1 $Cambridge: exim/doc/doc-txt/ChangeLog,v 1.457 2007/01/22 16:29:54 ph10 Exp $
3 Change log file for Exim from version 4.21
4 -------------------------------------------
9 MH/01 Fix for bug #448, segfault in Dovecot authenticator when interface_address
10 is unset (happens when testing with -bh and -oMi isn't used). Thanks to
13 PH/01 Added a new log selector smtp_no_mail, to log SMTP sessions that do not
16 PH/02 In an ACL statement such as
18 deny dnslists = X!=127.0.0.2 : X=127.0.0.2
20 if a client was not listed at all, or was listed with a value other than
21 127.0.0.2, in the X list, but was listed with 127.0.0.2 in the Y list,
22 the condition was not true (as it should be), so access was not denied.
23 The bug was that the ! inversion was incorrectly passed on to the second
24 item. This has been fixed.
26 PH/03 Added additional dnslists conditions == and =& which are different from
27 = and & when the dns lookup returns more than one IP address.
29 PH/04 Added gnutls_require_{kx,mac,protocols} to give more control over the
30 cipher suites used by GnuTLS. These options are ignored by OpenSSL.
32 PH/05 After discussion on the list, added a compile time option ENABLE_DISABLE_
33 FSYNC, which compiles an option called disable_fsync that allows for
34 bypassing fsync(). The documentation is heavily laced with warnings.
36 SC/01 Updated eximstats to collate all SpamAssassin rejects into one bucket.
42 PH/01 Two more bugs that were introduced by 4.64/PH/07, in addition to the one
43 fixed by 4.65/MH/01 (is this a record?) are fixed:
45 (i) An empty string was always treated as zero by the numeric comparison
46 operators. This behaviour has been restored.
48 (ii) It is documented that the numeric comparison operators always treat
49 their arguments as decimal numbers. This was broken in that numbers
50 starting with 0 were being interpreted as octal.
52 While fixing these problems I realized that there was another issue that
53 hadn't been noticed. Values of message_size_limit (both the global option
54 and the transport option) were treated as octal if they started with 0.
55 The documentation was vague. These values are now always treated as
56 decimal, and I will make that clear in the documentation.
62 TK/01 Disable default definition of HAVE_LINUX_SENDFILE. Clashes with
63 Linux large file support (_FILE_OFFSET_BITS=64) on older glibc
66 MH/01 Don't check that the operands of numeric comparison operators are
67 integers when their expansion is in "skipping" mode (fixes bug
68 introduced by 4.64-PH/07).
70 PH/01 If a system filter or a router generates more than SHRT_MAX (32767)
71 child addresses, Exim now panics and dies. Previously, because the count
72 is held in a short int, deliveries were likely to be lost. As such a
73 large number of recipients for a single message is ridiculous
74 (performance will be very, very poor), I have chosen to impose a limit
75 rather than extend the field.
81 TK/01 Bugzilla #401. Fix DK spooling code so that it can overwrite a
82 leftover -K file (the existence of which was triggered by #402).
83 While we were at it, introduced process PID as part of the -K
84 filename. This should rule out race conditions when creating
87 TK/02 Bugzilla #402. Apply patch from Simon Arlott, speeding up DK signing
88 processing considerably. Previous code took too long for large mails,
89 triggering a timeout which in turn triggers #401.
91 TK/03 Introduced HAVE_LINUX_SENDFILE to os.h-Linux. Currently only used
92 in the DK code in transports.c. sendfile() is not really portable,
93 hence the _LINUX specificness.
95 TF/01 In the add_headers option to the mail command in an Exim filter,
96 there was a bug that Exim would claim a syntax error in any
97 header after the first one which had an odd number of characters
100 PH/01 If a server that rejects MAIL FROM:<> was the target of a sender
101 callout verification, Exim cached a "reject" for the entire domain. This
102 is correct for most verifications, but it is not correct for a recipient
103 verification with use_sender or use_postmaster set, because in that case
104 the callout does not use MAIL FROM:<>. Exim now distinguishes the special
105 case of MAIL FROM:<> rejection from other early rejections (e.g.
106 rejection of HELO). When verifying a recipient using a non-null MAIL
107 address, the cache is ignored if it shows MAIL FROM:<> rejection.
108 Whatever the result of the callout, the value of the domain cache is
109 left unchanged (for any other kind of callout, getting as far as trying
110 RCPT means that the domain itself is ok).
112 PH/02 Tidied a number of unused variable and signed/unsigned warnings that
115 PH/03 On Solaris, an unexpectedly close socket (dropped connection) can
116 manifest itself as EPIPE rather than ECONNECT. When tidying away a
117 session, the daemon ignores ECONNECT errors and logs others; it now
118 ignores EPIPE as well.
120 PH/04 Applied Nico Erfurth's refactoring patch to tidy up mime.c
121 (quoted-printable decoding).
123 PH/05 Applied Nico Erfurth's refactoring patch to tidy up spool_mbox.c, and
124 later the small subsequent patch to fix an introduced bug.
126 PH/06 Installed the latest Cygwin Makefile from the Cygwin maintainer.
128 PH/07 There was no check for overflow in expansions such as ${if >{1}{4096M}}.
130 PH/08 An error is now given if message_size_limit is specified negative.
132 PH/09 Applied and tidied up Jakob Hirsch's patch for allowing ACL variables
133 to be given (somewhat) arbitrary names.
135 JJ/01 exipick 20060919.0, allow for arbitrary acl_ variables introduced
138 JJ/02 exipick 20060919.0, --show-vars args can now be regular expressions,
139 miscellaneous code fixes
141 PH/10 Added the log_reject_target ACL modifier to specify where to log
144 PH/11 Callouts were setting the name used for EHLO/HELO from $smtp_active_
145 hostname. This is wrong, because it relates to the incoming message (and
146 probably the interface on which it is arriving) and not to the outgoing
147 callout (which could be using a different interface). This has been
148 changed to use the value of the helo_data option from the smtp transport
149 instead - this is what is used when a message is actually being sent. If
150 there is no remote transport (possible with a router that sets up host
151 addresses), $smtp_active_hostname is used.
153 PH/12 Installed Andrey Panin's patch to add a dovecot authenticator. Various
154 tweaks were necessary in order to get it to work (see also 21 below):
155 (a) The code assumed that strncpy() returns a negative number on buffer
156 overflow, which isn't the case. Replaced with Exim's string_format()
158 (b) There were several signed/unsigned issues. I just did the minimum
159 hacking in of casts. There is scope for a larger refactoring.
160 (c) The code used strcasecmp() which is not a standard C function.
161 Replaced with Exim's strcmpic() function.
162 (d) The code set only $1; it now sets $auth1 as well.
163 (e) A simple test gave the error "authentication client didn't specify
164 service in request". It would seem that Dovecot has changed its
165 interface. Fortunately there's a specification; I followed it and
166 changed what the client sends and it appears to be working now.
168 PH/13 Added $message_headers_raw to provide the headers without RFC 2047
171 PH/14 Corrected misleading output from -bv when -v was also used. Suppose the
172 address A is aliased to B and C, where B exists and C does not. Without
173 -v the output is "A verified" because verification stops after a
174 successful redirection if more than one address is generated. However,
175 with -v the child addresses are also verified. Exim was outputting "A
176 failed to verify" and then showing the successful verification for C,
177 with its parentage. It now outputs "B failed to verify", showing B's
178 parentage before showing the successful verification of C.
180 PH/15 Applied Michael Deutschmann's patch to allow DNS black list processing to
181 look up a TXT record in a specific list after matching in a combined
184 PH/16 It seems that the options setting for the resolver (RES_DEFNAMES and
185 RES_DNSRCH) can affect the behaviour of gethostbyname() and friends when
186 they consult the DNS. I had assumed they would set it the way they
187 wanted; and indeed my experiments on Linux seem to show that in some
188 cases they do (I could influence IPv6 lookups but not IPv4 lookups).
189 To be on the safe side, however, I have now made the interface to
190 host_find_byname() similar to host_find_bydns(), with an argument
191 containing the DNS resolver options. The host_find_byname() function now
192 sets these options at its start, just as host_find_bydns() does. The smtp
193 transport options dns_qualify_single and dns_search_parents are passed to
194 host_find_byname() when gethostbyname=TRUE in this transport. Other uses
195 of host_find_byname() use the default settings of RES_DEFNAMES
196 (qualify_single) but not RES_DNSRCH (search_parents).
198 PH/17 Applied (a modified version of) Nico Erfurth's patch to make
199 spool_read_header() do less string testing, by means of a preliminary
200 switch on the second character of optional "-foo" lines. (This is
201 overdue, caused by the large number of possibilities that now exist.
202 Originally there were few.) While I was there, I also converted the
203 str(n)cmp tests so they don't re-test the leading "-" and the first
204 character, in the hope this might squeeze out yet more improvement.
206 PH/18 Two problems with "group" syntax in header lines when verifying: (1) The
207 flag allowing group syntax was set by the header_syntax check but not
208 turned off, possible causing trouble later; (2) The flag was not being
209 set at all for the header_verify test, causing "group"-style headers to
210 be rejected. I have now set it in this case, and also caused header_
211 verify to ignore an empty address taken from a group. While doing this, I
212 came across some other cases where the code for allowing group syntax
213 while scanning a header line wasn't quite right (mostly, not resetting
214 the flag correctly in the right place). These bugs could have caused
215 trouble for malformed header lines. I hope it is now all correct.
217 PH/19 The functions {pwcheck,saslauthd}_verify_password() are always called
218 with the "reply" argument non-NULL. The code, however (which originally
219 came from elsewhere) had *some* tests for NULL when it wrote to *reply,
220 but it didn't always do it. This confused somebody who was copying the
221 code for some other use. I have removed all the tests.
223 PH/20 It was discovered that the GnuTLS code had support for RSA_EXPORT, a
224 feature that was used to support insecure browsers during the U.S. crypto
225 embargo. It requires special client support, and Exim is probably the
226 only MTA that supported it -- and would never use it because real RSA is
227 always available. This code has been removed, because it had the bad
228 effect of slowing Exim down by computing (never used) parameters for the
229 RSA_EXPORT functionality.
231 PH/21 On the advice of Timo Sirainen, added a check to the dovecot
232 authenticator to fail if there's a tab character in the incoming data
233 (there should never be unless someone is messing about, as it's supposed
234 to be base64-encoded). Also added, on Timo's advice, the "secured" option
235 if the connection is using TLS or if the remote IP is the same as the
236 local IP, and the "valid-client-cert option" if a client certificate has
239 PH/22 As suggested by Dennis Davis, added a server_condition option to *all*
240 authenticators. This can be used for authorization after authentication
241 succeeds. (In the case of plaintext, it servers for both authentication
244 PH/23 Testing for tls_required and lost_connection in a retry rule didn't work
245 if any retry times were supplied.
247 PH/24 Exim crashed if verify=helo was activated during an incoming -bs
248 connection, where there is no client IP address to check. In this
249 situation, the verify now always succeeds.
251 PH/25 Applied John Jetmore's -Mset patch.
253 PH/26 Added -bem to be like -Mset, but loading a message from a file.
255 PH/27 In a string expansion for a processed (not raw) header when multiple
256 headers of the same name were present, leading whitespace was being
257 removed from all of them, but trailing whitespace was being removed only
258 from the last one. Now trailing whitespace is removed from each header
259 before concatenation. Completely empty headers in a concatenation (as
262 PH/28 Fixed bug in backwards-compatibility feature of PH/09 (thanks to John
263 Jetmore). It would have mis-read ACL variables from pre-4.61 spool files.
265 PH/29 [Removed. This was a change that I later backed out, and forgot to
266 correct the ChangeLog entry (that I had efficiently created) before
267 committing the later change.]
269 PH/30 Exim was sometimes attempting to deliver messages that had suffered
270 address errors (4xx response to RCPT) over the same connection as other
271 messages routed to the same hosts. Such deliveries are always "forced",
272 so retry times are not inspected. This resulted in far too many retries
273 for the affected addresses. The effect occurred only when there were more
274 hosts than the hosts_max_try setting in the smtp transport when it had
275 the 4xx errors. Those hosts that it had tried were not added to the list
276 of hosts for which the message was waiting, so if all were tried, there
277 was no problem. Two fixes have been applied:
279 (i) If there are any address or message errors in an SMTP delivery, none
280 of the hosts (tried or untried) are now added to the list of hosts
281 for which the message is waiting, so the message should not be a
282 candidate for sending over the same connection that was used for a
283 successful delivery of some other message. This seems entirely
284 reasonable: after all the message is NOT "waiting for some host".
285 This is so "obvious" that I'm not sure why it wasn't done
286 previously. Hope I haven't missed anything, but it can't do any
287 harm, as the worst effect is to miss an optimization.
289 (ii) If, despite (i), such a delivery is accidentally attempted, the
290 routing retry time is respected, so at least it doesn't keep
291 hammering the server.
293 PH/31 Installed Andrew Findlay's patch to close the writing end of the socket
294 in ${readsocket because some servers need this prod.
296 PH/32 Added some extra debug output when updating a wait-xxx database.
298 PH/33 The hint "could be header name not terminated by colon", which has been
299 given for certain expansion errors for a long time, was not being given
300 for the ${if def:h_colon_omitted{... case.
302 PH/34 The spec says: "With one important exception, whenever a domain list is
303 being scanned, $domain contains the subject domain." There was at least
304 one case where this was not true.
306 PH/35 The error "getsockname() failed: connection reset by peer" was being
307 written to the panic log as well as the main log, but it isn't really
308 panic-worthy as it just means the connection died rather early on. I have
309 removed the panic log writing for the ECONNRESET error when getsockname()
312 PH/36 After a 4xx response to a RCPT error, that address was delayed (in queue
313 runs only) independently of the message's sender address. This meant
314 that, if the 4xx error was in fact related to the sender, a different
315 message to the same recipient with a different sender could confuse
316 things. In particualar, this can happen when sending to a greylisting
317 server, but other circumstances could also provoke similar problems.
318 I have changed the default so that the retry time for these errors is now
319 based a combination of the sender and recipient addresses. This change
320 can be overridden by setting address_retry_include_sender=false in the
323 PH/37 For LMTP over TCP/IP (the smtp transport), error responses from the
324 remote server are returned as part of bounce messages. This was not
325 happening for LMTP over a pipe (the lmtp transport), but now it is the
326 same for both kinds of LMTP.
328 PH/38 Despite being documented as not happening, Exim was rewriting addresses
329 in header lines that were in fact CNAMEs. This is no longer the case.
331 PH/39 If -R or -S was given with -q<time>, the effect of -R or -S was ignored,
332 and queue runs started by the daemon processed all messages. This has
333 been fixed so that -R and -S can now usefully be given with -q<time>.
335 PH/40 Import PCRE release 6.7 (fixes some bugs).
337 PH/41 Add bitwise logical operations to eval (courtesy Brad Jorsch).
339 PH/42 Give an error if -q is specified more than once.
341 PH/43 Renamed the variables $interface_address and $interface_port as
342 $received_ip_address and $received_port, to make it clear that these
343 values apply to message reception, and not to the outgoing interface when
344 a message is delivered. (The old names remain recognized, of course.)
346 PH/44 There was no timeout on the connect() call when using a Unix domain
347 socket in the ${readsocket expansion. There now is.
349 PH/45 Applied a modified version of Brad Jorsch's patch to allow "message" to
350 be meaningful with "accept".
352 SC/01 Eximstats V1.43
353 Bug fix for V1.42 with -h0 specified. Spotted by Chris Lear.
355 SC/02 Eximstats V1.44
356 Use a glob alias rather than an array ref in the generated
357 parser. This improves both readability and performance.
359 SC/03 Eximstats V1.45 (Marco Gaiarin / Steve Campbell)
360 Collect SpamAssassin and rejection statistics.
361 Don't display local sender or destination tables unless
362 there is data to show.
363 Added average volumes into the top table text output.
365 SC/04 Eximstats V1.46
366 Collect data on the number of addresses (recipients)
367 as well as the number of messages.
369 SC/05 Eximstats V1.47
370 Added 'Message too big' to the list of mail rejection
371 reasons (thanks to Marco Gaiarin).
373 SC/06 Eximstats V1.48
374 Mainlog lines which have GMT offsets and are too short to
375 have a flag are now skipped.
377 SC/07 Eximstats V1.49 (Alain Williams)
378 Added the -emptyok flag.
380 SC/08 Eximstats V1.50
381 Fixes for obtaining the IP address from reject messages.
383 JJ/03 exipick.20061117.2, made header handling as similar to exim as possible
384 (added [br]h_ prefixes, implemented RFC2047 decoding. Fixed
385 whitesspace changes from 4.64-PH/27
387 JJ/04 exipick.20061117.2, fixed format and added $message_headers_raw to
390 JJ/05 exipick.20061117.2, bug fixes (error out sooner when invalid criteria
391 are found, allow negative numbers in numeric criteria)
393 JJ/06 exipick.20061117.2, added new $message_body_missing variable
395 JJ/07 exipick.20061117.2, added $received_ip_address and $received_port
396 to match changes made in 4.64-PH/43
398 PH/46 Applied Jori Hamalainen's patch to add features to exiqsumm.
400 PH/47 Put in an explicit test for a DNS lookup of an address record where the
401 "domain" is actually an IP address, and force a failure. This locks out
402 those revolvers/nameservers that support "A-for-A" lookups, in
403 contravention of the specifications.
405 PH/48 When a host name was looked up from an IP address, and the subsequent
406 forward lookup of the name timed out, the host name was left in
407 $sender_host_name, contrary to the specification.
409 PH/49 Although default lookup types such as lsearch* or cdb*@ have always been
410 restricted to single-key lookups, Exim was not diagnosing an error if
411 * or *@ was used with a query-style lookup.
413 PH/50 Increased the value of DH_BITS in tls-gnu.c from 768 to 1024.
415 MH/01 local_scan ABI version incremented to 1.1. It should have been updated
416 long ago, but noone interested enough thought of it. Let's just say that
417 the "1.1" means that there are some new functions that weren't there at
418 some point in the past.
420 PH/51 Error processing for expansion failure of helo_data from an smtp
421 transport during callout processing was broken.
423 PH/52 Applied John Jetmore's patch to allow tls-on-connect and STARTTLS to be
424 tested/used via the -bh/-bhc/-bs options.
426 PH/53 Added missing "#include <time.h>" to pcre/pcretest.c (this was a PCRE
427 bug, fixed in subsequent PCRE releases).
429 PH/54 Applied Robert Bannocks' patch to avoid a problem with references that
430 arises when using the Solaris LDAP libraries (but not with OpenLDAP).
432 PH/55 Check for a ridiculously long file name in exim_dbmbuild.
438 SC/01 Use a glob alias rather than an array ref in eximstats generated
439 parser. This improves both readability and performance.
441 SC/02 Collect SpamAssassin and rejection statistics in eximstats.
442 Don't display local sender or destination tables in eximstats unless
443 there is data to show.
444 Added average volumes into the eximstats top table text output.
446 SC/03 Collect data on the number of addresses (recipients) as well
447 as the number of messages in eximstats.
449 TF/01 Correct an error in the documentation for the redirect router. Exim
450 does (usually) call initgroups() when daemonizing.
452 TF/02 Call initgroups() when dropping privilege in exim.c, so that Exim runs
453 with consistent privilege compared to when running as a daemon.
455 TF/03 Note in the spec that $authenticated_id is not set for local
456 submissions from trusted users.
458 TF/04 The ratelimit per_rcpt option now works correctly in acl_not_smtp.
459 Thanks to Dean Brooks <dean@iglou.com> for the patch.
461 TF/05 Make it easier to get SMTP authentication and TLS/SSL support working
462 by adding some example configuration directives to the default
463 configuration file. A little bit of work is required to uncomment the
464 directives and define how usernames and passwords are checked, but
465 there is now a framework to start from.
467 PH/01 Added #define LDAP_DEPRECATED 1 to ldap.c because some of the "old"
468 functions that Exim currently uses aren't defined in ldap.h for OpenLDAP
469 without this. I don't know how relevant this is to other LDAP libraries.
471 PH/02 Add the verb name to the "unknown ACL verb" error.
473 PH/03 Magnus Holmgren's patch for filter_prepend_home.
475 PH/03 Fixed Bugzilla #101: macro definition between ACLs doesn't work.
477 PH/04 Applied Magnus Holmgren's patch to fix Bugzilla #98: transport's home
478 directory not expanded when it should be if an expanded home directory
479 was set for the address (which is overridden by the transport).
481 PH/05 Applied Alex Kiernan's patch to fix Bugzilla #99: a problem with
484 PH/06 Added acl_not_smtp_start, based on Johannes Berg's patch, and set the
485 bit to forbid control=suppress_local_fixups in the acl_not_smtp ACL,
486 because it is too late at that time, and has no effect.
488 PH/07 Changed ${quote_pgsql to quote ' as '' instead of \' because of a
489 security issue with \' (bugzilla #107). I could not use the
490 PQescapeStringConn() function, because it needs a PGconn value as one of
493 PH/08 When testing addresses using -bt, indicate those final addresses that
494 are duplicates that would not cause an additional delivery. At least one
495 person was confused, thinking that -bt output corresponded to deliveries.
496 (Suppressing duplicates isn't a good idea as you lose the information
497 about possibly different redirections that led to the duplicates.)
499 PH/09 Applied patch from Erik to use select() instead of poll() in spam.c on
500 systems where poll() doesn't work, in particular OS X.
502 PH/10 Added more information to debugging output for retry time not reached.
504 PH/11 Applied patch from Arkadiusz Miskiewicz to apply a timeout to read
505 operations in malware.c.
507 PH/12 Applied patch from Magnus Holmgren to include the "h" tag in Domain Keys
510 PH/13 If write_rejectlog was set false when logging was sent to syslog with
511 syslog_duplication set false, log lines that would normally be written
512 both the the main log and to the reject log were not written to syslog at
515 PH/14 In the default configuration, change the use of "message" in ACL warn
516 statements to "add_header".
518 PH/15 Diagnose a filter syntax error for "seen", "unseen", or "noerror" if not
519 not followed by a command (e.g. "seen endif").
521 PH/16 Recognize SMTP codes at the start of "message" in ACLs and after :fail:
522 and :defer: in a redirect router. Add forbid_smtp_code to suppress the
525 PH/17 Added extra conditions to the default value of delay_warning_condition
529 { !eq{$h_list-id:$h_list-post:$h_list-subscribe:}{} } \
530 { match{$h_precedence:}{(?i)bulk|list|junk} } \
531 { match{$h_auto-submitted:}{(?i)auto-generated|auto-replied} } \
534 The Auto-Submitted: and various List- headers are standardised, whereas I
535 don't think Precedence: ever was.
537 PH/18 Refactored debugging code in route_finduser() to show more information,
538 in particular, the error code if getpwnam() issues one.
540 PH/19 Added PQsetClientEncoding(conn, "SQL_ASCII") to the pgsql code module.
541 This is apparently needed in addition to the PH/07 change above to avoid
542 any possible encoding problems.
544 PH/20 Perl can change the locale. Exim was resetting it after a ${perl call,
545 but not after initializing Perl.
547 PH/21 Added a call to PQsetNoticeProcessor() to catch pgsql "notices" and
548 output them only if debugging. By default they are written stderr,
549 apparently, which is not desirable.
551 PH/22 Added Alain Williams' LDAP patch to support setting REFERRALS=off on
554 JJ/01 exipick: added --reverse (and -R synonym), --random, --size, --sort and
557 JJ/02 exipick: rewrote --help documentation to hopefully make more clear.
559 PH/23 Made -oMaa and -oMt work with -bh and -bs to pretend the connection is
560 authenticated or an ident call has been made. Suppress the default
561 values for $authenticated_id and $authenticated_sender (but permit -oMai
562 and -oMas) when testing with -bh.
564 PH/24 Re-jigged the order of the tests in the default configuration so that the
565 tests for valid domains and recipients precede the DNS black list and CSA
566 tests, on the grounds that those ones are more expensive.
568 PH/25 Exim was not testing for a space following SMTP commands such as EHLO
569 that require one. Thus, EHLORHUBARB was interpreted as a valid command.
570 This bug exists in every version of Exim that I still have, right back to
573 PH/26 (n)wildlsearch lookups are documented as being done case-insensitively.
574 However, an attempt to turn on case-sensitivity in a regex key by
575 including (?-i) didn't work because the subject string was already
576 lowercased, and the effects were non-intuitive. It turns out that a
577 one-line patch can be used to allow (?-i) to work as expected.
583 TF/01 Fix the add_header change below (4.61 PH/55) which had a bug that (amongst
584 other effects) broke the use of negated acl sub-conditions.
586 PH/01 ${readsocket now supports Internet domain sockets (modified John Jetmore
589 PH/02 When tcp-wrappers is called from Exim, it returns only "deny" or "allow".
590 "Deny" causes Exim to reject the incoming connection with a 554 error.
591 Unfortunately, if there is a major crisis, such as a disk failure,
592 tcp-wrappers gives "deny", whereas what one would like would be some
593 kind of temporary error. A kludge has been added to help with this.
594 Before calling hosts_ctl(), errno is set zero. If the result is "deny", a
595 554 error is used if errno is still zero or contains ENOENT (which occurs
596 if either of the /etc/hosts.{allow,deny} files is missing). Otherwise, a
599 PH/03 Add -lutil to the default FreeBSD LIBS setting.
601 PH/04 Change PH/19 for 4.61 was too wide. It should not be applied to host
602 errors. Otherwise a message that provokes a temporary error (when other
603 messages do not) can cause a whole host to time out.
605 PH/05 Batch deliveries by appendfile and pipe transports did not work when the
606 addresses were routed directly to files or pipes from a redirect router.
607 File deliveries just didn't batch; pipe deliveries might have suffered
610 PH/06 A failure to get a lock for a hints database would erroneously always say
611 "Failed to get write lock", even when it was really a read lock.
613 PH/07 The appendfile transport was creating MBX lock files with a fixed mode
614 of 0600. This has been changed to use the value of the lockfile_mode
615 option (which defaults to 0600).
617 PH/08 Applied small patch from the Sieve maintainer.
619 PH/09 If maildir_quota_directory_regex was set to exclude (say) the .Trash
620 folder from quota calculations, a direct delivery into this folder messed
621 up the contents of the maildirsize file. This was because the regex was
622 used only to exclude .Trash (or whatever) when the size of the mailbox
623 was calculated. There was no check that a delivery was happening into an
624 excluded directory. This bug has been fixed by ignoring all quota
625 processing for deliveries into excluded directories.
627 PH/10 Added the maildirfolder_create_regex option to appendfile.
633 PH/01 The code for finding all the local interface addresses on a FreeBSD
634 system running IPv6 was broken. This may well have applied to all BSD
635 systems, as well as to others that have similar system calls. The broken
636 code found IPv4 interfaces correctly, but gave incorrect values for the
637 IPv6 interfaces. In particular, ::1 was not found. The effect in Exim was
638 that it would not match correctly against @[] and not recognize the IPv6
641 PH/02 The ipliteral router was not recognizing addresses of the form user@
642 [ipv6:....] because it didn't know about the "ipv6:" prefix.
644 PH/03 Added disable_ipv6.
646 PH/04 Changed $reply_address to use the raw form of the headers instead of the
647 decoded form, because it is most often used to construct To: headers
648 lines in autoreplies, and the decoded form may well be syntactically
649 invalid. However, $reply_address has leading white space removed, and all
650 newlines turned into spaces so that the autoreply transport does not
653 PH/05 If group was specified without a user on a router, and no group or user
654 was specified on a transport, the group from the router was ignored.
656 PH/06 Increased the number of ACL variables to 20 of each type, and arranged
657 for visible compile-time settings that can be used to change these
658 numbers, for those that want even more. Backwards compatibility with old
659 spool files has been maintained. However, going back to a previous Exim
660 release will lost any variables that are in spool files.
662 PH/07 Two small changes when running in the test harness: increase delay when
663 passing a TCP/IP connection to a new process, in case the original
664 process has to generate a bounce, and remove special handling of
665 127.0.0.2 (sic), which is no longer necessary.
667 PH/08 Changed debug output of dbfn_open() flags from numbers to names, so as to
668 be the same on different OS.
670 PH/09 Moved a debug statement in filter processing to avoid a race problem when
673 JJ/01 exipick: fixed bug where -b (brief) output option showed "Vars:"
674 whether --show-vars was specified or not
676 JJ/02 exipick: Added support for new ACL variable spool format introduced
679 PH/10 Fixed another bug related to PH/04 above: if an incoming message had a
680 syntactically invalid From: or Reply-to: line, and a filter used this to
681 generate an autoreply, and therefore failed to obtain an address for the
682 autoreply, Exim could try to deliver to a non-existent relative file
683 name, causing unrelated and misleading errors. What now happens is that
684 it logs this as a hard delivery error, but does not attempt to create a
687 PH/11 The exinext utility has a -C option for testing purposes, but although
688 the given file was scanned by exinext itself; it wasn't being passed on
689 when Exim was called.
691 PH/12 In the smtp transport, treat an explicit ECONNRESET error the same as
692 an end-of-file indication when reading a command response.
694 PH/13 Domain literals for IPv6 were not recognized unless IPv6 support was
695 compiled. In many other places in Exim, IPv6 addresses are always
696 recognized, so I have changed this. It also means that IPv4 domain
697 literals of the form [IPV4:n.n.n.n] are now always recognized.
699 PH/14 When a uid/gid is specified for the queryprogram router, it cannot be
700 used if the router is not running as root, for example, when verifying at
701 ACL time, or when using -bh. The debugging output from this situation was
702 non-existent - all you got was a failure to exec. I have made two
705 (a) Failures to set uid/gid, the current directory, or a process leader
706 in a subprocess such as that created by queryprogram now generate
707 suitable debugging ouput when -d is set.
709 (b) The queryprogram router detects when it is not running as root,
710 outputs suitable debugging information if -d is set, and then runs
711 the subprocess without attempting to change uid/gid.
713 PH/15 Minor change to Makefile for building test_host (undocumented testing
716 PH/16 As discussed on the list in Nov/Dec: Exim no longer looks at the
717 additional section of a DNS packet that returns MX or SRV records.
718 Instead, it always explicitly searches for A/AAAA records. This avoids
719 major problems that occur when a DNS server includes only records of one
720 type (A or AAAA) in an MX/SRV packet. A byproduct of this change has
721 fixed another bug: if SRV records were looked up and the corresponding
722 address records were *not* found in the additional section, the port
723 values from the SRV records were lost.
725 PH/17 If a delivery to a pipe, file, or autoreply was deferred, Exim was not
726 using the correct key (the original address) when searching the retry
727 rules in order to find which one to use for generating the retry hint.
729 PH/18 If quota_warn_message contains a From: header, Exim now refrains from
730 adding the default one. Similarly, if it contains a Reply-To: header, the
731 errors_reply_to option, if set, is not used.
733 PH/19 When calculating a retry time, Exim used to measure the "time since
734 failure" by looking at the "first failed" field in the retry record. Now
735 it does not use this if it is later than than the arrival time of the
736 message. Instead it uses the arrival time. This makes for better
737 behaviour in cases where some deliveries succeed, thus re-setting the
738 "first failed" field. An example is a quota failure for a huge message
739 when small messages continue to be delivered. Without this change, the
740 "time since failure" will always be short, possible causing more frequent
741 delivery attempts for the huge message than are intended.
742 [Note: This change was subsequently modified - see PH/04 for 4.62.]
744 PH/20 Added $auth1, $auth2, $auth3 to contain authentication data (as well as
745 $1, $2, $3) because the numerical variables can be reset during some
746 expansion items (e.g. "match"), thereby losing the authentication data.
748 PH/21 Make -bV show the size of off_t variables so that the test suite can
749 decide whether to run tests for quotas > 2G.
751 PH/22 Test the values given for quota, quota_filecount, quota_warn_threshold,
752 mailbox_size, and mailbox_filecount in the appendfile transport. If a
753 filecount value is greater than 2G or if a quota value is greater than 2G
754 on a system where the size of off_t is not greater than 4, a panic error
757 PH/23 When a malformed item such as 1.2.3/24 appears in a host list, it can
758 never match. The debug and -bh output now contains an explicit error
759 message indicating a malformed IPv4 address or mask.
761 PH/24 An host item such as 1.2.3.4/abc was being treated as the IP address
762 1.2.3.4 without a mask. Now it is not recognized as an IP address, and
765 PH/25 Do not write to syslog when running in the test harness. The only
766 occasion when this arises is a failure to open the main or panic logs
767 (for which there is an explicit test).
769 PH/26 Added the /no_tell option to "control=freeze".
771 PH/27 If a host name lookup failed very early in a connection, for example, if
772 the IP address matched host_lookup and the reverse lookup yielded a name
773 that did not have a forward lookup, an error message of the form "no IP
774 address found for host xxx.xxx.xxx (during SMTP connection from NULL)"
775 could be logged. Now it outputs the IP address instead of "NULL".
777 PH/28 An enabling patch from MH: add new function child_open_exim2() which
778 allows the sender and the authenticated sender to be set when
779 submitting a message from within Exim. Since child_open_exim() is
780 documented for local_scan(), the new function should be too.
782 PH/29 In GnuTLS, a forced expansion failure for tls_privatekey was not being
783 ignored. In both GnuTLS and OpenSSL, an expansion of tls_privatekey that
784 results in an empty string is now treated as unset.
786 PH/30 Fix eximon buffer overflow bug (Bugzilla #73).
788 PH/31 Added sender_verify_fail logging option.
790 PH/32 In November 2003, the code in Exim that added an empty Bcc: header when
791 needed by RFC 822 but not by RFC 2822 was commented out. I have now
792 tidied the source and removed it altogether.
794 PH/33 When a queue run was abandoned because the load average was too high, a
795 log line was always written; now it is written only if the queue_run log
796 selector is set. In addition, the log line for abandonment now contains
797 information about the queue run such as the pid. This is always present
798 in "start" and "stop" lines but was omitted from the "abandon" line.
800 PH/34 Omit spaces between a header name and the colon in the error message that
801 is given when verify = headers_syntax fails (if there are lots of them,
802 the message gets confusing).
804 PH/35 Change the default for dns_check_names_pattern to allow slashes within
805 names, as there are now some PTR records that contain slashes. This check
806 is only to protect against broken name servers that fall over on strange
807 characters, so the fact that it applies to all lookups doesn't matter.
809 PH/36 Now that the new test suite is complete, we can remove some of the
810 special code in Exim that was needed for the old test suite. For example,
811 sorting DNS records because real resolvers return them in an arbitrary
812 order. The new test suite's fake resolver always returns records in the
815 PH/37 When running in the test harness, use -odi for submitted messages (e.g.
816 bounces) except when queue_only is set, to avoid logging races between
817 the different processes.
819 PH/38 Panic-die if .include specifies a non-absolute path.
821 PH/39 A tweak to the "H" retry rule from its user.
823 JJ/03 exipick: Removed parentheses from 'next' and 'last' calls that specified
824 a label. They prevented compilation on older perls.
826 JJ/04 exipick: Refactored code to prevent implicit split to @_ which caused
827 a warning to be raised on newish perls.
829 JJ/05 exipick: Fixed bug where -bpc always showed a count of all messages
830 on queue. Changes to match documented behaviour of showing count of
831 messages matching specified criteria.
833 PH/40 Changed the default ident timeout from 30s to 5s.
835 PH/41 Added support for the use of login_cap features, on those BSD systems
836 that have them, for controlling the resources used by pipe deliveries.
838 PH/42 The content-scanning code uses fopen() to create files in which to put
839 message data. Previously it was not paying any attention to the mode of
840 the files. Exim runs with umask(0) because the rest of the code creates
841 files with open(), and sets the required mode explicitly. Thus, these
842 files were ending up world-writeable. This was not a big issue, because,
843 being within the spool directory, they were not world-accessible. I have
844 created a function called modefopen, which takes an additional mode
845 argument. It sets umask(777), creates the file, chmods it to the required
846 mode, then resets the umask. All the relevant calls to fopen() in the
847 content scanning code have been changed to use this function.
849 PH/43 If retry_interval_max is set greater than 24 hours, it is quietly reset
850 to 24 hours. This avoids potential overflow problems when processing G
851 and H retry rules. I suspect nobody ever tinkers with this value.
853 PH/44 Added STRIP_COMMAND=/usr/bin/strip to the FreeBSD Makefile.
855 PH/45 When the plaintext authenticator is running as a client, the server's
856 challenges are checked to ensure they are valid base64 strings. By
857 default, the authentication attempt is cancelled if an invalid string is
858 received. Setting client_ignore_invalid_base64 true ignores these errors.
859 The decoded challenge strings are now placed in $auth1, $auth2, etc. as
860 they are received. Thus, the responses can be made to depend on the
861 challenges. If an invalid string is ignored, an empty string is placed in
864 PH/46 Messages that are created by the autoreply transport now contains a
865 References: header, in accordance with RFCs 2822 and 3834.
867 PH/47 Added authenticated_sender_force to the smtp transport.
869 PH/48 The ${prvs expansion was broken on systems where time_t was long long.
871 PH/49 Installed latest patch from the Sieve maintainer.
873 PH/50 When an Exim quota was set without a file count quota, and mailbox_size
874 was also set, the appendfile transport was unnecessarily scanning a
875 directory of message files (e.g. for maildir delivery) to find the count
876 of files (along with the size), even though it did not need this
877 information. It now does the scan only if it needs to find either the
878 size of the count of files.
880 PH/51 Added ${time_eval: to convert Exim time strings into seconds.
882 PH/52 Two bugs concerned with error handling when the smtp transport is
885 (i) Exim was not creating retry information for temporary errors given
886 for individual recipients after the DATA command when the smtp transport
887 was used in LMTP mode. This meant that they could be retried too
888 frequently, and not timed out correctly.
890 (ii) Exim was setting the flag that allows error details to be returned
891 for LMTP errors on RCPT commands, but not for LMTP errors for individual
892 recipients that were returned after the DATA command.
894 PH/53 This is related to PH/52, but is more general: for any failing address,
895 when detailed error information was permitted to be returned to the
896 sender, but the error was temporary, then after the final timeout, only
897 "retry timeout exceeded" was returned. Now it returns the full error as
898 well as "retry timeout exceeded".
900 PH/54 Added control=allow_auth_unadvertised, as it seems there are clients that
901 do this, and (what is worse) MTAs that accept it.
903 PH/55 Added the add_header modified to ACLs. The use of "message" with "warn"
904 will now be deprecated.
906 PH/56 New os.c-cygwin from the Cygwin maintainer.
908 JJ/06 exipick: added --unsorted option to allow unsorted output in all output
909 formats (previously only available in exim formats via -bpr, -bpru,
910 and -bpra. Now also available in native and exiqgrep formats)
912 JJ/07 exipick: added --freeze and --thaw options to allow faster interaction
913 with very large, slow to parse queues
915 JJ/08 exipick: added ! as generic prefix to negate any criteria format
917 JJ/09 exipick: miscellaneous performance enhancements (~24% improvements)
919 PH/57 Tidies in SMTP dialogue display in debug output: (i) It was not showing
920 responses to authentication challenges, though it was showing the
921 challenges; (ii) I've removed the CR characters from the debug output for
924 PH/58 Allow for the insertion of a newline as well as a space when a string
925 is turned into more than one encoded-word during RFC 2047 encoding. The
926 Sieve code now uses this.
928 PH/59 Added the following errors that can be detected in retry rules: mail_4xx,
929 data_4xx, lost_connection, tls_required.
931 PH/60 When a VRFY deferred or FAILED, the log message rather than the user
932 message was being sent as an SMTP response.
934 PH/61 Add -l and -k options to exicyclog.
936 PH/62 When verifying, if an address was redirected to one new address, so that
937 verification continued, and the new address failed or deferred after
938 having set something in $address_data, the value of $address_data was not
939 passed back to the ACL. This was different to the case when no
940 redirection occurred. The value is now passed back in both cases.
942 PH/63 Changed the macro HAVE_LOGIN_CAP (see PH/41 for this release above) to
943 HAVE_SETCLASSRESOURCES because there are different APIs in use that all
944 use login_cap.h, so on its own it isn't the distinguishing feature. The
945 new name refers directly to the setclassresources() function.
947 PH/65 Added configuration files for NetBSD3.
949 PH/66 Updated OS/Makefile-HP-UX for gcc 4.1.0 with HP-UX 11.
951 PH/67 Fixed minor infelicity in the sorting of addresses to ensure that IPv6
952 is preferred over IPv4.
954 PH/68 The bounce_return_message and bounce_return_body options were not being
955 honoured for bounces generated during the reception of non-SMTP messages.
956 In particular, this applied to messages rejected by the ACL. This bug has
957 been fixed. However, if bounce_return_message is true and bounce_return_
958 body is false, the headers that are returned for a non-SMTP message
959 include only those that have been read before the error was detected.
960 (In the case of an ACL rejection, they have all been read.)
962 PH/69 The HTML version of the specification is now built in a directory called
963 spec_html instead of spec.html, because the latter looks like a path with
964 a MIME-type, and this confuses some software.
966 PH/70 Catch two compiler warnings in sieve.c.
968 PH/71 Fixed an obscure and subtle bug (thanks Alexander & Matthias). The
969 function verify_get_ident() calls ip_connect() to connect a socket, but
970 if the "connect()" function timed out, ip_connect() used to close the
971 socket. However, verify_get_ident() also closes the socket later, and in
972 between Exim writes to the log, which may get opened at this point. When
973 the socket was closed in ip_connect(), the log could get the same file
974 descriptor number as the socket. This naturally causes chaos. The fix is
975 not to close the socket in ip_connect(); the socket should be closed by
976 the function that creates it. There was only one place in the code where
977 this was missing, in the iplookup router, which I don't think anybody now
978 uses, but I've fixed it anyway.
980 PH/72 Make dns_again_means_nonexist apply to lookups using gethostbyname() as
981 well as to direct DNS lookups. Otherwise the handling of names in host
982 lists is inconsistent and therefore confusing.
988 PH/01 Two changes to the default runtime configuration:
990 (1) Move the checks for relay_from_hosts and authenticated clients from
991 after to before the (commented out) DNS black list checks.
993 (2) Add control=submission to the relay_from_hosts and authenticated
994 clients checks, on the grounds that messages accepted by these
995 statements are most likely to be submissions.
997 PH/02 Several tidies to the handling of ${prvs and ${prvscheck:
999 (1) Generate an error if the third argument for the ${prvs expansion is
1002 (2) Treat a missing third argument of ${prvscheck as if it were an empty
1005 (3) Reset the variables that are obtained from the first argument of
1006 ${prvscheck and used in the second argument before leaving the code,
1007 because their memory is reclaimed, so using them afterwards may do
1010 (4) Tidy up the code for expanding the arguments of ${prvscheck one by
1011 one (it's much easier than Tom thought :-).
1013 (5) Because of (4), we can now allow for the use of $prvscheck_result
1014 inside the third argument.
1016 PH/03 For some reason, the default setting of PATH when running a command from
1017 a pipe transport was just "/usr/bin". I have changed it to
1020 PH/04 SUPPORT_TRANSLATE_IP_ADDRESS and MOVE_FROZEN_MESSAGES did not cause
1021 anything to be listed in the output from -bV.
1023 PH/05 When a filter generated an autoreply, the entire To: header line was
1024 quoted in the delivery log line, like this:
1026 => >A.N.Other <ano@some.domain> <original@ddress> ...
1028 This has been changed so that it extracts the operative address. There
1029 may be more than one such address. If so, they are comma-separated, like
1032 => >ano@some.domain,ona@other.domain <original@ddress> ...
1034 PH/06 When a client host used a correct literal IP address in a HELO or EHLO
1035 command, (for example, EHLO [1.2.3.4]) and the client's IP address was
1036 not being looked up in the rDNS to get a host name, Exim was showing the
1037 IP address twice in Received: lines, even though the IP addresses were
1038 identical. For example:
1040 Received: from [1.2.3.4] (helo=[1.2.3.4])
1042 However, if the real host name was known, it was omitting the HELO data
1043 if it matched the actual IP address. This has been tidied up so that it
1044 doesn't show the same IP address twice.
1046 PH/07 When both +timestamp and +memory debugging was on, the value given by
1047 $tod_xxx expansions could be wrong, because the tod_stamp() function was
1048 called by the debug printing, thereby overwriting the timestamp buffer.
1049 Debugging no longer uses the tod_stamp() function when +timestamp is set.
1051 PH/08 When the original message was included in an autoreply transport, it
1052 always said "this is a copy of the message, including all the headers",
1053 even if body_only or headers_only was set. It now gives an appropriate
1056 PH/09 Applied a patch from the Sieve maintainer which:
1058 o fixes some comments
1059 o adds the (disabled) notify extension core
1060 o adds some debug output for the result of if/elsif tests
1061 o points to the current vacation draft in the documentation
1062 and documents the missing references header update
1066 o fixes a bug in processing the envelope test (when testing
1067 multiple envelope elements, the last element determinted the
1070 PH/10 Exim was violating RFC 3834 ("Recommendations for Automatic Responses to
1071 Electronic Mail") by including:
1073 Auto-submitted: auto-generated
1075 in the messages that it generates (bounce messages and others, such as
1076 warnings). In the case of bounce messages for non-SMTP mesages, there was
1077 also a typo: it was using "Auto_submitted" (underscore instead of
1078 hyphen). Since every message generated by Exim is necessarily in response
1079 to another message, thes have all been changed to:
1081 Auto-Submitted: auto-replied
1083 in accordance with these statements in the RFC:
1085 The auto-replied keyword:
1087 - SHOULD be used on messages sent in direct response to another
1088 message by an automatic process,
1090 - MUST NOT be used on manually-generated messages,
1092 - MAY be used on Delivery Status Notifications (DSNs) and Message
1093 Disposition Notifications (MDNs),
1095 - MUST NOT be used on messages generated by automatic or periodic
1096 processes, except for messages which are automatic responses to
1099 PH/11 Added "${if def:sender_address {(envelope-from <$sender_address>)\n\t}}"
1100 to the default Received: header definition.
1102 PH/12 Added log selector acl_warn_skipped (default on).
1104 PH/13 After a successful wildlsearch lookup, discard the values of numeric
1105 variables because (a) they are in the wrong storage pool and (b) even if
1106 they were copied, it wouldn't work properly because of the caching.
1108 PH/14 Add check_rfc2047_length to disable enforcement of RFC 2047 length
1109 checking when decoding. Apparently there are clients that generate
1110 overlong encoded strings. Why am I not surprised?
1112 PH/15 If the first argument of "${if match_address" was not empty, but did not
1113 contain an "@" character, Exim crashed. Now it writes a panic log message
1114 and treats the condition as false.
1116 PH/16 In autoreply, treat an empty string for "once" the same as unset.
1118 PH/17 A further patch from the Sieve maintainer: "Introduce the new Sieve
1119 extension "envelope-auth". The code is finished and in agreement with
1120 other implementations, but there is no documentation so far and in fact,
1121 nobody wrote the draft yet. This extension is currently #undef'ed, thus
1122 not changing the active code.
1124 Print executed "if" and "elsif" statements when debugging is used. This
1125 helps a great deal to understand what a filter does.
1127 Document more things not specified clearly in RFC3028. I had all this
1128 sorted out, when out of a sudden new issues came to my mind. Oops."
1130 PH/18 Exim was not recognizing the "net-" search type prefix in match_ip lists
1133 PH/19 Exim expands the IPv6 address given to -bh to its full non-abbreviated
1134 canonical form (as documented). However, after a host name lookup from
1135 the IP address, check_host() was doing a simple string comparison with
1136 addresses acquired from the DNS when checking that the found name did
1137 have the original IP as one of its addresses. Since any found IPv6
1138 addresses are likely to be in abbreviated form, the comparison could
1139 fail. Luckily, there already exists a function for doing the comparison
1140 by converting both addresses to binary, so now that is used instead of
1141 the text comparison.
1143 PH/20 There was another similar case to PH/19, when a complete host name was
1144 given in a host list; looking up its IP address could give an abbreviated
1145 form, whereas the current host's name might or might not be abbreviated.
1146 The same fix has been applied.
1152 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
1153 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
1156 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
1157 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
1159 PH/03 Typo: missing ".o" in src/pcre/Makefile.
1161 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
1162 header line, restrict the check to what is listed in RFCs 2369 and 2929.
1163 Also, for "Auto-Submitted", treat anything other than "no" as
1164 non-personal, in accordance with RFC 3834. (Previously it treated
1165 anything starting "auto-" as non-personal.)
1167 TF/01 The control=submission/name=... option had a problem with syntax
1168 errors if the name included a slash character. The /name= option
1169 now slurps the rest of the string, so it can include any characters
1170 but it must come last in the list of options (after /sender_retain
1173 PH/05 Some modifications to the interface to the fake nameserver for the new
1181 TK/01 Added the "success_on_redirect" address verification option. See
1182 NewStuff for rationale and an example.
1184 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
1186 PH/02 Patch to exigrep to allow it to work on syslog lines.
1188 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
1189 fread() to skip over the body file's header line, because in Cygwin the
1190 header line is locked and is inaccessible.
1192 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
1193 co-exist for some time) to make it clear that it is the Exim ID that is
1194 referenced, not the Message-ID: header line.
1196 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
1197 string_format() function, because snprintf() does not exist on all
1200 PH/06 The use of forbid_filter_existstest now also locks out the use of the
1201 ${stat: expansion item.
1203 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
1204 protocol synchronization error", to keep the pedants happy.
1206 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
1207 well as for IRIX systems, when gcc is being used. See the host.c source
1210 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
1212 PH/10 Named domain lists were not working if used in a queue_smtp_domains
1215 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
1216 transport and to the smtp transport in LMTP mode.
1218 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
1220 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
1221 run a filter in a subprocess. This could lead to confusion in subsequent
1222 lookups in the parent process. There should also be a search_tidyup() at
1223 the end of the subprocess.
1225 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
1226 only if the host matched helo_try_verify_hosts, which caused the
1227 verification to occur when the EHLO/HELO command was issued. The ACL just
1228 tested the remembered result. Now, if a previous verification attempt has
1229 not happened, "verify = helo" does it there and then.
1231 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
1233 TK/03 Fix log output including CR from clamd.
1235 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
1236 exist provoked a memory error which could cause a segfault.
1238 PH/15 Installed PCRE 6.2
1240 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
1242 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
1243 of the problem. Specifically, suggested +O2 rather than +O1 for the
1246 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
1248 PH/20 If a delivery was routed to a non-standard port by means of an SRV
1249 record, the port was not correctly logged when the outgoing_port log
1250 selector was set (it logged the transort's default port).
1252 PH/21 Added support for host-specific ports to manualroute, queryprogram,
1253 fallback_hosts, and "hosts" in the smtp transport.
1255 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
1256 host errors such as "Connection refused".
1258 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
1259 authentication with radiusclient 0.4.9:
1261 - Error returned from rc_read_config was caught wrongly
1262 - Username/password not passed on to radius server due to wrong length.
1264 The presumption is that some radiusclient API changes for 4.51/PH/17
1265 were not taken care of correctly. The code is still untested by me (my
1266 Linux distribution still has 0.3.2 of radiusclient), but it was
1267 contributed by a Radius user.
1269 PH/24 When doing a callout, the value of $domain wasn't set correctly when
1270 expanding the "port" option of the smtp transport.
1272 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
1273 while reading a MIME header. Thanks to Tom Hughes for a patch.
1275 PH/24 Include config.h inside local_scan.h so that configuration settings are
1278 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
1279 that in an ACL for RCPT (for example), you can examine exactly what was
1282 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
1283 commands, but it was not correctly comparing the address with the actual
1284 client host address. Thus, it would show the EHLO address in Received:
1285 header lines when this was not necessary.
1287 PH/27 Added the % operator to ${eval:}.
1289 PH/28 Exim tries to create and chdir to its spool directory when it starts;
1290 it should be ignoring failures (because with -C, for example, it has lost
1291 privilege). It wasn't ignoring creation failures other than "already
1294 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
1297 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
1298 because an input line was too long, either on its own, or by virtue of
1299 too many continuations, the temporary file was not being removed, and the
1300 return code was incorrect.
1302 PH/31 Missing "BOOL" in function definition in filtertest.c.
1304 PH/32 Applied Sieve patches from the maintainer.
1306 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
1308 PH/33 Added "verify = not_blind".
1310 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
1311 Local/Makefile (with some defaults set). These are used in built scripts
1312 such as exicyclog, but they have never been used in the exim_install
1313 script (though there are many overriding facilities there). I have
1314 arranged that the exim_install script now takes note of these two
1317 PH/35 Installed configuration files for Dragonfly.
1319 PH/36 When a locally submitted message by a trusted user did not contain a
1320 From: header, and the sender address was obtained from -f or from an SMTP
1321 MAIL command, and the trusted user did not use -F to supply a sender
1322 name, $originator_name was incorrectly used when constructing a From:
1323 header. Furthermore, $originator_name was used for submission mode
1324 messages from external hosts without From: headers in a similar way,
1325 which is clearly wrong.
1327 PH/37 Added control=suppress_local_fixups.
1329 PH/38 When log_selector = +received_sender was set, and the addition of the
1330 sender made the log line's construction buffer exactly full, or one byte
1331 less than full, an overflow happened when the terminating "\n" was
1334 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
1335 when the result of a list match is failure because a DNS lookup failed.
1337 PH/40 RM_COMMAND is now used in the building process.
1339 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
1340 the "build-* directories that it finds.
1342 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
1343 address in a domain literal was a prefix of an interface address.
1345 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
1346 when verifying a sender address, unless rewrite_headers is false.
1348 PH/44 Wrote a long comment about why errors_to addresses are verified as
1349 recipients, not senders.
1351 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
1352 the ratelimit ACL was added.
1354 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
1356 PH/46 Added extra information about PostgreSQL errors to the error string.
1358 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
1359 avoiding the need to install special zones in a real server. This is
1360 backwards compatible; if it can't find the fake resolver, it drops back.
1361 Thus, both old and new test suites can be run.
1363 TF/02 Added util/ratelimit.pl
1365 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
1366 clock is set back in time.
1368 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
1369 Candler <B.Candler@pobox.com>.
1371 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
1372 OK for addresses that are the result of redirections.
1374 PH/48 A number of further additions for the benefit of the new test suite,
1375 including a fake gethostbyname() that interfaces to the fake DNS resolver
1378 TF/06 The fix for widen_domains has also been applied to qualify_single and
1379 search_parents which are the other dnslookup options that can cause
1382 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
1385 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
1387 TF/07 Exim produced the error message "an SRV record indicated no SMTP
1388 service" if it encountered an MX record with an empty target hostname.
1389 The message is now "an MX or SRV record indicated no SMTP service".
1391 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
1392 if the DNS of the sending site is misconfigured. This is quite a
1393 common situation. This change restores the behaviour of treating a
1394 helo verification defer as a failure.
1396 PH/51 If self=fail was set on a router, the bounce message did not include the
1397 actual error message.
1403 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
1405 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
1406 command itself ended in error, the underlying message about the transport
1407 filter timeout was being overwritten with the pipe command error. Now the
1408 underlying error message should be appended to the second error message.
1410 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
1412 PH/02 Reduce the amount of output that "make" produces by default. Full output
1413 can still be requested.
1415 PH/03 The warning log line about a condition test deferring for a "warn" verb
1416 was being output only once per connection, rather than after each
1417 occurrence (because it was using the same function as for successful
1418 "warn" verbs). This seems wrong, so I have changed it.
1420 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
1421 it should not have, which might have caused a crash in the right
1422 circumstances, but probably never did.
1424 PH/04 Installed a modified version of Tony Finch's patch to make submission
1425 mode fix the return path as well as the Sender: header line, and to
1426 add a /name= option so that you can make the user's friendly name appear
1429 TF/03 Added the control = fakedefer ACL modifier.
1431 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
1432 Mark Lowes for thorough testing.
1434 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
1436 TK/03 Merged latest SRS patch from Miles Wilton.
1438 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
1439 with the definition in sysexits.h (which is #included earlier).
1440 Fortunately, Exim does not actually use EX_OK. The code used to try to
1441 preserve the sysexits.h value, by assumimg that macro definitions were
1442 scanned for macro replacements. I have been disabused of this notion,
1443 so now the code just undefines EX_OK before #including unistd.h.
1445 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
1446 in the smtp transport. When a block could not be written in a single
1447 write() function, the timeout was being re-applied to each part-write.
1448 This seems wrong - if the receiver was accepting one byte at a time it
1449 would take for ever. The timeout is now adjusted when this happens. It
1450 doesn't have to be particularly precise.
1452 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
1453 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
1455 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
1456 without a domain if the check to <postmaster@domain> fails.
1458 SC/01 Eximstats: added -xls and the ability to specify output files
1459 (patch written by Frank Heydlauf).
1461 SC/02 Eximstats: use FileHandles for outputing results.
1463 SC/03 Eximstats: allow any combination of xls, txt, and html output.
1465 SC/04 Eximstats: fixed display of large numbers with -nvr option
1467 SC/05 Eximstats: fixed merging of reports with empty tables.
1469 SC/06 Eximstats: added the -include_original_destination flag
1471 SC/07 Eximstats: removed tabs and trailing whitespace.
1473 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
1475 TK/06 MBOX spool code: Add real "From " MBOX separator line
1476 so the .eml file is really in mbox format (even though
1477 most programs do not really care). Patch from Alex Miller.
1479 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
1480 The latter is generated from $received_to and is only set if the
1481 message has one envelope recipient. SA can use these headers,
1482 obviously out-of-the-box. Patch from Alex Miller.
1484 PH/08 The ${def test on a variable was returning false if the variable's
1485 value was "0", contrary to what the specification has always said!
1486 The result should be true unless the variable is empty.
1488 PH/09 The syntax error of a character other than { following "${if
1489 def:variable_name" (after optional whitespace) was not being diagnosed.
1490 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
1491 accidental colon was present, for example, could give incorrect results.
1493 PH/10 Tidied the code in a number of places where the st_size field of a stat()
1494 result is used (not including appendfile, where other changes are about
1497 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
1498 This involved changing a lot of size variables from int to off_t. It
1499 should work with maildirs and everything.
1501 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
1502 spamd dying while we are connected to it.
1504 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
1507 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
1508 function for BDB 4.3.
1510 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
1512 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
1513 the sources for PCRE have been re-arranged and more files are now
1516 PH/15 The code I had for printing potentially long long variables in PH/11
1517 above was not the best (it lost precision). The length of off_t variables
1518 is now inspected at build time, and an appropriate printing format (%ld
1519 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
1520 to be "long long int" or "long int". This is needed for the internal
1521 formatting function string_vformat().
1523 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
1524 the configuration file to be ":syslog", then the script "guesses" where
1525 the logs files are, rather than using the compiled in default. In our
1526 case the guess is not the same as the compiled default, so the script
1527 suddenly stopped working when I started to use syslog. The patch checks
1528 to see if log_file_path is "". If so, it attempts to read it from exim
1529 with no configuration file to get the compiled in version, before it
1530 falls back to the previous guessing code."
1532 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
1533 implementing BATV in an Exim configuration. See NewStuff for the gory
1536 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
1537 Makefile that are specific to HP-UX.
1539 PH/18 If the "use_postmaster" option was set for a recipient callout together
1540 with the "random" option, the postmaster address was used as the MAIL
1541 FROM address for the random test, but not for the subsequent recipient
1542 test. It is now used for both.
1544 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
1545 patch removes a few documentation additions to RFC 3028, because the
1546 latest draft now contains them. It adds the new en;ascii-case comparator
1547 and a new error check for 8bit text in MIME parts. Comparator and
1548 require names are now matched exactly. I enabled the subaddress
1549 extension, but it is not well tested yet (read: it works for me)."
1551 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
1552 rework some of the code of TK/09 above to avoid the hardwired use of
1553 "%lld" and "long long". Replaced the call to snprintf() with a call to
1556 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
1557 records point to non-existent hosts", "retry timeout exceeded", and
1558 "retry time not reached for any host after a long failure period".
1560 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
1561 experimental DomainKeys support:
1563 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
1564 (2) On an error such as an illegally used "control", the wrong name for
1565 the control was given.
1567 These problems did NOT occur unless DomainKeys support was compiled.
1569 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
1571 PH/24 Added ${if match_ip condition.
1573 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
1574 databases so that it will be absolutely obvious if a crash occurs in the
1575 DB library. This is a regular occurrence (often caused by mis-matched
1578 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
1579 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
1580 picked up on a user's system that detects such things. There doesn't seem
1581 to be a gcc warning option for this - only an attribute that has to be
1582 put on the function's prototype. It seems that in Fedora Core 4 they have
1583 set this on a number of new functions. No doubt there will be more in due
1586 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
1587 specify a transport. However, if an address that was verified by such a
1588 router was the subject of a callout, Exim crashed because it tried to
1589 read the rcpt_include_affixes from the non-existent transport. Now it
1590 just assumes that the setting of that option is false. This bug was
1591 introduced by 4.51/PH/31.
1593 PH/28 Changed -d+all to exclude +memory, because that information is very
1594 rarely of interest, but it makes the output a lot bigger. People tend to
1595 do -d+all out of habit.
1597 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
1598 code in os-type was giving problems when libc.so lives in lib64, like on
1601 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
1602 aren't the modern standard, and it seems that some systems' include files
1603 don't always have them. Exim was already checking for some of the newer
1604 ones like T_AAAA, and defining it itself. I've added checks for all the
1605 record types that Exim uses.
1607 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
1608 not automatically generating a new one, as it is supposed to. This
1609 prevented TLS from working. If the file did exist, but contained invalid
1610 data, a new version was generated, as expected. It was only the case of a
1611 non-existent file that was broken.
1613 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
1614 with a change in libdomainkeys > 0.64.
1616 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
1617 from DNS. If the selector record carries the flag, it now has
1618 precedence over the domain-wide flag.
1620 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
1622 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
1623 the use of a port name that isn't defined in /etc/services) occurred, the
1624 message was deferred as in a normal delivery, and thus remained on the
1625 spool, instead of being failed because of the mua_wrapper setting. This
1626 is now fixed, and I tidied up some of the mua_wrapper messages at the
1629 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
1630 the messages in a hash of arrays rather than using individual hashes.
1631 This is a bit cleaner and results in dramatic memory savings, albeit
1632 at a slight CPU cost.
1634 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
1635 as requested by Marc Sherman.
1637 SC/10 Eximstats: added histograms for user specified patterns as requested
1640 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
1642 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
1643 fopen() in the content-scanning modules that did not already have it.
1649 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
1650 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
1652 TK/02 Fix ACL "control" statement not being available in MIME ACL.
1654 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
1656 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
1657 to test Sieve filters that use "vacation".
1659 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
1660 that changes the way the GnuTLS parameters are stored in the cache file.
1661 The new format can be generated externally. For backward compatibility,
1662 if the data in the cache doesn't make sense, Exim assumes it has read an
1663 old-format file, and it generates new data and writes a new file. This
1664 means that you can't go back to an older release without removing the
1667 PH/03 A redirect router that has both "unseen" and "one_time" set does not
1668 work if there are any delivery delays because "one_time" forces the
1669 parent to be marked "delivered", so its unseen clone is never tried
1670 again. For this reason, Exim now forbids the simultaneous setting of
1673 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
1674 redirected to themselves ("homonym" addresses). Read the long ChangeLog
1675 entry if you want to know the details. The fix, however, neglected to
1676 consider the case when local delivery batching is involved. The test for
1677 "previously delivered" was not happening when checking to see if an
1678 address could be batched with a previous (undelivered) one; under
1679 certain circumstances this could lead to multiple deliveries to the same
1682 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
1683 in its include files, and this causes problems building Exim.
1685 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
1686 header_syntax) but Exim was just ignoring anything given after a slash.
1687 In particular, this caused confusion with an attempt to use "verify =
1688 reverse_host_lookup/defer_ok". An error is now given when options are
1689 supplied for verify items that do not have them. (Maybe reverse_host_
1690 lookup should have a defer_ok option, but that's a different point.)
1692 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
1693 defined by RFC 821) to 2048, because there were problems with some AUTH
1694 commands, and RFC 1869 says the size should be increased for extended
1695 SMTP commands that take arguments.
1697 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
1700 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
1701 "unknown" error; now it says that the functionality isn't in the binary.
1703 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
1704 an address' error message when a string expansion fails (syntax or
1705 whatever). Otherwise the password may appear in the log. Following change
1706 PH/42 below, there is no longer a chance of it appearing in a bounce
1709 PH/11 Installed exipick version 20050225.0 from John Jetmore.
1711 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
1712 of its addresses was ever tried. (Bugzilla bug #2.)
1714 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
1715 the result incorrectly in the debug output. (It correctly added a newline
1716 to what was transported.)
1718 TF/01 Added $received_time.
1720 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
1721 commented out examples of how to interface to a virus scanner and to
1722 SpamAssassin. Also added commented examples of av_scanner and
1723 spamd_address settings.
1725 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
1726 and controls are allowed in which ACLs. There were a couple of minor
1727 errors. Some of the entries in the conditions table (which is a table of
1728 where they are NOT allowed) were getting very unwieldy; rewrote them as a
1729 negation of where the condition IS allowed.
1731 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
1733 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
1734 header file does not have a version number, so I've had to invent a new
1735 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
1736 API. The code is untested by me (my Linux distribution still has 0.3.2 of
1737 radiusclient), but it was contributed by a Radius user.
1739 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
1740 files or directories, for OpenSSL.
1742 PH/19 When an Exim process that is running as root has to create an Exim log
1743 file, it does so in a subprocess that runs as exim:exim so as to get the
1744 ownership right at creation (otherwise, other Exim processes might see
1745 the file with the wrong ownership). There was no test for failure of this
1746 fork() call, which would lead to the process getting stuck as it waited
1747 for a non-existent subprocess. Forks do occasionally fail when resources
1748 run out. I reviewed all the other calls to fork(); they all seem to check
1751 PH/20 When checking for unexpected SMTP input at connect time (before writing
1752 the banner), Exim was not dealing correctly with a non-positive return
1753 from the read() function. If the client had disconnected by this time,
1754 the result was a log entry for a synchronization error with an empty
1755 string after "input=" when read() returned zero. If read() returned -1
1756 (an event I could not check), uninitialized data bytes were printed.
1757 There were reports of junk text (parts of files, etc) appearing after
1760 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
1762 PH/22 Added support for macro redefinition, and (re)definition in between
1763 driver and ACL definitions.
1765 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
1766 forgetting to use the resulting value; it was using the unexpanded value.
1768 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
1769 hadn't been configured. The fix is from Juergen Kreileder, who
1770 understands it better than I do:
1772 "Here's what I see happening with three configured cyrus_sasl
1773 authenticators configured (plain, login, cram-md5):
1775 On startup auth_cyrus_sasl_init() gets called for each of these.
1776 This means three calls to sasl_listmech() without a specified mech_list.
1777 => SASL tests which mechs of all available mechs actually work
1778 => three warnings about OTP not working
1779 => the returned list contains: plain, login, cram-md5, digest-md5, ...
1781 With the patch, sasl_listmech() also gets called three times. But now
1782 SASL's mech_list option is set to the server_mech specified in the the
1783 authenticator. Or in other words, the answer from sasl_listmech()
1784 gets limited to just the mech you're testing for (which is different
1786 => the return list contains just 'plain' or 'login', 'cram-md5' or
1787 nothing depending on the value of ob->server_mech.
1789 I've just tested the patch: Authentication still works fine,
1790 unavailable mechs specified in the exim configuration are still
1791 caught, and the auth.log warnings about OTP are gone."
1793 PH/25 When debugging is enabled, the contents of the command line are added
1794 to the debugging output, even when log_selector=+arguments is not
1797 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
1798 answer is "GNU", and only if the return is "GNU/something" is the answer
1801 PH/27 $acl_verify_message is now set immediately after the failure of a
1802 verification in an ACL, and so is available in subsequent modifiers. In
1803 particular, the message can be preserved by coding like this:
1805 warn !verify = sender
1806 set acl_m0 = $acl_verify_message
1808 Previously, $acl_verify_message was set only while expanding "message"
1809 and "log_message" when a very denied access.
1811 PH/28 Modified OS/os.c-Linux with
1813 -#ifndef OS_LOAD_AVERAGE
1814 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
1816 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
1817 nomenclature these days.)
1819 PH/29 Installed patch from the Sieve maintainer that adds the options
1820 sieve_useraddress and sieve_subaddress to the redirect router.
1822 PH/30 In these circumstances:
1823 . Two addresses routed to the same list of hosts;
1824 . First host does not offer TLS;
1825 . First host accepts first address;
1826 . First host gives temporary error to second address;
1827 . Second host offers TLS and a TLS session is established;
1828 . Second host accepts second address.
1829 Exim incorrectly logged both deliveries with the TLS parameters (cipher
1830 and peerdn, if requested) that were in fact used only for the second
1833 PH/31 When doing a callout as part of verifying an address, Exim was not paying
1834 attention to any local part prefix or suffix that was matched by the
1835 router that accepted the address. It now behaves in the same way as it
1836 does for delivery: the affixes are removed from the local part unless
1837 rcpt_include_affixes is set on the transport.
1839 PH/32 Add the sender address, as F=<...>, to the log line when logging a
1840 timeout during the DATA phase of an incoming message.
1842 PH/33 Sieve envelope tests were broken for match types other than :is. I have
1843 applied a patch sanctioned by the Sieve maintainer.
1845 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
1846 the uid or gid is negative. A case of a negative gid caused this to be
1847 noticed. The fix allows for either to be negative.
1849 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
1850 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
1853 PH/36 The change PH/12 above was broken. Fixed it.
1855 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
1856 the grounds that routing the same address twice would always produce the
1857 same answer. This might have been true once, but it is certainly no
1858 longer true now. Routing a child address may depend on the previous
1859 routing that produced that child. Some complicated redirection strategies
1860 went wrong when messages had multiple recipients, and made Exim's
1861 behaviour dependent on the order in which the addresses were given.
1863 I have moved the duplicate checking until after the routing is complete.
1864 Exim scans the addresses that are assigned to local and remote
1865 transports, and removes any duplicates. This means that more work will be
1866 done, as duplicates will always all be routed, but duplicates are
1867 presumably rare, so I don't expect this is of any significance.
1869 For deliveries to pipes, files, and autoreplies, the duplicate checking
1870 still happens during the routing process, since they are not going to be
1873 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
1874 It corrects a timeout issue with spamd. This is Ian's comment: "The
1875 background is that sometimes spamd either never reads data from a
1876 connection it has accepted, or it never writes response data. The exiscan
1877 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
1878 blindly assumes that writes won't block so it may never time out."
1880 PH/39 Allow G after quota size as well as K and M.
1882 PH/40 The value set for $authenticated_id in an authenticator may not contain
1883 binary zeroes or newlines because the value is written to log lines and
1884 to spool files. There was no check on this. Now the value is run through
1885 the string_printing() function so that such characters are converted to
1886 printable escape sequences.
1888 PH/41 $message_linecount is a new variable that contains the total number of
1889 lines in the message. Compare $body_linecount, which is the count for the
1892 PH/42 Exim no longer gives details of delivery errors for specific addresses in
1893 bounce and delay warning messages, except in certain special cases, which
1896 (a) An SMTP error message from a remote host;
1897 (b) A message specified in a :fail: redirection;
1898 (c) A message specified in a "fail" command in a system filter;
1899 (d) A message specified in a FAIL return from the queryprogram router;
1900 (e) A message specified by the cannot_route_message router option.
1902 In these cases only, Exim does include the error details in bounce and
1903 warning messages. There are also a few cases where bland messages such
1904 as "unrouteable address" or "local delivery error" are given.
1906 PH/43 $value is now also set for the "else" part of a ${run expansion.
1908 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
1909 being worked on, but at least Exim now implements the latest version to
1912 PH/45 In a pipe transport, although a timeout while waiting for the pipe
1913 process to complete was treated as a delivery failure, a timeout while
1914 writing the message to the pipe was logged, but erroneously treated as a
1915 successful delivery. Such timeouts include transport filter timeouts. For
1916 consistency with the overall process timeout, these timeouts are now
1917 treated as errors, giving rise to delivery failures by default. However,
1918 there is now a new Boolean option for the pipe transport called
1919 timeout_defer, which, if set TRUE, converts the failures into defers for
1920 both kinds of timeout. A transport filter timeout is now identified in
1923 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
1924 systems where "make" and "gmake" are different, calling "gmake" at top
1925 level broke things. I've arranged for the value of $(MAKE) to be passed
1926 from the Makefile to this script so that it can call the same version of
1930 A note about Exim versions 4.44 and 4.50
1931 ----------------------------------------
1933 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
1934 changes of various kinds. As a consequence, a big documentation update was
1935 needed. This delayed the release for rather longer than seemed good, especially
1936 in the light of a couple of (minor) security issues. Therefore, the changes
1937 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
1938 release. So 4.44 and 4.50 are in effect two different branches that both start
1941 I have left the 4.50 change log unchanged; it contains all the changes since
1942 4.43. The change log for 4.44 is below; many of its items are identical to
1943 those for 4.50. This seems to be the most sensible way to preserve the
1944 historical information.
1950 1. Minor wording change to the doc/README.SIEVE file.
1952 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
1953 computation of the current number of files was incorrect.
1955 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
1956 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
1959 4. Give more explanation in the error message when the command for a transport
1960 filter fails to execute.
1962 5. There are several places where Exim runs a non-Exim command in a
1963 subprocess. The SIGUSR1 signal should be disabled for these processes. This
1964 was being done only for the command run by the queryprogram router. It is
1965 now done for all such subprocesses. The other cases are: ${run, transport
1966 filters, and the commands run by the lmtp and pipe transports.
1968 6. Added CONFIGURE_GROUP build-time option.
1970 7. Some older OS have a limit of 256 on the maximum number of file
1971 descriptors. Exim was using setrlimit() to set 1000 as a large value
1972 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
1973 systems. I've change it so that if it can't get 1000, it tries for 256.
1975 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
1976 was an oversight, and furthermore, ever since the addition of extra
1977 controls (e.g. 4.43/32), the checks on when to allow different forms of
1978 "control" were broken. There should now be diagnostics for all cases when a
1979 control that does not make sense is encountered.
1981 9. Added the /retain_sender option to "control=submission".
1983 10. $recipients is now available in the predata ACL (oversight).
1985 11. Tidy the search cache before the fork to do a delivery from a message
1986 received from the command line. Otherwise the child will trigger a lookup
1987 failure and thereby defer the delivery if it tries to use (for example) a
1988 cached ldap connection that the parent has called unbind on.
1990 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
1991 of $address_data from the recipient verification was clobbered by the
1992 sender verification.
1994 13. The value of address_data from a sender verification is now available in
1995 $sender_address_data in subsequent conditions in the ACL statement.
1997 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
1999 15. Added a new option "connect=<time>" to callout options, to set a different
2002 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
2003 was its contents. (It was OK if the option was not defined at all.)
2005 17. A "Completed" log line is now written for messages that are removed from
2006 the spool by the -Mrm option.
2008 18. New variables $sender_verify_failure and $recipient_verify_failure contain
2009 information about exactly what failed.
2011 19. Added -dd to debug only the daemon process.
2013 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
2014 handles timeouts, both on the server side and network timeouts. Renamed the
2015 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
2017 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
2018 It is now set to "smtps".
2020 22. $host_address is now set to the target address during the checking of
2021 ignore_target_hosts.
2023 23. When checking ignore_target_hosts for an ipliteral router, no host name was
2024 being passed; this would have caused $sender_host_name to have been used if
2025 matching the list had actually called for a host name (not very likely,
2026 since this list is usually IP addresses). A host name is now passed as
2029 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
2030 code that specifies a non-restarting handler (typically sigaction() in
2031 modern systems) in an attempt to fix a rare and obscure crash bug.
2033 25. Narrowed the window for a race in the daemon that could cause it to ignore
2034 SIGCHLD signals. This is not a major problem, because they are used only to
2035 wake it up if nothing else does.
2037 26. A malformed maildirsize file could cause Exim to calculate negative values
2038 for the mailbox size or file count. Odd effects could occur as a result.
2039 The maildirsize information is now recalculated if the size or filecount
2042 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
2043 support for a long time. Removed HAVE_SYS_VFS_H.
2045 28. Installed the latest version of exipick from John Jetmore.
2047 29. In an address list, if the pattern was not a regular expression, an empty
2048 subject address (from a bounce message) matched only if the pattern was an
2049 empty string. Non-empty patterns were not even tested. This was the wrong
2050 because it is perfectly reasonable to use an empty address as part of a
2051 database query. An empty address is now tested by patterns that are
2052 lookups. However, all the other forms of pattern expect the subject to
2053 contain a local part and a domain, and therefore, for them, an empty
2054 address still always fails if the pattern is not itself empty.
2056 30. Exim went into a mad DNS loop when attempting to do a callout where the
2057 host was specified on an smtp transport, and looking it up yielded more
2058 than one IP address.
2060 31. Re-factored the code for checking spool and log partition space into a
2061 function that finds that data and another that does the check. The former
2062 is then used to implement four new variables: $spool_space, $log_space,
2063 $spool_inodes, and $log_inodes.
2065 32. The RFC2047 encoding function was originally intended for short strings
2066 such as real names; it was not keeping to the 75-character limit for
2067 encoded words that the RFC imposes. It now respects the limit, and
2068 generates multiple encoded words if necessary. To be on the safe side, I
2069 have increased the buffer size for the ${rfc2047: expansion operator from
2072 33. It is now permitted to omit both strings after an "if" condition; if the
2073 condition is true, the result is "true". As before, when the second string
2074 is omitted, a false condition yields an empty string. This makes it less
2075 cumbersome to write custom ACL and router conditions.
2077 34. Failure to deliver a bounce message always caused it to be frozen, even if
2078 there was an errors_to setting on the router. The errors_to setting is now
2081 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
2082 canonical form (fully expanded) before being placed in
2083 $sender_host_address.
2085 36. The table in the code that translates DNS record types into text (T_A to
2086 "A" for instance) was missing entries for NS and CNAME. It is just possible
2087 that this could have caused confusion if both these types were looked up
2088 for the same domain, because the text type is used as part of Exim's
2089 per-process caching. But the chance of anyone hitting this buglet seems
2092 37. The dnsdb lookup has been extended in a number of ways.
2094 (1) There is a new type, "zns", which walks up the domain tree until it
2095 finds some nameserver records. It should be used with care.
2097 (2) There is a new type, "mxh", which is like "mx" except that it returns
2098 just the host names, not the priorities.
2100 (3) It is now possible to give a list of domains (or IP addresses) to be
2101 looked up. The behaviour when one of the lookups defers can be
2102 controlled by a keyword.
2104 (4) It is now possible to specify the separator character for use when
2105 multiple records are returned.
2107 38. The dnslists ACL condition has been extended: it is now possible to supply
2108 a list of IP addresses and/or domains to be looked up in a particular DNS
2111 39. Added log_selector=+queue_time_overall.
2113 40. When running the queue in the test harness, wait just a tad after forking a
2114 delivery process, to get repeatability of debugging output.
2116 41. Include certificate and key file names in error message when GnuTLS fails
2117 to set them up, because the GnuTLS error message doesn't include the name
2118 of the failing file when there is a problem reading it.
2120 42. Allow both -bf and -bF in the same test run.
2122 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
2124 44. The "Exiscan patch" is now merged into the mainline Exim source.
2126 45. Sometimes the final signoff response after QUIT could fail to get
2127 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
2128 < 0 before doing a fflush(). This bug looks as though it goes back to the
2129 introduction of TLS in release 3.20, but "sometimes" must have been rare
2130 because the tests only now provoked it.
2132 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
2133 (this can affect the format of dates).
2135 47. exim_tidydb, when checking for the continued existence of a message for
2136 which it has found a message-specific retry record, was not finding
2137 messages that were in split spool directories. Consequently, it was
2138 deleting retry records that should have stayed in existence.
2140 48. Steve fixed some bugs in eximstats.
2142 49. The SPA authentication driver was not abandoning authentication and moving
2143 on to the next authenticator when an expansion was forced to fail,
2144 contradicting the general specification for all authenticators. Instead it
2145 was generating a temporary error. It now behaves as specified.
2147 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
2148 (the order specifies the preference for clients). The order is now AES256,
2149 AES128, 3DES, ARCFOUR128.
2151 51. Small patch to Sieve code - explicitly set From: when generating an
2154 52. Exim crashed if a remote delivery caused a very long error message to be
2155 recorded - for instance if somebody sent an entire SpamAssassin report back
2156 as a large number of 550 error lines. This bug was coincidentally fixed by
2157 increasing the size of one of Exim's internal buffers (big_buffer) that
2158 happened as part of the Exiscan merge. However, to be on the safe side, I
2159 have made the code more robust (and fixed the comments that describe what
2162 53. Now that there can be additional text after "Completed" in log lines (if
2163 the queue_time_overall log selector is set), a one-byte patch to exigrep
2164 was needed to allow it to recognize "Completed" as not the last thing in
2167 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
2168 patch that reportedly fixes this has been added. I am not expert enough to
2169 create a test for it. This is what the patch creator wrote:
2171 "I found a little strange behaviour of ldap code when working with
2172 Windows 2003 AD Domain, where users was placed in more than one
2173 Organization Units. When I tried to give exim partial DN, the exit code
2174 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
2175 But simultaneously result of request was absolutely normal ldap result,
2176 so I produce this patch..."
2178 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
2179 REFERENCE, so I have modified the code to exclude the patch when that macro
2182 55. Some experimental protocols are using DNS PTR records for new purposes. The
2183 keys for these records are domain names, not reversed IP addresses. The
2184 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
2185 leaves it alone. Component reversal etc. now happens only for IP addresses.
2188 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
2190 57. Double the size of the debug message buffer (to 2048) so that more of very
2191 long debug lines gets shown.
2193 58. The exicyclog utility now does better if the number of log files to keep
2194 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
2196 59. Two changes related to the smtp_active_hostname option:
2198 (1) $smtp_active_hostname is now available as a variable.
2199 (2) The default for smtp_banner uses $smtp_active_hostname instead
2200 of $primary_hostname.
2202 60. The host_aton() function is supposed to be passed a string that is known
2203 to be a valid IP address. However, in the case of IPv6 addresses, it was
2204 not checking this. This is a hostage to fortune. Exim now panics and dies
2205 if the condition is not met. A case was found where this could be provoked
2206 from a dnsdb PTR lookup with an IPv6 address that had more than 8
2207 components; fortuitously, this particular loophole had already been fixed
2208 by change 4.50/55 above.
2210 If there are any other similar loopholes, the new check in host_aton()
2211 itself should stop them being exploited. The report I received stated that
2212 data on the command line could provoke the exploit when Exim was running as
2213 exim, but did not say which command line option was involved. All I could
2214 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
2215 running as the user.
2218 61. There was a buffer overflow vulnerability in the SPA authentication code
2219 (which came originally from the Samba project). I have added a test to the
2220 spa_base64_to_bits() function which I hope fixes it.
2223 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
2224 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
2226 63. The daemon start-up calls getloadavg() while still root for those OS that
2227 need the first call to be done as root, but it missed one case: when
2228 deliver_queue_load_max is set with deliver_drop_privilege. This is
2229 necessary for the benefit of the queue runner, because there is no re-exec
2230 when deliver_drop_privilege is set.
2232 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
2233 This has been fixed.
2235 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
2236 was in use, was not putting the data itself into the right store pool;
2237 consequently, it could be overwritten for a subsequent message in the same
2238 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
2241 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
2243 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
2244 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
2245 the function were treating the return as a boolean value, which happened to
2246 work because 0=false and not-0=true, but is not correct code.
2248 68. The host_aton() function was not handling scoped IPv6 addresses (those
2249 with, for example, "%eth0" on the end) correctly.
2251 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
2252 negated items (that is, ~something) in unsigned ints. Some compilers
2253 apparently mutter when there is no cast.
2255 70. If an address verification called from an ACL failed, and did not produce a
2256 user-specific message (i.e. there was only a "system" message), nothing was
2257 put in $acl_verify_message. In this situation, it now puts the system
2260 71. Change 4.23/11 added synchronization checking at the start of an SMTP
2261 session; change 4.31/43 added the unwanted input to the log line - except
2262 that it did not do this in the start of session case. It now does.
2264 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
2265 This is wrong and can cause the other end to generate a synchronization
2266 error if it is another Exim or anything else that does the synchronization
2267 check. A QUIT command is no longer sent after a timeout.
2269 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
2270 during host lookups.
2272 74. The defer_ok option of callout verification was not working if it was used
2273 when verifying addresses in header lines, that is, for this case:
2275 verify = header_sender/callout=defer_ok
2277 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
2278 those file descriptors could be used for SMTP connections. If anything
2279 wrote to stderr (the example that came up was "warn" in embedded Perl), it
2280 could be sent to the SMTP client, causing chaos. The daemon now opens
2281 stdin, stdout, and stderr to /dev/null when it puts itself into the
2284 76. Arrange for output from Perl's "warn" command to be written to Exim's main
2285 log by default. The user can override this with suitable Perl magic.
2287 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
2288 the log message when discard triggers, was not working for the DATA ACL or
2289 for the non-SMTP ACL.
2291 78. Error message wording change in sieve.c.
2293 79. If smtp_accept_max_per_host was set, the number of connections could be
2294 restricted to fewer than expected, because the daemon was trying to set up
2295 a new connection before checking whether the processes handling previous
2296 connections had finished. The check for completed processes is now done
2297 earlier. On busy systems, this bug wouldn't be noticed because something
2298 else would have woken the daemon, and it would have reaped the completed
2301 80. If a message was submitted locally by a user whose login name contained one
2302 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
2303 It caused a spool format error. I have fixed the spool reading code. A
2304 related problem was that the "from" clause in the Received: line became
2305 illegal because of the space(s). It is now covered by ${quote_local_part.
2307 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
2310 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
2312 83. Patch from Sieve maintainer to fix unterminated string problem in
2313 "vacation" handling.
2315 84. Some minor changes to the Linux configuration files to help with other
2316 OS variants using glibc.
2318 85. One more patch for Sieve to update vacation handling to latest spec.
2321 ----------------------------------------------------
2322 See the note above about the 4.44 and 4.50 releases.
2323 ----------------------------------------------------
2329 1. Change 4.43/35 introduced a bug that caused file counts to be
2330 incorrectly computed when quota_filecount was set in an appendfile
2333 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
2334 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
2337 3. Give more explanation in the error message when the command for a transport
2338 filter fails to execute.
2340 4. There are several places where Exim runs a non-Exim command in a
2341 subprocess. The SIGUSR1 signal should be disabled for these processes. This
2342 was being done only for the command run by the queryprogram router. It is
2343 now done for all such subprocesses. The other cases are: ${run, transport
2344 filters, and the commands run by the lmtp and pipe transports.
2346 5. Some older OS have a limit of 256 on the maximum number of file
2347 descriptors. Exim was using setrlimit() to set 1000 as a large value
2348 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
2349 systems. I've change it so that if it can't get 1000, it tries for 256.
2351 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
2352 was an oversight, and furthermore, ever since the addition of extra
2353 controls (e.g. 4.43/32), the checks on when to allow different forms of
2354 "control" were broken. There should now be diagnostics for all cases when a
2355 control that does not make sense is encountered.
2357 7. $recipients is now available in the predata ACL (oversight).
2359 8. Tidy the search cache before the fork to do a delivery from a message
2360 received from the command line. Otherwise the child will trigger a lookup
2361 failure and thereby defer the delivery if it tries to use (for example) a
2362 cached ldap connection that the parent has called unbind on.
2364 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
2365 of $address_data from the recipient verification was clobbered by the
2366 sender verification.
2368 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
2369 was its contents. (It was OK if the option was not defined at all.)
2371 11. A "Completed" log line is now written for messages that are removed from
2372 the spool by the -Mrm option.
2374 12. $host_address is now set to the target address during the checking of
2375 ignore_target_hosts.
2377 13. When checking ignore_target_hosts for an ipliteral router, no host name was
2378 being passed; this would have caused $sender_host_name to have been used if
2379 matching the list had actually called for a host name (not very likely,
2380 since this list is usually IP addresses). A host name is now passed as
2383 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
2384 code that specifies a non-restarting handler (typically sigaction() in
2385 modern systems) in an attempt to fix a rare and obscure crash bug.
2387 15. Narrowed the window for a race in the daemon that could cause it to ignore
2388 SIGCHLD signals. This is not a major problem, because they are used only to
2389 wake it up if nothing else does.
2391 16. A malformed maildirsize file could cause Exim to calculate negative values
2392 for the mailbox size or file count. Odd effects could occur as a result.
2393 The maildirsize information is now recalculated if the size or filecount
2396 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
2397 support for a long time. Removed HAVE_SYS_VFS_H.
2399 18. Updated exipick to current release from John Jetmore.
2401 19. Allow an empty sender to be matched against a lookup in an address list.
2402 Previously the only cases considered were a regular expression, or an
2405 20. Exim went into a mad DNS lookup loop when doing a callout where the
2406 host was specified on the transport, if the DNS lookup yielded more than
2409 21. The RFC2047 encoding function was originally intended for short strings
2410 such as real names; it was not keeping to the 75-character limit for
2411 encoded words that the RFC imposes. It now respects the limit, and
2412 generates multiple encoded words if necessary. To be on the safe side, I
2413 have increased the buffer size for the ${rfc2047: expansion operator from
2416 22. Failure to deliver a bounce message always caused it to be frozen, even if
2417 there was an errors_to setting on the router. The errors_to setting is now
2420 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
2421 canonical form (fully expanded) before being placed in
2422 $sender_host_address.
2424 24. Updated eximstats to version 1.33
2426 25. Include certificate and key file names in error message when GnuTLS fails
2427 to set them up, because the GnuTLS error message doesn't include the name
2428 of the failing file when there is a problem reading it.
2430 26. Expand error message when OpenSSL has problems setting up cert/key files.
2433 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
2434 (this can affect the format of dates).
2436 28. exim_tidydb, when checking for the continued existence of a message for
2437 which it has found a message-specific retry record, was not finding
2438 messages that were in split spool directories. Consequently, it was
2439 deleting retry records that should have stayed in existence.
2441 29. eximstats updated to version 1.35
2442 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
2443 1.35 - bugfix such that pie charts by volume are generated correctly
2445 30. The SPA authentication driver was not abandoning authentication and moving
2446 on to the next authenticator when an expansion was forced to fail,
2447 contradicting the general specification for all authenticators. Instead it
2448 was generating a temporary error. It now behaves as specified.
2450 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
2451 (the order specifies the preference for clients). The order is now AES256,
2452 AES128, 3DES, ARCFOUR128.
2454 31. Small patch to Sieve code - explicitly set From: when generating an
2457 32. Exim crashed if a remote delivery caused a very long error message to be
2458 recorded - for instance if somebody sent an entire SpamAssassin report back
2459 as a large number of 550 error lines. This bug was coincidentally fixed by
2460 increasing the size of one of Exim's internal buffers (big_buffer) that
2461 happened as part of the Exiscan merge. However, to be on the safe side, I
2462 have made the code more robust (and fixed the comments that describe what
2465 33. Some experimental protocols are using DNS PTR records for new purposes. The
2466 keys for these records are domain names, not reversed IP addresses. The
2467 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
2468 leaves it alone. Component reversal etc. now happens only for IP addresses.
2471 34. The host_aton() function is supposed to be passed a string that is known
2472 to be a valid IP address. However, in the case of IPv6 addresses, it was
2473 not checking this. This is a hostage to fortune. Exim now panics and dies
2474 if the condition is not met. A case was found where this could be provoked
2475 from a dnsdb PTR lookup with an IPv6 address that had more than 8
2476 components; fortuitously, this particular loophole had already been fixed
2477 by change 4.50/55 or 4.44/33 above.
2479 If there are any other similar loopholes, the new check in host_aton()
2480 itself should stop them being exploited. The report I received stated that
2481 data on the command line could provoke the exploit when Exim was running as
2482 exim, but did not say which command line option was involved. All I could
2483 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
2484 running as the user.
2487 35. There was a buffer overflow vulnerability in the SPA authentication code
2488 (which came originally from the Samba project). I have added a test to the
2489 spa_base64_to_bits() function which I hope fixes it.
2492 36. The daemon start-up calls getloadavg() while still root for those OS that
2493 need the first call to be done as root, but it missed one case: when
2494 deliver_queue_load_max is set with deliver_drop_privilege. This is
2495 necessary for the benefit of the queue runner, because there is no re-exec
2496 when deliver_drop_privilege is set.
2498 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
2499 was in use, was not putting the data itself into the right store pool;
2500 consequently, it could be overwritten for a subsequent message in the same
2501 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
2504 38. Sometimes the final signoff response after QUIT could fail to get
2505 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
2506 < 0 before doing a fflush(). This bug looks as though it goes back to the
2507 introduction of TLS in release 3.20, but "sometimes" must have been rare
2508 because the tests only now provoked it.
2514 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
2515 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
2516 Now it returns an integer. A number of calls were still expecting a T/F
2517 return. Fortuitously, in all cases, the tests worked in OK situations,
2518 which is the norm. However, things would have gone wrong on any write
2519 failures on the smtp file descriptor. This function is used when sending
2520 messages over SMTP and also when doing verify callouts.
2522 2. When Exim is called to do synchronous delivery of a locally submitted
2523 message (the -odf or -odi options), it no longer closes stderr before doing
2526 3. Implemented the mua_wrapper option.
2528 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
2530 5. Implemented the functions header_remove(), header_testname(),
2531 header_add_at_position(), and receive_remove_recipient(), and exported them
2534 6. If an ACL "warn" statement specified the addition of headers, Exim already
2535 inserted X-ACL-Warn: at the start if there was no header name. However, it
2536 was not making this test for the second and subsequent header lines if
2537 there were newlines in the string. This meant that an invalid header could
2538 be inserted if Exim was badly configured.
2540 7. Allow an ACL "warn" statement to add header lines at the start or after all
2541 the Received: headers, as well as at the end.
2543 8. Added the rcpt_4xx retry error code.
2545 9. Added postmaster_mailfrom=xxx to callout verification option.
2547 10. Added mailfrom=xxxx to the callout verification option, for verify=
2550 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
2551 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
2553 12. Inserted some casts to stop certain compilers warning when using pointer
2554 differences as field lengths or precisions in printf-type calls (mostly
2555 affecting debugging statements).
2557 13. Added optional readline() support for -be (dynamically loaded).
2559 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
2560 same clock tick as a message's arrival, so that its received time was the
2561 same as the "first fail" time on the retry record, and that message
2562 remained on the queue past the ultimate address timeout, every queue runner
2563 would try a delivery (because it was past the ultimate address timeout) but
2564 after another failure, the ultimate address timeout, which should have then
2565 bounced the address, did not kick in. This was a "< instead of <=" error;
2566 in most cases the first failure would have been in the next clock tick
2567 after the received time, and all would be well.
2569 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
2570 being recognized when the domain list was tested by the match_domain
2571 condition in an expansion string.
2573 16. Added the ${str2b64: operator.
2575 17. Exim was always calling setrlimit() to set a large limit for the number of
2576 processes, without checking whether the existing limit was already
2577 adequate. (It did check for the limit on file descriptors.) Furthermore,
2578 errors from getrlimit() and setrlimit() were being ignored. Now they are
2579 logged to the main and panic logs, but Exim does carry on, to try to do its
2580 job under whatever limits there are.
2582 18. Imported PCRE 5.0.
2584 19. Trivial typo in log message " temporarily refused connection" (the leading
2587 20. If the log selector return_path_on_delivery was set and an address was
2588 redirected to /dev/null, the delivery process crashed because it assumed
2589 that a return path would always be set for a "successful" delivery. In this
2590 case, the whole delivery is bypassed as an optimization, and therefore no
2593 21. Internal re-arrangement: the function for sending a challenge and reading
2594 a response while authentication was assuming a zero-terminated challenge
2595 string. It's now changed to take a pointer and a length, to allow for
2596 binary data in such strings.
2598 22. Added the cyrus_sasl authenticator (code supplied by MBM).
2600 23. Exim was not respecting finduser_retries when seeking the login of the
2601 uid under which it was called; it was always trying 10 times. (The default
2602 setting of finduser_retries is zero.) Also, it was sleeping after the final
2603 failure, which is pointless.
2605 24. Implemented tls_on_connect_ports.
2607 25. Implemented acl_smtp_predata.
2609 26. If the domain in control=submission is set empty, Exim assumes that the
2610 authenticated id is a complete email address when it generates From: or
2611 Sender: header lines.
2613 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
2614 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
2615 chown and chgrp in /bin and hostname in /usr/bin.
2617 28. Exim was keeping the "process log" file open after each use, just as it
2618 does for the main log. This opens the possibility of it remaining open for
2619 long periods when the USR1 signal hits a daemon. Occasional processlog
2620 errors were reported, that could have been caused by this. Anyway, it seems
2621 much more sensible not to leave this file open at all, so that is what now
2624 29. The long-running daemon process does not normally write to the log once it
2625 has entered its main loop, and it closes the log before doing so. This is
2626 so that log files can straightforwardly be renamed and moved. However,
2627 there are a couple of unusual error situations where the daemon does write
2628 log entries, and I had neglected to close the log afterwards.
2630 30. The text of an SMTP error response that was received during a remote
2631 delivery was being truncated at 512 bytes. This is too short for some of
2632 the long messages that one sometimes sees. I've increased the limit to
2635 31. It is now possible to make retry rules that apply only when a message has a
2636 specific sender, in particular, an empty sender.
2638 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
2639 it possible to be selective about when SMTP synchronization is enforced.
2641 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
2643 32. Implemented hosts_connection_nolog.
2645 33. Added an ACL for QUIT.
2647 34. Setting "delay_warning=" to disable warnings was not working; it gave a
2650 35. Added mailbox_size and mailbox_filecount to appendfile.
2652 36. Added control = no_multiline_responses to ACLs.
2654 37. There was a bug in the logic of the code that waits for the clock to tick
2655 in the case where the clock went backwards by a substantial amount such
2656 that the microsecond fraction of "now" was more than the microsecond
2657 fraction of "then" (but the whole seconds number was less).
2659 38. Added support for the libradius Radius client library this is found on
2660 FreeBSD (previously only the radiusclient library was supported).
2666 1. When certain lookups returned multiple values in the form name=value, the
2667 quoting of the values was not always being done properly. Specifically:
2668 (a) If the value started with a double quote, but contained no whitespace,
2670 (b) If the value contained whitespace other than a space character (i.e.
2671 tabs or newlines or carriage returns) it was not quoted.
2672 This fix has been applied to the mysql and pgsql lookups by writing a
2673 separate quoting function and calling it from the lookup code. The fix
2674 should probably also be applied to nisplus, ibase and oracle lookups, but
2675 since I cannot test any of those, I have not disturbed their existing code.
2677 2. A hit in the callout cache for a specific address caused a log line with no
2678 reason for rejecting RCPT. Now it says "Previous (cached) callout
2679 verification failure".
2681 3. There was an off-by-one bug in the queryprogram router. An over-long
2682 return line was truncated at 256 instead of 255 characters, thereby
2683 overflowing its buffer with the terminating zero. As well as fixing this, I
2684 have increased the buffer size to 1024 (and made a note to document this).
2686 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
2687 when Exim is waiting for an SMTP response from a remote server, Exim
2688 restarts its select() call on the socket, thereby resetting its timeout.
2689 This is not a problem when such interrupts are rare. Somebody set up a cron
2690 job to run exiwhat every 2 minutes, which is less than the normal select()
2691 timeout (5 or 10 minutes). This meant that the select() timeout never
2692 kicked in because it was always reset. I have fixed this by comparing the
2693 time when an interrupt arrives with the time at the start of the first call
2694 to select(). If more time than the timeout has elapsed, the interrupt is
2695 treated as a timeout.
2697 5. Some internal re-factoring in preparation for the addition of Sieve
2698 extensions (by MH). In particular, the "personal" test is moved to a
2699 separate function, and given an option for scanning Cc: and Bcc: (which is
2700 not set for Exim filters).
2702 6. When Exim created an email address using the login of the caller as the
2703 local part (e.g. when creating a From: or Sender: header line), it was not
2704 quoting the local part when it contained special characters such as @.
2706 7. Installed new OpenBSD configuration files.
2708 8. Reworded some messages for syntax errors in "and" and "or" conditions to
2709 try to make them clearer.
2711 9. Callout options, other than the timeout value, were being ignored when
2712 verifying sender addresses in header lines. For example, when using
2714 verify = header_sender/callout=no_cache
2716 the cache was (incorrectly) being used.
2718 10. Added a missing instance of ${EXE} to the exim_install script; this affects
2719 only the Cygwin environment.
2721 11. When return_path_on_delivery was set as a log selector, if different remote
2722 addresses in the same message used different return paths and parallel
2723 remote delivery occurred, the wrong values would sometimes be logged.
2724 (Whenever a remote delivery process finished, the return path value from
2725 the most recently started remote delivery process was logged.)
2727 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
2728 lines when AUTH and/or TLS are in use. This is the "received protocol"
2729 field. Exim used to use "asmtp" for authenticated SMTP, without any
2730 indication (in the protocol name) for TLS use. Now it follows the RFC and
2731 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
2732 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
2733 names appear in log lines as well as in Received: header lines.
2735 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
2736 extensions, and comparison tests, and to fix some bugs.
2738 14. Changes to the "personal" filter test:
2740 (1) The test was buggy in that it was just doing the equivalent of
2741 "contains" tests on header lines. For example, if a user's address was
2742 anne@some.where, the "personal" test would incorrectly be true for
2744 To: susanne@some.where
2746 This test is now done by extracting each address from the header in turn,
2747 and checking the entire address. Other tests that are part of "personal"
2748 are now done using regular expressions (for example, to check local parts
2749 of addresses in From: header lines).
2751 (2) The list of non-personal local parts in From: addresses has been
2752 extended to include "listserv", "majordomo", "*-request", and "owner-*",
2753 taken from the Sieve specification recommendations.
2755 (3) If the message contains any header line starting with "List-" it is
2756 treated as non-personal.
2758 (4) The test for "circular" in the Subject: header line has been removed
2759 because it now seems ill-conceived.
2761 15. Minor typos in src/EDITME comments corrected.
2763 16. Installed latest exipick from John Jetmore.
2765 17. If headers_add on a router specified a text string that was too long for
2766 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
2767 of string_sprintf() is now avoided.
2769 18. $message_body_size was not set (it was always zero) when running the DATA
2770 ACL and the local_scan() function.
2772 19. For the "mail" command in an Exim filter, no default was being set for
2773 the once_repeat time, causing a random time value to be used if "once" was
2774 specified. (If the value happened to be <= 0, no repeat happened.) The
2775 default is now 0s, meaning "never repeat". The "vacation" command was OK
2776 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
2777 (I found it when inspecting the code).
2779 20. There is now an overall timeout for performing a callout verification. It
2780 defaults to 4 times the callout timeout, which applies to individual SMTP
2781 commands during the callout. The overall timeout applies when there is more
2782 than one host that can be tried. The timeout is checked before trying the
2783 next host. This prevents very long delays if there are a large number of
2784 hosts and all are timing out (e.g. when the network connections are timing
2785 out). The value of the overall timeout can be changed by specifying an
2786 additional sub-option for "callout", called "maxwait". For example:
2788 verify = sender/callout=5s,maxwait=20s
2790 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
2791 to the end before writing, but this should make it even safer).
2793 22. Exim was forgetting that it had advertised PIPELINING for the second and
2794 subsequent messages on an SMTP connection. It was also not resetting its
2795 memory on STARTTLS and an internal HELO.
2797 23. When Exim logs an SMTP synchronization error within a session, it now
2798 records whether PIPELINING has been advertised or not.
2800 24. Added 3 instances of "(long int)" casts to time_t variables that were being
2801 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
2802 rather than long int.
2804 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
2806 26. Added the never_mail option to autoreply.
2812 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
2813 crash if the getsockname() call failed; this can happen if a connection is
2814 closed very soon after it is established. The problem was simply in the
2815 order in which certain operations were done, causing Exim to try to write
2816 to the SMTP stream before it had set up the file descriptor. The bug has
2817 been fixed by making things happen in the correct order.
2823 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
2824 before the connection was closed, thus losing the rejection response.
2826 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
2827 some early Solaris releases, but causes trouble in current releases where
2828 socklen_t is defined.
2830 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
2833 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
2836 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
2837 that releasing the top of it at the end releases what was used for sub-
2838 expansions (unless the block got too big). However, discard this block if
2839 the first thing is a variable or header, so that we can use its block when
2840 it is dynamic (useful for very large $message_headers, for example).
2842 6. Lookups now cache *every* query, not just the most recent. A new, separate
2843 store pool is used for this. It can be recovered when all lookup caches are
2844 flushed. Lookups now release memory at the end of their result strings.
2845 This has involved some general refactoring of the lookup sources.
2847 7. Some code has been added to the store_xxx() functions to reduce the amount
2848 of flapping under certain conditions.
2850 8. log_incoming_interface used to affect only the <= reception log lines. Now
2851 it causes the local interface and port to be added to several more SMTP log
2852 lines, for example "SMTP connection from", and rejection lines.
2854 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
2856 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
2858 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
2859 could be overwritten at the end of the current message (or the start of a
2860 new message if it was set in a HELO ACL). The value is now preserved for
2861 the duration of the SMTP connection.
2863 12. If a transport had a headers_rewrite setting, and a matching header line
2864 contained an unqualified address, that address was qualified, even if it
2865 did not match any rewriting rules. The underlying bug was that the values
2866 of the flags that permit the existence of unqualified sender and recipient
2867 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
2868 non-local messages, and by -bnq for local messages) were not being
2869 preserved with the message after it was received.
2871 13. When Exim was logging an SMTP synchronization error, it could sometimes log
2872 "next input=" as part of the text comprising the host identity instead of
2873 the correct text. The code was using the same buffer for two different
2874 strings. However, depending on which order the printing function evaluated
2875 its arguments, the bug did not always show up. Under Linux, for example, my
2876 test suite worked just fine.
2878 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
2879 doesn't work with some older versions of Perl. It has been changed to "my",
2880 which in any case is probably the better facility to use.
2882 15. A really picky compiler found some instances of statements for creating
2883 error messages that either had too many or two few arguments for the format
2886 16. The size of the buffer for calls to the DNS resolver has been increased
2887 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
2888 for addresses that have a lot of PTR records. This alleviates a problem; it
2889 does not fully solve it.
2891 17. A dnsdb lookup for PTR records that receives more data than will fit in the
2892 buffer now truncates the list and logs the incident, which is the same
2893 action as happens when Exim is looking up a host name and its aliases.
2894 Previously in this situation something unpredictable would happen;
2895 sometimes it was "internal error: store_reset failed".
2897 18. If a server dropped the connection unexpectedly when an Exim client was
2898 using GnuTLS and trying to read a response, the client delivery process
2899 crashed while trying to generate an error log message.
2901 19. If a "warn" verb in an ACL added multiple headers to a message in a single
2902 string, for example:
2904 warn message = H1: something\nH2: something
2906 the text was added as a single header line from Exim's point of view
2907 though it ended up OK in the delivered message. However, searching for the
2908 second and subsequent header lines using $h_h2: did not work. This has been
2909 fixed. Similarly, if a system filter added multiple headers in this way,
2910 the routers could not see them.
2912 20. Expanded the error message when iplsearch is called with an invalid key to
2913 suggest using net-iplsearch in a host list.
2915 21. When running tests using -bh, any delays imposed by "delay" modifiers in
2916 ACLs are no longer actually imposed (and a message to that effect is
2919 22. If a "gecos" field in a passwd entry contained escaped characters, in
2920 particular, if it contained a \" sequence, Exim got it wrong when building
2921 a From: or a Sender: header from that name. A second bug also caused
2922 incorrect handling when an unquoted " was present following a character
2923 that needed quoting.
2925 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
2926 was not being matched caselessly.
2928 24. Arranged for all hyphens in the exim.8 source to be escaped with
2931 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
2932 properly. Recipient callout cache records were still being keyed to include
2933 the sender, even when use_sender was set false. This led to far more
2934 callouts that were necessary. The sender is no longer included in the key
2935 when use_sender is false.
2937 26. Added "control = submission" modifier to ACLs.
2939 27. Added the ${base62d: operator to decode base 62 numbers.
2941 28. dnsdb lookups can now access SRV records.
2943 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
2944 the configuration file.
2946 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
2947 (-v) mode. This makes the output for a verbose queue run more intelligible.
2949 31. Added a use_postmaster feature to recipient callouts.
2951 32. Added the $body_zerocount variable, containing the number of binary zero
2952 bytes in the message body.
2954 33. The time of last modification of the "new" subdirectory is now used as the
2955 "mailbox time last read" when there is a quota error for a maildir
2958 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
2960 35. Added +ignore_unknown as a special item in host lists.
2962 36. Code for decoding IPv6 addresses in host lists is now included, even if
2963 IPv6 support is not being compiled. This fixes a bug in which an IPv6
2964 address was recognized as an IP address, but was then not correctly decoded
2965 into binary, causing unexpected and incorrect effects when compared with
2972 1. Very minor rewording of debugging text in manualroute to say "list of
2973 hosts" instead of "hostlist".
2975 2. If verify=header_syntax was set, and a header line with an unqualified
2976 address (no domain) and a large number of spaces between the end of the
2977 name and the colon was received, the reception process suffered a buffer
2978 overflow, and (when I tested it) crashed. This was caused by some obsolete
2979 code that should have been removed. The fix is to remove it!
2981 3. When running in the test harness, delay a bit after writing a bounce
2982 message to get a bit more predictability in the log output.
2984 4. Added a call to search_tidyup() just before forking a reception process. In
2985 theory, someone could use a lookup in the expansion of smtp_accept_max_
2986 per_host which, without the tidyup, could leave open a database connection.
2988 5. Added the variables $recipient_data and $sender_data which get set from a
2989 lookup success in an ACL "recipients" or "senders" condition, or a router
2990 "senders" option, similar to $domain_data and $local_part_data.
2992 6. Moved the writing of debug_print from before to after the "senders" test
2995 7. Change 4.31/66 (moving the time when the Received: is generated) caused
2996 problems for message scanning, either using a data ACL, or using
2997 local_scan() because the Received: header was not generated till after they
2998 were called (in order to set the time as the time of reception completion).
2999 I have revised the way this works. The header is now generated after the
3000 body is received, but before the ACL or local_scan() are called. After they
3001 are run, the timestamp in the header is updated.
3007 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
3008 before starting a queue runner without re-exec. This happened only when
3009 deliver_drop_privilege was set or when the Exim user was set to root. The
3010 effect of the bug was that timeouts during subsequent deliveries caused
3011 crashes instead of being properly handled. The handler is now left at its
3012 default (and expected) setting.
3014 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
3015 message, again when deliver_drop_privilege is set or Exim is run as root.
3016 The bug described in (1) was not present in this case, but the tidying up
3017 of the other signals was missing. I have made the two cases consistent.
3019 3. The ignore_target_hosts setting on a manualroute router was being ignored
3020 for hosts that were looked up using the /MX notation.
3022 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
3025 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
3026 operated on the sender address. After changing the $sender_address to <>
3027 for the sender address verify, Exim was re-instated it as the original
3028 (before rewriting) address, but remembering that it had rewritten it, so it
3029 wasn't rewriting it again. This bug also had the effect of breaking the
3030 sender address verification caching when the sender address was rewritten.
3032 6. The ignore_target_hosts option was being ignored by the ipliteral router.
3033 This has been changed so that if the ip literal address matches
3034 ignore_target_hosts, the router declines.
3036 7. Added expansion conditions match_domain, match_address, and match_local_
3037 part (NOT match_host).
3039 8. The placeholder for the Received: header didn't have a length field set.
3041 9. Added code to Exim itself and to exim_lock to test for a specific race
3042 condition that could lead to file corruption when using MBX delivery. The
3043 issue is with the lockfile that is created in /tmp. If this file is removed
3044 after a process has opened it but before that process has acquired a lock,
3045 there is the potential for a second process to recreate the file and also
3046 acquire a lock. This could lead to two Exim processes writing to the file
3047 at the same time. The added code performs the same test as UW imapd; it
3048 checks after acquiring the lock that its file descriptor still refers to
3049 the same named file.
3051 10. The buffer for building added header lines was of fixed size, 8192 bytes.
3052 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
3055 11. Added the smtp_active_hostname option. If used, this will typically be made
3056 to depend on the incoming interface address. Because $interface_address is
3057 not set up until the daemon has forked a reception process, error responses
3058 that can happen earlier (such as "too many connections") no longer contain
3061 12. If an expansion in a condition on a "warn" statement fails because a lookup
3062 defers, the "warn" statement is abandoned, and the next ACL statement is
3063 processed. Previously this caused the whole ACL to be aborted.
3065 13. Added the iplsearch lookup type.
3067 14. Added ident_timeout as a log selector.
3069 15. Added tls_certificate_verified as a log selector.
3071 16. Added a global option tls_require_ciphers (compare the smtp transport
3072 option of the same name). This controls incoming TLS connections.
3074 17. I finally figured out how to make tls_require_ciphers do a similar thing
3075 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
3076 before starting the TLS session.
3078 18. Tabs are now shown as \t in -bP output.
3080 19. If the log selector return_path_on_delivery was set, Exim crashed when
3081 bouncing a message because it had too many Received: header lines.
3083 20. If two routers both had headers_remove settings, and the first one included
3084 a superfluous trailing colon, the final name in the first list and the
3085 first name in the second list were incorrectly joined into one item (with a
3086 colon in the middle).
3092 1. Added -C and -D options to the exinext utility, mainly to make it easier
3093 to include in the automated testing, but these could be helpful when
3094 multiple configurations are in use.
3096 2. The exinext utility was not formatting the output nicely when there was
3097 an alternate port involved in the retry record key, nor when there was a
3098 message id as well (for retries that were specific to a specific message
3099 and a specific host). It was also confused by IPv6 addresses, because of
3100 the additional colons they contain. I have fixed the IPv4 problem, and
3101 patched it up to do a reasonable job for IPv6.
3103 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
3104 delivery, the log line now contains "pipelined" if PIPELINING was used.
3106 4. An SMTP transport process used to panic and die if the bind() call to set
3107 an explicit outgoing interface failed. This has been changed; it is now
3108 treated in the same way as a connect() failure.
3110 5. A reference to $sender_host_name in the part of a conditional expansion
3111 that was being skipped was still causing a DNS lookup. This no longer
3114 6. The def: expansion condition was not recognizing references to header lines
3115 that used bh_ and bheader_.
3117 7. Added the _cache feature to named lists.
3119 8. The code for checking quota_filecount in the appendfile transport was
3120 allowing one more file than it should have been.
3122 9. For compatibility with Sendmail, the command line option
3130 and sets the incoming protocol and host name (for trusted callers). The
3131 host name and its colon can be omitted when only the protocol is to be set.
3132 Note the Exim already has two private options, -pd and -ps, that refer to
3133 embedded Perl. It is therefore impossible to set a protocol value of "d" or
3134 "s", but I don't think that's a major issue.
3136 10. A number of refactoring changes to the code, none of which should affect
3139 (a) The number of logging options was getting close to filling up the
3140 32-bit word that was used as a bit map. I have split them into two classes:
3141 those that are passed in the argument to log_write(), and those that are
3142 only ever tested independently outside of that function. These are now in
3143 separate 32-bit words, so there is plenty of room for expansion again.
3144 There is no change in the user interface or the logging behaviour.
3146 (b) When building, for example, log lines, the code previously used a
3147 macro that called string_cat() twice, in order to add two strings. This is
3148 not really sufficiently general. Furthermore, there was one instance where
3149 it was actually wrong because one of the argument was used twice, and in
3150 one call a function was used. (As it happened, calling the function twice
3151 did not affect the overall behaviour.) The macro has been replaced by a
3152 function that can join an arbitrary number of extra strings onto a growing
3155 (c) The code for expansion conditions now uses a table and a binary chop
3156 instead of a serial search (which was left over from when there were very
3157 few conditions). Also, it now recognizes conditions like "pam" even when
3158 the relevant support is not compiled in: a suitably worded error message is
3159 given if an attempt is made to use such a condition.
3161 11. Added ${time_interval:xxxxx}.
3163 12. A bug was causing one of the ddress fields not to be passed back correctly
3164 from remote delivery subprocesses. The field in question was not being
3165 subsequently used, so this caused to problems in practice.
3167 13. Added new log selectors queue_time and deliver_time.
3169 14. Might have fixed a bug in maildirsizefile handling that threw up
3170 "unexpected character" debug warnings, and recalculated the data
3171 unnecessarily. In any case, I expanded the warning message to give more
3174 15. Added the message "Restricted characters in address" to the statements in
3175 the default ACL that block characters like @ and % in local parts.
3177 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
3178 Three changes have been made:
3180 (a) There was a serious bug; a negative response to MAIL caused the whole
3181 recipient domain to be cached as invalid, thereby blocking all messages
3182 to all local parts at the same domain, from all senders. This bug has
3183 been fixed. The domain is no longer cached after a negative response to
3184 MAIL if the sender used is not empty.
3186 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
3189 (c) A new callout option, "use_sender" has been added for people who want
3190 the modified behaviour.
3196 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
3199 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
3200 indeed breaks things for older releases.
3202 3. Added additional logging to the case where there is a problem reading data
3203 from a filter that is running in a subprocess using a pipe, in order to
3204 try to track down a specific problem.
3206 4. Testing facility fudge: when running in the test harness and attempting
3207 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
3208 getting "No route to host". Convert this to a timeout.
3210 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
3213 6. Some OS don't have socklen_t but use size_t instead. This affects the
3214 fifth argument of getsockopt() amongst other things. This is now
3215 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
3216 can be set for individual OS. I have set it for SunOS5, OSF1, and
3217 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
3218 some earlier ones do not.
3220 7. Change 4.30/15 was not doing the test caselessly.
3222 8. The standard form for an IPv6 address literal was being rejected by address
3223 parsing in, for example, MAIL and RCPT commands. An example of this kind of
3224 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
3225 this, as well as the form without the "IPv6" on the front (but only when
3226 address literals are enabled, of course).
3228 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
3230 10. Exim crashed if a message with an empty sender address specified by -f
3231 encountered a router with an errors_to setting. This could be provoked only
3232 by a command such as
3236 where an empty string was supplied; "<>" did not hit this bug.
3238 11. Installed PCRE release 4.5.
3240 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
3241 remained set. It is now erased.
3243 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
3244 times from message ids (which are base 36 rather than the normal 62).
3246 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
3247 were being counted as actual protocol errors, and logged if the log
3248 selector +smtp_protocol_error was set. One cannot be perfect in this test,
3249 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
3250 and DATA following a set of rejected RCPTs do not count as protocol errors.
3251 In other words, Exim assumes they were pipelined, though this may not
3252 actually be the case. Of course, in all cases the client gets an
3253 appropriate error code.
3255 15. If a lookup fails in an ACL condition, a message about the failure may
3256 be available; it is used if testing the ACL cannot continue, because most
3257 such messages specify what the cause of the deferral is. However, some
3258 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
3259 that caused an old message to be retained and used if a later statement
3260 caused a defer, replacing the real cause of the deferral.
3262 16. If an IP address had so many PTR records that the DNS lookup buffer
3263 was not large enough to hold them, Exim could crash while trying to process
3264 the truncated data. It now detects and logs this case.
3266 17. Further to 4.21/58, another change has been made: if (and only if) the
3267 first line of a message (the first header line) ends with CRLF, a bare LF
3268 in a subsequent header line has a space inserted after it, so as not to
3269 terminate the header.
3271 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
3272 unnecessarily, used atoi() instead of strtol(), and didn't check the
3273 termination when getting file sizes from file names by regex.
3275 19. Completely re-implemented the support for maildirsize files, in the light
3276 of a number of problems with the previous contributed implementation
3277 (4.30/29). In particular:
3279 . If the quota is zero, the maildirsize file is maintained, but no quota is
3282 . If the maildir directory does not exist, it is created before any attempt
3283 to write a maildirsize file.
3285 . The quota value in the file is just a cache; if the quota is changed in
3286 the transport, the new value overrides.
3288 . A regular expression is available for excluding directories from the
3291 20. The autoreply transport checks the characters in options that define the
3292 message's headers; it allows continued headers, but it was checking with
3293 isspace() after an embedded newline instead of explicitly looking for a
3296 21. If all the "regular" hosts to which an address was routed had passed their
3297 expiry times, and had not reached their retry times, the address was
3298 bounced, even if fallback hosts were defined. Now Exim should go on to try
3301 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
3302 equivalent code in the SMTP transport. Some hosts send humungous responses
3303 to HELO/EHLO, more than 1024 it seems.
3305 23. Refactoring: code in filter.c used (void *) for "any old type" but this
3306 gives compiler warnings in some environments. I've now done it "properly",
3309 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
3310 (because of problems with the built-in one) was declared to return uschar *
3311 instead of char *, causing compiler failure.
3313 25. Fixed a file descriptor leak when processing alias/forward files.
3315 26. Fixed a minor format string issue in dbfn.c.
3317 27. Typo in exim.c: ("dmbnz" for "dbmnz").
3319 28. If a filter file refered to $h_xxx or $message_headers, and the headers
3320 contained RFC 2047 "words", Exim's memory could, under certain conditions,
3323 29. When a sender address is verified, it is cached, to save repeating the test
3324 when there is more than one recipient in a message. However, when the
3325 verification involves a callout, it is possible for different callout
3326 options to be set for different recipients. It is too complicated to keep
3327 track of this in the cache, so now Exim always runs a verification when a
3328 callout is required, relying on the callout cache for the optimization.
3329 The overhead is duplication of the address routing, but this should not be
3332 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
3333 to be verified with callout=postmaster, and the main callout worked but the
3334 postmaster check failed, the verification correctly failed. However, if a
3335 subsequent RCPT command asked for sender verification *without* the
3336 postmaster check, incorrect caching caused this verification also to fail,
3339 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
3340 it was not caching the DNS options (qualify_single, search_parents) that
3341 were used when the lookup failed. A subsequent lookup with different
3342 options therefore always gave the same answer, though there were cases
3343 where it should not have. (Example: a "domains = !$mx_any" option on a
3344 dnslookup router: the "domains" option is always processed without any
3345 widening, but the router might have qualify_single set.) Now Exim uses the
3346 cached value only when the same options are set.
3348 32. Added John Jetmore's "exipick" utility to the distribution.
3350 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
3351 than a timeout (e.g. a certificate is required, and is not provided), an
3352 Exim server now closes the connection immediately. Previously it waited for
3353 the client to close - but if the client is SSL, it seems that they each
3354 wait for each other, leading to a delay before one of them times out.
3356 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
3357 maintained 0.8.x compatibility because I don't think many are using it, and
3358 it is clearly obsolete.
3360 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
3363 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
3364 client certificate was expired. A simple patch fixes this, though I don't
3365 understand the full logic of why the verify callback is called multiple
3368 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
3369 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
3370 which causes problems with some clients (such as the Certicom SSL Plus
3371 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
3372 disables the coutermeasure allowing Eudora to connect."
3374 38. Exim was not checking that a write() to a log file succeeded. This could
3375 lead to Bad Things if a log got too big, in particular if it hit a file
3376 size limit. Exim now panics and dies if it cannot write to a log file, just
3377 as it does if it cannot open a log file.
3379 39. Modified OS/Makefile-Linux so that it now contains
3381 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
3383 The two -D definitions ensure that Exim is compiled with large file
3384 support, which makes it possible to handle log files that are bigger than
3387 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
3388 instance) a domain was checked against a named list that involved a lookup,
3389 causing $domain_data to be set, then another domain was checked against the
3390 same list, then the first domain was re-checked, the value of $domain_data
3391 after the final check could be wrong. In particular, if the second check
3392 failed, it could be set empty. This bug probably also applied to
3395 41. The strip_trailing_dot option was not being applied to the address given
3396 with the -f command-line option.
3398 42. The code for reading a message's header from the spool was incrementing
3399 $received_count, but never initializing it. This meant that the value was
3400 incorrect (doubled) while delivering a message in the same process in which
3401 it was received. In the most common configuration of Exim, this never
3402 happens - a fresh exec is done - but it can happen when
3403 deliver_drop_privilege is set.
3405 43. When Exim logs an SMTP synchronization error - client data sent too soon -
3406 it now includes up to 150 characters of the unexpected data in the log
3409 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
3410 and building data strings. The size of both of these buffers was 10 000
3411 bytes - far larger than anybody would *ever* want, thought I. Needless to
3412 say, somebody hit the limit. I have increased the maximum line length to
3413 20 000 and the maximum data length of concatenated lines to 100 000. I have
3414 also fixed two bugs, because there was no checking on these buffers. Tsk,
3415 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
3416 buffer is too small.
3418 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
3419 lsearch lookups. Now it does.
3421 46. When parsing a route_list item in a manualroute router, a fixed-length
3422 buffer was used for the list of hosts. I made this 1024 bytes long,
3423 thinking that nobody would ever have a list of hosts that long. Wrong.
3424 Somebody had a whole pile of complicated expansion conditions, and the
3425 string was silently truncated, leading to an expansion error. It turns out
3426 that it is easier to change to an unlimited length (owing to other changes
3427 that have happened since this code was originally written) than to build
3428 structure for giving a limitation error. The length of the item that
3429 expands into the list of hosts is now unlimited.
3431 47. The lsearch lookup could not handle data where the length of text line was
3432 more than 4095 characters. Such lines were truncated, leading to shortened
3433 data being returned. It should now handle lines of any length.
3435 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
3436 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
3439 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
3440 replacing strings such as BIN_DIRECTORY in a source file: the replacement
3441 no longer happens in comment lines. A list of replacements is now placed
3442 at the head of all of the source files, except those whose only change is
3443 to replace PERL_COMMAND in the very first #! line.
3445 50. Replaced the slow insertion sort in queue.c, for sorting the list of
3446 messages on the queue, with a bottom-up merge sort, using code contributed
3447 by Michael Haardt. This should make operations like -bp somewhat faster on
3448 large queues. It won't affect queue runners, except when queue_run_in_order
3451 51. Installed eximstats 1.31 in the distribution.
3453 52. Added support for SRV lookups to the dnslookup router.
3455 53. If an ACL referred to $message_body or $message_body_end, the value was not
3456 reset for any messages that followed in the same SMTP session.
3458 54. The store-handling optimization for building very long strings was not
3459 differentiating between the different store pools. I don't think this
3460 actually made any difference in practice, but I've tidied it.
3462 55. While running the routers to verify a sender address, $sender_address
3463 was still set to the sender address. This is wrong, because when routing to
3464 send a bounce to the sender, it would be empty. Therefore, I have changed
3465 it so that, while verifying a sender address, $sender_address is set to <>.
3466 (There is no change to what happens when verifying a recipient address.)
3468 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
3469 target A or AAAA records (if not already returned) without resetting the
3470 qualify_single or search_parents options of the DNS resolver. These are
3471 inappropriate in this case because the targets of MX and SRV records must
3472 be FQDNs. A broken DNS record could cause trouble if it happened to have a
3473 target that, when qualified, matched something in the local domain. These
3474 two options are now turned off when doing these lookups.
3476 57. It seems that at least some releases of Reiserfs (which does not have the
3477 concept of a fixed number of inodes) returns zero and not -1 for the
3478 number of available inodes. This interacted badly with check_spool_inodes,
3479 which assumed that -1 was the "no such thing" setting. What I have done is
3480 to check that the total number of inodes is greater than zero before doing
3481 the test of how many are available.
3483 58. When a "warn" ACL statement has a log_message modifier, the message is
3484 remembered, and not repeated. This is to avoid a lot of repetition when a
3485 message has many recipients that cause the same warning to be written.
3486 Howewer, Exim was preserving the list of already written lines for an
3487 entire SMTP session, which doesn't seem right. The memory is now reset if a
3488 new message is started.
3490 59. The "rewrite" debugging flag was not showing the result of rewriting in the
3491 debugging output unless log_rewrite was also set.
3493 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
3494 of (int)(handle) when we know that handle contains (void *)(-1).
3496 61. The Exim daemon panic-logs an error return when it closes the incoming
3497 connection. However "connection reset by peer" seems to be common, and
3498 isn't really an error worthy of noting specially, so that particular error
3501 62. When Exim is trying to find all the local interfaces, it used to panic and
3502 die if the ioctl to get the interface flags failed. However, it seems that
3503 on at least one OS (Solaris 9) it is possible to have an interface that is
3504 included in the list of interfaces, but for which you get a failure error
3505 for this call. This happens when the interface is not "plumbed" into a
3506 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
3507 failure of the "get flags" call assumes that the interface is down.
3509 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
3510 makes life easier for people who are doing arithmetic on fields extracted
3511 from dates, where you often get leading zeros that should not be
3512 interpreted as octal.
3514 64. Added qualify_domain to the redirect router, to override the global
3517 65. If a pathologically long header line contained very many addresses (the
3518 report of this problem mentioned 10 000) and each of them was rewritten,
3519 Exim could use up a very large amount of memory. (It kept on making new
3520 copies of the header line as it rewrote, and never released the old ones.)
3521 At the expense of a bit more processing, the header rewriting function has
3522 been changed so that it no longer eats memory in this way.
3524 66. The generation of the Received: header has been moved from the time that a
3525 message starts to be received, to the time that it finishes. The timestamp
3526 in the Received: header should now be very close to that of the <= log
3527 line. There are two side-effects of this change:
3529 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
3530 logged header lines no longer include the local Received: line, because
3531 it has not yet been created. The same applies to a copy of the message
3532 that is returned to a non-SMTP sender when a message is rejected.
3534 (b) When a filter file is tested using -bf, no additional Received: header
3535 is added to the test message. After some thought, I decided that this
3538 This change does not affect the value of $received_for. It is still set
3539 after address rewriting, but before local_scan() is called.
3541 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
3543 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
3544 gave an unhelpful panic error message, and a defer error. I have managed to
3545 change this behaviour so that it now rejects any supplied certificate,
3546 which seems right, as the list of acceptable certificates is empty.
3548 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
3549 gave an unhelpful defer error. I have not managed to make this reject any
3550 supplied certificates, but the error message it gives is "no certificate
3551 supplied", which is not helpful.
3553 70. exigrep's output now also includes lines that are not associated with any
3554 message, but which match the given pattern. Implemented by a patch from
3555 Martin Sluka, which also tidied up the Perl a bit.
3557 71. Recipient callout verification, like sender verification, was using <> in
3558 the MAIL FROM command. This isn't really the right thing, since the actual
3559 sender may affect whether the remote host accepts the recipient or not. I
3560 have changed it to use the actual sender in the callout; this means that
3561 the cache record is now keyed on a recipient/sender pair, not just the
3562 recipient address. There doesn't seem to be a real danger of callout loops,
3563 since a callout by the remote host to check the sender would use <>.
3564 [SEE ABOVE: changed after hitting problems.]
3566 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
3567 temporary errors. However, in the case of such a code being given after
3568 the end of a data transmission (i.e. after ".") Exim was failing to write
3569 a retry record for the message. (Yes, there was some broken host that was
3570 actually sending 8xx at this point.)
3572 73. An unknown lookup type in a host list could cause Exim to panic-die when
3573 the list was checked. (An example that provoked this was putting <; in the
3574 middle of a list instead of at the start.) If this happened during a DATA
3575 ACL check, a -D file could be left lying around. This kind of configuration
3576 error no longer causes Exim to die; instead it causes a defer errror. The
3577 incident is still logged to the main and panic logs.
3579 74. Buglet left over from Exim 3 conversion. The message "too many messages
3580 in one connection" was written to the rejectlog but not the mainlog, except
3581 when address rewriting (yes!) was being logged.
3583 75. Added write_rejectlog option.
3585 76. When a system filter was run not as root (that is, when system_filter_user
3586 was set), the values of the $n variables were not being returned to the
3587 main process; thus, they were not subsequently available in the $sn
3590 77. Added +return_path_on_delivery log selector.
3592 78. A connection timeout was being treated differently from recipients deferred
3593 when testing hosts_max_try with a message that was older than the host's
3594 retry timeout. (The host should not be counted, thus allowing all hosts to
3595 be tried at least once before bouncing.) This may have been the cause of an
3596 occasionally reported bug whereby a message would remain on the queue
3597 longer than the retry timeout, but would be bounced if a delivery was
3598 forced. I say "may" because I never totally pinned down the problem;
3599 setting up timeout/retry tests is difficult. See also the next item.
3601 79. The ultimate address timeout was not being applied to errors that involved
3602 a combination of host plus message (for example, a timeout on a MAIL
3603 command). When an address resolved to a number of possible hosts, and they
3604 were not all tried for each delivery (e.g. because of hosts_max_try), a
3605 message could remain on the queue longer than the retry timeout.
3607 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
3610 81. Fixed an obscure SMTP outgoing bug which required at least the following
3611 conditions: (a) there was another message waiting for the same server;
3612 (b) the server returned 5xx to all RCPT commands in the first message so
3613 that the message was not completed; (c) the server dropped the connection
3614 or gave a negative response to the RSET that Exim sends to abort the
3615 transaction. The observed case was a dropped connection after DATA that had
3616 been sent in pipelining mode. That is, the server had advertised PIPELINING
3617 but was not implementing it correctly. The effect of the bug was incorrect
3618 behaviour, such as trying another host, and this could lead to a crash.
3624 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
3625 and daemon.c were passed as pointers to ints; they should have been
3626 pointers to socklen_t variables (which are typically unsigned ints).
3628 2. Some signed/unsigned type warnings in the os.c file for Linux have been
3631 3. Fixed a really odd bug that affected only the testing scheme; patching a
3632 certain fixed string in the binary changed the value of another string that
3633 happened to be identical to the end of the original first string.
3635 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
3636 name", it returns that address as the IP address. On some operating
3637 systems (e.g. Solaris), it also passes back the IP address string as the
3638 "host name". However, on others (e.g. Linux), it passes back an empty
3639 string. Exim wasn't checking for this, and was changing the host name to an
3640 empty string, assuming it had been canonicized.
3642 5. Although rare, it is permitted to have more than one PTR record for a given
3643 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
3644 all the names associated with an address, because they do in Solaris.
3645 However, it seems that they do not in Linux for data that comes from the
3646 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
3647 I found this out when I moved to a new Linux workstation and tried to run
3648 the Exim test suite.
3650 To get round this problem I have changed the code so that it now does its
3651 own call to the DNS to look up PTR records when searching for a host name.
3652 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
3653 addresses that are only in /etc/hosts are still found.
3655 This behaviour is, however, controlled by an option called host_lookup_
3656 order, which defaults to "bydns:byaddr". If people want to use the other
3657 order, or indeed, just use one or the other means of lookup, they can
3658 specify it in this variable.
3660 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
3661 some operating systems, this comes back from gethostbyaddr() as an empty
3662 string, and this is what Exim used to test for. However, it seems that in
3663 other systems, "." is yielded. Exim now tests for this case too.
3665 7. The values of check_spool_space and check_log_space are now held internally
3666 as a number of kilobytes instead of an absolute number of bytes. If a
3667 numbers is specified without 'K' or 'M', it is rounded up to the nearest
3668 kilobyte. This means that much larger values can be stored.
3670 8. Exim monitor: an attempt to get the action menu when not actually pointing
3671 at a message produces an empty menu entitled "No message selected". This
3672 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
3673 no entries in it ("Shell widget menu has zero width and/or height"). So I
3674 have added a single, blank menu entry in this case.
3676 9. Added ${quote_local_part.
3678 10. MIME decoding is now applied to the contents of Subject: header lines when
3681 11. Now that a reference to $sender_host_address automatically causes a reverse
3682 lookup to occur if necessary (4.13/18), there is no need to arrange for a
3683 host lookup before query-style lookups in lists that might use this
3684 variable. This has therefore been abolished, and the "net-" prefix is no
3685 longer necessary for query-style lookups.
3687 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
3688 have been a typo for LFLAGS, so it has been changed.
3690 13. The install script calls Exim with "-C /dev/null" in order to find the
3691 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
3692 to be output. Howeve, since Exim outputs its version number before the
3693 error, it didn't break the script. It just looked ugly. I fixed this by
3694 always allowing "-C /dev/null" if the caller is root.
3696 14. Ignore overlarge ACL variable number when reading spool file - insurance
3697 against a later release with more variables having written the file.
3699 15. The standard form for an IPv6 address literal was being rejected by EHLO.
3700 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
3701 this, as well as the form without the "IPv6" on the front.
3703 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
3704 OS/Makefile-Darwin file.
3706 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
3707 only to LDAP libraries that do not have LDAP_OPT_DEREF.
3709 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
3710 variable. However, some OS use ints for inodes. I've added cast to long int
3711 to get rid of the compiler warning.
3713 19. I had forgotten to lock out "/../" in configuration file names when
3714 ALT_CONFIG_PREFIX was set.
3716 20. Routers used for verification do not need to specify transports. However,
3717 if such a router generated a host list, and callout was configured, Exim
3718 crashed, because it could not find a port number from the (non-existent)
3719 transport. It now assumes port 25 in this circumstance.
3721 21. Added the -t option to exigrep.
3723 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
3724 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
3725 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
3726 them is set, LOOKUP_LSEARCH is forced.
3728 23. "exim -bV" now outputs a list of lookups that are included in the binary.
3730 24. Added sender and host information to the "rejected by local_scan()" log
3731 line; previously there was no indication of these.
3733 25. Added .include_if_exists.
3735 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
3736 Linux. It turns out that not all file systems support this. Apparently some
3737 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
3738 do it.) To cope with this, the error EINVAL, which means that sync-ing is
3739 not supported on the file descriptor, is now ignored when Exim is trying to
3740 sync a directory. This applies only to Linux.
3742 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
3744 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
3745 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
3748 29. Added support for maildirsize files from supplied patch (modified a bit).
3750 30. The use of :fail: followed by an empty string could lead Exim to respond to
3751 sender verification failures with (e.g.):
3753 550 Verification failed for <xxx>
3754 550 Sender verify failed
3756 where the first response line was missing the '-' that indicates it is not
3757 the final line of the response.
3759 31. The loop for finding the name of the user that called Exim had a hardwired
3760 limit of 10; it now uses the value of finduser_retries, which is used for
3761 all other user lookups.
3763 32. Added $received_count variable, available in data and not_smtp ACLs, and at
3766 33. Exim was neglecting to zero errno before one call of strtol() when
3767 expanding a string and expecting an integer value. On some systems this
3768 resulted in spurious "integer overflow" errors. Also, it was casting the
3769 result into an int without checking.
3771 34. Testing for a connection timeout using "timeout_connect" in the retry rules
3772 did not work. The code looks as if it has *never* worked, though it appears
3773 to have been documented since at least releast 1.62. I have made it work.
3775 35. The "timeout_DNS" error in retry rules, also documented since at least
3776 1.62, also never worked. As it isn't clear exactly what this means, and
3777 clearly it isn't a major issue, I have abolished the feature by treating it
3778 as "timeout", and writing a warning to the main and panic logs.
3780 36. The display of retry rules for -brt wasn't always showing the error code
3783 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
3784 timeout_connect_A, timeout_connect_MX.
3786 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
3787 to the empty sender.
3789 39. The daemon was not analysing the content of -oX till after it had closed
3790 stderr and disconnected from the controlling terminal. This meant that any
3791 syntax errors were only noted on the panic log, and the return code from
3792 the command was 0. By re-arranging the code a little, I've made the
3793 decoding happen first, so such errors now appear on stderr, and the return
3794 code is 1. However, the actual setting up of the sockets still happens in
3795 the disconnected process, so errors there are still only recorded on the
3798 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
3799 connections (as happens on some IP stacks) was logged at start up time as
3800 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
3801 it from "IPv6 and IPv4", which means that two separate sockets are being
3804 41. The debug output for gethostbyname2() or getipnodebyname() failures now
3805 says whether AF_INET or AF_INET6 was passed as an argument.
3807 42. Exiwhat output was messed up when time zones were included in log
3810 43. Exiwhat now gives more information about the daemon's listening ports,
3811 and whether -tls-on-connect was used.
3813 44. The "port" option of the smtp transport is now expanded.
3815 45. A "message" modifier in a "warn" statement in a non-message ACL was being
3816 silently ignored. Now an error message is written to the main and panic
3819 46. There's a new ACL modifier called "logwrite" which writes to a log file
3820 as soon as it is encountered.
3822 47. Added $local_user_uid and $local_user_gid at routing time.
3824 48. Exim crashed when trying to verify a sender address that was being
3827 49. Exim was recognizing only a space character after ".include". It now also
3828 recognizes a tab character.
3830 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
3831 extracting the relevant information from the specification. The man page no
3832 longer contains scrambled data for the -d option, and I've added a section
3833 at the front about calling Exim under different names.
3835 51. Added "extra_headers" argument to the "mail" command in filter files.
3837 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
3840 53. Installed eximstats 1.29.
3842 54. Added transport_filter_timeout as a generic transport option.
3844 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
3845 Cc: header lines. This was required by RFC 822, but it not required by RFC
3848 56. Exim used to add From:, Date:, and Message-Id: header lines to any
3849 incoming messages that did not have them. Now it does so only if the
3850 message originates locally, that is, if there is no associated remote host
3851 address. When Resent- header lines are present, this applies to the Resent-
3852 lines rather than the non-Resent- lines.
3854 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
3855 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
3857 58. Messages for configuration errors now include the name of the main
3858 configuration file - useful now that there may be more than one file in a
3859 list (.included file names were always shown).
3861 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
3862 for those rare installations that do not start the daemon as root or run it
3863 setuid root. I've cut out the call to initgroups() if the daemon is not
3866 60. The Exim user and group can now be bound into the binary as text strings
3867 that are looked up at the start of Exim's processing.
3869 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
3871 62. Added $mailstore_basename variable.
3873 63. Installed patch to sieve.c from Michael Haardt.
3875 64. When Exim failed to open the panic log after failing to open the main log,
3876 the original message it was trying to log was written to stderr and debug
3877 output, but if they were not available (the usual case in production), it
3878 was lost. Now it is written to syslog before the two lines that record the
3879 failures to open the logs.
3881 65. Users' Exim filters run in subprocesses under the user's uid. It is
3882 possible for a "deliver" command or an alias in a "personal" command to
3883 provoke an address rewrite. If logging of address rewriting is configured,
3884 this fails because the process is not running as root or exim. There may be
3885 a better way of dealing with this, but for the moment (because 4.30 needs
3886 to be released), I have disabled address rewrite logging when running a
3887 filter in a non-root, non-exim process.
3893 1. The buildconfig auxiliary program wasn't quoting the value set for
3894 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
3895 not defined. This bug was masked in 4.22 by the effect that was fixed in
3898 2. Some messages that were rejected after a message id was allocated were
3899 shown as "incomplete" by exigrep. It no longer does this for messages that
3900 are rejected by local_scan() or the DATA or non-SMTP ACLs.
3902 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
3903 have allow_domain_literals set, the ID did not get logged in the <= line.
3904 Domain literals are now always recognized in Message-ID: header lines.
3906 4. The first argument for a ${extract expansion item is the key name or field
3907 number. Leading and trailing spaces in this item were not being ignored,
3908 causing some misleading effects.
3910 5. When deliver_drop_privilege was set, single queue runner processes started
3911 manually (i.e. by the command "exim -q") or by the daemon (which uses the
3912 same command in the process it spins off) were not dropping privilege.
3914 6. When the daemon running as "exim" started a queue runner, it always
3915 re-executed Exim in the spun-off process. This is a waste of effort when
3916 deliver_drop_privilege is set. The new process now just calls the
3917 queue-runner function directly.
3923 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
3926 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
3927 ignored. Though the use of -oP was forcing the writing of a pid file, it
3928 was always written to the default place.
3930 3. If the message "no IP address found for host xxxx" is generated during
3931 incoming verification, it is now followed by identification of the incoming
3932 connection (so you can more easily find what provoked it).
3934 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
3936 5. Added some features to "harden" Exim a bit more against certain attacks:
3938 (a) There is now a build-time option called FIXED_NEVER_USERS that can
3939 be put in Local/Makefile. This is like the never_users runtime option,
3940 but it cannot be overridden. The default setting is "root".
3942 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
3943 prefix string with which any file named in a -C command line option
3946 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
3947 is retained for -C and -D only if the caller of Exim is root. Without
3948 it, the exim user may also use -C and -D and retain privilege.
3950 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
3951 command line option is disabled.
3953 6. Macro names set by the -D option must start with an upper case letter, just
3954 like macro names defined in the configuration file.
3956 7. Added "dereference=" facility to LDAP.
3958 8. Two instances of the typo "uknown" in the source files are fixed.
3960 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
3961 the Configure-Makefile script screwed up while processing it.
3963 10. Incorporated PCRE 4.4.
3965 11. The SMTP synchronization check was not operating right at the start of an
3966 SMTP session. For example, it could not catch a HELO sent before the client
3967 waited for the greeting. There is now a check for outstanding input at the
3968 point when the greeting is written. Because of the duplex, asynchronous
3969 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
3970 way, but not yet received, when the check is performed.
3972 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
3973 on TCP/IP sockets, because this apparently causes some broken clients to
3976 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
3977 unchanged) from the Cygwin maintainer.
3979 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
3980 was supported instead of testing for mbx. Effectively a typo.
3982 15. The spa authenticator server code was not checking that the input it
3983 received was valid base64.
3985 16. The debug output line for the "set" modifier in ACLs was not showing the
3986 name of the variable that was being set.
3988 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
3990 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
3991 lookup on its own. Something else was needed to trigger the lookup. For
3992 example, a match in host_lookup or the need for a host name in a host list.
3993 Now, if $sender_host_name is referenced and the host name has not yet been
3994 looked up, a lookup is performed. If the lookup fails, the variable remains
3995 empty, and $host_lookup_failed is set to "1".
3997 19. Added "eqi" as a case-independent comparison operator.
3999 20. The saslauthd authentication condition could segfault if neither service
4000 nor realm was specified.
4002 21. If an overflowing value such as "2048M" was set for message_size_limit, the
4003 error message that was logged was misleading, and incoming SMTP
4004 connections were dropped. The message is now more accurate, and temporary
4005 errors are given to SMTP connections.
4007 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
4008 (except RSET) with a 421 error, until QUIT is received. However, it was
4009 failing to send a response to QUIT.
4011 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
4012 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
4013 ACL is now run after the helo_try_verify_hosts code.
4015 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
4016 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
4017 case-independent, so other case variants are also recognized). Apparently
4018 some systems use these upper case variants.
4020 25. If more than two messages were waiting for the same host, and a transport
4021 filter was specified for the transport, Exim sent two messages over the
4022 same TCP/IP connection, and then failed with "socket operation on non-
4023 socket" when it tried to send the third.
4025 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
4027 27. The extern definition of crypt16() in expand.c was not being excluded when
4028 the OS had its own crypt16() function.
4030 28. Added bounce_return_body as a new option, and bounce_return_size_limit
4031 as a preferred synonym for return_size_limit, both as an option and as an
4034 29. Added LIBS=-liconv to OS/Makefile-OSF1.
4036 30. Changed the default configuration ACL to relax the local part checking rule
4037 for addresses that are not in any local domains. For these addresses,
4038 slashes and pipe symbols are allowed within local parts, but the sequence
4039 /../ is explicitly forbidden.
4041 31. SPA server authentication was not clearing the challenge buffer before
4044 32. log_message in a "warn" ACL statement was writing to the reject log as
4045 well as to the main log, which contradicts the documentation and doesn't
4046 seem right (because no rejection is happening). So I have stopped it.
4048 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
4049 However, I am unable to do any testing of this.
4051 34. Fixed an infelicity in the appendfile transport. When checking directories
4052 for a mailbox, to see if any needed to be created, it was accidentally
4053 using path names with one or more superfluous leading slashes; tracing
4054 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
4056 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
4057 contents are added to the log line that is written for every discarded
4058 recipient. (Previously a log_message setting was ignored.)
4060 36. The ${quote: operator now quotes the string if it is empty.
4062 37. The install script runs exim in order to find its version number. If for
4063 some reason other than non-existence or emptiness, which it checks, it
4064 could not run './exim', it was installing it with an empty version number,
4065 i.e. as "exim-". This error state is now caught, and the installation is
4068 38. An argument was missing from the function that creates an error message
4069 when Exim fails to connect to the socket for saslauthd authentication.
4070 This could cause Exim to crash, or give a corrupted message.
4072 39. Added isip, isip4, and isip6 to ${if conditions.
4074 40. The ACL variables $acl_xx are now saved with the message, and can be
4075 accessed later in routers, transports, and filters.
4077 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
4078 strings in the file are not string-expanded.
4080 42. If a MAIL command specified a SIZE value that was too large to fit into an
4081 int variable, the check against message_size_limit failed. Such values are
4082 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
4083 day this will have to be increased, but I don't think I want to be around
4084 when emails are that large.
4091 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
4092 iconv() is not standard in FreeBSD.
4094 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
4095 IPv6 enabled. The observed symptom was a segmentation fault on return from
4096 the function os_common_find_running_interfaces() in src/os.c.
4098 3. In the check_special_case() function in daemon.c I had used "errno" as an
4099 argument name, which causes warnings on some systems. This was basically a
4100 typo, since it was named "eno" in the comments!
4102 4. The code that waits for the clock to tick (at a resolution of some fraction
4103 of a second) so as to ensure message-id uniqueness was always waiting for
4104 at least one whole tick, when it could have waited for less. [This is
4105 almost certainly not relevant at current processor speeds, where it is
4106 unlikely to ever wait at all. But we try to future-proof.]
4108 5. The function that sleeps for a time interval that includes fractions of a
4109 second contained a race. It did not block SIGALRM between setting the
4110 timer, and suspending (a couple of lines later). If the interval was short
4111 and the sigsuspend() was delayed until after it had expired, the suspension
4112 never ended. On busy systems this could lead to processes getting stuck for
4115 6. Some uncommon configurations may cause a lookup to happen in a queue runner
4116 process, before it forks any delivery processes. The open lookup caching
4117 mechanism meant that the open file or database connection was passed into
4118 the delivery process. The problem was that delivery processes always tidy
4119 up cached lookup data. This could cause a problem for the next delivery
4120 process started by the queue runner, because the external queue runner
4121 process does not know about the closure. So the next delivery process
4122 still has data in the lookup cache. In the case of a file lookup, there was
4123 no problem because closing a file descriptor in a subprocess doesn't affect
4124 the parent. However, if the lookup was caching a connection to a database,
4125 the connection was closed, and the second delivery process was likely to
4126 see errors such as "PGSQL: query failed: server closed the connection
4127 unexpectedly". The problem has been fixed by closing all cached lookups
4128 in a queue runner before running a delivery process.
4130 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
4131 seem to have the "const" qualifier which it has on other OS. I've
4134 8. Change 4.21/2 was too strict. It is only if there are two authenticators
4135 *of the same type* (client or server) with the same public name that an
4136 error should be diagnosed.
4138 9. When Exim looked up a host name for an IP address, but failed to find the
4139 original IP address when looking up the host name (a safety check), it
4140 output the message "<ip address> does not match any IP for NULL", which was
4141 confusing, to say the least. The bug was that the host name should have
4142 appeared instead of "NULL".
4144 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
4145 user that is built into the binary, and the -C or -D options is used, root
4146 privilege is dropped before the configuration file is read. In addition,
4147 logging is switched to stderr instead of the normal log files. If the
4148 configuration then re-defines the Exim user, the unprivileged environment
4149 is probably not what is expected, so Exim logs a panic warning message (but
4152 However, if deliver_drop_privilege is set, the unprivileged state may well
4153 be exactly what is intended, so the warning has been cut out in that case,
4154 and Exim is allowed to try to write to its normal log files.
4160 1. smtp_return_error_details was not giving details for temporary sender
4161 or receiver verification errors.
4163 2. Diagnose a configuration error if two authenticators have the same public
4166 3. Exim used not to create the message log file for a message until the first
4167 delivery attempt. This could be confusing when incoming messages were held
4168 for policy or load reasons. The message log file is now created at the time
4169 the message is received, and an initial "Received" line is written to it.
4171 4. The automatically generated man page for command line options had a minor
4172 bug that caused no ill effects; however, a more serious problem was that
4173 the procedure for building the man page automatically didn't always
4174 operate. Consequently, release 4.20 contains an out-of-date version. This
4175 shouldn't happen again.
4177 5. When building Exim with embedded Perl support, the script that builds the
4178 Makefile was calling 'perl' to find its compile-time parameters, ignoring
4179 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
4181 6. The freeze_tell option was not being used for messages that were frozen on
4182 arrival, either by an ACL or by local_scan().
4184 7. Added the smtp_incomplete_transaction log selector.
4186 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
4187 was accepting AUTH without a new EHLO.
4189 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
4190 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
4193 10. Logging of TCP/IP connections (when configured) now happens in the main
4194 daemon process instead of the child process, so that the TCP/IP connection
4195 count is more accurate (but it can never be perfect).
4197 11. The use of "drop" in a nested ACL was not being handled correctly in the
4198 outer ACL. Now, if condition failure induced by the nested "drop" causes
4199 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
4200 or "require"), the connection is dropped.
4202 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
4203 that yield "discard" can now be used with an "accept" or a "discard" verb,
4204 but an error is generated for any others (because I can't see a useful way
4205 to define what should happen).
4207 13. When an ACL is read dynamically from a file (or anywhere else), the lines
4208 are now processed in the same way as lines in the Exim configuration file.
4209 In particular, continuation lines are supported.
4211 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
4213 15. Added -ti meaning -t -i.
4215 16. Check for letters, digits, hyphens, and dots in the names of dnslist
4216 domains, and warn by logging if others are found.
4218 17. At least on BSD, alignment is not guarenteed for the array of ifreq's
4219 returned from GIFCONF when Exim is trying to find the list of interfaces on
4220 a host. The code in os.c has been modified to copy each ifreq to an aligned
4221 structure in all cases.
4223 Also, in some cases, the returned ifreq's were being copied to a 'struct
4224 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
4225 means the last couple of bytes of an IPv6 address could be chopped if the
4226 ifreq contained only a normal sockaddr (14 bytes storage).
4228 18. Named domain lists were not supported in the hosts_treat_as_local option.
4229 An entry such as +xxxx was not recognized, and was treated as a literal
4232 19. Ensure that header lines added by a DATA ACL are included in the reject log
4233 if the ACL subsequently rejects the message.
4235 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
4236 MD5 (which is deprecated).
4238 21. When testing a filter file using -bf, Exim was writing a message when it
4239 took the sender from a "From " line in the message, but it was not doing so
4240 when it took $return_path from a Return-Path: header line. It now does.
4242 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
4243 with a valid header line field name (a series of printing characters
4244 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
4246 23. Changed "disc" in the source to "disk" to conform to the documentation and
4247 the book and for uniformity.
4249 24. Ignore Sendmail's -Ooption=value command line item.
4251 25. When execve() failed while trying to run a command in a pipe transport,
4252 Exim was returning EX_UNAVAILBLE (69) from the subprocess. However, this
4253 could be confused with a return value of 69 from the command itself. This
4254 has been changed to 127, the value the shell returns if it is asked to run
4255 a non-existent command. The wording for the related log line suggests a
4256 non-existent command as the problem.
4258 26. If received_header_text expands to an empty string, do not add a Received:
4259 header line to the message. (Well, it adds a token one on the spool, but
4260 marks it "old" so that it doesn't get used or transmitted.)
4262 27. Installed eximstats 1.28 (addition of -nt option).
4264 28. There was no check for failure on the call to getsockname() in the daemon
4265 code. This can fail if there is a shortage of resources on the system, with
4266 ENOMEM, for example. A temporary error is now given on failure.
4268 29. Contrary to the C standard, it seems that in some environments, the
4269 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
4270 program. Exim now does this explicitly; it affects the formatting of
4271 timestamps using strftime().
4273 30. If exiqsumm was given junk data, it threw up some uninitialized variable
4274 complaints. I've now initialized all the variables, to avoid this.
4276 32. Header lines added by a system filter were not being "seen" during
4277 transport-time rewrites.
4279 33. The info_callback() function passed to OpenSSL is set up with type void
4280 (*)(SSL *, int, int), as described somewhere. However, when calling the
4281 function (actually a macro) that sets it up, the type void(*)() is
4282 expected. I've put in a cast to prevent warnings from picky compilers.
4284 34. If a DNS black list lookup found a CNAME record, but there were no A
4285 records associated with the domain it pointed at, Exim crashed.
4287 35. If a DNS black list lookup returned more than one A record, Exim ignored
4288 all but the first. It now scans all returned addresses if a particular IP
4289 value is being sought. In this situation, the contents of the
4290 $dnslist_value variable are a list of all the addresses, separated by a
4293 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
4294 to accept a host name and all its aliases if the forward lookup for any of
4295 them yielded the IP address of the incoming connection. Now it accepts only
4296 those names whose forward lookup yields the correct IP address. Any other
4297 names are discarded. This closes a loophole whereby a rogue DNS
4298 administrator could create reverse DNS records to break through a
4299 wildcarded host restriction in an ACL.
4301 37. If a user filter or a system filter that ran in a subprocess used any of
4302 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
4303 the wrong values were passed to the pipe command ($thisaddress had the
4304 value of $0, $0 had the value of $1, etc). This bug was introduced by
4305 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
4307 38. Improved the line breaking for long SMTP error messages from ACLs.
4308 Previously, if there was no break point between 40 and 75 characters, Exim
4309 left the rest of the message alone. Two changes have been made: (a) I've
4310 reduced the minimum length to 35 characters; (b) if it can't find a break
4311 point between 35 and 75 characters, it looks ahead and uses the first one
4312 that it finds. This may give the occasional overlong line, but at least the
4313 remaining text gets split now.
4315 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
4316 file descriptors might be low, and that setting 1000 would always raise it.
4317 It turns out that in some environments, the limit is already over 1000 and
4318 that lowering it causes trouble. So now Exim takes care not to decrease it.
4320 40. When delivering a message, the value of $return_path is set to $sender_
4321 address at the start of routing (routers may change the value). By an
4322 oversight, this default was not being set up when an address was tested by
4323 -bt or -bv, which affected the outcome if any router or filter referred to
4326 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
4327 log only when "message" or "log_message" are set. However, if one of the
4328 conditions was an address verification, or a call to a nested ACL, the
4329 messages generated by the underlying test were being passed through. This
4330 no longer happens. The underlying message is available in $acl_verify_
4331 message for both "message" and "log_message" expansions, so it can be
4332 passed through if needed.
4334 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
4335 new expansion $bh_ to give the encoded byte string without charset
4336 translation. Translation happens only if iconv() is available; HAVE_ICONV
4337 indicates this at build time. HEADERS_CHARSET gives the charset to
4338 translate to; headers_charset can change it in the configuration, and
4339 "headers charset" can change it in an individual filter file.
4341 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
4342 that creates RFC 2047 encoded "words" labels them as that charset instead
4343 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
4344 expansion operator; (ii) when Exim creates a From: line for a local
4345 message; (iii) when a header line is rewritten to include a "phrase" part.
4347 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
4348 buggy, causing it to skip the first lines of messages whose message ID
4349 ended in 'D'. This would not have bitten before Exim release 4.14, because
4350 message IDs were unlikely to end in 'D' before then. The effect was to have
4351 incorrect size information for certain domains.
4353 45. #include "config.h" was missing at the start of the crypt16.c module. This
4354 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
4357 46. If there was a timeout during a "random" callout check, Exim treated it as
4358 a failure of the random address, and carried on sending RSET and the real
4359 address. If the delay was just some slowness somewhere, the response to the
4360 original RCPT would be taken as a response to RSET and so on, causing
4361 mayhem of various kinds.
4363 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
4364 when I implemented it. It didn't allow for the fact that some option values
4365 may legitimatetly be negative (e.g. size_addition), and it didn't even do
4366 the right test for positive values.
4368 48. Domain names in DNS records are case-independent. Exim always looks them up
4369 in lower case. Some resolvers return domain names in exactly the case they
4370 appear in the zone file, that is, they may contain uppercase letters. Not
4371 all resolvers do this - some return always lower case. Exim was treating a
4372 change of case by a resolver as a change of domain, similar to a widening
4373 of a domain abbreviation. This triggered its re-routing code and so it was
4374 trying to route what was effectively the same domain again. This normally
4375 caused routing to fail (because the router wouldn't handle the domain
4376 twice). Now Exim checks for this case specially, and just changes the
4377 casing of the domain that it ultimately uses when it transmits the message
4380 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
4383 50. If a filter generated a file delivery with a non-absolute name (possible if
4384 no home directory exists for the router), the forbid_file option was not
4387 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
4388 the existing equality matching.
4390 52. Exim was using ints instead of ino_t variables in some places where it was
4391 dealing with inode numbers.
4393 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
4394 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
4395 called TMPDIR, and if it finds it is different, it changes its value.
4397 54. The smtp_printf() function is now made available to local_scan() so
4398 additional output lines can be written before returning. There is also an
4399 smtp_fflush() function to enable the detection of a dropped connection.
4400 The variables smtp_input and smtp_batched_input are exported to
4403 55. Changed the default runtime configuration: the message "Unknown user"
4404 has been removed from the ACL, and instead placed on the localuser router,
4405 using the cannot_route_message feature. This means that any verification
4406 failures that generate their own messages won't get overridden. Similarly,
4407 the "Unrouteable address" message that was in the ACL for unverifiable
4408 relay addresses has also been removed.
4410 56. Added hosts_avoid_esmtp to the smtp transport.
4412 57. The exicyclog script was not checking for the esoteric option
4413 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
4414 will work only if exicyclog is run under the appropriate euid.
4416 58. Following a discussion on the list, the rules by which Exim recognises line
4417 endings on incoming messages have been changed. The -dropcr and drop_cr
4418 options are now no-ops, retained only for backwards compatibility. The
4419 following line terminators are recognized: LF CRLF CR. However, special
4420 processing applies to CR:
4422 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
4423 nor a local message in the state where . is a terminator.
4425 (ii) If a bare CR is encountered in a header line, an extra space is added
4426 after the line terminator so as not to end the header. The reasoning
4427 behind this is that bare CRs in header lines are most likely either
4428 to be mistakes, or people trying to play silly games.
4430 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
4431 was being incorrectly given as 18 bytes larger than it should have been.
4432 This is a VOB (very old bug).
4434 60. This may never have affected anything current, but just in case it has:
4435 When the local host is found other than at the start of a list of hosts,
4436 the local host, those with the same MX, and any that follow, are discarded.
4437 When the list in question was part of a longer list of hosts, the following
4438 hosts (not currently being processed) were also being discarded. This no
4439 longer happens. I'm not sure if this situation could ever has previously
4442 61. Added the "/MX" feature to lists of hosts in the manualroute and query
4445 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
4446 header. This is something that is recommended in a new Internet Draft, and
4447 is something that is documented as being done by Sendmail. There are two
4448 possible values. For messages generated by the autoreply transport, Exim
4451 Auto-Submitted: auto-replied
4453 whereas for all other generated messages (e.g. bounces) it adds
4455 Auto-Submitted: auto-generated
4457 63. The "personal" condition in filters now includes a test for the
4458 Auto-Submitted: header. If it contains the string "auto-" the message it
4459 not considered personal.
4461 64. Added rcpt_include_affixes as a generic transport option.
4463 65. Added queue_only_override (default true).
4465 66. Added the syslog_duplication option.
4467 67. If what should have been the first header line of a message consisted of
4468 a space followed by a colon, Exim was mis-interpreting it as a header line.
4469 It isn't of course - it is syntactically invalid and should therefore be
4470 treated as the start of the message body. The misbehaviour could have
4471 caused a number of strange effects, including loss of data in subsequent
4472 header lines, and spool format errors.
4474 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
4475 client host had authenticated. This control can now be exercised by an ACL
4476 for more flexibility.
4478 69. By default, callouts do not happen when testing with -bh. There is now a
4479 variant, -bhc, which does actually run the callout code, including
4480 consulting and updating the callout cache.
4482 70. Added support for saslauthd authentication, courtesy of Alexander
4485 71. If statvfs() failed on the spool or log directories while checking their
4486 size for availability, Exim confusingly gave the error "space shortage".
4487 Furthermore, in debugging mode it crashed with a floating point exception.
4488 These checks are done if check_{spool,log}_{space,inodes} are set, and when
4489 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
4490 serious problem, Exim now writes to the main and panic logs when this
4491 happens, with details of the failure. It then refuses to accept the
4492 incoming message, giving the message "spool directory problem" or "log
4493 directory problem" with a 421 code for SMTP messages.
4495 72. When Exim is about to re-exec itself, it ensures that the file descriptors
4496 0, 1, and 2 exist, because some OS complain for execs without them (see
4497 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
4498 descriptors. However, the code omitted to check that the open succeeded,
4499 causing mysterious errors if for some reason the permissions on /dev/null
4500 got screwed. Now Exim writes a message to the main and panic logs, and
4501 bombs out if it can't open /dev/null.
4503 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
4504 interact so that it is all more flexible. It is supposed to remain
4505 backwards compatible. Also added extra_local_interfaces.
4507 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
4508 to bomb out with an assertion failure - to the client this appears as a
4509 connection drop. This problem occurs in the part of the code that was taken
4510 from the Samba project. Fortunately, the assertion is in a very simple
4511 function, so I have fixed this by reproducing the function inline in the
4512 one place where it is called, and arranging for authentication to fail
4513 instead of killing the process with assert().
4515 75. The SPA client code was not working when the server requested OEM rather
4516 than Unicode encoding.
4518 76. Added code to make require_files with a specific uid setting more usable in
4519 the case where statting the file as root fails - usually a non-root-mounted
4520 NFS file system. When this happens and the failure is EACCES, Exim now
4521 forks a subprocess and does the per-uid checking as the relevant uid.
4523 77. Added process_log_path.
4525 78. If log_file_path was not explicitly set, a setting of check_log_space or
4526 check_log_inodes was ignored.
4528 79. If a space check for the spool or log partitions fails, the incident is now
4529 logged. Of course, in the latter case the data may get lost...
4531 80. Added the %p formatting code to string_format() so that it can be used to
4532 print addresses in debug_print(). Adjusted all the address printing in the
4533 debugging in store.c to use %p rather than %d.
4535 81. There was a concern that a line of code in smtp_in.c could overflow a
4536 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
4537 initially expressed, the concern was not well-founded, because trailing
4538 spaces are removed early. However, if the trailing spaces were followed by
4539 a NULL, they did not get removed, so the overflow was possible. Two fixes
4542 (a) I re-wrote the offending code in a cleaner fashion.
4543 (b) If an incoming SMTP command contains a NULL character, it is rejected
4546 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
4547 runs initgroups(), so that if the Exim user is in any additional groups,
4548 they will be used during message reception.
4554 The change log for 4.20 and earlier releases has been archived.
projects / exim.git / blob