1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.178 2005/06/28 20:50:56 steve Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
10 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
12 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
13 command itself ended in error, the underlying message about the transport
14 filter timeout was being overwritten with the pipe command error. Now the
15 underlying error message should be appended to the second error message.
17 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
19 PH/02 Reduce the amount of output that "make" produces by default. Full output
20 can still be requested.
22 PH/03 The warning log line about a condition test deferring for a "warn" verb
23 was being output only once per connection, rather than after each
24 occurrence (because it was using the same function as for successful
25 "warn" verbs). This seems wrong, so I have changed it.
27 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
28 it should not have, which might have caused a crash in the right
29 circumstances, but probably never did.
31 PH/04 Installed a modified version of Tony Finch's patch to make submission
32 mode fix the return path as well as the Sender: header line, and to
33 add a /name= option so that you can make the user's friendly name appear
36 TF/03 Added the control = fakedefer ACL modifier.
38 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
39 Mark Lowes for thorough testing.
41 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
43 TK/03 Merged latest SRS patch from Miles Wilton.
45 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
46 with the definition in sysexits.h (which is #included earlier).
47 Fortunately, Exim does not actually use EX_OK. The code used to try to
48 preserve the sysexits.h value, by assumimg that macro definitions were
49 scanned for macro replacements. I have been disabused of this notion,
50 so now the code just undefines EX_OK before #including unistd.h.
52 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
53 in the smtp transport. When a block could not be written in a single
54 write() function, the timeout was being re-applied to each part-write.
55 This seems wrong - if the receiver was accepting one byte at a time it
56 would take for ever. The timeout is now adjusted when this happens. It
57 doesn't have to be particularly precise.
59 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
60 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
62 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
63 without a domain if the check to <postmaster@domain> fails.
65 SC/01 Eximstats: added -xls and the ability to specify output files
66 (patch written by Frank Heydlauf).
68 SC/02 Eximstats: use FileHandles for outputing results.
70 SC/03 Eximstats: allow any combination of xls, txt, and html output.
72 SC/04 Eximstats: fixed display of large numbers with -nvr option
74 SC/05 Eximstats: fixed merging of reports with empty tables.
76 SC/06 Eximstats: added the -include_original_destination flag
78 SC/07 Eximstats: removed tabs and trailing whitespace.
80 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
82 TK/06 MBOX spool code: Add real "From " MBOX separator line
83 so the .eml file is really in mbox format (even though
84 most programs do not really care). Patch from Alex Miller.
86 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
87 The latter is generated from $received_to and is only set if the
88 message has one envelope recipient. SA can use these headers,
89 obviously out-of-the-box. Patch from Alex Miller.
91 PH/08 The ${def test on a variable was returning false if the variable's
92 value was "0", contrary to what the specification has always said!
93 The result should be true unless the variable is empty.
95 PH/09 The syntax error of a character other than { following "${if
96 def:variable_name" (after optional whitespace) was not being diagnosed.
97 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
98 accidental colon was present, for example, could give incorrect results.
100 PH/10 Tidied the code in a number of places where the st_size field of a stat()
101 result is used (not including appendfile, where other changes are about
104 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
105 This involved changing a lot of size variables from int to off_t. It
106 should work with maildirs and everything.
108 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
109 spamd dying while we are connected to it.
111 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
114 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
115 function for BDB 4.3.
117 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
119 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
120 the sources for PCRE have been re-arranged and more files are now
123 PH/15 The code I had for printing potentially long long variables in PH/11
124 above was not the best (it lost precision). The length of off_t variables
125 is now inspected at build time, and an appropriate printing format (%ld
126 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
127 to be "long long int" or "long int". This is needed for the internal
128 formatting function string_vformat().
130 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
131 the configuration file to be ":syslog", then the script "guesses" where
132 the logs files are, rather than using the compiled in default. In our
133 case the guess is not the same as the compiled default, so the script
134 suddenly stopped working when I started to use syslog. The patch checks
135 to see if log_file_path is "". If so, it attempts to read it from exim
136 with no configuration file to get the compiled in version, before it
137 falls back to the previous guessing code."
139 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
140 implementing BATV in an Exim configuration. See NewStuff for the gory
143 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
144 Makefile that are specific to HP-UX.
146 PH/18 If the "use_postmaster" option was set for a recipient callout together
147 with the "random" option, the postmaster address was used as the MAIL
148 FROM address for the random test, but not for the subsequent recipient
149 test. It is now used for both.
151 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
152 patch removes a few documentation additions to RFC 3028, because the
153 latest draft now contains them. It adds the new en;ascii-case comparator
154 and a new error check for 8bit text in MIME parts. Comparator and
155 require names are now matched exactly. I enabled the subaddress
156 extension, but it is not well tested yet (read: it works for me)."
158 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
159 rework some of the code of TK/09 above to avoid the hardwired use of
160 "%lld" and "long long". Replaced the call to snprintf() with a call to
163 PH/21 Added another message to those in 4.51/PH/42, namely "All relevant MX
164 records point to non-existent hosts".
166 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
167 experimental DomainKeys support:
169 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
170 (2) On an error such as an illegally used "control", the wrong name for
171 the control was given.
173 These problems did NOT occur unless DomainKeys support was compiled.
175 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
177 PH/24 Added ${if match_ip condition.
179 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
180 databases so that it will be absolutely obvious if a crash occurs in the
181 DB library. This is a regular occurrence (often caused by mis-matched
184 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
185 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
186 picked up on a user's system that detects such things. There doesn't seem
187 to be a gcc warning option for this - only an attribute that has to be
188 put on the function's prototype. It seems that in Fedora Core 4 they have
189 set this on a number of new functions. No doubt there will be more in due
192 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
193 specify a transport. However, if an address that was verified by such a
194 router was the subject of a callout, Exim crashed because it tried to
195 read the rcpt_include_affixes from the non-existent transport. Now it
196 just assumes that the setting of that option is false. This bug was
197 introduced by 4.51/PH/31.
199 PH/28 Changed -d+all to exclude +memory, because that information is very
200 rarely of interest, but it makes the output a lot bigger. People tend to
201 do -d+all out of habit.
203 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
204 code in os-type was giving problems when libc.so lives in lib64, like on
207 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
208 aren't the modern standard, and it seems that some systems' include files
209 don't always have them. Exim was already checking for some of the newer
210 ones like T_AAAA, and defining it itself. I've added checks for all the
211 record types that Exim uses.
213 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
214 not automatically generating a new one, as it is supposed to. This
215 prevented TLS from working. If the file did exist, but contained invalid
216 data, a new version was generated, as expected. It was only the case of a
217 non-existent file that was broken.
219 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
220 with a change in libdomainkeys > 0.64.
222 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
223 from DNS. If the selector record carries the flag, it now has
224 precedence over the domain-wide flag.
226 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
228 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
229 the use of a port name that isn't defined in /etc/services) occurred, the
230 message was deferred as in a normal delivery, and thus remained on the
231 spool, instead of being failed because of the mua_wrapper setting. This
232 is now fixed, and I tidied up some of the mua_wrapper messages at the
235 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
236 the messages in a hash of arrays rather than using individual hashes.
237 This is a bit cleaner and results in dramatic memory savings, albeit
238 at a slight CPU cost.
240 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
241 as requested by Marc Sherman.
243 SC/10 Eximstats: added histograms for user specified patterns as requested
250 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
251 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
253 TK/02 Fix ACL "control" statement not being available in MIME ACL.
255 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
257 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
258 to test Sieve filters that use "vacation".
260 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
261 that changes the way the GnuTLS parameters are stored in the cache file.
262 The new format can be generated externally. For backward compatibility,
263 if the data in the cache doesn't make sense, Exim assumes it has read an
264 old-format file, and it generates new data and writes a new file. This
265 means that you can't go back to an older release without removing the
268 PH/03 A redirect router that has both "unseen" and "one_time" set does not
269 work if there are any delivery delays because "one_time" forces the
270 parent to be marked "delivered", so its unseen clone is never tried
271 again. For this reason, Exim now forbids the simultaneous setting of
274 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
275 redirected to themselves ("homonym" addresses). Read the long ChangeLog
276 entry if you want to know the details. The fix, however, neglected to
277 consider the case when local delivery batching is involved. The test for
278 "previously delivered" was not happening when checking to see if an
279 address could be batched with a previous (undelivered) one; under
280 certain circumstances this could lead to multiple deliveries to the same
283 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
284 in its include files, and this causes problems building Exim.
286 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
287 header_syntax) but Exim was just ignoring anything given after a slash.
288 In particular, this caused confusion with an attempt to use "verify =
289 reverse_host_lookup/defer_ok". An error is now given when options are
290 supplied for verify items that do not have them. (Maybe reverse_host_
291 lookup should have a defer_ok option, but that's a different point.)
293 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
294 defined by RFC 821) to 2048, because there were problems with some AUTH
295 commands, and RFC 1869 says the size should be increased for extended
296 SMTP commands that take arguments.
298 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
301 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
302 "unknown" error; now it says that the functionality isn't in the binary.
304 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
305 an address' error message when a string expansion fails (syntax or
306 whatever). Otherwise the password may appear in the log. Following change
307 PH/42 below, there is no longer a chance of it appearing in a bounce
310 PH/11 Installed exipick version 20050225.0 from John Jetmore.
312 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
313 of its addresses was ever tried. (Bugzilla bug #2.)
315 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
316 the result incorrectly in the debug output. (It correctly added a newline
317 to what was transported.)
319 TF/01 Added $received_time.
321 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
322 commented out examples of how to interface to a virus scanner and to
323 SpamAssassin. Also added commented examples of av_scanner and
324 spamd_address settings.
326 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
327 and controls are allowed in which ACLs. There were a couple of minor
328 errors. Some of the entries in the conditions table (which is a table of
329 where they are NOT allowed) were getting very unwieldy; rewrote them as a
330 negation of where the condition IS allowed.
332 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
334 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
335 header file does not have a version number, so I've had to invent a new
336 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
337 API. The code is untested by me (my Linux distribution still has 0.3.2 of
338 radiusclient), but it was contributed by a Radius user.
340 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
341 files or directories, for OpenSSL.
343 PH/19 When an Exim process that is running as root has to create an Exim log
344 file, it does so in a subprocess that runs as exim:exim so as to get the
345 ownership right at creation (otherwise, other Exim processes might see
346 the file with the wrong ownership). There was no test for failure of this
347 fork() call, which would lead to the process getting stuck as it waited
348 for a non-existent subprocess. Forks do occasionally fail when resources
349 run out. I reviewed all the other calls to fork(); they all seem to check
352 PH/20 When checking for unexpected SMTP input at connect time (before writing
353 the banner), Exim was not dealing correctly with a non-positive return
354 from the read() function. If the client had disconnected by this time,
355 the result was a log entry for a synchronization error with an empty
356 string after "input=" when read() returned zero. If read() returned -1
357 (an event I could not check), uninitialized data bytes were printed.
358 There were reports of junk text (parts of files, etc) appearing after
361 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
363 PH/22 Added support for macro redefinition, and (re)definition in between
364 driver and ACL definitions.
366 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
367 forgetting to use the resulting value; it was using the unexpanded value.
369 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
370 hadn't been configured. The fix is from Juergen Kreileder, who
371 understands it better than I do:
373 "Here's what I see happening with three configured cyrus_sasl
374 authenticators configured (plain, login, cram-md5):
376 On startup auth_cyrus_sasl_init() gets called for each of these.
377 This means three calls to sasl_listmech() without a specified mech_list.
378 => SASL tests which mechs of all available mechs actually work
379 => three warnings about OTP not working
380 => the returned list contains: plain, login, cram-md5, digest-md5, ...
382 With the patch, sasl_listmech() also gets called three times. But now
383 SASL's mech_list option is set to the server_mech specified in the the
384 authenticator. Or in other words, the answer from sasl_listmech()
385 gets limited to just the mech you're testing for (which is different
387 => the return list contains just 'plain' or 'login', 'cram-md5' or
388 nothing depending on the value of ob->server_mech.
390 I've just tested the patch: Authentication still works fine,
391 unavailable mechs specified in the exim configuration are still
392 caught, and the auth.log warnings about OTP are gone."
394 PH/25 When debugging is enabled, the contents of the command line are added
395 to the debugging output, even when log_selector=+arguments is not
398 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
399 answer is "GNU", and only if the return is "GNU/something" is the answer
402 PH/27 $acl_verify_message is now set immediately after the failure of a
403 verification in an ACL, and so is available in subsequent modifiers. In
404 particular, the message can be preserved by coding like this:
406 warn !verify = sender
407 set acl_m0 = $acl_verify_message
409 Previously, $acl_verify_message was set only while expanding "message"
410 and "log_message" when a very denied access.
412 PH/28 Modified OS/os.c-Linux with
414 -#ifndef OS_LOAD_AVERAGE
415 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
417 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
418 nomenclature these days.)
420 PH/29 Installed patch from the Sieve maintainer that adds the options
421 sieve_useraddress and sieve_subaddress to the redirect router.
423 PH/30 In these circumstances:
424 . Two addresses routed to the same list of hosts;
425 . First host does not offer TLS;
426 . First host accepts first address;
427 . First host gives temporary error to second address;
428 . Second host offers TLS and a TLS session is established;
429 . Second host accepts second address.
430 Exim incorrectly logged both deliveries with the TLS parameters (cipher
431 and peerdn, if requested) that were in fact used only for the second
434 PH/31 When doing a callout as part of verifying an address, Exim was not paying
435 attention to any local part prefix or suffix that was matched by the
436 router that accepted the address. It now behaves in the same way as it
437 does for delivery: the affixes are removed from the local part unless
438 rcpt_include_affixes is set on the transport.
440 PH/32 Add the sender address, as F=<...>, to the log line when logging a
441 timeout during the DATA phase of an incoming message.
443 PH/33 Sieve envelope tests were broken for match types other than :is. I have
444 applied a patch sanctioned by the Sieve maintainer.
446 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
447 the uid or gid is negative. A case of a negative gid caused this to be
448 noticed. The fix allows for either to be negative.
450 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
451 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
454 PH/36 The change PH/12 above was broken. Fixed it.
456 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
457 the grounds that routing the same address twice would always produce the
458 same answer. This might have been true once, but it is certainly no
459 longer true now. Routing a child address may depend on the previous
460 routing that produced that child. Some complicated redirection strategies
461 went wrong when messages had multiple recipients, and made Exim's
462 behaviour dependent on the order in which the addresses were given.
464 I have moved the duplicate checking until after the routing is complete.
465 Exim scans the addresses that are assigned to local and remote
466 transports, and removes any duplicates. This means that more work will be
467 done, as duplicates will always all be routed, but duplicates are
468 presumably rare, so I don't expect this is of any significance.
470 For deliveries to pipes, files, and autoreplies, the duplicate checking
471 still happens during the routing process, since they are not going to be
474 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
475 It corrects a timeout issue with spamd. This is Ian's comment: "The
476 background is that sometimes spamd either never reads data from a
477 connection it has accepted, or it never writes response data. The exiscan
478 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
479 blindly assumes that writes won't block so it may never time out."
481 PH/39 Allow G after quota size as well as K and M.
483 PH/40 The value set for $authenticated_id in an authenticator may not contain
484 binary zeroes or newlines because the value is written to log lines and
485 to spool files. There was no check on this. Now the value is run through
486 the string_printing() function so that such characters are converted to
487 printable escape sequences.
489 PH/41 $message_linecount is a new variable that contains the total number of
490 lines in the message. Compare $body_linecount, which is the count for the
493 PH/42 Exim no longer gives details of delivery errors for specific addresses in
494 bounce and delay warning messages, except in certain special cases, which
497 (a) An SMTP error message from a remote host;
498 (b) A message specified in a :fail: redirection;
499 (c) A message specified in a "fail" command in a system filter;
500 (d) A message specified in a FAIL return from the queryprogram router;
501 (e) A message specified by the cannot_route_message router option.
503 In these cases only, Exim does include the error details in bounce and
504 warning messages. There are also a few cases where bland messages such
505 as "unrouteable address" or "local delivery error" are given.
507 PH/43 $value is now also set for the "else" part of a ${run expansion.
509 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
510 being worked on, but at least Exim now implements the latest version to
513 PH/45 In a pipe transport, although a timeout while waiting for the pipe
514 process to complete was treated as a delivery failure, a timeout while
515 writing the message to the pipe was logged, but erroneously treated as a
516 successful delivery. Such timeouts include transport filter timeouts. For
517 consistency with the overall process timeout, these timeouts are now
518 treated as errors, giving rise to delivery failures by default. However,
519 there is now a new Boolean option for the pipe transport called
520 timeout_defer, which, if set TRUE, converts the failures into defers for
521 both kinds of timeout. A transport filter timeout is now identified in
524 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
525 systems where "make" and "gmake" are different, calling "gmake" at top
526 level broke things. I've arranged for the value of $(MAKE) to be passed
527 from the Makefile to this script so that it can call the same version of
531 A note about Exim versions 4.44 and 4.50
532 ----------------------------------------
534 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
535 changes of various kinds. As a consequence, a big documentation update was
536 needed. This delayed the release for rather longer than seemed good, especially
537 in the light of a couple of (minor) security issues. Therefore, the changes
538 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
539 release. So 4.44 and 4.50 are in effect two different branches that both start
542 I have left the 4.50 change log unchanged; it contains all the changes since
543 4.43. The change log for 4.44 is below; many of its items are identical to
544 those for 4.50. This seems to be the most sensible way to preserve the
545 historical information.
551 1. Minor wording change to the doc/README.SIEVE file.
553 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
554 computation of the current number of files was incorrect.
556 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
557 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
560 4. Give more explanation in the error message when the command for a transport
561 filter fails to execute.
563 5. There are several places where Exim runs a non-Exim command in a
564 subprocess. The SIGUSR1 signal should be disabled for these processes. This
565 was being done only for the command run by the queryprogram router. It is
566 now done for all such subprocesses. The other cases are: ${run, transport
567 filters, and the commands run by the lmtp and pipe transports.
569 6. Added CONFIGURE_GROUP build-time option.
571 7. Some older OS have a limit of 256 on the maximum number of file
572 descriptors. Exim was using setrlimit() to set 1000 as a large value
573 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
574 systems. I've change it so that if it can't get 1000, it tries for 256.
576 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
577 was an oversight, and furthermore, ever since the addition of extra
578 controls (e.g. 4.43/32), the checks on when to allow different forms of
579 "control" were broken. There should now be diagnostics for all cases when a
580 control that does not make sense is encountered.
582 9. Added the /retain_sender option to "control=submission".
584 10. $recipients is now available in the predata ACL (oversight).
586 11. Tidy the search cache before the fork to do a delivery from a message
587 received from the command line. Otherwise the child will trigger a lookup
588 failure and thereby defer the delivery if it tries to use (for example) a
589 cached ldap connection that the parent has called unbind on.
591 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
592 of $address_data from the recipient verification was clobbered by the
595 13. The value of address_data from a sender verification is now available in
596 $sender_address_data in subsequent conditions in the ACL statement.
598 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
600 15. Added a new option "connect=<time>" to callout options, to set a different
603 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
604 was its contents. (It was OK if the option was not defined at all.)
606 17. A "Completed" log line is now written for messages that are removed from
607 the spool by the -Mrm option.
609 18. New variables $sender_verify_failure and $recipient_verify_failure contain
610 information about exactly what failed.
612 19. Added -dd to debug only the daemon process.
614 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
615 handles timeouts, both on the server side and network timeouts. Renamed the
616 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
618 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
619 It is now set to "smtps".
621 22. $host_address is now set to the target address during the checking of
624 23. When checking ignore_target_hosts for an ipliteral router, no host name was
625 being passed; this would have caused $sender_host_name to have been used if
626 matching the list had actually called for a host name (not very likely,
627 since this list is usually IP addresses). A host name is now passed as
630 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
631 code that specifies a non-restarting handler (typically sigaction() in
632 modern systems) in an attempt to fix a rare and obscure crash bug.
634 25. Narrowed the window for a race in the daemon that could cause it to ignore
635 SIGCHLD signals. This is not a major problem, because they are used only to
636 wake it up if nothing else does.
638 26. A malformed maildirsize file could cause Exim to calculate negative values
639 for the mailbox size or file count. Odd effects could occur as a result.
640 The maildirsize information is now recalculated if the size or filecount
643 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
644 support for a long time. Removed HAVE_SYS_VFS_H.
646 28. Installed the latest version of exipick from John Jetmore.
648 29. In an address list, if the pattern was not a regular expression, an empty
649 subject address (from a bounce message) matched only if the pattern was an
650 empty string. Non-empty patterns were not even tested. This was the wrong
651 because it is perfectly reasonable to use an empty address as part of a
652 database query. An empty address is now tested by patterns that are
653 lookups. However, all the other forms of pattern expect the subject to
654 contain a local part and a domain, and therefore, for them, an empty
655 address still always fails if the pattern is not itself empty.
657 30. Exim went into a mad DNS loop when attempting to do a callout where the
658 host was specified on an smtp transport, and looking it up yielded more
661 31. Re-factored the code for checking spool and log partition space into a
662 function that finds that data and another that does the check. The former
663 is then used to implement four new variables: $spool_space, $log_space,
664 $spool_inodes, and $log_inodes.
666 32. The RFC2047 encoding function was originally intended for short strings
667 such as real names; it was not keeping to the 75-character limit for
668 encoded words that the RFC imposes. It now respects the limit, and
669 generates multiple encoded words if necessary. To be on the safe side, I
670 have increased the buffer size for the ${rfc2047: expansion operator from
673 33. It is now permitted to omit both strings after an "if" condition; if the
674 condition is true, the result is "true". As before, when the second string
675 is omitted, a false condition yields an empty string. This makes it less
676 cumbersome to write custom ACL and router conditions.
678 34. Failure to deliver a bounce message always caused it to be frozen, even if
679 there was an errors_to setting on the router. The errors_to setting is now
682 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
683 canonical form (fully expanded) before being placed in
684 $sender_host_address.
686 36. The table in the code that translates DNS record types into text (T_A to
687 "A" for instance) was missing entries for NS and CNAME. It is just possible
688 that this could have caused confusion if both these types were looked up
689 for the same domain, because the text type is used as part of Exim's
690 per-process caching. But the chance of anyone hitting this buglet seems
693 37. The dnsdb lookup has been extended in a number of ways.
695 (1) There is a new type, "zns", which walks up the domain tree until it
696 finds some nameserver records. It should be used with care.
698 (2) There is a new type, "mxh", which is like "mx" except that it returns
699 just the host names, not the priorities.
701 (3) It is now possible to give a list of domains (or IP addresses) to be
702 looked up. The behaviour when one of the lookups defers can be
703 controlled by a keyword.
705 (4) It is now possible to specify the separator character for use when
706 multiple records are returned.
708 38. The dnslists ACL condition has been extended: it is now possible to supply
709 a list of IP addresses and/or domains to be looked up in a particular DNS
712 39. Added log_selector=+queue_time_overall.
714 40. When running the queue in the test harness, wait just a tad after forking a
715 delivery process, to get repeatability of debugging output.
717 41. Include certificate and key file names in error message when GnuTLS fails
718 to set them up, because the GnuTLS error message doesn't include the name
719 of the failing file when there is a problem reading it.
721 42. Allow both -bf and -bF in the same test run.
723 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
725 44. The "Exiscan patch" is now merged into the mainline Exim source.
727 45. Sometimes the final signoff response after QUIT could fail to get
728 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
729 < 0 before doing a fflush(). This bug looks as though it goes back to the
730 introduction of TLS in release 3.20, but "sometimes" must have been rare
731 because the tests only now provoked it.
733 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
734 (this can affect the format of dates).
736 47. exim_tidydb, when checking for the continued existence of a message for
737 which it has found a message-specific retry record, was not finding
738 messages that were in split spool directories. Consequently, it was
739 deleting retry records that should have stayed in existence.
741 48. Steve fixed some bugs in eximstats.
743 49. The SPA authentication driver was not abandoning authentication and moving
744 on to the next authenticator when an expansion was forced to fail,
745 contradicting the general specification for all authenticators. Instead it
746 was generating a temporary error. It now behaves as specified.
748 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
749 (the order specifies the preference for clients). The order is now AES256,
750 AES128, 3DES, ARCFOUR128.
752 51. Small patch to Sieve code - explicitly set From: when generating an
755 52. Exim crashed if a remote delivery caused a very long error message to be
756 recorded - for instance if somebody sent an entire SpamAssassin report back
757 as a large number of 550 error lines. This bug was coincidentally fixed by
758 increasing the size of one of Exim's internal buffers (big_buffer) that
759 happened as part of the Exiscan merge. However, to be on the safe side, I
760 have made the code more robust (and fixed the comments that describe what
763 53. Now that there can be additional text after "Completed" in log lines (if
764 the queue_time_overall log selector is set), a one-byte patch to exigrep
765 was needed to allow it to recognize "Completed" as not the last thing in
768 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
769 patch that reportedly fixes this has been added. I am not expert enough to
770 create a test for it. This is what the patch creator wrote:
772 "I found a little strange behaviour of ldap code when working with
773 Windows 2003 AD Domain, where users was placed in more than one
774 Organization Units. When I tried to give exim partial DN, the exit code
775 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
776 But simultaneously result of request was absolutely normal ldap result,
777 so I produce this patch..."
779 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
780 REFERENCE, so I have modified the code to exclude the patch when that macro
783 55. Some experimental protocols are using DNS PTR records for new purposes. The
784 keys for these records are domain names, not reversed IP addresses. The
785 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
786 leaves it alone. Component reversal etc. now happens only for IP addresses.
789 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
791 57. Double the size of the debug message buffer (to 2048) so that more of very
792 long debug lines gets shown.
794 58. The exicyclog utility now does better if the number of log files to keep
795 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
797 59. Two changes related to the smtp_active_hostname option:
799 (1) $smtp_active_hostname is now available as a variable.
800 (2) The default for smtp_banner uses $smtp_active_hostname instead
801 of $primary_hostname.
803 60. The host_aton() function is supposed to be passed a string that is known
804 to be a valid IP address. However, in the case of IPv6 addresses, it was
805 not checking this. This is a hostage to fortune. Exim now panics and dies
806 if the condition is not met. A case was found where this could be provoked
807 from a dnsdb PTR lookup with an IPv6 address that had more than 8
808 components; fortuitously, this particular loophole had already been fixed
809 by change 4.50/55 above.
811 If there are any other similar loopholes, the new check in host_aton()
812 itself should stop them being exploited. The report I received stated that
813 data on the command line could provoke the exploit when Exim was running as
814 exim, but did not say which command line option was involved. All I could
815 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
819 61. There was a buffer overflow vulnerability in the SPA authentication code
820 (which came originally from the Samba project). I have added a test to the
821 spa_base64_to_bits() function which I hope fixes it.
824 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
825 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
827 63. The daemon start-up calls getloadavg() while still root for those OS that
828 need the first call to be done as root, but it missed one case: when
829 deliver_queue_load_max is set with deliver_drop_privilege. This is
830 necessary for the benefit of the queue runner, because there is no re-exec
831 when deliver_drop_privilege is set.
833 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
836 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
837 was in use, was not putting the data itself into the right store pool;
838 consequently, it could be overwritten for a subsequent message in the same
839 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
842 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
844 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
845 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
846 the function were treating the return as a boolean value, which happened to
847 work because 0=false and not-0=true, but is not correct code.
849 68. The host_aton() function was not handling scoped IPv6 addresses (those
850 with, for example, "%eth0" on the end) correctly.
852 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
853 negated items (that is, ~something) in unsigned ints. Some compilers
854 apparently mutter when there is no cast.
856 70. If an address verification called from an ACL failed, and did not produce a
857 user-specific message (i.e. there was only a "system" message), nothing was
858 put in $acl_verify_message. In this situation, it now puts the system
861 71. Change 4.23/11 added synchronization checking at the start of an SMTP
862 session; change 4.31/43 added the unwanted input to the log line - except
863 that it did not do this in the start of session case. It now does.
865 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
866 This is wrong and can cause the other end to generate a synchronization
867 error if it is another Exim or anything else that does the synchronization
868 check. A QUIT command is no longer sent after a timeout.
870 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
873 74. The defer_ok option of callout verification was not working if it was used
874 when verifying addresses in header lines, that is, for this case:
876 verify = header_sender/callout=defer_ok
878 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
879 those file descriptors could be used for SMTP connections. If anything
880 wrote to stderr (the example that came up was "warn" in embedded Perl), it
881 could be sent to the SMTP client, causing chaos. The daemon now opens
882 stdin, stdout, and stderr to /dev/null when it puts itself into the
885 76. Arrange for output from Perl's "warn" command to be written to Exim's main
886 log by default. The user can override this with suitable Perl magic.
888 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
889 the log message when discard triggers, was not working for the DATA ACL or
890 for the non-SMTP ACL.
892 78. Error message wording change in sieve.c.
894 79. If smtp_accept_max_per_host was set, the number of connections could be
895 restricted to fewer than expected, because the daemon was trying to set up
896 a new connection before checking whether the processes handling previous
897 connections had finished. The check for completed processes is now done
898 earlier. On busy systems, this bug wouldn't be noticed because something
899 else would have woken the daemon, and it would have reaped the completed
902 80. If a message was submitted locally by a user whose login name contained one
903 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
904 It caused a spool format error. I have fixed the spool reading code. A
905 related problem was that the "from" clause in the Received: line became
906 illegal because of the space(s). It is now covered by ${quote_local_part.
908 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
911 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
913 83. Patch from Sieve maintainer to fix unterminated string problem in
916 84. Some minor changes to the Linux configuration files to help with other
917 OS variants using glibc.
919 85. One more patch for Sieve to update vacation handling to latest spec.
922 ----------------------------------------------------
923 See the note above about the 4.44 and 4.50 releases.
924 ----------------------------------------------------
930 1. Change 4.43/35 introduced a bug that caused file counts to be
931 incorrectly computed when quota_filecount was set in an appendfile
934 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
935 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
938 3. Give more explanation in the error message when the command for a transport
939 filter fails to execute.
941 4. There are several places where Exim runs a non-Exim command in a
942 subprocess. The SIGUSR1 signal should be disabled for these processes. This
943 was being done only for the command run by the queryprogram router. It is
944 now done for all such subprocesses. The other cases are: ${run, transport
945 filters, and the commands run by the lmtp and pipe transports.
947 5. Some older OS have a limit of 256 on the maximum number of file
948 descriptors. Exim was using setrlimit() to set 1000 as a large value
949 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
950 systems. I've change it so that if it can't get 1000, it tries for 256.
952 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
953 was an oversight, and furthermore, ever since the addition of extra
954 controls (e.g. 4.43/32), the checks on when to allow different forms of
955 "control" were broken. There should now be diagnostics for all cases when a
956 control that does not make sense is encountered.
958 7. $recipients is now available in the predata ACL (oversight).
960 8. Tidy the search cache before the fork to do a delivery from a message
961 received from the command line. Otherwise the child will trigger a lookup
962 failure and thereby defer the delivery if it tries to use (for example) a
963 cached ldap connection that the parent has called unbind on.
965 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
966 of $address_data from the recipient verification was clobbered by the
969 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
970 was its contents. (It was OK if the option was not defined at all.)
972 11. A "Completed" log line is now written for messages that are removed from
973 the spool by the -Mrm option.
975 12. $host_address is now set to the target address during the checking of
978 13. When checking ignore_target_hosts for an ipliteral router, no host name was
979 being passed; this would have caused $sender_host_name to have been used if
980 matching the list had actually called for a host name (not very likely,
981 since this list is usually IP addresses). A host name is now passed as
984 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
985 code that specifies a non-restarting handler (typically sigaction() in
986 modern systems) in an attempt to fix a rare and obscure crash bug.
988 15. Narrowed the window for a race in the daemon that could cause it to ignore
989 SIGCHLD signals. This is not a major problem, because they are used only to
990 wake it up if nothing else does.
992 16. A malformed maildirsize file could cause Exim to calculate negative values
993 for the mailbox size or file count. Odd effects could occur as a result.
994 The maildirsize information is now recalculated if the size or filecount
997 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
998 support for a long time. Removed HAVE_SYS_VFS_H.
1000 18. Updated exipick to current release from John Jetmore.
1002 19. Allow an empty sender to be matched against a lookup in an address list.
1003 Previously the only cases considered were a regular expression, or an
1006 20. Exim went into a mad DNS lookup loop when doing a callout where the
1007 host was specified on the transport, if the DNS lookup yielded more than
1010 21. The RFC2047 encoding function was originally intended for short strings
1011 such as real names; it was not keeping to the 75-character limit for
1012 encoded words that the RFC imposes. It now respects the limit, and
1013 generates multiple encoded words if necessary. To be on the safe side, I
1014 have increased the buffer size for the ${rfc2047: expansion operator from
1017 22. Failure to deliver a bounce message always caused it to be frozen, even if
1018 there was an errors_to setting on the router. The errors_to setting is now
1021 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
1022 canonical form (fully expanded) before being placed in
1023 $sender_host_address.
1025 24. Updated eximstats to version 1.33
1027 25. Include certificate and key file names in error message when GnuTLS fails
1028 to set them up, because the GnuTLS error message doesn't include the name
1029 of the failing file when there is a problem reading it.
1031 26. Expand error message when OpenSSL has problems setting up cert/key files.
1034 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
1035 (this can affect the format of dates).
1037 28. exim_tidydb, when checking for the continued existence of a message for
1038 which it has found a message-specific retry record, was not finding
1039 messages that were in split spool directories. Consequently, it was
1040 deleting retry records that should have stayed in existence.
1042 29. eximstats updated to version 1.35
1043 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
1044 1.35 - bugfix such that pie charts by volume are generated correctly
1046 30. The SPA authentication driver was not abandoning authentication and moving
1047 on to the next authenticator when an expansion was forced to fail,
1048 contradicting the general specification for all authenticators. Instead it
1049 was generating a temporary error. It now behaves as specified.
1051 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
1052 (the order specifies the preference for clients). The order is now AES256,
1053 AES128, 3DES, ARCFOUR128.
1055 31. Small patch to Sieve code - explicitly set From: when generating an
1058 32. Exim crashed if a remote delivery caused a very long error message to be
1059 recorded - for instance if somebody sent an entire SpamAssassin report back
1060 as a large number of 550 error lines. This bug was coincidentally fixed by
1061 increasing the size of one of Exim's internal buffers (big_buffer) that
1062 happened as part of the Exiscan merge. However, to be on the safe side, I
1063 have made the code more robust (and fixed the comments that describe what
1066 33. Some experimental protocols are using DNS PTR records for new purposes. The
1067 keys for these records are domain names, not reversed IP addresses. The
1068 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
1069 leaves it alone. Component reversal etc. now happens only for IP addresses.
1072 34. The host_aton() function is supposed to be passed a string that is known
1073 to be a valid IP address. However, in the case of IPv6 addresses, it was
1074 not checking this. This is a hostage to fortune. Exim now panics and dies
1075 if the condition is not met. A case was found where this could be provoked
1076 from a dnsdb PTR lookup with an IPv6 address that had more than 8
1077 components; fortuitously, this particular loophole had already been fixed
1078 by change 4.50/55 or 4.44/33 above.
1080 If there are any other similar loopholes, the new check in host_aton()
1081 itself should stop them being exploited. The report I received stated that
1082 data on the command line could provoke the exploit when Exim was running as
1083 exim, but did not say which command line option was involved. All I could
1084 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
1085 running as the user.
1088 35. There was a buffer overflow vulnerability in the SPA authentication code
1089 (which came originally from the Samba project). I have added a test to the
1090 spa_base64_to_bits() function which I hope fixes it.
1093 36. The daemon start-up calls getloadavg() while still root for those OS that
1094 need the first call to be done as root, but it missed one case: when
1095 deliver_queue_load_max is set with deliver_drop_privilege. This is
1096 necessary for the benefit of the queue runner, because there is no re-exec
1097 when deliver_drop_privilege is set.
1099 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
1100 was in use, was not putting the data itself into the right store pool;
1101 consequently, it could be overwritten for a subsequent message in the same
1102 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
1105 38. Sometimes the final signoff response after QUIT could fail to get
1106 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
1107 < 0 before doing a fflush(). This bug looks as though it goes back to the
1108 introduction of TLS in release 3.20, but "sometimes" must have been rare
1109 because the tests only now provoked it.
1115 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
1116 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
1117 Now it returns an integer. A number of calls were still expecting a T/F
1118 return. Fortuitously, in all cases, the tests worked in OK situations,
1119 which is the norm. However, things would have gone wrong on any write
1120 failures on the smtp file descriptor. This function is used when sending
1121 messages over SMTP and also when doing verify callouts.
1123 2. When Exim is called to do synchronous delivery of a locally submitted
1124 message (the -odf or -odi options), it no longer closes stderr before doing
1127 3. Implemented the mua_wrapper option.
1129 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
1131 5. Implemented the functions header_remove(), header_testname(),
1132 header_add_at_position(), and receive_remove_recipient(), and exported them
1135 6. If an ACL "warn" statement specified the addition of headers, Exim already
1136 inserted X-ACL-Warn: at the start if there was no header name. However, it
1137 was not making this test for the second and subsequent header lines if
1138 there were newlines in the string. This meant that an invalid header could
1139 be inserted if Exim was badly configured.
1141 7. Allow an ACL "warn" statement to add header lines at the start or after all
1142 the Received: headers, as well as at the end.
1144 8. Added the rcpt_4xx retry error code.
1146 9. Added postmaster_mailfrom=xxx to callout verification option.
1148 10. Added mailfrom=xxxx to the callout verification option, for verify=
1151 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
1152 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
1154 12. Inserted some casts to stop certain compilers warning when using pointer
1155 differences as field lengths or precisions in printf-type calls (mostly
1156 affecting debugging statements).
1158 13. Added optional readline() support for -be (dynamically loaded).
1160 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
1161 same clock tick as a message's arrival, so that its received time was the
1162 same as the "first fail" time on the retry record, and that message
1163 remained on the queue past the ultimate address timeout, every queue runner
1164 would try a delivery (because it was past the ultimate address timeout) but
1165 after another failure, the ultimate address timeout, which should have then
1166 bounced the address, did not kick in. This was a "< instead of <=" error;
1167 in most cases the first failure would have been in the next clock tick
1168 after the received time, and all would be well.
1170 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
1171 being recognized when the domain list was tested by the match_domain
1172 condition in an expansion string.
1174 16. Added the ${str2b64: operator.
1176 17. Exim was always calling setrlimit() to set a large limit for the number of
1177 processes, without checking whether the existing limit was already
1178 adequate. (It did check for the limit on file descriptors.) Furthermore,
1179 errors from getrlimit() and setrlimit() were being ignored. Now they are
1180 logged to the main and panic logs, but Exim does carry on, to try to do its
1181 job under whatever limits there are.
1183 18. Imported PCRE 5.0.
1185 19. Trivial typo in log message " temporarily refused connection" (the leading
1188 20. If the log selector return_path_on_delivery was set and an address was
1189 redirected to /dev/null, the delivery process crashed because it assumed
1190 that a return path would always be set for a "successful" delivery. In this
1191 case, the whole delivery is bypassed as an optimization, and therefore no
1194 21. Internal re-arrangement: the function for sending a challenge and reading
1195 a response while authentication was assuming a zero-terminated challenge
1196 string. It's now changed to take a pointer and a length, to allow for
1197 binary data in such strings.
1199 22. Added the cyrus_sasl authenticator (code supplied by MBM).
1201 23. Exim was not respecting finduser_retries when seeking the login of the
1202 uid under which it was called; it was always trying 10 times. (The default
1203 setting of finduser_retries is zero.) Also, it was sleeping after the final
1204 failure, which is pointless.
1206 24. Implemented tls_on_connect_ports.
1208 25. Implemented acl_smtp_predata.
1210 26. If the domain in control=submission is set empty, Exim assumes that the
1211 authenticated id is a complete email address when it generates From: or
1212 Sender: header lines.
1214 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
1215 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
1216 chown and chgrp in /bin and hostname in /usr/bin.
1218 28. Exim was keeping the "process log" file open after each use, just as it
1219 does for the main log. This opens the possibility of it remaining open for
1220 long periods when the USR1 signal hits a daemon. Occasional processlog
1221 errors were reported, that could have been caused by this. Anyway, it seems
1222 much more sensible not to leave this file open at all, so that is what now
1225 29. The long-running daemon process does not normally write to the log once it
1226 has entered its main loop, and it closes the log before doing so. This is
1227 so that log files can straightforwardly be renamed and moved. However,
1228 there are a couple of unusual error situations where the daemon does write
1229 log entries, and I had neglected to close the log afterwards.
1231 30. The text of an SMTP error response that was received during a remote
1232 delivery was being truncated at 512 bytes. This is too short for some of
1233 the long messages that one sometimes sees. I've increased the limit to
1236 31. It is now possible to make retry rules that apply only when a message has a
1237 specific sender, in particular, an empty sender.
1239 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
1240 it possible to be selective about when SMTP synchronization is enforced.
1242 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
1244 32. Implemented hosts_connection_nolog.
1246 33. Added an ACL for QUIT.
1248 34. Setting "delay_warning=" to disable warnings was not working; it gave a
1251 35. Added mailbox_size and mailbox_filecount to appendfile.
1253 36. Added control = no_multiline_responses to ACLs.
1255 37. There was a bug in the logic of the code that waits for the clock to tick
1256 in the case where the clock went backwards by a substantial amount such
1257 that the microsecond fraction of "now" was more than the microsecond
1258 fraction of "then" (but the whole seconds number was less).
1260 38. Added support for the libradius Radius client library this is found on
1261 FreeBSD (previously only the radiusclient library was supported).
1267 1. When certain lookups returned multiple values in the form name=value, the
1268 quoting of the values was not always being done properly. Specifically:
1269 (a) If the value started with a double quote, but contained no whitespace,
1271 (b) If the value contained whitespace other than a space character (i.e.
1272 tabs or newlines or carriage returns) it was not quoted.
1273 This fix has been applied to the mysql and pgsql lookups by writing a
1274 separate quoting function and calling it from the lookup code. The fix
1275 should probably also be applied to nisplus, ibase and oracle lookups, but
1276 since I cannot test any of those, I have not disturbed their existing code.
1278 2. A hit in the callout cache for a specific address caused a log line with no
1279 reason for rejecting RCPT. Now it says "Previous (cached) callout
1280 verification failure".
1282 3. There was an off-by-one bug in the queryprogram router. An over-long
1283 return line was truncated at 256 instead of 255 characters, thereby
1284 overflowing its buffer with the terminating zero. As well as fixing this, I
1285 have increased the buffer size to 1024 (and made a note to document this).
1287 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
1288 when Exim is waiting for an SMTP response from a remote server, Exim
1289 restarts its select() call on the socket, thereby resetting its timeout.
1290 This is not a problem when such interrupts are rare. Somebody set up a cron
1291 job to run exiwhat every 2 minutes, which is less than the normal select()
1292 timeout (5 or 10 minutes). This meant that the select() timeout never
1293 kicked in because it was always reset. I have fixed this by comparing the
1294 time when an interrupt arrives with the time at the start of the first call
1295 to select(). If more time than the timeout has elapsed, the interrupt is
1296 treated as a timeout.
1298 5. Some internal re-factoring in preparation for the addition of Sieve
1299 extensions (by MH). In particular, the "personal" test is moved to a
1300 separate function, and given an option for scanning Cc: and Bcc: (which is
1301 not set for Exim filters).
1303 6. When Exim created an email address using the login of the caller as the
1304 local part (e.g. when creating a From: or Sender: header line), it was not
1305 quoting the local part when it contained special characters such as @.
1307 7. Installed new OpenBSD configuration files.
1309 8. Reworded some messages for syntax errors in "and" and "or" conditions to
1310 try to make them clearer.
1312 9. Callout options, other than the timeout value, were being ignored when
1313 verifying sender addresses in header lines. For example, when using
1315 verify = header_sender/callout=no_cache
1317 the cache was (incorrectly) being used.
1319 10. Added a missing instance of ${EXE} to the exim_install script; this affects
1320 only the Cygwin environment.
1322 11. When return_path_on_delivery was set as a log selector, if different remote
1323 addresses in the same message used different return paths and parallel
1324 remote delivery occurred, the wrong values would sometimes be logged.
1325 (Whenever a remote delivery process finished, the return path value from
1326 the most recently started remote delivery process was logged.)
1328 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
1329 lines when AUTH and/or TLS are in use. This is the "received protocol"
1330 field. Exim used to use "asmtp" for authenticated SMTP, without any
1331 indication (in the protocol name) for TLS use. Now it follows the RFC and
1332 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
1333 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
1334 names appear in log lines as well as in Received: header lines.
1336 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
1337 extensions, and comparison tests, and to fix some bugs.
1339 14. Changes to the "personal" filter test:
1341 (1) The test was buggy in that it was just doing the equivalent of
1342 "contains" tests on header lines. For example, if a user's address was
1343 anne@some.where, the "personal" test would incorrectly be true for
1345 To: susanne@some.where
1347 This test is now done by extracting each address from the header in turn,
1348 and checking the entire address. Other tests that are part of "personal"
1349 are now done using regular expressions (for example, to check local parts
1350 of addresses in From: header lines).
1352 (2) The list of non-personal local parts in From: addresses has been
1353 extended to include "listserv", "majordomo", "*-request", and "owner-*",
1354 taken from the Sieve specification recommendations.
1356 (3) If the message contains any header line starting with "List-" it is
1357 treated as non-personal.
1359 (4) The test for "circular" in the Subject: header line has been removed
1360 because it now seems ill-conceived.
1362 15. Minor typos in src/EDITME comments corrected.
1364 16. Installed latest exipick from John Jetmore.
1366 17. If headers_add on a router specified a text string that was too long for
1367 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
1368 of string_sprintf() is now avoided.
1370 18. $message_body_size was not set (it was always zero) when running the DATA
1371 ACL and the local_scan() function.
1373 19. For the "mail" command in an Exim filter, no default was being set for
1374 the once_repeat time, causing a random time value to be used if "once" was
1375 specified. (If the value happened to be <= 0, no repeat happened.) The
1376 default is now 0s, meaning "never repeat". The "vacation" command was OK
1377 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
1378 (I found it when inspecting the code).
1380 20. There is now an overall timeout for performing a callout verification. It
1381 defaults to 4 times the callout timeout, which applies to individual SMTP
1382 commands during the callout. The overall timeout applies when there is more
1383 than one host that can be tried. The timeout is checked before trying the
1384 next host. This prevents very long delays if there are a large number of
1385 hosts and all are timing out (e.g. when the network connections are timing
1386 out). The value of the overall timeout can be changed by specifying an
1387 additional sub-option for "callout", called "maxwait". For example:
1389 verify = sender/callout=5s,maxwait=20s
1391 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
1392 to the end before writing, but this should make it even safer).
1394 22. Exim was forgetting that it had advertised PIPELINING for the second and
1395 subsequent messages on an SMTP connection. It was also not resetting its
1396 memory on STARTTLS and an internal HELO.
1398 23. When Exim logs an SMTP synchronization error within a session, it now
1399 records whether PIPELINING has been advertised or not.
1401 24. Added 3 instances of "(long int)" casts to time_t variables that were being
1402 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
1403 rather than long int.
1405 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
1407 26. Added the never_mail option to autoreply.
1413 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
1414 crash if the getsockname() call failed; this can happen if a connection is
1415 closed very soon after it is established. The problem was simply in the
1416 order in which certain operations were done, causing Exim to try to write
1417 to the SMTP stream before it had set up the file descriptor. The bug has
1418 been fixed by making things happen in the correct order.
1424 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
1425 before the connection was closed, thus losing the rejection response.
1427 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
1428 some early Solaris releases, but causes trouble in current releases where
1429 socklen_t is defined.
1431 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
1434 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
1437 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
1438 that releasing the top of it at the end releases what was used for sub-
1439 expansions (unless the block got too big). However, discard this block if
1440 the first thing is a variable or header, so that we can use its block when
1441 it is dynamic (useful for very large $message_headers, for example).
1443 6. Lookups now cache *every* query, not just the most recent. A new, separate
1444 store pool is used for this. It can be recovered when all lookup caches are
1445 flushed. Lookups now release memory at the end of their result strings.
1446 This has involved some general refactoring of the lookup sources.
1448 7. Some code has been added to the store_xxx() functions to reduce the amount
1449 of flapping under certain conditions.
1451 8. log_incoming_interface used to affect only the <= reception log lines. Now
1452 it causes the local interface and port to be added to several more SMTP log
1453 lines, for example "SMTP connection from", and rejection lines.
1455 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
1457 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
1459 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
1460 could be overwritten at the end of the current message (or the start of a
1461 new message if it was set in a HELO ACL). The value is now preserved for
1462 the duration of the SMTP connection.
1464 12. If a transport had a headers_rewrite setting, and a matching header line
1465 contained an unqualified address, that address was qualified, even if it
1466 did not match any rewriting rules. The underlying bug was that the values
1467 of the flags that permit the existence of unqualified sender and recipient
1468 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
1469 non-local messages, and by -bnq for local messages) were not being
1470 preserved with the message after it was received.
1472 13. When Exim was logging an SMTP synchronization error, it could sometimes log
1473 "next input=" as part of the text comprising the host identity instead of
1474 the correct text. The code was using the same buffer for two different
1475 strings. However, depending on which order the printing function evaluated
1476 its arguments, the bug did not always show up. Under Linux, for example, my
1477 test suite worked just fine.
1479 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
1480 doesn't work with some older versions of Perl. It has been changed to "my",
1481 which in any case is probably the better facility to use.
1483 15. A really picky compiler found some instances of statements for creating
1484 error messages that either had too many or two few arguments for the format
1487 16. The size of the buffer for calls to the DNS resolver has been increased
1488 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
1489 for addresses that have a lot of PTR records. This alleviates a problem; it
1490 does not fully solve it.
1492 17. A dnsdb lookup for PTR records that receives more data than will fit in the
1493 buffer now truncates the list and logs the incident, which is the same
1494 action as happens when Exim is looking up a host name and its aliases.
1495 Previously in this situation something unpredictable would happen;
1496 sometimes it was "internal error: store_reset failed".
1498 18. If a server dropped the connection unexpectedly when an Exim client was
1499 using GnuTLS and trying to read a response, the client delivery process
1500 crashed while trying to generate an error log message.
1502 19. If a "warn" verb in an ACL added multiple headers to a message in a single
1503 string, for example:
1505 warn message = H1: something\nH2: something
1507 the text was added as a single header line from Exim's point of view
1508 though it ended up OK in the delivered message. However, searching for the
1509 second and subsequent header lines using $h_h2: did not work. This has been
1510 fixed. Similarly, if a system filter added multiple headers in this way,
1511 the routers could not see them.
1513 20. Expanded the error message when iplsearch is called with an invalid key to
1514 suggest using net-iplsearch in a host list.
1516 21. When running tests using -bh, any delays imposed by "delay" modifiers in
1517 ACLs are no longer actually imposed (and a message to that effect is
1520 22. If a "gecos" field in a passwd entry contained escaped characters, in
1521 particular, if it contained a \" sequence, Exim got it wrong when building
1522 a From: or a Sender: header from that name. A second bug also caused
1523 incorrect handling when an unquoted " was present following a character
1524 that needed quoting.
1526 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
1527 was not being matched caselessly.
1529 24. Arranged for all hyphens in the exim.8 source to be escaped with
1532 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
1533 properly. Recipient callout cache records were still being keyed to include
1534 the sender, even when use_sender was set false. This led to far more
1535 callouts that were necessary. The sender is no longer included in the key
1536 when use_sender is false.
1538 26. Added "control = submission" modifier to ACLs.
1540 27. Added the ${base62d: operator to decode base 62 numbers.
1542 28. dnsdb lookups can now access SRV records.
1544 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
1545 the configuration file.
1547 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
1548 (-v) mode. This makes the output for a verbose queue run more intelligible.
1550 31. Added a use_postmaster feature to recipient callouts.
1552 32. Added the $body_zerocount variable, containing the number of binary zero
1553 bytes in the message body.
1555 33. The time of last modification of the "new" subdirectory is now used as the
1556 "mailbox time last read" when there is a quota error for a maildir
1559 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
1561 35. Added +ignore_unknown as a special item in host lists.
1563 36. Code for decoding IPv6 addresses in host lists is now included, even if
1564 IPv6 support is not being compiled. This fixes a bug in which an IPv6
1565 address was recognized as an IP address, but was then not correctly decoded
1566 into binary, causing unexpected and incorrect effects when compared with
1573 1. Very minor rewording of debugging text in manualroute to say "list of
1574 hosts" instead of "hostlist".
1576 2. If verify=header_syntax was set, and a header line with an unqualified
1577 address (no domain) and a large number of spaces between the end of the
1578 name and the colon was received, the reception process suffered a buffer
1579 overflow, and (when I tested it) crashed. This was caused by some obsolete
1580 code that should have been removed. The fix is to remove it!
1582 3. When running in the test harness, delay a bit after writing a bounce
1583 message to get a bit more predictability in the log output.
1585 4. Added a call to search_tidyup() just before forking a reception process. In
1586 theory, someone could use a lookup in the expansion of smtp_accept_max_
1587 per_host which, without the tidyup, could leave open a database connection.
1589 5. Added the variables $recipient_data and $sender_data which get set from a
1590 lookup success in an ACL "recipients" or "senders" condition, or a router
1591 "senders" option, similar to $domain_data and $local_part_data.
1593 6. Moved the writing of debug_print from before to after the "senders" test
1596 7. Change 4.31/66 (moving the time when the Received: is generated) caused
1597 problems for message scanning, either using a data ACL, or using
1598 local_scan() because the Received: header was not generated till after they
1599 were called (in order to set the time as the time of reception completion).
1600 I have revised the way this works. The header is now generated after the
1601 body is received, but before the ACL or local_scan() are called. After they
1602 are run, the timestamp in the header is updated.
1608 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
1609 before starting a queue runner without re-exec. This happened only when
1610 deliver_drop_privilege was set or when the Exim user was set to root. The
1611 effect of the bug was that timeouts during subsequent deliveries caused
1612 crashes instead of being properly handled. The handler is now left at its
1613 default (and expected) setting.
1615 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
1616 message, again when deliver_drop_privilege is set or Exim is run as root.
1617 The bug described in (1) was not present in this case, but the tidying up
1618 of the other signals was missing. I have made the two cases consistent.
1620 3. The ignore_target_hosts setting on a manualroute router was being ignored
1621 for hosts that were looked up using the /MX notation.
1623 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
1626 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
1627 operated on the sender address. After changing the $sender_address to <>
1628 for the sender address verify, Exim was re-instated it as the original
1629 (before rewriting) address, but remembering that it had rewritten it, so it
1630 wasn't rewriting it again. This bug also had the effect of breaking the
1631 sender address verification caching when the sender address was rewritten.
1633 6. The ignore_target_hosts option was being ignored by the ipliteral router.
1634 This has been changed so that if the ip literal address matches
1635 ignore_target_hosts, the router declines.
1637 7. Added expansion conditions match_domain, match_address, and match_local_
1638 part (NOT match_host).
1640 8. The placeholder for the Received: header didn't have a length field set.
1642 9. Added code to Exim itself and to exim_lock to test for a specific race
1643 condition that could lead to file corruption when using MBX delivery. The
1644 issue is with the lockfile that is created in /tmp. If this file is removed
1645 after a process has opened it but before that process has acquired a lock,
1646 there is the potential for a second process to recreate the file and also
1647 acquire a lock. This could lead to two Exim processes writing to the file
1648 at the same time. The added code performs the same test as UW imapd; it
1649 checks after acquiring the lock that its file descriptor still refers to
1650 the same named file.
1652 10. The buffer for building added header lines was of fixed size, 8192 bytes.
1653 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
1656 11. Added the smtp_active_hostname option. If used, this will typically be made
1657 to depend on the incoming interface address. Because $interface_address is
1658 not set up until the daemon has forked a reception process, error responses
1659 that can happen earlier (such as "too many connections") no longer contain
1662 12. If an expansion in a condition on a "warn" statement fails because a lookup
1663 defers, the "warn" statement is abandoned, and the next ACL statement is
1664 processed. Previously this caused the whole ACL to be aborted.
1666 13. Added the iplsearch lookup type.
1668 14. Added ident_timeout as a log selector.
1670 15. Added tls_certificate_verified as a log selector.
1672 16. Added a global option tls_require_ciphers (compare the smtp transport
1673 option of the same name). This controls incoming TLS connections.
1675 17. I finally figured out how to make tls_require_ciphers do a similar thing
1676 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
1677 before starting the TLS session.
1679 18. Tabs are now shown as \t in -bP output.
1681 19. If the log selector return_path_on_delivery was set, Exim crashed when
1682 bouncing a message because it had too many Received: header lines.
1684 20. If two routers both had headers_remove settings, and the first one included
1685 a superfluous trailing colon, the final name in the first list and the
1686 first name in the second list were incorrectly joined into one item (with a
1687 colon in the middle).
1693 1. Added -C and -D options to the exinext utility, mainly to make it easier
1694 to include in the automated testing, but these could be helpful when
1695 multiple configurations are in use.
1697 2. The exinext utility was not formatting the output nicely when there was
1698 an alternate port involved in the retry record key, nor when there was a
1699 message id as well (for retries that were specific to a specific message
1700 and a specific host). It was also confused by IPv6 addresses, because of
1701 the additional colons they contain. I have fixed the IPv4 problem, and
1702 patched it up to do a reasonable job for IPv6.
1704 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
1705 delivery, the log line now contains "pipelined" if PIPELINING was used.
1707 4. An SMTP transport process used to panic and die if the bind() call to set
1708 an explicit outgoing interface failed. This has been changed; it is now
1709 treated in the same way as a connect() failure.
1711 5. A reference to $sender_host_name in the part of a conditional expansion
1712 that was being skipped was still causing a DNS lookup. This no longer
1715 6. The def: expansion condition was not recognizing references to header lines
1716 that used bh_ and bheader_.
1718 7. Added the _cache feature to named lists.
1720 8. The code for checking quota_filecount in the appendfile transport was
1721 allowing one more file than it should have been.
1723 9. For compatibility with Sendmail, the command line option
1731 and sets the incoming protocol and host name (for trusted callers). The
1732 host name and its colon can be omitted when only the protocol is to be set.
1733 Note the Exim already has two private options, -pd and -ps, that refer to
1734 embedded Perl. It is therefore impossible to set a protocol value of "d" or
1735 "s", but I don't think that's a major issue.
1737 10. A number of refactoring changes to the code, none of which should affect
1740 (a) The number of logging options was getting close to filling up the
1741 32-bit word that was used as a bit map. I have split them into two classes:
1742 those that are passed in the argument to log_write(), and those that are
1743 only ever tested independently outside of that function. These are now in
1744 separate 32-bit words, so there is plenty of room for expansion again.
1745 There is no change in the user interface or the logging behaviour.
1747 (b) When building, for example, log lines, the code previously used a
1748 macro that called string_cat() twice, in order to add two strings. This is
1749 not really sufficiently general. Furthermore, there was one instance where
1750 it was actually wrong because one of the argument was used twice, and in
1751 one call a function was used. (As it happened, calling the function twice
1752 did not affect the overall behaviour.) The macro has been replaced by a
1753 function that can join an arbitrary number of extra strings onto a growing
1756 (c) The code for expansion conditions now uses a table and a binary chop
1757 instead of a serial search (which was left over from when there were very
1758 few conditions). Also, it now recognizes conditions like "pam" even when
1759 the relevant support is not compiled in: a suitably worded error message is
1760 given if an attempt is made to use such a condition.
1762 11. Added ${time_interval:xxxxx}.
1764 12. A bug was causing one of the ddress fields not to be passed back correctly
1765 from remote delivery subprocesses. The field in question was not being
1766 subsequently used, so this caused to problems in practice.
1768 13. Added new log selectors queue_time and deliver_time.
1770 14. Might have fixed a bug in maildirsizefile handling that threw up
1771 "unexpected character" debug warnings, and recalculated the data
1772 unnecessarily. In any case, I expanded the warning message to give more
1775 15. Added the message "Restricted characters in address" to the statements in
1776 the default ACL that block characters like @ and % in local parts.
1778 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
1779 Three changes have been made:
1781 (a) There was a serious bug; a negative response to MAIL caused the whole
1782 recipient domain to be cached as invalid, thereby blocking all messages
1783 to all local parts at the same domain, from all senders. This bug has
1784 been fixed. The domain is no longer cached after a negative response to
1785 MAIL if the sender used is not empty.
1787 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
1790 (c) A new callout option, "use_sender" has been added for people who want
1791 the modified behaviour.
1797 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
1800 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
1801 indeed breaks things for older releases.
1803 3. Added additional logging to the case where there is a problem reading data
1804 from a filter that is running in a subprocess using a pipe, in order to
1805 try to track down a specific problem.
1807 4. Testing facility fudge: when running in the test harness and attempting
1808 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
1809 getting "No route to host". Convert this to a timeout.
1811 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
1814 6. Some OS don't have socklen_t but use size_t instead. This affects the
1815 fifth argument of getsockopt() amongst other things. This is now
1816 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
1817 can be set for individual OS. I have set it for SunOS5, OSF1, and
1818 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
1819 some earlier ones do not.
1821 7. Change 4.30/15 was not doing the test caselessly.
1823 8. The standard form for an IPv6 address literal was being rejected by address
1824 parsing in, for example, MAIL and RCPT commands. An example of this kind of
1825 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
1826 this, as well as the form without the "IPv6" on the front (but only when
1827 address literals are enabled, of course).
1829 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
1831 10. Exim crashed if a message with an empty sender address specified by -f
1832 encountered a router with an errors_to setting. This could be provoked only
1833 by a command such as
1837 where an empty string was supplied; "<>" did not hit this bug.
1839 11. Installed PCRE release 4.5.
1841 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
1842 remained set. It is now erased.
1844 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
1845 times from message ids (which are base 36 rather than the normal 62).
1847 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
1848 were being counted as actual protocol errors, and logged if the log
1849 selector +smtp_protocol_error was set. One cannot be perfect in this test,
1850 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
1851 and DATA following a set of rejected RCPTs do not count as protocol errors.
1852 In other words, Exim assumes they were pipelined, though this may not
1853 actually be the case. Of course, in all cases the client gets an
1854 appropriate error code.
1856 15. If a lookup fails in an ACL condition, a message about the failure may
1857 be available; it is used if testing the ACL cannot continue, because most
1858 such messages specify what the cause of the deferral is. However, some
1859 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
1860 that caused an old message to be retained and used if a later statement
1861 caused a defer, replacing the real cause of the deferral.
1863 16. If an IP address had so many PTR records that the DNS lookup buffer
1864 was not large enough to hold them, Exim could crash while trying to process
1865 the truncated data. It now detects and logs this case.
1867 17. Further to 4.21/58, another change has been made: if (and only if) the
1868 first line of a message (the first header line) ends with CRLF, a bare LF
1869 in a subsequent header line has a space inserted after it, so as not to
1870 terminate the header.
1872 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
1873 unnecessarily, used atoi() instead of strtol(), and didn't check the
1874 termination when getting file sizes from file names by regex.
1876 19. Completely re-implemented the support for maildirsize files, in the light
1877 of a number of problems with the previous contributed implementation
1878 (4.30/29). In particular:
1880 . If the quota is zero, the maildirsize file is maintained, but no quota is
1883 . If the maildir directory does not exist, it is created before any attempt
1884 to write a maildirsize file.
1886 . The quota value in the file is just a cache; if the quota is changed in
1887 the transport, the new value overrides.
1889 . A regular expression is available for excluding directories from the
1892 20. The autoreply transport checks the characters in options that define the
1893 message's headers; it allows continued headers, but it was checking with
1894 isspace() after an embedded newline instead of explicitly looking for a
1897 21. If all the "regular" hosts to which an address was routed had passed their
1898 expiry times, and had not reached their retry times, the address was
1899 bounced, even if fallback hosts were defined. Now Exim should go on to try
1902 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
1903 equivalent code in the SMTP transport. Some hosts send humungous responses
1904 to HELO/EHLO, more than 1024 it seems.
1906 23. Refactoring: code in filter.c used (void *) for "any old type" but this
1907 gives compiler warnings in some environments. I've now done it "properly",
1910 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
1911 (because of problems with the built-in one) was declared to return uschar *
1912 instead of char *, causing compiler failure.
1914 25. Fixed a file descriptor leak when processing alias/forward files.
1916 26. Fixed a minor format string issue in dbfn.c.
1918 27. Typo in exim.c: ("dmbnz" for "dbmnz").
1920 28. If a filter file refered to $h_xxx or $message_headers, and the headers
1921 contained RFC 2047 "words", Exim's memory could, under certain conditions,
1924 29. When a sender address is verified, it is cached, to save repeating the test
1925 when there is more than one recipient in a message. However, when the
1926 verification involves a callout, it is possible for different callout
1927 options to be set for different recipients. It is too complicated to keep
1928 track of this in the cache, so now Exim always runs a verification when a
1929 callout is required, relying on the callout cache for the optimization.
1930 The overhead is duplication of the address routing, but this should not be
1933 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
1934 to be verified with callout=postmaster, and the main callout worked but the
1935 postmaster check failed, the verification correctly failed. However, if a
1936 subsequent RCPT command asked for sender verification *without* the
1937 postmaster check, incorrect caching caused this verification also to fail,
1940 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
1941 it was not caching the DNS options (qualify_single, search_parents) that
1942 were used when the lookup failed. A subsequent lookup with different
1943 options therefore always gave the same answer, though there were cases
1944 where it should not have. (Example: a "domains = !$mx_any" option on a
1945 dnslookup router: the "domains" option is always processed without any
1946 widening, but the router might have qualify_single set.) Now Exim uses the
1947 cached value only when the same options are set.
1949 32. Added John Jetmore's "exipick" utility to the distribution.
1951 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
1952 than a timeout (e.g. a certificate is required, and is not provided), an
1953 Exim server now closes the connection immediately. Previously it waited for
1954 the client to close - but if the client is SSL, it seems that they each
1955 wait for each other, leading to a delay before one of them times out.
1957 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
1958 maintained 0.8.x compatibility because I don't think many are using it, and
1959 it is clearly obsolete.
1961 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
1964 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
1965 client certificate was expired. A simple patch fixes this, though I don't
1966 understand the full logic of why the verify callback is called multiple
1969 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
1970 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
1971 which causes problems with some clients (such as the Certicom SSL Plus
1972 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
1973 disables the coutermeasure allowing Eudora to connect."
1975 38. Exim was not checking that a write() to a log file succeeded. This could
1976 lead to Bad Things if a log got too big, in particular if it hit a file
1977 size limit. Exim now panics and dies if it cannot write to a log file, just
1978 as it does if it cannot open a log file.
1980 39. Modified OS/Makefile-Linux so that it now contains
1982 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
1984 The two -D definitions ensure that Exim is compiled with large file
1985 support, which makes it possible to handle log files that are bigger than
1988 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
1989 instance) a domain was checked against a named list that involved a lookup,
1990 causing $domain_data to be set, then another domain was checked against the
1991 same list, then the first domain was re-checked, the value of $domain_data
1992 after the final check could be wrong. In particular, if the second check
1993 failed, it could be set empty. This bug probably also applied to
1996 41. The strip_trailing_dot option was not being applied to the address given
1997 with the -f command-line option.
1999 42. The code for reading a message's header from the spool was incrementing
2000 $received_count, but never initializing it. This meant that the value was
2001 incorrect (doubled) while delivering a message in the same process in which
2002 it was received. In the most common configuration of Exim, this never
2003 happens - a fresh exec is done - but it can happen when
2004 deliver_drop_privilege is set.
2006 43. When Exim logs an SMTP synchronization error - client data sent too soon -
2007 it now includes up to 150 characters of the unexpected data in the log
2010 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
2011 and building data strings. The size of both of these buffers was 10 000
2012 bytes - far larger than anybody would *ever* want, thought I. Needless to
2013 say, somebody hit the limit. I have increased the maximum line length to
2014 20 000 and the maximum data length of concatenated lines to 100 000. I have
2015 also fixed two bugs, because there was no checking on these buffers. Tsk,
2016 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
2017 buffer is too small.
2019 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
2020 lsearch lookups. Now it does.
2022 46. When parsing a route_list item in a manualroute router, a fixed-length
2023 buffer was used for the list of hosts. I made this 1024 bytes long,
2024 thinking that nobody would ever have a list of hosts that long. Wrong.
2025 Somebody had a whole pile of complicated expansion conditions, and the
2026 string was silently truncated, leading to an expansion error. It turns out
2027 that it is easier to change to an unlimited length (owing to other changes
2028 that have happened since this code was originally written) than to build
2029 structure for giving a limitation error. The length of the item that
2030 expands into the list of hosts is now unlimited.
2032 47. The lsearch lookup could not handle data where the length of text line was
2033 more than 4095 characters. Such lines were truncated, leading to shortened
2034 data being returned. It should now handle lines of any length.
2036 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
2037 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
2040 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
2041 replacing strings such as BIN_DIRECTORY in a source file: the replacement
2042 no longer happens in comment lines. A list of replacements is now placed
2043 at the head of all of the source files, except those whose only change is
2044 to replace PERL_COMMAND in the very first #! line.
2046 50. Replaced the slow insertion sort in queue.c, for sorting the list of
2047 messages on the queue, with a bottom-up merge sort, using code contributed
2048 by Michael Haardt. This should make operations like -bp somewhat faster on
2049 large queues. It won't affect queue runners, except when queue_run_in_order
2052 51. Installed eximstats 1.31 in the distribution.
2054 52. Added support for SRV lookups to the dnslookup router.
2056 53. If an ACL referred to $message_body or $message_body_end, the value was not
2057 reset for any messages that followed in the same SMTP session.
2059 54. The store-handling optimization for building very long strings was not
2060 differentiating between the different store pools. I don't think this
2061 actually made any difference in practice, but I've tidied it.
2063 55. While running the routers to verify a sender address, $sender_address
2064 was still set to the sender address. This is wrong, because when routing to
2065 send a bounce to the sender, it would be empty. Therefore, I have changed
2066 it so that, while verifying a sender address, $sender_address is set to <>.
2067 (There is no change to what happens when verifying a recipient address.)
2069 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
2070 target A or AAAA records (if not already returned) without resetting the
2071 qualify_single or search_parents options of the DNS resolver. These are
2072 inappropriate in this case because the targets of MX and SRV records must
2073 be FQDNs. A broken DNS record could cause trouble if it happened to have a
2074 target that, when qualified, matched something in the local domain. These
2075 two options are now turned off when doing these lookups.
2077 57. It seems that at least some releases of Reiserfs (which does not have the
2078 concept of a fixed number of inodes) returns zero and not -1 for the
2079 number of available inodes. This interacted badly with check_spool_inodes,
2080 which assumed that -1 was the "no such thing" setting. What I have done is
2081 to check that the total number of inodes is greater than zero before doing
2082 the test of how many are available.
2084 58. When a "warn" ACL statement has a log_message modifier, the message is
2085 remembered, and not repeated. This is to avoid a lot of repetition when a
2086 message has many recipients that cause the same warning to be written.
2087 Howewer, Exim was preserving the list of already written lines for an
2088 entire SMTP session, which doesn't seem right. The memory is now reset if a
2089 new message is started.
2091 59. The "rewrite" debugging flag was not showing the result of rewriting in the
2092 debugging output unless log_rewrite was also set.
2094 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
2095 of (int)(handle) when we know that handle contains (void *)(-1).
2097 61. The Exim daemon panic-logs an error return when it closes the incoming
2098 connection. However "connection reset by peer" seems to be common, and
2099 isn't really an error worthy of noting specially, so that particular error
2102 62. When Exim is trying to find all the local interfaces, it used to panic and
2103 die if the ioctl to get the interface flags failed. However, it seems that
2104 on at least one OS (Solaris 9) it is possible to have an interface that is
2105 included in the list of interfaces, but for which you get a failure error
2106 for this call. This happens when the interface is not "plumbed" into a
2107 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
2108 failure of the "get flags" call assumes that the interface is down.
2110 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
2111 makes life easier for people who are doing arithmetic on fields extracted
2112 from dates, where you often get leading zeros that should not be
2113 interpreted as octal.
2115 64. Added qualify_domain to the redirect router, to override the global
2118 65. If a pathologically long header line contained very many addresses (the
2119 report of this problem mentioned 10 000) and each of them was rewritten,
2120 Exim could use up a very large amount of memory. (It kept on making new
2121 copies of the header line as it rewrote, and never released the old ones.)
2122 At the expense of a bit more processing, the header rewriting function has
2123 been changed so that it no longer eats memory in this way.
2125 66. The generation of the Received: header has been moved from the time that a
2126 message starts to be received, to the time that it finishes. The timestamp
2127 in the Received: header should now be very close to that of the <= log
2128 line. There are two side-effects of this change:
2130 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
2131 logged header lines no longer include the local Received: line, because
2132 it has not yet been created. The same applies to a copy of the message
2133 that is returned to a non-SMTP sender when a message is rejected.
2135 (b) When a filter file is tested using -bf, no additional Received: header
2136 is added to the test message. After some thought, I decided that this
2139 This change does not affect the value of $received_for. It is still set
2140 after address rewriting, but before local_scan() is called.
2142 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
2144 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
2145 gave an unhelpful panic error message, and a defer error. I have managed to
2146 change this behaviour so that it now rejects any supplied certificate,
2147 which seems right, as the list of acceptable certificates is empty.
2149 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
2150 gave an unhelpful defer error. I have not managed to make this reject any
2151 supplied certificates, but the error message it gives is "no certificate
2152 supplied", which is not helpful.
2154 70. exigrep's output now also includes lines that are not associated with any
2155 message, but which match the given pattern. Implemented by a patch from
2156 Martin Sluka, which also tidied up the Perl a bit.
2158 71. Recipient callout verification, like sender verification, was using <> in
2159 the MAIL FROM command. This isn't really the right thing, since the actual
2160 sender may affect whether the remote host accepts the recipient or not. I
2161 have changed it to use the actual sender in the callout; this means that
2162 the cache record is now keyed on a recipient/sender pair, not just the
2163 recipient address. There doesn't seem to be a real danger of callout loops,
2164 since a callout by the remote host to check the sender would use <>.
2165 [SEE ABOVE: changed after hitting problems.]
2167 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
2168 temporary errors. However, in the case of such a code being given after
2169 the end of a data transmission (i.e. after ".") Exim was failing to write
2170 a retry record for the message. (Yes, there was some broken host that was
2171 actually sending 8xx at this point.)
2173 73. An unknown lookup type in a host list could cause Exim to panic-die when
2174 the list was checked. (An example that provoked this was putting <; in the
2175 middle of a list instead of at the start.) If this happened during a DATA
2176 ACL check, a -D file could be left lying around. This kind of configuration
2177 error no longer causes Exim to die; instead it causes a defer errror. The
2178 incident is still logged to the main and panic logs.
2180 74. Buglet left over from Exim 3 conversion. The message "too many messages
2181 in one connection" was written to the rejectlog but not the mainlog, except
2182 when address rewriting (yes!) was being logged.
2184 75. Added write_rejectlog option.
2186 76. When a system filter was run not as root (that is, when system_filter_user
2187 was set), the values of the $n variables were not being returned to the
2188 main process; thus, they were not subsequently available in the $sn
2191 77. Added +return_path_on_delivery log selector.
2193 78. A connection timeout was being treated differently from recipients deferred
2194 when testing hosts_max_try with a message that was older than the host's
2195 retry timeout. (The host should not be counted, thus allowing all hosts to
2196 be tried at least once before bouncing.) This may have been the cause of an
2197 occasionally reported bug whereby a message would remain on the queue
2198 longer than the retry timeout, but would be bounced if a delivery was
2199 forced. I say "may" because I never totally pinned down the problem;
2200 setting up timeout/retry tests is difficult. See also the next item.
2202 79. The ultimate address timeout was not being applied to errors that involved
2203 a combination of host plus message (for example, a timeout on a MAIL
2204 command). When an address resolved to a number of possible hosts, and they
2205 were not all tried for each delivery (e.g. because of hosts_max_try), a
2206 message could remain on the queue longer than the retry timeout.
2208 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
2211 81. Fixed an obscure SMTP outgoing bug which required at least the following
2212 conditions: (a) there was another message waiting for the same server;
2213 (b) the server returned 5xx to all RCPT commands in the first message so
2214 that the message was not completed; (c) the server dropped the connection
2215 or gave a negative response to the RSET that Exim sends to abort the
2216 transaction. The observed case was a dropped connection after DATA that had
2217 been sent in pipelining mode. That is, the server had advertised PIPELINING
2218 but was not implementing it correctly. The effect of the bug was incorrect
2219 behaviour, such as trying another host, and this could lead to a crash.
2225 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
2226 and daemon.c were passed as pointers to ints; they should have been
2227 pointers to socklen_t variables (which are typically unsigned ints).
2229 2. Some signed/unsigned type warnings in the os.c file for Linux have been
2232 3. Fixed a really odd bug that affected only the testing scheme; patching a
2233 certain fixed string in the binary changed the value of another string that
2234 happened to be identical to the end of the original first string.
2236 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
2237 name", it returns that address as the IP address. On some operating
2238 systems (e.g. Solaris), it also passes back the IP address string as the
2239 "host name". However, on others (e.g. Linux), it passes back an empty
2240 string. Exim wasn't checking for this, and was changing the host name to an
2241 empty string, assuming it had been canonicized.
2243 5. Although rare, it is permitted to have more than one PTR record for a given
2244 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
2245 all the names associated with an address, because they do in Solaris.
2246 However, it seems that they do not in Linux for data that comes from the
2247 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
2248 I found this out when I moved to a new Linux workstation and tried to run
2249 the Exim test suite.
2251 To get round this problem I have changed the code so that it now does its
2252 own call to the DNS to look up PTR records when searching for a host name.
2253 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
2254 addresses that are only in /etc/hosts are still found.
2256 This behaviour is, however, controlled by an option called host_lookup_
2257 order, which defaults to "bydns:byaddr". If people want to use the other
2258 order, or indeed, just use one or the other means of lookup, they can
2259 specify it in this variable.
2261 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
2262 some operating systems, this comes back from gethostbyaddr() as an empty
2263 string, and this is what Exim used to test for. However, it seems that in
2264 other systems, "." is yielded. Exim now tests for this case too.
2266 7. The values of check_spool_space and check_log_space are now held internally
2267 as a number of kilobytes instead of an absolute number of bytes. If a
2268 numbers is specified without 'K' or 'M', it is rounded up to the nearest
2269 kilobyte. This means that much larger values can be stored.
2271 8. Exim monitor: an attempt to get the action menu when not actually pointing
2272 at a message produces an empty menu entitled "No message selected". This
2273 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
2274 no entries in it ("Shell widget menu has zero width and/or height"). So I
2275 have added a single, blank menu entry in this case.
2277 9. Added ${quote_local_part.
2279 10. MIME decoding is now applied to the contents of Subject: header lines when
2282 11. Now that a reference to $sender_host_address automatically causes a reverse
2283 lookup to occur if necessary (4.13/18), there is no need to arrange for a
2284 host lookup before query-style lookups in lists that might use this
2285 variable. This has therefore been abolished, and the "net-" prefix is no
2286 longer necessary for query-style lookups.
2288 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
2289 have been a typo for LFLAGS, so it has been changed.
2291 13. The install script calls Exim with "-C /dev/null" in order to find the
2292 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
2293 to be output. Howeve, since Exim outputs its version number before the
2294 error, it didn't break the script. It just looked ugly. I fixed this by
2295 always allowing "-C /dev/null" if the caller is root.
2297 14. Ignore overlarge ACL variable number when reading spool file - insurance
2298 against a later release with more variables having written the file.
2300 15. The standard form for an IPv6 address literal was being rejected by EHLO.
2301 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
2302 this, as well as the form without the "IPv6" on the front.
2304 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
2305 OS/Makefile-Darwin file.
2307 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
2308 only to LDAP libraries that do not have LDAP_OPT_DEREF.
2310 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
2311 variable. However, some OS use ints for inodes. I've added cast to long int
2312 to get rid of the compiler warning.
2314 19. I had forgotten to lock out "/../" in configuration file names when
2315 ALT_CONFIG_PREFIX was set.
2317 20. Routers used for verification do not need to specify transports. However,
2318 if such a router generated a host list, and callout was configured, Exim
2319 crashed, because it could not find a port number from the (non-existent)
2320 transport. It now assumes port 25 in this circumstance.
2322 21. Added the -t option to exigrep.
2324 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
2325 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
2326 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
2327 them is set, LOOKUP_LSEARCH is forced.
2329 23. "exim -bV" now outputs a list of lookups that are included in the binary.
2331 24. Added sender and host information to the "rejected by local_scan()" log
2332 line; previously there was no indication of these.
2334 25. Added .include_if_exists.
2336 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
2337 Linux. It turns out that not all file systems support this. Apparently some
2338 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
2339 do it.) To cope with this, the error EINVAL, which means that sync-ing is
2340 not supported on the file descriptor, is now ignored when Exim is trying to
2341 sync a directory. This applies only to Linux.
2343 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
2345 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
2346 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
2349 29. Added support for maildirsize files from supplied patch (modified a bit).
2351 30. The use of :fail: followed by an empty string could lead Exim to respond to
2352 sender verification failures with (e.g.):
2354 550 Verification failed for <xxx>
2355 550 Sender verify failed
2357 where the first response line was missing the '-' that indicates it is not
2358 the final line of the response.
2360 31. The loop for finding the name of the user that called Exim had a hardwired
2361 limit of 10; it now uses the value of finduser_retries, which is used for
2362 all other user lookups.
2364 32. Added $received_count variable, available in data and not_smtp ACLs, and at
2367 33. Exim was neglecting to zero errno before one call of strtol() when
2368 expanding a string and expecting an integer value. On some systems this
2369 resulted in spurious "integer overflow" errors. Also, it was casting the
2370 result into an int without checking.
2372 34. Testing for a connection timeout using "timeout_connect" in the retry rules
2373 did not work. The code looks as if it has *never* worked, though it appears
2374 to have been documented since at least releast 1.62. I have made it work.
2376 35. The "timeout_DNS" error in retry rules, also documented since at least
2377 1.62, also never worked. As it isn't clear exactly what this means, and
2378 clearly it isn't a major issue, I have abolished the feature by treating it
2379 as "timeout", and writing a warning to the main and panic logs.
2381 36. The display of retry rules for -brt wasn't always showing the error code
2384 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
2385 timeout_connect_A, timeout_connect_MX.
2387 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
2388 to the empty sender.
2390 39. The daemon was not analysing the content of -oX till after it had closed
2391 stderr and disconnected from the controlling terminal. This meant that any
2392 syntax errors were only noted on the panic log, and the return code from
2393 the command was 0. By re-arranging the code a little, I've made the
2394 decoding happen first, so such errors now appear on stderr, and the return
2395 code is 1. However, the actual setting up of the sockets still happens in
2396 the disconnected process, so errors there are still only recorded on the
2399 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
2400 connections (as happens on some IP stacks) was logged at start up time as
2401 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
2402 it from "IPv6 and IPv4", which means that two separate sockets are being
2405 41. The debug output for gethostbyname2() or getipnodebyname() failures now
2406 says whether AF_INET or AF_INET6 was passed as an argument.
2408 42. Exiwhat output was messed up when time zones were included in log
2411 43. Exiwhat now gives more information about the daemon's listening ports,
2412 and whether -tls-on-connect was used.
2414 44. The "port" option of the smtp transport is now expanded.
2416 45. A "message" modifier in a "warn" statement in a non-message ACL was being
2417 silently ignored. Now an error message is written to the main and panic
2420 46. There's a new ACL modifier called "logwrite" which writes to a log file
2421 as soon as it is encountered.
2423 47. Added $local_user_uid and $local_user_gid at routing time.
2425 48. Exim crashed when trying to verify a sender address that was being
2428 49. Exim was recognizing only a space character after ".include". It now also
2429 recognizes a tab character.
2431 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
2432 extracting the relevant information from the specification. The man page no
2433 longer contains scrambled data for the -d option, and I've added a section
2434 at the front about calling Exim under different names.
2436 51. Added "extra_headers" argument to the "mail" command in filter files.
2438 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
2441 53. Installed eximstats 1.29.
2443 54. Added transport_filter_timeout as a generic transport option.
2445 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
2446 Cc: header lines. This was required by RFC 822, but it not required by RFC
2449 56. Exim used to add From:, Date:, and Message-Id: header lines to any
2450 incoming messages that did not have them. Now it does so only if the
2451 message originates locally, that is, if there is no associated remote host
2452 address. When Resent- header lines are present, this applies to the Resent-
2453 lines rather than the non-Resent- lines.
2455 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
2456 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
2458 58. Messages for configuration errors now include the name of the main
2459 configuration file - useful now that there may be more than one file in a
2460 list (.included file names were always shown).
2462 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
2463 for those rare installations that do not start the daemon as root or run it
2464 setuid root. I've cut out the call to initgroups() if the daemon is not
2467 60. The Exim user and group can now be bound into the binary as text strings
2468 that are looked up at the start of Exim's processing.
2470 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
2472 62. Added $mailstore_basename variable.
2474 63. Installed patch to sieve.c from Michael Haardt.
2476 64. When Exim failed to open the panic log after failing to open the main log,
2477 the original message it was trying to log was written to stderr and debug
2478 output, but if they were not available (the usual case in production), it
2479 was lost. Now it is written to syslog before the two lines that record the
2480 failures to open the logs.
2482 65. Users' Exim filters run in subprocesses under the user's uid. It is
2483 possible for a "deliver" command or an alias in a "personal" command to
2484 provoke an address rewrite. If logging of address rewriting is configured,
2485 this fails because the process is not running as root or exim. There may be
2486 a better way of dealing with this, but for the moment (because 4.30 needs
2487 to be released), I have disabled address rewrite logging when running a
2488 filter in a non-root, non-exim process.
2494 1. The buildconfig auxiliary program wasn't quoting the value set for
2495 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
2496 not defined. This bug was masked in 4.22 by the effect that was fixed in
2499 2. Some messages that were rejected after a message id was allocated were
2500 shown as "incomplete" by exigrep. It no longer does this for messages that
2501 are rejected by local_scan() or the DATA or non-SMTP ACLs.
2503 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
2504 have allow_domain_literals set, the ID did not get logged in the <= line.
2505 Domain literals are now always recognized in Message-ID: header lines.
2507 4. The first argument for a ${extract expansion item is the key name or field
2508 number. Leading and trailing spaces in this item were not being ignored,
2509 causing some misleading effects.
2511 5. When deliver_drop_privilege was set, single queue runner processes started
2512 manually (i.e. by the command "exim -q") or by the daemon (which uses the
2513 same command in the process it spins off) were not dropping privilege.
2515 6. When the daemon running as "exim" started a queue runner, it always
2516 re-executed Exim in the spun-off process. This is a waste of effort when
2517 deliver_drop_privilege is set. The new process now just calls the
2518 queue-runner function directly.
2524 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
2527 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
2528 ignored. Though the use of -oP was forcing the writing of a pid file, it
2529 was always written to the default place.
2531 3. If the message "no IP address found for host xxxx" is generated during
2532 incoming verification, it is now followed by identification of the incoming
2533 connection (so you can more easily find what provoked it).
2535 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
2537 5. Added some features to "harden" Exim a bit more against certain attacks:
2539 (a) There is now a build-time option called FIXED_NEVER_USERS that can
2540 be put in Local/Makefile. This is like the never_users runtime option,
2541 but it cannot be overridden. The default setting is "root".
2543 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
2544 prefix string with which any file named in a -C command line option
2547 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
2548 is retained for -C and -D only if the caller of Exim is root. Without
2549 it, the exim user may also use -C and -D and retain privilege.
2551 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
2552 command line option is disabled.
2554 6. Macro names set by the -D option must start with an upper case letter, just
2555 like macro names defined in the configuration file.
2557 7. Added "dereference=" facility to LDAP.
2559 8. Two instances of the typo "uknown" in the source files are fixed.
2561 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
2562 the Configure-Makefile script screwed up while processing it.
2564 10. Incorporated PCRE 4.4.
2566 11. The SMTP synchronization check was not operating right at the start of an
2567 SMTP session. For example, it could not catch a HELO sent before the client
2568 waited for the greeting. There is now a check for outstanding input at the
2569 point when the greeting is written. Because of the duplex, asynchronous
2570 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
2571 way, but not yet received, when the check is performed.
2573 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
2574 on TCP/IP sockets, because this apparently causes some broken clients to
2577 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
2578 unchanged) from the Cygwin maintainer.
2580 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
2581 was supported instead of testing for mbx. Effectively a typo.
2583 15. The spa authenticator server code was not checking that the input it
2584 received was valid base64.
2586 16. The debug output line for the "set" modifier in ACLs was not showing the
2587 name of the variable that was being set.
2589 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
2591 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
2592 lookup on its own. Something else was needed to trigger the lookup. For
2593 example, a match in host_lookup or the need for a host name in a host list.
2594 Now, if $sender_host_name is referenced and the host name has not yet been
2595 looked up, a lookup is performed. If the lookup fails, the variable remains
2596 empty, and $host_lookup_failed is set to "1".
2598 19. Added "eqi" as a case-independent comparison operator.
2600 20. The saslauthd authentication condition could segfault if neither service
2601 nor realm was specified.
2603 21. If an overflowing value such as "2048M" was set for message_size_limit, the
2604 error message that was logged was misleading, and incoming SMTP
2605 connections were dropped. The message is now more accurate, and temporary
2606 errors are given to SMTP connections.
2608 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
2609 (except RSET) with a 421 error, until QUIT is received. However, it was
2610 failing to send a response to QUIT.
2612 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
2613 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
2614 ACL is now run after the helo_try_verify_hosts code.
2616 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
2617 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
2618 case-independent, so other case variants are also recognized). Apparently
2619 some systems use these upper case variants.
2621 25. If more than two messages were waiting for the same host, and a transport
2622 filter was specified for the transport, Exim sent two messages over the
2623 same TCP/IP connection, and then failed with "socket operation on non-
2624 socket" when it tried to send the third.
2626 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
2628 27. The extern definition of crypt16() in expand.c was not being excluded when
2629 the OS had its own crypt16() function.
2631 28. Added bounce_return_body as a new option, and bounce_return_size_limit
2632 as a preferred synonym for return_size_limit, both as an option and as an
2635 29. Added LIBS=-liconv to OS/Makefile-OSF1.
2637 30. Changed the default configuration ACL to relax the local part checking rule
2638 for addresses that are not in any local domains. For these addresses,
2639 slashes and pipe symbols are allowed within local parts, but the sequence
2640 /../ is explicitly forbidden.
2642 31. SPA server authentication was not clearing the challenge buffer before
2645 32. log_message in a "warn" ACL statement was writing to the reject log as
2646 well as to the main log, which contradicts the documentation and doesn't
2647 seem right (because no rejection is happening). So I have stopped it.
2649 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
2650 However, I am unable to do any testing of this.
2652 34. Fixed an infelicity in the appendfile transport. When checking directories
2653 for a mailbox, to see if any needed to be created, it was accidentally
2654 using path names with one or more superfluous leading slashes; tracing
2655 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
2657 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
2658 contents are added to the log line that is written for every discarded
2659 recipient. (Previously a log_message setting was ignored.)
2661 36. The ${quote: operator now quotes the string if it is empty.
2663 37. The install script runs exim in order to find its version number. If for
2664 some reason other than non-existence or emptiness, which it checks, it
2665 could not run './exim', it was installing it with an empty version number,
2666 i.e. as "exim-". This error state is now caught, and the installation is
2669 38. An argument was missing from the function that creates an error message
2670 when Exim fails to connect to the socket for saslauthd authentication.
2671 This could cause Exim to crash, or give a corrupted message.
2673 39. Added isip, isip4, and isip6 to ${if conditions.
2675 40. The ACL variables $acl_xx are now saved with the message, and can be
2676 accessed later in routers, transports, and filters.
2678 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
2679 strings in the file are not string-expanded.
2681 42. If a MAIL command specified a SIZE value that was too large to fit into an
2682 int variable, the check against message_size_limit failed. Such values are
2683 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
2684 day this will have to be increased, but I don't think I want to be around
2685 when emails are that large.
2692 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
2693 iconv() is not standard in FreeBSD.
2695 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
2696 IPv6 enabled. The observed symptom was a segmentation fault on return from
2697 the function os_common_find_running_interfaces() in src/os.c.
2699 3. In the check_special_case() function in daemon.c I had used "errno" as an
2700 argument name, which causes warnings on some systems. This was basically a
2701 typo, since it was named "eno" in the comments!
2703 4. The code that waits for the clock to tick (at a resolution of some fraction
2704 of a second) so as to ensure message-id uniqueness was always waiting for
2705 at least one whole tick, when it could have waited for less. [This is
2706 almost certainly not relevant at current processor speeds, where it is
2707 unlikely to ever wait at all. But we try to future-proof.]
2709 5. The function that sleeps for a time interval that includes fractions of a
2710 second contained a race. It did not block SIGALRM between setting the
2711 timer, and suspending (a couple of lines later). If the interval was short
2712 and the sigsuspend() was delayed until after it had expired, the suspension
2713 never ended. On busy systems this could lead to processes getting stuck for
2716 6. Some uncommon configurations may cause a lookup to happen in a queue runner
2717 process, before it forks any delivery processes. The open lookup caching
2718 mechanism meant that the open file or database connection was passed into
2719 the delivery process. The problem was that delivery processes always tidy
2720 up cached lookup data. This could cause a problem for the next delivery
2721 process started by the queue runner, because the external queue runner
2722 process does not know about the closure. So the next delivery process
2723 still has data in the lookup cache. In the case of a file lookup, there was
2724 no problem because closing a file descriptor in a subprocess doesn't affect
2725 the parent. However, if the lookup was caching a connection to a database,
2726 the connection was closed, and the second delivery process was likely to
2727 see errors such as "PGSQL: query failed: server closed the connection
2728 unexpectedly". The problem has been fixed by closing all cached lookups
2729 in a queue runner before running a delivery process.
2731 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
2732 seem to have the "const" qualifier which it has on other OS. I've
2735 8. Change 4.21/2 was too strict. It is only if there are two authenticators
2736 *of the same type* (client or server) with the same public name that an
2737 error should be diagnosed.
2739 9. When Exim looked up a host name for an IP address, but failed to find the
2740 original IP address when looking up the host name (a safety check), it
2741 output the message "<ip address> does not match any IP for NULL", which was
2742 confusing, to say the least. The bug was that the host name should have
2743 appeared instead of "NULL".
2745 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
2746 user that is built into the binary, and the -C or -D options is used, root
2747 privilege is dropped before the configuration file is read. In addition,
2748 logging is switched to stderr instead of the normal log files. If the
2749 configuration then re-defines the Exim user, the unprivileged environment
2750 is probably not what is expected, so Exim logs a panic warning message (but
2753 However, if deliver_drop_privilege is set, the unprivileged state may well
2754 be exactly what is intended, so the warning has been cut out in that case,
2755 and Exim is allowed to try to write to its normal log files.
2761 1. smtp_return_error_details was not giving details for temporary sender
2762 or receiver verification errors.
2764 2. Diagnose a configuration error if two authenticators have the same public
2767 3. Exim used not to create the message log file for a message until the first
2768 delivery attempt. This could be confusing when incoming messages were held
2769 for policy or load reasons. The message log file is now created at the time
2770 the message is received, and an initial "Received" line is written to it.
2772 4. The automatically generated man page for command line options had a minor
2773 bug that caused no ill effects; however, a more serious problem was that
2774 the procedure for building the man page automatically didn't always
2775 operate. Consequently, release 4.20 contains an out-of-date version. This
2776 shouldn't happen again.
2778 5. When building Exim with embedded Perl support, the script that builds the
2779 Makefile was calling 'perl' to find its compile-time parameters, ignoring
2780 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
2782 6. The freeze_tell option was not being used for messages that were frozen on
2783 arrival, either by an ACL or by local_scan().
2785 7. Added the smtp_incomplete_transaction log selector.
2787 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
2788 was accepting AUTH without a new EHLO.
2790 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
2791 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
2794 10. Logging of TCP/IP connections (when configured) now happens in the main
2795 daemon process instead of the child process, so that the TCP/IP connection
2796 count is more accurate (but it can never be perfect).
2798 11. The use of "drop" in a nested ACL was not being handled correctly in the
2799 outer ACL. Now, if condition failure induced by the nested "drop" causes
2800 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
2801 or "require"), the connection is dropped.
2803 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
2804 that yield "discard" can now be used with an "accept" or a "discard" verb,
2805 but an error is generated for any others (because I can't see a useful way
2806 to define what should happen).
2808 13. When an ACL is read dynamically from a file (or anywhere else), the lines
2809 are now processed in the same way as lines in the Exim configuration file.
2810 In particular, continuation lines are supported.
2812 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
2814 15. Added -ti meaning -t -i.
2816 16. Check for letters, digits, hyphens, and dots in the names of dnslist
2817 domains, and warn by logging if others are found.
2819 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
2820 returned from GIFCONF when Exim is trying to find the list of interfaces on
2821 a host. The code in os.c has been modified to copy each ifreq to an aligned
2822 structure in all cases.
2824 Also, in some cases, the returned ifreq's were being copied to a 'struct
2825 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
2826 means the last couple of bytes of an IPv6 address could be chopped if the
2827 ifreq contained only a normal sockaddr (14 bytes storage).
2829 18. Named domain lists were not supported in the hosts_treat_as_local option.
2830 An entry such as +xxxx was not recognized, and was treated as a literal
2833 19. Ensure that header lines added by a DATA ACL are included in the reject log
2834 if the ACL subsequently rejects the message.
2836 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
2837 MD5 (which is deprecated).
2839 21. When testing a filter file using -bf, Exim was writing a message when it
2840 took the sender from a "From " line in the message, but it was not doing so
2841 when it took $return_path from a Return-Path: header line. It now does.
2843 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
2844 with a valid header line field name (a series of printing characters
2845 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
2847 23. Changed "disc" in the source to "disk" to conform to the documentation and
2848 the book and for uniformity.
2850 24. Ignore Sendmail's -Ooption=value command line item.
2852 25. When execve() failed while trying to run a command in a pipe transport,
2853 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
2854 could be confused with a return value of 69 from the command itself. This
2855 has been changed to 127, the value the shell returns if it is asked to run
2856 a non-existent command. The wording for the related log line suggests a
2857 non-existent command as the problem.
2859 26. If received_header_text expands to an empty string, do not add a Received:
2860 header line to the message. (Well, it adds a token one on the spool, but
2861 marks it "old" so that it doesn't get used or transmitted.)
2863 27. Installed eximstats 1.28 (addition of -nt option).
2865 28. There was no check for failure on the call to getsockname() in the daemon
2866 code. This can fail if there is a shortage of resources on the system, with
2867 ENOMEM, for example. A temporary error is now given on failure.
2869 29. Contrary to the C standard, it seems that in some environments, the
2870 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
2871 program. Exim now does this explicitly; it affects the formatting of
2872 timestamps using strftime().
2874 30. If exiqsumm was given junk data, it threw up some uninitialized variable
2875 complaints. I've now initialized all the variables, to avoid this.
2877 32. Header lines added by a system filter were not being "seen" during
2878 transport-time rewrites.
2880 33. The info_callback() function passed to OpenSSL is set up with type void
2881 (*)(SSL *, int, int), as described somewhere. However, when calling the
2882 function (actually a macro) that sets it up, the type void(*)() is
2883 expected. I've put in a cast to prevent warnings from picky compilers.
2885 34. If a DNS black list lookup found a CNAME record, but there were no A
2886 records associated with the domain it pointed at, Exim crashed.
2888 35. If a DNS black list lookup returned more than one A record, Exim ignored
2889 all but the first. It now scans all returned addresses if a particular IP
2890 value is being sought. In this situation, the contents of the
2891 $dnslist_value variable are a list of all the addresses, separated by a
2894 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
2895 to accept a host name and all its aliases if the forward lookup for any of
2896 them yielded the IP address of the incoming connection. Now it accepts only
2897 those names whose forward lookup yields the correct IP address. Any other
2898 names are discarded. This closes a loophole whereby a rogue DNS
2899 administrator could create reverse DNS records to break through a
2900 wildcarded host restriction in an ACL.
2902 37. If a user filter or a system filter that ran in a subprocess used any of
2903 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
2904 the wrong values were passed to the pipe command ($thisaddress had the
2905 value of $0, $0 had the value of $1, etc). This bug was introduced by
2906 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
2908 38. Improved the line breaking for long SMTP error messages from ACLs.
2909 Previously, if there was no break point between 40 and 75 characters, Exim
2910 left the rest of the message alone. Two changes have been made: (a) I've
2911 reduced the minimum length to 35 characters; (b) if it can't find a break
2912 point between 35 and 75 characters, it looks ahead and uses the first one
2913 that it finds. This may give the occasional overlong line, but at least the
2914 remaining text gets split now.
2916 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
2917 file descriptors might be low, and that setting 1000 would always raise it.
2918 It turns out that in some environments, the limit is already over 1000 and
2919 that lowering it causes trouble. So now Exim takes care not to decrease it.
2921 40. When delivering a message, the value of $return_path is set to $sender_
2922 address at the start of routing (routers may change the value). By an
2923 oversight, this default was not being set up when an address was tested by
2924 -bt or -bv, which affected the outcome if any router or filter referred to
2927 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
2928 log only when "message" or "log_message" are set. However, if one of the
2929 conditions was an address verification, or a call to a nested ACL, the
2930 messages generated by the underlying test were being passed through. This
2931 no longer happens. The underlying message is available in $acl_verify_
2932 message for both "message" and "log_message" expansions, so it can be
2933 passed through if needed.
2935 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
2936 new expansion $bh_ to give the encoded byte string without charset
2937 translation. Translation happens only if iconv() is available; HAVE_ICONV
2938 indicates this at build time. HEADERS_CHARSET gives the charset to
2939 translate to; headers_charset can change it in the configuration, and
2940 "headers charset" can change it in an individual filter file.
2942 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
2943 that creates RFC 2047 encoded "words" labels them as that charset instead
2944 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
2945 expansion operator; (ii) when Exim creates a From: line for a local
2946 message; (iii) when a header line is rewritten to include a "phrase" part.
2948 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
2949 buggy, causing it to skip the first lines of messages whose message ID
2950 ended in 'D'. This would not have bitten before Exim release 4.14, because
2951 message IDs were unlikely to end in 'D' before then. The effect was to have
2952 incorrect size information for certain domains.
2954 45. #include "config.h" was missing at the start of the crypt16.c module. This
2955 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
2958 46. If there was a timeout during a "random" callout check, Exim treated it as
2959 a failure of the random address, and carried on sending RSET and the real
2960 address. If the delay was just some slowness somewhere, the response to the
2961 original RCPT would be taken as a response to RSET and so on, causing
2962 mayhem of various kinds.
2964 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
2965 when I implemented it. It didn't allow for the fact that some option values
2966 may legitimatetly be negative (e.g. size_addition), and it didn't even do
2967 the right test for positive values.
2969 48. Domain names in DNS records are case-independent. Exim always looks them up
2970 in lower case. Some resolvers return domain names in exactly the case they
2971 appear in the zone file, that is, they may contain uppercase letters. Not
2972 all resolvers do this - some return always lower case. Exim was treating a
2973 change of case by a resolver as a change of domain, similar to a widening
2974 of a domain abbreviation. This triggered its re-routing code and so it was
2975 trying to route what was effectively the same domain again. This normally
2976 caused routing to fail (because the router wouldn't handle the domain
2977 twice). Now Exim checks for this case specially, and just changes the
2978 casing of the domain that it ultimately uses when it transmits the message
2981 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
2984 50. If a filter generated a file delivery with a non-absolute name (possible if
2985 no home directory exists for the router), the forbid_file option was not
2988 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
2989 the existing equality matching.
2991 52. Exim was using ints instead of ino_t variables in some places where it was
2992 dealing with inode numbers.
2994 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
2995 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
2996 called TMPDIR, and if it finds it is different, it changes its value.
2998 54. The smtp_printf() function is now made available to local_scan() so
2999 additional output lines can be written before returning. There is also an
3000 smtp_fflush() function to enable the detection of a dropped connection.
3001 The variables smtp_input and smtp_batched_input are exported to
3004 55. Changed the default runtime configuration: the message "Unknown user"
3005 has been removed from the ACL, and instead placed on the localuser router,
3006 using the cannot_route_message feature. This means that any verification
3007 failures that generate their own messages won't get overridden. Similarly,
3008 the "Unrouteable address" message that was in the ACL for unverifiable
3009 relay addresses has also been removed.
3011 56. Added hosts_avoid_esmtp to the smtp transport.
3013 57. The exicyclog script was not checking for the esoteric option
3014 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
3015 will work only if exicyclog is run under the appropriate euid.
3017 58. Following a discussion on the list, the rules by which Exim recognises line
3018 endings on incoming messages have been changed. The -dropcr and drop_cr
3019 options are now no-ops, retained only for backwards compatibility. The
3020 following line terminators are recognized: LF CRLF CR. However, special
3021 processing applies to CR:
3023 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
3024 nor a local message in the state where . is a terminator.
3026 (ii) If a bare CR is encountered in a header line, an extra space is added
3027 after the line terminator so as not to end the header. The reasoning
3028 behind this is that bare CRs in header lines are most likely either
3029 to be mistakes, or people trying to play silly games.
3031 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
3032 was being incorrectly given as 18 bytes larger than it should have been.
3033 This is a VOB (very old bug).
3035 60. This may never have affected anything current, but just in case it has:
3036 When the local host is found other than at the start of a list of hosts,
3037 the local host, those with the same MX, and any that follow, are discarded.
3038 When the list in question was part of a longer list of hosts, the following
3039 hosts (not currently being processed) were also being discarded. This no
3040 longer happens. I'm not sure if this situation could ever has previously
3043 61. Added the "/MX" feature to lists of hosts in the manualroute and query
3046 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
3047 header. This is something that is recommended in a new Internet Draft, and
3048 is something that is documented as being done by Sendmail. There are two
3049 possible values. For messages generated by the autoreply transport, Exim
3052 Auto-Submitted: auto-replied
3054 whereas for all other generated messages (e.g. bounces) it adds
3056 Auto-Submitted: auto-generated
3058 63. The "personal" condition in filters now includes a test for the
3059 Auto-Submitted: header. If it contains the string "auto-" the message it
3060 not considered personal.
3062 64. Added rcpt_include_affixes as a generic transport option.
3064 65. Added queue_only_override (default true).
3066 66. Added the syslog_duplication option.
3068 67. If what should have been the first header line of a message consisted of
3069 a space followed by a colon, Exim was mis-interpreting it as a header line.
3070 It isn't of course - it is syntactically invalid and should therefore be
3071 treated as the start of the message body. The misbehaviour could have
3072 caused a number of strange effects, including loss of data in subsequent
3073 header lines, and spool format errors.
3075 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
3076 client host had authenticated. This control can now be exercised by an ACL
3077 for more flexibility.
3079 69. By default, callouts do not happen when testing with -bh. There is now a
3080 variant, -bhc, which does actually run the callout code, including
3081 consulting and updating the callout cache.
3083 70. Added support for saslauthd authentication, courtesy of Alexander
3086 71. If statvfs() failed on the spool or log directories while checking their
3087 size for availability, Exim confusingly gave the error "space shortage".
3088 Furthermore, in debugging mode it crashed with a floating point exception.
3089 These checks are done if check_{spool,log}_{space,inodes} are set, and when
3090 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
3091 serious problem, Exim now writes to the main and panic logs when this
3092 happens, with details of the failure. It then refuses to accept the
3093 incoming message, giving the message "spool directory problem" or "log
3094 directory problem" with a 421 code for SMTP messages.
3096 72. When Exim is about to re-exec itself, it ensures that the file descriptors
3097 0, 1, and 2 exist, because some OS complain for execs without them (see
3098 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
3099 descriptors. However, the code omitted to check that the open succeeded,
3100 causing mysterious errors if for some reason the permissions on /dev/null
3101 got screwed. Now Exim writes a message to the main and panic logs, and
3102 bombs out if it can't open /dev/null.
3104 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
3105 interact so that it is all more flexible. It is supposed to remain
3106 backwards compatible. Also added extra_local_interfaces.
3108 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
3109 to bomb out with an assertion failure - to the client this appears as a
3110 connection drop. This problem occurs in the part of the code that was taken
3111 from the Samba project. Fortunately, the assertion is in a very simple
3112 function, so I have fixed this by reproducing the function inline in the
3113 one place where it is called, and arranging for authentication to fail
3114 instead of killing the process with assert().
3116 75. The SPA client code was not working when the server requested OEM rather
3117 than Unicode encoding.
3119 76. Added code to make require_files with a specific uid setting more usable in
3120 the case where statting the file as root fails - usually a non-root-mounted
3121 NFS file system. When this happens and the failure is EACCES, Exim now
3122 forks a subprocess and does the per-uid checking as the relevant uid.
3124 77. Added process_log_path.
3126 78. If log_file_path was not explicitly set, a setting of check_log_space or
3127 check_log_inodes was ignored.
3129 79. If a space check for the spool or log partitions fails, the incident is now
3130 logged. Of course, in the latter case the data may get lost...
3132 80. Added the %p formatting code to string_format() so that it can be used to
3133 print addresses in debug_print(). Adjusted all the address printing in the
3134 debugging in store.c to use %p rather than %d.
3136 81. There was a concern that a line of code in smtp_in.c could overflow a
3137 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
3138 initially expressed, the concern was not well-founded, because trailing
3139 spaces are removed early. However, if the trailing spaces were followed by
3140 a NULL, they did not get removed, so the overflow was possible. Two fixes
3143 (a) I re-wrote the offending code in a cleaner fashion.
3144 (b) If an incoming SMTP command contains a NULL character, it is rejected
3147 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
3148 runs initgroups(), so that if the Exim user is in any additional groups,
3149 they will be used during message reception.
3155 The change log for 4.20 and earlier releases has been archived.
git://git.exim.org / users / heiko / exim.git / blob