool_space and check_log_space are now held internally
2281 as a number of kilobytes instead of an absolute number of bytes. If a
2282 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2283 kilobyte. This means that much larger values can be stored.
2285 8. Exim monitor: an attempt to get the action menu when not actually pointing
2286 at a message produces an empty menu entitled "No message selected". This
2287 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2288 no entries in it ("Shell widget menu has zero width and/or height"). So I
2289 have added a single, blank menu entry in this case.
2291 9. Added ${quote_local_part.
2293 10. MIME decoding is now applied to the contents of Subject: header lines when
2296 11. Now that a reference to $sender_host_address automatically causes a reverse
2297 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2298 host lookup before query-style lookups in lists that might use this
2299 variable. This has therefore been abolished, and the "net-" prefix is no
2300 longer necessary for query-style lookups.
2302 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2303 have been a typo for LFLAGS, so it has been changed.
2305 13. The install script calls Exim with "-C /dev/null" in order to find the
2306 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2307 to be output. Howeve, since Exim outputs its version number before the
2308 error, it didn't break the script. It just looked ugly. I fixed this by
2309 always allowing "-C /dev/null" if the caller is root.
2311 14. Ignore overlarge ACL variable number when reading spool file - insurance
2312 against a later release with more variables having written the file.
2314 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2315 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2316 this, as well as the form without the "IPv6" on the front.
2318 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2319 OS/Makefile-Darwin file.
2321 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2322 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2324 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2325 variable. However, some OS use ints for inodes. I've added cast to long int
2326 to get rid of the compiler warning.
2328 19. I had forgotten to lock out "/../" in configuration file names when
2329 ALT_CONFIG_PREFIX was set.
2331 20. Routers used for verification do not need to specify transports. However,
2332 if such a router generated a host list, and callout was configured, Exim
2333 crashed, because it could not find a port number from the (non-existent)
2334 transport. It now assumes port 25 in this circumstance.
2336 21. Added the -t option to exigrep.
2338 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2339 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2340 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2341 them is set, LOOKUP_LSEARCH is forced.
2343 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2345 24. Added sender and host information to the "rejected by local_scan()" log
2346 line; previously there was no indication of these.
2348 25. Added .include_if_exists.
2350 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2351 Linux. It turns out that not all file systems support this. Apparently some
2352 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2353 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2354 not supported on the file descriptor, is now ignored when Exim is trying to
2355 sync a directory. This applies only to Linux.
2357 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2359 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2360 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2363 29. Added support for maildirsize files from supplied patch (modified a bit).
2365 30. The use of :fail: followed by an empty string could lead Exim to respond to
2366 sender verification failures with (e.g.):
2368 550 Verification failed for <xxx>
2369 550 Sender verify failed
2371 where the first response line was missing the '-' that indicates it is not
2372 the final line of the response.
2374 31. The loop for finding the name of the user that called Exim had a hardwired
2375 limit of 10; it now uses the value of finduser_retries, which is used for
2376 all other user lookups.
2378 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2381 33. Exim was neglecting to zero errno before one call of strtol() when
2382 expanding a string and expecting an integer value. On some systems this
2383 resulted in spurious "integer overflow" errors. Also, it was casting the
2384 result into an int without checking.
2386 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2387 did not work. The code looks as if it has *never* worked, though it appears
2388 to have been documented since at least releast 1.62. I have made it work.
2390 35. The "timeout_DNS" error in retry rules, also documented since at least
2391 1.62, also never worked. As it isn't clear exactly what this means, and
2392 clearly it isn't a major issue, I have abolished the feature by treating it
2393 as "timeout", and writing a warning to the main and panic logs.
2395 36. The display of retry rules for -brt wasn't always showing the error code
2398 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2399 timeout_connect_A, timeout_connect_MX.
2401 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2402 to the empty sender.
2404 39. The daemon was not analysing the content of -oX till after it had closed
2405 stderr and disconnected from the controlling terminal. This meant that any
2406 syntax errors were only noted on the panic log, and the return code from
2407 the command was 0. By re-arranging the code a little, I've made the
2408 decoding happen first, so such errors now appear on stderr, and the return
2409 code is 1. However, the actual setting up of the sockets still happens in
2410 the disconnected process, so errors there are still only recorded on the
2413 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2414 connections (as happens on some IP stacks) was logged at start up time as
2415 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2416 it from "IPv6 and IPv4", which means that two separate sockets are being
2419 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2420 says whether AF_INET or AF_INET6 was passed as an argument.
2422 42. Exiwhat output was messed up when time zones were included in log
2425 43. Exiwhat now gives more information about the daemon's listening ports,
2426 and whether -tls-on-connect was used.
2428 44. The "port" option of the smtp transport is now expanded.
2430 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2431 silently ignored. Now an error message is written to the main and panic
2434 46. There's a new ACL modifier called "logwrite" which writes to a log file
2435 as soon as it is encountered.
2437 47. Added $local_user_uid and $local_user_gid at routing time.
2439 48. Exim crashed when trying to verify a sender address that was being
2442 49. Exim was recognizing only a space character after ".include". It now also
2443 recognizes a tab character.
2445 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2446 extracting the relevant information from the specification. The man page no
2447 longer contains scrambled data for the -d option, and I've added a section
2448 at the front about calling Exim under different names.
2450 51. Added "extra_headers" argument to the "mail" command in filter files.
2452 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2455 53. Installed eximstats 1.29.
2457 54. Added transport_filter_timeout as a generic transport option.
2459 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2460 Cc: header lines. This was required by RFC 822, but it not required by RFC
2463 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2464 incoming messages that did not have them. Now it does so only if the
2465 message originates locally, that is, if there is no associated remote host
2466 address. When Resent- header lines are present, this applies to the Resent-
2467 lines rather than the non-Resent- lines.
2469 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2470 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2472 58. Messages for configuration errors now include the name of the main
2473 configuration file - useful now that there may be more than one file in a
2474 list (.included file names were always shown).
2476 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2477 for those rare installations that do not start the daemon as root or run it
2478 setuid root. I've cut out the call to initgroups() if the daemon is not
2481 60. The Exim user and group can now be bound into the binary as text strings
2482 that are looked up at the start of Exim's processing.
2484 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2486 62. Added $mailstore_basename variable.
2488 63. Installed patch to sieve.c from Michael Haardt.
2490 64. When Exim failed to open the panic log after failing to open the main log,
2491 the original message it was trying to log was written to stderr and debug
2492 output, but if they were not available (the usual case in production), it
2493 was lost. Now it is written to syslog before the two lines that record the
2494 failures to open the logs.
2496 65. Users' Exim filters run in subprocesses under the user's uid. It is
2497 possible for a "deliver" command or an alias in a "personal" command to
2498 provoke an address rewrite. If logging of address rewriting is configured,
2499 this fails because the process is not running as root or exim. There may be
2500 a better way of dealing with this, but for the moment (because 4.30 needs
2501 to be released), I have disabled address rewrite logging when running a
2502 filter in a non-root, non-exim process.
2508 1. The buildconfig auxiliary program wasn't quoting the value set for
2509 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2510 not defined. This bug was masked in 4.22 by the effect that was fixed in
2513 2. Some messages that were rejected after a message id was allocated were
2514 shown as "incomplete" by exigrep. It no longer does this for messages that
2515 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2517 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2518 have allow_domain_literals set, the ID did not get logged in the <= line.
2519 Domain literals are now always recognized in Message-ID: header lines.
2521 4. The first argument for a ${extract expansion item is the key name or field
2522 number. Leading and trailing spaces in this item were not being ignored,
2523 causing some misleading effects.
2525 5. When deliver_drop_privilege was set, single queue runner processes started
2526 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2527 same command in the process it spins off) were not dropping privilege.
2529 6. When the daemon running as "exim" started a queue runner, it always
2530 re-executed Exim in the spun-off process. This is a waste of effort when
2531 deliver_drop_privilege is set. The new process now just calls the
2532 queue-runner function directly.
2538 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2541 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2542 ignored. Though the use of -oP was forcing the writing of a pid file, it
2543 was always written to the default place.
2545 3. If the message "no IP address found for host xxxx" is generated during
2546 incoming verification, it is now followed by identification of the incoming
2547 connection (so you can more easily find what provoked it).
2549 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2551 5. Added some features to "harden" Exim a bit more against certain attacks:
2553 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2554 be put in Local/Makefile. This is like the never_users runtime option,
2555 but it cannot be overridden. The default setting is "root".
2557 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2558 prefix string with which any file named in a -C command line option
2561 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2562 is retained for -C and -D only if the caller of Exim is root. Without
2563 it, the exim user may also use -C and -D and retain privilege.
2565 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2566 command line option is disabled.
2568 6. Macro names set by the -D option must start with an upper case letter, just
2569 like macro names defined in the configuration file.
2571 7. Added "dereference=" facility to LDAP.
2573 8. Two instances of the typo "uknown" in the source files are fixed.
2575 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2576 the Configure-Makefile script screwed up while processing it.
2578 10. Incorporated PCRE 4.4.
2580 11. The SMTP synchronization check was not operating right at the start of an
2581 SMTP session. For example, it could not catch a HELO sent before the client
2582 waited for the greeting. There is now a check for outstanding input at the
2583 point when the greeting is written. Because of the duplex, asynchronous
2584 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2585 way, but not yet received, when the check is performed.
2587 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2588 on TCP/IP sockets, because this apparently causes some broken clients to
2591 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2592 unchanged) from the Cygwin maintainer.
2594 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2595 was supported instead of testing for mbx. Effectively a typo.
2597 15. The spa authenticator server code was not checking that the input it
2598 received was valid base64.
2600 16. The debug output line for the "set" modifier in ACLs was not showing the
2601 name of the variable that was being set.
2603 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2605 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2606 lookup on its own. Something else was needed to trigger the lookup. For
2607 example, a match in host_lookup or the need for a host name in a host list.
2608 Now, if $sender_host_name is referenced and the host name has not yet been
2609 looked up, a lookup is performed. If the lookup fails, the variable remains
2610 empty, and $host_lookup_failed is set to "1".
2612 19. Added "eqi" as a case-independent comparison operator.
2614 20. The saslauthd authentication condition could segfault if neither service
2615 nor realm was specified.
2617 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2618 error message that was logged was misleading, and incoming SMTP
2619 connections were dropped. The message is now more accurate, and temporary
2620 errors are given to SMTP connections.
2622 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2623 (except RSET) with a 421 error, until QUIT is received. However, it was
2624 failing to send a response to QUIT.
2626 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2627 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2628 ACL is now run after the helo_try_verify_hosts code.
2630 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2631 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2632 case-independent, so other case variants are also recognized). Apparently
2633 some systems use these upper case variants.
2635 25. If more than two messages were waiting for the same host, and a transport
2636 filter was specified for the transport, Exim sent two messages over the
2637 same TCP/IP connection, and then failed with "socket operation on non-
2638 socket" when it tried to send the third.
2640 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2642 27. The extern definition of crypt16() in expand.c was not being excluded when
2643 the OS had its own crypt16() function.
2645 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2646 as a preferred synonym for return_size_limit, both as an option and as an
2649 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2651 30. Changed the default configuration ACL to relax the local part checking rule
2652 for addresses that are not in any local domains. For these addresses,
2653 slashes and pipe symbols are allowed within local parts, but the sequence
2654 /../ is explicitly forbidden.
2656 31. SPA server authentication was not clearing the challenge buffer before
2659 32. log_message in a "warn" ACL statement was writing to the reject log as
2660 well as to the main log, which contradicts the documentation and doesn't
2661 seem right (because no rejection is happening). So I have stopped it.
2663 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2664 However, I am unable to do any testing of this.
2666 34. Fixed an infelicity in the appendfile transport. When checking directories
2667 for a mailbox, to see if any needed to be created, it was accidentally
2668 using path names with one or more superfluous leading slashes; tracing
2669 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2671 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2672 contents are added to the log line that is written for every discarded
2673 recipient. (Previously a log_message setting was ignored.)
2675 36. The ${quote: operator now quotes the string if it is empty.
2677 37. The install script runs exim in order to find its version number. If for
2678 some reason other than non-existence or emptiness, which it checks, it
2679 could not run './exim', it was installing it with an empty version number,
2680 i.e. as "exim-". This error state is now caught, and the installation is
2683 38. An argument was missing from the function that creates an error message
2684 when Exim fails to connect to the socket for saslauthd authentication.
2685 This could cause Exim to crash, or give a corrupted message.
2687 39. Added isip, isip4, and isip6 to ${if conditions.
2689 40. The ACL variables $acl_xx are now saved with the message, and can be
2690 accessed later in routers, transports, and filters.
2692 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2693 strings in the file are not string-expanded.
2695 42. If a MAIL command specified a SIZE value that was too large to fit into an
2696 int variable, the check against message_size_limit failed. Such values are
2697 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2698 day this will have to be increased, but I don't think I want to be around
2699 when emails are that large.
2706 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2707 iconv() is not standard in FreeBSD.
2709 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2710 IPv6 enabled. The observed symptom was a segmentation fault on return from
2711 the function os_common_find_running_interfaces() in src/os.c.
2713 3. In the check_special_case() function in daemon.c I had used "errno" as an
2714 argument name, which causes warnings on some systems. This was basically a
2715 typo, since it was named "eno" in the comments!
2717 4. The code that waits for the clock to tick (at a resolution of some fraction
2718 of a second) so as to ensure message-id uniqueness was always waiting for
2719 at least one whole tick, when it could have waited for less. [This is
2720 almost certainly not relevant at current processor speeds, where it is
2721 unlikely to ever wait at all. But we try to future-proof.]
2723 5. The function that sleeps for a time interval that includes fractions of a
2724 second contained a race. It did not block SIGALRM between setting the
2725 timer, and suspending (a couple of lines later). If the interval was short
2726 and the sigsuspend() was delayed until after it had expired, the suspension
2727 never ended. On busy systems this could lead to processes getting stuck for
2730 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2731 process, before it forks any delivery processes. The open lookup caching
2732 mechanism meant that the open file or database connection was passed into
2733 the delivery process. The problem was that delivery processes always tidy
2734 up cached lookup data. This could cause a problem for the next delivery
2735 process started by the queue runner, because the external queue runner
2736 process does not know about the closure. So the next delivery process
2737 still has data in the lookup cache. In the case of a file lookup, there was
2738 no problem because closing a file descriptor in a subprocess doesn't affect
2739 the parent. However, if the lookup was caching a connection to a database,
2740 the connection was closed, and the second delivery process was likely to
2741 see errors such as "PGSQL: query failed: server closed the connection
2742 unexpectedly". The problem has been fixed by closing all cached lookups
2743 in a queue runner before running a delivery process.
2745 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2746 seem to have the "const" qualifier which it has on other OS. I've
2749 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2750 *of the same type* (client or server) with the same public name that an
2751 error should be diagnosed.
2753 9. When Exim looked up a host name for an IP address, but failed to find the
2754 original IP address when looking up the host name (a safety check), it
2755 output the message "<ip address> does not match any IP for NULL", which was
2756 confusing, to say the least. The bug was that the host name should have
2757 appeared instead of "NULL".
2759 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2760 user that is built into the binary, and the -C or -D options is used, root
2761 privilege is dropped before the configuration file is read. In addition,
2762 logging is switched to stderr instead of the normal log files. If the
2763 configuration then re-defines the Exim user, the unprivileged environment
2764 is probably not what is expected, so Exim logs a panic warning message (but
2767 However, if deliver_drop_privilege is set, the unprivileged state may well
2768 be exactly what is intended, so the warning has been cut out in that case,
2769 and Exim is allowed to try to write to its normal log files.
2775 1. smtp_return_error_details was not giving details for temporary sender
2776 or receiver verification errors.
2778 2. Diagnose a configuration error if two authenticators have the same public
2781 3. Exim used not to create the message log file for a message until the first
2782 delivery attempt. This could be confusing when incoming messages were held
2783 for policy or load reasons. The message log file is now created at the time
2784 the message is received, and an initial "Received" line is written to it.
2786 4. The automatically generated man page for command line options had a minor
2787 bug that caused no ill effects; however, a more serious problem was that
2788 the procedure for building the man page automatically didn't always
2789 operate. Consequently, release 4.20 contains an out-of-date version. This
2790 shouldn't happen again.
2792 5. When building Exim with embedded Perl support, the script that builds the
2793 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2794 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2796 6. The freeze_tell option was not being used for messages that were frozen on
2797 arrival, either by an ACL or by local_scan().
2799 7. Added the smtp_incomplete_transaction log selector.
2801 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2802 was accepting AUTH without a new EHLO.
2804 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2805 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2808 10. Logging of TCP/IP connections (when configured) now happens in the main
2809 daemon process instead of the child process, so that the TCP/IP connection
2810 count is more accurate (but it can never be perfect).
2812 11. The use of "drop" in a nested ACL was not being handled correctly in the
2813 outer ACL. Now, if condition failure induced by the nested "drop" causes
2814 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2815 or "require"), the connection is dropped.
2817 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2818 that yield "discard" can now be used with an "accept" or a "discard" verb,
2819 but an error is generated for any others (because I can't see a useful way
2820 to define what should happen).
2822 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2823 are now processed in the same way as lines in the Exim configuration file.
2824 In particular, continuation lines are supported.
2826 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2828 15. Added -ti meaning -t -i.
2830 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2831 domains, and warn by logging if others are found.
2833 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2834 returned from GIFCONF when Exim is trying to find the list of interfaces on
2835 a host. The code in os.c has been modified to copy each ifreq to an aligned
2836 structure in all cases.
2838 Also, in some cases, the returned ifreq's were being copied to a 'struct
2839 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2840 means the last couple of bytes of an IPv6 address could be chopped if the
2841 ifreq contained only a normal sockaddr (14 bytes storage).
2843 18. Named domain lists were not supported in the hosts_treat_as_local option.
2844 An entry such as +xxxx was not recognized, and was treated as a literal
2847 19. Ensure that header lines added by a DATA ACL are included in the reject log
2848 if the ACL subsequently rejects the message.
2850 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2851 MD5 (which is deprecated).
2853 21. When testing a filter file using -bf, Exim was writing a message when it
2854 took the sender from a "From " line in the message, but it was not doing so
2855 when it took $return_path from a Return-Path: header line. It now does.
2857 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2858 with a valid header line field name (a series of printing characters
2859 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2861 23. Changed "disc" in the source to "disk" to conform to the documentation and
2862 the book and for uniformity.
2864 24. Ignore Sendmail's -Ooption=value command line item.
2866 25. When execve() failed while trying to run a command in a pipe transport,
2867 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2868 could be confused with a return value of 69 from the command itself. This
2869 has been changed to 127, the value the shell returns if it is asked to run
2870 a non-existent command. The wording for the related log line suggests a
2871 non-existent command as the problem.
2873 26. If received_header_text expands to an empty string, do not add a Received:
2874 header line to the message. (Well, it adds a token one on the spool, but
2875 marks it "old" so that it doesn't get used or transmitted.)
2877 27. Installed eximstats 1.28 (addition of -nt option).
2879 28. There was no check for failure on the call to getsockname() in the daemon
2880 code. This can fail if there is a shortage of resources on the system, with
2881 ENOMEM, for example. A temporary error is now given on failure.
2883 29. Contrary to the C standard, it seems that in some environments, the
2884 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2885 program. Exim now does this explicitly; it affects the formatting of
2886 timestamps using strftime().
2888 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2889 complaints. I've now initialized all the variables, to avoid this.
2891 32. Header lines added by a system filter were not being "seen" during
2892 transport-time rewrites.
2894 33. The info_callback() function passed to OpenSSL is set up with type void
2895 (*)(SSL *, int, int), as described somewhere. However, when calling the
2896 function (actually a macro) that sets it up, the type void(*)() is
2897 expected. I've put in a cast to prevent warnings from picky compilers.
2899 34. If a DNS black list lookup found a CNAME record, but there were no A
2900 records associated with the domain it pointed at, Exim crashed.
2902 35. If a DNS black list lookup returned more than one A record, Exim ignored
2903 all but the first. It now scans all returned addresses if a particular IP
2904 value is being sought. In this situation, the contents of the
2905 $dnslist_value variable are a list of all the addresses, separated by a
2908 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2909 to accept a host name and all its aliases if the forward lookup for any of
2910 them yielded the IP address of the incoming connection. Now it accepts only
2911 those names whose forward lookup yields the correct IP address. Any other
2912 names are discarded. This closes a loophole whereby a rogue DNS
2913 administrator could create reverse DNS records to break through a
2914 wildcarded host restriction in an ACL.
2916 37. If a user filter or a system filter that ran in a subprocess used any of
2917 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2918 the wrong values were passed to the pipe command ($thisaddress had the
2919 value of $0, $0 had the value of $1, etc). This bug was introduced by
2920 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2922 38. Improved the line breaking for long SMTP error messages from ACLs.
2923 Previously, if there was no break point between 40 and 75 characters, Exim
2924 left the rest of the message alone. Two changes have been made: (a) I've
2925 reduced the minimum length to 35 characters; (b) if it can't find a break
2926 point between 35 and 75 characters, it looks ahead and uses the first one
2927 that it finds. This may give the occasional overlong line, but at least the
2928 remaining text gets split now.
2930 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2931 file descriptors might be low, and that setting 1000 would always raise it.
2932 It turns out that in some environments, the limit is already over 1000 and
2933 that lowering it causes trouble. So now Exim takes care not to decrease it.
2935 40. When delivering a message, the value of $return_path is set to $sender_
2936 address at the start of routing (routers may change the value). By an
2937 oversight, this default was not being set up when an address was tested by
2938 -bt or -bv, which affected the outcome if any router or filter referred to
2941 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2942 log only when "message" or "log_message" are set. However, if one of the
2943 conditions was an address verification, or a call to a nested ACL, the
2944 messages generated by the underlying test were being passed through. This
2945 no longer happens. The underlying message is available in $acl_verify_
2946 message for both "message" and "log_message" expansions, so it can be
2947 passed through if needed.
2949 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2950 new expansion $bh_ to give the encoded byte string without charset
2951 translation. Translation happens only if iconv() is available; HAVE_ICONV
2952 indicates this at build time. HEADERS_CHARSET gives the charset to
2953 translate to; headers_charset can change it in the configuration, and
2954 "headers charset" can change it in an individual filter file.
2956 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2957 that creates RFC 2047 encoded "words" labels them as that charset instead
2958 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2959 expansion operator; (ii) when Exim creates a From: line for a local
2960 message; (iii) when a header line is rewritten to include a "phrase" part.
2962 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2963 buggy, causing it to skip the first lines of messages whose message ID
2964 ended in 'D'. This would not have bitten before Exim release 4.14, because
2965 message IDs were unlikely to end in 'D' before then. The effect was to have
2966 incorrect size information for certain domains.
2968 45. #include "config.h" was missing at the start of the crypt16.c module. This
2969 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2972 46. If there was a timeout during a "random" callout check, Exim treated it as
2973 a failure of the random address, and carried on sending RSET and the real
2974 address. If the delay was just some slowness somewhere, the response to the
2975 original RCPT would be taken as a response to RSET and so on, causing
2976 mayhem of various kinds.
2978 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2979 when I implemented it. It didn't allow for the fact that some option values
2980 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2981 the right test for positive values.
2983 48. Domain names in DNS records are case-independent. Exim always looks them up
2984 in lower case. Some resolvers return domain names in exactly the case they
2985 appear in the zone file, that is, they may contain uppercase letters. Not
2986 all resolvers do this - some return always lower case. Exim was treating a
2987 change of case by a resolver as a change of domain, similar to a widening
2988 of a domain abbreviation. This triggered its re-routing code and so it was
2989 trying to route what was effectively the same domain again. This normally
2990 caused routing to fail (because the router wouldn't handle the domain
2991 twice). Now Exim checks for this case specially, and just changes the
2992 casing of the domain that it ultimately uses when it transmits the message
2995 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2998 50. If a filter generated a file delivery with a non-absolute name (possible if
2999 no home directory exists for the router), the forbid_file option was not
3002 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
3003 the existing equality matching.
3005 52. Exim was using ints instead of ino_t variables in some places where it was
3006 dealing with inode numbers.
3008 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
3009 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
3010 called TMPDIR, and if it finds it is different, it changes its value.
3012 54. The smtp_printf() function is now made available to local_scan() so
3013 additional output lines can be written before returning. There is also an
3014 smtp_fflush() function to enable the detection of a dropped connection.
3015 The variables smtp_input and smtp_batched_input are exported to
3018 55. Changed the default runtime configuration: the message "Unknown user"
3019 has been removed from the ACL, and instead placed on the localuser router,
3020 using the cannot_route_message feature. This means that any verification
3021 failures that generate their own messages won't get overridden. Similarly,
3022 the "Unrouteable address" message that was in the ACL for unverifiable
3023 relay addresses has also been removed.
3025 56. Added hosts_avoid_esmtp to the smtp transport.
3027 57. The exicyclog script was not checking for the esoteric option
3028 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3029 will work only if exicyclog is run under the appropriate euid.
3031 58. Following a discussion on the list, the rules by which Exim recognises line
3032 endings on incoming messages have been changed. The -dropcr and drop_cr
3033 options are now no-ops, retained only for backwards compatibility. The
3034 following line terminators are recognized: LF CRLF CR. However, special
3035 processing applies to CR:
3037 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3038 nor a local message in the state where . is a terminator.
3040 (ii) If a bare CR is encountered in a header line, an extra space is added
3041 after the line terminator so as not to end the header. The reasoning
3042 behind this is that bare CRs in header lines are most likely either
3043 to be mistakes, or people trying to play silly games.
3045 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3046 was being incorrectly given as 18 bytes larger than it should have been.
3047 This is a VOB (very old bug).
3049 60. This may never have affected anything current, but just in case it has:
3050 When the local host is found other than at the start of a list of hosts,
3051 the local host, those with the same MX, and any that follow, are discarded.
3052 When the list in question was part of a longer list of hosts, the following
3053 hosts (not currently being processed) were also being discarded. This no
3054 longer happens. I'm not sure if this situation could ever has previously
3057 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3060 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3061 header. This is something that is recommended in a new Internet Draft, and
3062 is something that is documented as being done by Sendmail. There are two
3063 possible values. For messages generated by the autoreply transport, Exim
3066 Auto-Submitted: auto-replied
3068 whereas for all other generated messages (e.g. bounces) it adds
3070 Auto-Submitted: auto-generated
3072 63. The "personal" condition in filters now includes a test for the
3073 Auto-Submitted: header. If it contains the string "auto-" the message it
3074 not considered personal.
3076 64. Added rcpt_include_affixes as a generic transport option.
3078 65. Added queue_only_override (default true).
3080 66. Added the syslog_duplication option.
3082 67. If what should have been the first header line of a message consisted of
3083 a space followed by a colon, Exim was mis-interpreting it as a header line.
3084 It isn't of course - it is syntactically invalid and should therefore be
3085 treated as the start of the message body. The misbehaviour could have
3086 caused a number of strange effects, including loss of data in subsequent
3087 header lines, and spool format errors.
3089 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3090 client host had authenticated. This control can now be exercised by an ACL
3091 for more flexibility.
3093 69. By default, callouts do not happen when testing with -bh. There is now a
3094 variant, -bhc, which does actually run the callout code, including
3095 consulting and updating the callout cache.
3097 70. Added support for saslauthd authentication, courtesy of Alexander
3100 71. If statvfs() failed on the spool or log directories while checking their
3101 size for availability, Exim confusingly gave the error "space shortage".
3102 Furthermore, in debugging mode it crashed with a floating point exception.
3103 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3104 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3105 serious problem, Exim now writes to the main and panic logs when this
3106 happens, with details of the failure. It then refuses to accept the
3107 incoming message, giving the message "spool directory problem" or "log
3108 directory problem" with a 421 code for SMTP messages.
3110 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3111 0, 1, and 2 exist, because some OS complain for execs without them (see
3112 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3113 descriptors. However, the code omitted to check that the open succeeded,
3114 causing mysterious errors if for some reason the permissions on /dev/null
3115 got screwed. Now Exim writes a message to the main and panic logs, and
3116 bombs out if it can't open /dev/null.
3118 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3119 interact so that it is all more flexible. It is supposed to remain
3120 backwards compatible. Also added extra_local_interfaces.
3122 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3123 to bomb out with an assertion failure - to the client this appears as a
3124 connection drop. This problem occurs in the part of the code that was taken
3125 from the Samba project. Fortunately, the assertion is in a very simple
3126 function, so I have fixed this by reproducing the function inline in the
3127 one place where it is called, and arranging for authentication to fail
3128 instead of killing the process with assert().
3130 75. The SPA client code was not working when the server requested OEM rather
3131 than Unicode encoding.
3133 76. Added code to make require_files with a specific uid setting more usable in
3134 the case where statting the file as root fails - usually a non-root-mounted
3135 NFS file system. When this happens and the failure is EACCES, Exim now
3136 forks a subprocess and does the per-uid checking as the relevant uid.
3138 77. Added process_log_path.
3140 78. If log_file_path was not explicitly set, a setting of check_log_space or
3141 check_log_inodes was ignored.
3143 79. If a space check for the spool or log partitions fails, the incident is now
3144 logged. Of course, in the latter case the data may get lost...
3146 80. Added the %p formatting code to string_format() so that it can be used to
3147 print addresses in debug_print(). Adjusted all the address printing in the
3148 debugging in store.c to use %p rather than %d.
3150 81. There was a concern that a line of code in smtp_in.c could overflow a
3151 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3152 initially expressed, the concern was not well-founded, because trailing
3153 spaces are removed early. However, if the trailing spaces were followed by
3154 a NULL, they did not get removed, so the overflow was possible. Two fixes
3157 (a) I re-wrote the offending code in a cleaner fashion.
3158 (b) If an incoming SMTP command contains a NULL character, it is rejected
3161 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3162 runs initgroups(), so that if the Exim user is in any additional groups,
3163 they will be used during message reception.
3169 The change log for 4.20 and earlier releases has been archived.
projects / exim.git / blob