1 This document describes *changes* to previous versions, that might
2 affect Exim's operation, with an unchanged configuration file. For new
3 options, and new features, see the NewStuff file next to this ChangeLog.
8 JH/01 The hosts_connection_nolog main option now also controls "no MAIL in
9 SMTP connection" log lines.
11 JH/02 Option default value updates:
12 - queue_fast_ramp (main) true (was false)
13 - remote_max_parallel (main) 4 (was 2)
15 JH/03 Cache static regex pattern compilations, for use by ACLs.
17 JH/04 Bug 2903: avoid exit on an attempt to rewrite a malformed address.
18 Make the rewrite never match and keep the logging. Trust the
19 admin to be using verify=header-syntax (to actually reject the message).
21 JH/05 Follow symlinks for placing a watch on TLS creds files. This means
22 (under Linux) we watch the dir containing the final file; previously
23 it would be the dir with the first symlink. We still do not monitor
26 JH/06 Check for bad chars in rDNS for sender_host_name. The OpenBSD (at least)
27 dn_expand() is happy to pass them through.
29 JH/07 OpenSSL Fix auto-reload of changed server OCSP proof. Previously, if
30 the file with the proof had an unchanged name, the new proof(s) were
31 loaded on top of the old ones (and nover used; the old ones were stapled).
33 JH/08 Bug 2915: Fix use-after-free for $regex<n> variables. Previously when
34 more than one message arrived in a single connection a reference from
35 the earlier message could be re-used. Often a sigsegv resulted.
36 These variables were introduced in Exim 4.87.
37 Debug help from Graeme Fowler.
39 JH/09 Fix ${filter } for conditions that modify $value. Previously the
40 modified version would be used in construction the result, and a memory
43 JH/10 GnuTLS: fix for (IOT?) clients offering no TLS extensions at all.
44 Find and fix by Jasen Betts.
46 JH/11 OpenSSL: fix for ancient clients needing TLS support for versions earlier
47 than TLSv1,2, Previously, more-recent versions of OpenSSL were permitting
48 the systemwide configuration to override the Exim config.
50 HS/01 Bug 2728: Introduce EDITME option "DMARC_API" to work around incompatible
51 API changes in libopendmarc.
53 JH/12 Bug 2930: Fix daemon startup. When started from any process apart from
54 pid 1, in the normal "background daemon" mode, having to drop process-
55 group leadership also lost track of needing to create listener sockets.
57 JH/13 Bug 2929: Fix using $recipients after ${run...}. A change made for 4.96
58 resulted in the variable appearing empty. Find and fix by Ruben Jenster.
60 JH/14 Bug 2933: Fix regex substring match variables for null matches. Since 4.96
61 a capture group which obtained no text (eg. "(abc)*" matching zero
62 occurrences) could cause a segfault if the corresponding $<n> was
65 JH/15 Fix argument parsing for ${run } expansion. Previously, when an argument
66 included a close-brace character (eg. it itself used an expansion) an
69 JH/16 Move running the smtp connect ACL to before, for TLS-on-connect ports,
70 starting TLS. Previously it was after, meaning that attackers on such
71 ports had to be screened using the host_reject_connection main config
72 option. The new sequence aligns better with the STARTTLS behaviour, and
73 permits defences against crypto-processing load attacks, even though it
74 is strictly an incompatible change.
75 Also, avoid sending any SMTP fail response for either the connect ACL
76 or host_reject_connection, for TLS-on-connect ports.
78 JH/17 Permit the ACL "encrypted" condition to be used in a HELO/EHLO ACL,
79 Previously this was not permitted, but it makes reasonable sense.
80 While there, restore a restriction on using it from a connect ACL; given
81 the change JH/16 it could only return false (and before 4.91 was not
84 JH/18 Fix a fencepost error in logging. Previously (since 4.92) when a log line
85 was exactly sized compared to the log buffer, a crash occurred with the
86 misleading message "bad memory reference; pool not found".
87 Found and traced by Jasen Betts.
89 JH/19 Bug 2911: Fix a recursion in DNS lookups. Previously, if the main option
90 dns_again_means_nonexist included an element causing a DNS lookup which
91 itself returned DNS_AGAIN, unbounded recursion occurred. Possible results
92 included (though probably not limited to) a process crash from stack
93 memory limit, or from excessive open files. Replace this with a paniclog
94 whine (as this is likely a configuration error), and returning
97 JH/20 Bug 2954: (OpenSSL) Fix setting of explicit EC curve/group. Previously
98 this always failed, probably leading to the usual downgrade to in-clear
101 JH/21 Fix TLSA lookups. Previously dns_again_means_nonexist would affect
102 SERVFAIL results, which breaks the downgrade resistance of DANE. Change
103 to not checking that list for these lookups.
105 JH/22 Bug 2434: Add connection-elapsed "D=" element to more connection
108 JH/23 Fix crash in string expansions. Previously, if an empty variable was
109 immediately followed by an expansion operator, a null-indirection read
110 was done, killing the process.
112 JH/24 Bug 2997: When built with EXPERIMENTAL_DSN_INFO, bounce messages can
113 include an SMTP response string which is longer than that supported
114 by the delivering transport. Alleviate by wrapping such lines before
117 JH/25 Bug 2827: Restrict size of References: header in bounce messages to 998
118 chars (RFC limit). Previously a limit of 12 items was made, which with
119 a not-impossible References: in the message being bounced could still
120 be over-large and get stopped in the transport.
122 JH/26 For a ${readsocket } in TLS mode, send a TLS Close Alert before the TCP
123 close. Previously a bare socket close was done.
125 JH/27 Fix ${srs_encode ..}. Previously it would give a bad result for one day
128 JH/28 Bug 2996: Fix a crash in the smtp transport. When finding that the
129 message being considered for delivery was already being handled by
130 another process, and having an SMTP connection already open, the function
131 to close it tried to use an uninitialized variable. This would afftect
132 high-volume sites more, especially when running mailing-list-style loads.
133 Pollution of logs was the major effect, as the other process delivered
134 the message. Found and partly investigated by Graeme Fowler.
136 JH/29 Change format of the internal ID used for message identification. The old
137 version only supported 31 bits for a PID element; the new 64 (on systems
138 which can use Base-62 encoding, which is all currently supported ones
139 but not Darwin (MacOS) or Cygwin, which have case-insensitive filesystems
140 and must use Base-36). The new ID is 23 characters rather than 16, and is
141 visible in various places - notably logs, message headers, and spool file
142 names. Various of the ancillary utilities also have to know the format.
143 As well as the expanded PID portion, the sub-second part of the time
144 recorded in the ID is expanded to support finer precision. Theoretically
145 this permits a receive rate from a single comms channel of better than the
147 The major timestamp part of the ID is not changed; at 6 characters it is
148 usable until about year 3700.
149 Updating from previously releases is fully supported: old-format spool
150 files are still usable, and the utilities support both formats. New
151 message will use the new format. The one hints-DB file type which uses
152 message-IDs (the transport wait- DB) will be discarded if an old-format ID
153 is seen; new ones will be built with only new-format IDs.
154 Optionally, a utility can be used to convert spool files from old to new,
155 but this is only an efficiency measure not a requirement for operation
156 Downgrading from new to old requires running a provided utility, having
157 first stopped all operations. This will convert any spool files from new
158 back to old (losing time-precision and PID information) and remove any
159 wait- hints databases.
161 JH/30 Bug 3006: Fix handling of JSON strings having embedded commas. Previously
162 we treated them as item separators when parsing for a list item, but they
163 need to be protected by the doublequotes. While there, add handling for
166 JH/31 Bug 2998: Fix ${utf8clean:...} to disallow UTF-16 surrogate codepoints.
167 Found and fixed by Jasen Betts. No testcase for this as my usual text
168 editor insists on emitting only valid UTF-8.
170 JH/32 Fix "tls_dhparam = none" under GnuTLS. At least with 3.7.9 this gave
171 a null-indirection SIGSEGV for the receive process.
173 JH/33 Fix free for live variable $value created by a ${run ...} expansion during
174 -bh use. Internal checking would spot this and take a panic.
176 JH/34 Bug 3013: Fix use of $recipients within arguments for ${run...}.
177 In 4.96 this would expand to empty.
179 JH/35 Bug 3014: GnuTLS: fix expiry date for an auto-generated server
180 certificate. Find and fix by Andreas Metzler.
182 JH/36 Add ARC info to DMARC hostory records.
184 JH/37 Bug 3016: Avoid sending DSN when message was accepted under fakereject
185 or fakedefer. Previously the sender could discover that the message
186 had in fact been accepted.
188 JH/38 Taint-track intermediate values from the peer in multi-stage authentation
189 sequences. Previously the input was not noted as being tainted; notably
190 this resulted in behaviour of LOGIN vs. PLAIN being inconsistent under
191 bad coding of authenticators.
193 JH/39 Bug 3023: Fix crash induced by some combinations of zero-length strings
194 and ${tr...}. Found and diagnosed by Heiko Schlichting.
200 JH/01 Move the wait-for-next-tick (needed for unique message IDs) from
201 after reception to before a subsequent reception. This should
202 mean slightly faster delivery, and also confirmation of reception
205 JH/02 Move from using the pcre library to pcre2. The former is no longer
206 being developed or supported (by the original developer).
208 JH/03 Constification work in the filters module required a major version
209 bump for the local-scan API. Specifically, the "headers_charset"
210 global which is visible via the API is now const and may therefore
211 not be modified by local-scan code.
213 JH/04 Fix ClamAV TCP use under FreeBSD. Previously the OS-specific shim for
214 sendfile() didi not account for the way the ClamAV driver code called it.
216 JH/05 Bug 2819: speed up command-line messages being read in. Previously a
217 time check was being done for every character; replace that with one
220 JH/06 Bug 2815: Fix ALPN sent by server under OpenSSL. Previously the string
221 sent was prefixed with a length byte.
223 JH/07 Change the SMTP feature name for pipelining connect to be compliant with
224 RFC 5321. Previously Dovecot (at least) would log errors during
227 JH/08 Remove stripping of the binaries from the FreeBSD build. This was added
228 in 4.61 without a reason logged. Binaries will be bigger, which might
229 matter on diskspace-constrained systems, but debug is easier.
231 JH/09 Fix macro-definition during "-be" expansion testing. The move to
232 write-protected store for macros had not accounted for these runtime
233 additions; fix by removing this protection for "-be" mode.
235 JH/10 Convert all uses of select() to poll(). FreeBSD 12.2 was found to be
236 handing out large-numbered file descriptors, violating the usual Unix
237 assumption (and required by Posix) that the lowest possible number will be
238 allocated by the kernel when a new one is needed. In the daemon, and any
239 child procesees, values higher than 1024 (being bigger than FD_SETSIZE)
240 are not useable for FD_SET() [and hence select()] and overwrite the stack.
241 Assorted crashes happen.
243 JH/11 Fix use of $sender_host_name in daemon process. When used in certain
244 main-section options or in a connect ACL, the value from the first ever
245 connection was never replaced for subsequent connections. Found by
248 JH/12 Bug 2838: Fix for i32lp64 hard-align platforms. Found for SPARC Linux,
249 though only once PCRE2 was introduced: the memory accounting used under
250 debug offset allocations by an int, giving a hard trap in early startup.
251 Change to using a size_t. Debug and fix by John Paul Adrian Glaubitz.
253 JH/13 Bug 2845: Fix handling of tls_require_ciphers for OpenSSL when a value
254 with underbars is given. The write-protection of configuration introduced
255 in 4.95 trapped when normalisation was applied to an option not needing
258 JH/14 Bug 1895: TLS: Deprecate RFC 5114 Diffie-Hellman parameters.
260 JH/15 Fix a resource leak in *BSD. An off-by-one error resulted in the daemon
261 failing to close the certificates directory, every hour or any time it
264 JH/16 Debugging initiated by an ACL control now continues through into routing
265 and transport processes. Previously debugging stopped any time Exim
266 re-execs, or for processing a queued message.
268 JH/17 The "expand" debug selector now gives more detail, specifically on the
269 result of expansion operators and items.
271 JH/18 Bug 2751: Fix include_directory in redirect routers. Previously a
272 bad comparison between the option value and the name of the file to
273 be included was done, and a mismatch was wrongly identified.
274 4.88 to 4.95 are affected.
276 JH/19 Support for Berkeley DB versions 1 and 2 is withdrawn.
278 JH/20 When built with NDBM for hints DB's check for nonexistence of a name
279 supplied as the db file-pair basename. Previously, if a directory
280 path was given, for example via the autoreply "once" option, the DB
281 file.pag and file.dir files would be created in that directory's
284 JH/21 Remove the "allow_insecure_tainted_data" main config option and the
285 "taint" log_selector. These were previously deprecated.
287 JH/22 Fix static address-list lookups to properly return the matched item.
288 Previously only the domain part was returned.
290 JH/23 Bug 2864: FreeBSD: fix transport hang after 4xx/5xx response. Previously
291 the call into OpenSSL to send a TLS Close was being repeated; this
292 resulted in the library waiting for the peer's Close. If that was never
293 sent we waited forever. Fix by tracking send calls.
295 JH/24 The ${run} expansion item now expands its command string elements after
296 splitting. Previously it was before; the new ordering makes handling
297 zero-length arguments simpler. The old ordering can be obtained by
298 appending a new option "preexpand", after a comma, to the "run".
300 JH/25 Taint-check exec arguments for transport-initiated external processes.
301 Previously, tainted values could be used. This affects "pipe", "lmtp" and
302 "queryprogram" transport, transport-filter, and ETRN commands.
303 The ${run} expansion is also affected: in "preexpand" mode no part of
304 the command line may be tainted, in default mode the executable name
307 JH/26 Fix CHUNKING on a continued-transport. Previously the usabliility of
308 the the facility was not passed across execs, and only the first message
309 passed over a connection could use BDAT; any further ones using DATA.
311 JH/27 Support the PIPECONNECT facility in the smtp transport when the helo_data
312 uses $sending_ip_address and an interface is specified.
313 Previously any use of the local address in the EHLO name disabled
314 PIPECONNECT, the common case being to use the rDNS of it.
316 JH/28 OpenSSL: fix transport-required OCSP stapling verification under session
317 resumption. Previously verify failed because no certificate status is
318 passed on the wire for the restarted session. Fix by using the recorded
319 ocsp status of the stored session for the new connection.
321 JH/29 TLS resumption: the key for session lookup in the client now includes
322 more info that a server could potentially use in configuring a TLS
323 session, avoiding oferring mismatching sessions to such a server.
324 Previously only the server IP was used.
326 JH/30 Fix string_copyn() for limit greater than actual string length.
327 Previously the copied amount was the limit, which could result in a
328 overlapping memcpy for newly allocated destination soon after a
329 source string shorter than the limit. Found/investigated by KM.
331 JH/31 Bug 2886: GnuTLS: Do not free the cached creds on transport connection
332 close; it may be needed for a subsequent connection. This caused a
333 SEGV on primary-MX defer. Found/investigated by Gedalya & Andreas.
335 JH/32 Fix CHUNKING for a second message on a connection when the first was
336 rejected. Previously we did not reset the chunking-offered state, and
337 erroneously rejected the BDAT command. Investigation help from
340 JH/33 Fis ${srs_encode ...} to handle an empty sender address, now returning
341 an empty address. Previously the expansion returned an error.
343 HS/01 Bug 2855: Handle a v4mapped sender address given us by a frontending
344 proxy. Previously these were misparsed, leading to paniclog entries.
350 JH/01 Bug 1329: Fix format of Maildir-format filenames to match other mail-
351 related applications. Previously an "H" was used where available info
352 says that "M" should be, so change to match.
354 JH/02 Bug 2587: Fix pam expansion condition. Tainted values are commonly used
355 as arguments, so an implementation trying to copy these into a local
356 buffer was taking a taint-enforcement trap. Fix by using dynamically
357 created buffers. Similar fix for radius expansion condition.
359 JH/03 Bug 2586: Fix listcount expansion operator. Using tainted arguments is
360 reasonable, eg. to count headers. Fix by using dynamically created
361 buffers rather than a local. Do similar fixes for ACL actions "dcc",
362 "log_reject_target", "malware" and "spam"; the arguments are expanded
363 so could be handling tainted values.
365 JH/04 Bug 2590: Fix -bi (newaliases). A previous code rearrangement had
366 broken the (no-op) support for this sendmail command. Restore it
367 to doing nothing, silently, and returning good status.
369 JH/05 Bug 2593: Fix "vacation" in Exim filter. Previously, when a "once"
370 record path was given (or the default used) without a leading directory
371 path, an error occurred on trying to open it. Use the transport's working
374 JH/06 Bug 2594: Change the name used for certificate name checks in the smtp
375 transport. Previously it was the name on the DNS A-record; use instead
376 the head of the CNAME chain leading there (if there is one). This seems
377 to align better with RFC 6125.
379 JH/07 Bug 2597: Fix a resource leak. Using a lookup in obtaining a value for
380 smtp_accept_max_per_host allocated resources which were not released
381 when the limit was exceeded. This eventually crashed the daemon. Fix
382 by adding a release action in that path.
384 JH/08 Bug 2598: Fix verify ACL condition. The options for the condition are
385 expanded; previously using tainted values was rejected. Fix by using
386 dynamically-created buffers.
388 JH/09 Relax restrictions on ACL verify condition needing access to message
389 headers. Previously they were only permitted in data and non-smtp ACLs;
390 permit also mime, dkim, prdr quit and notquit. Applies to header-syntax,
391 not_blind, header_sender and header_names_ascii verification.
393 JH/10 Bug 2603: Fix coding of string copying to only evaluate arguments once.
394 Previously a macro used one argument twice; when called with the
395 argument as an expression having side-effects, incorrect operation
396 resulted. Use an inlineable function.
398 JH/11 Bug 2604: Fix request to cutthrough-deliver when a connection is already
399 held open for a verify callout. Previously this wan not accounted for
400 and a corrupt onward SMTP conversation resulted.
402 JH/12 Bug 2607: Fix the ${srs_encode } expansion to handle quoted local_parts.
403 Previously they were embedded naively in the constructed address; when
404 needed, strip the quoting and quote the entire local_part.
405 Also make the inbound_srs expansion condition handle quoting.
407 JH/13 Fix dsearch "subdir" filter to ignore ".". Previously only ".." was
408 excluded, not matching the documentation.
410 JH/14 Bug 2606: Fix a segfault in sqlite lookups. When no, or a bad, filename
411 was given for the sqlite_dbfile a trap resulted.
413 JH/15 Bug 2620: Fix "spam" ACL condition. Previously, tainted values for the
414 "name" argument resulted in a trap. There is no reason to disallow such;
415 this was a coding error.
417 JH/16 Bug 2615: Fix pause during message reception, on systems that have been
418 suspended/resumed. The Linux CLOCK_MONOTONIC does not account for time
419 spent suspended, ignoring the POSIX definition. Previously we assumed
420 it did and a constant offset from real time could be used as a correction.
421 Change to using the same clock source for the start-of-message and the
422 post-message next-tick-wait. Also change to using CLOCK_BOOTTIME if it
423 exists, just to get a clock slightly more aligned to reality.
425 JH/17 Bug 2295: Fix DKIM signing to always semicolon-terminate. Although the
426 RFC says it is optional some validators care. The missing char was not
427 intended but triggered by a line-wrap alignment. Discovery and fix by
428 Guillaume Outters, hacked on by JH.
430 JH/18 Bug 2617: Fix a taint trap in parse_fix_phrase(). Previously when the
431 name being quoted was tainted a trap would be taken. Fix by using
432 dynamically created buffers. The routine could have been called by a
433 rewrite with the "h" flag, by using the "-F" command-line option, or
434 by using a "name=" option on a control=submission ACL modifier.
436 JH/19 SPF: change the Authentication-Results expansion component to give
437 smtp.helo when the sender domain is empty. Previously it gave
440 JH/20 Bug 2631: ACL dnslist conditions now ignore and log any lookups returns
441 not in 127.0.0.0/8 to help in spotting list domains taken over by a
442 domain-parking registrar.
444 JH/21 Bug 2630: Fix eol-replacement string for the ${readsocket } expansion.
445 Previously when a whitespace character was specified it was not inserted
446 after removing the newline.
448 JH/22 Bug 2265: Force SNI usage for smtp transport DANE'd connections, to be
449 the domain part of the recipient address. This overrides any tls_sni
450 option set, which was previously used.
452 JH/23 Logging: with the +tls_sni log_selector, do not wrap the received SNI
455 JH/24 Bug 2634: Fix a taint trap seen on NetBSD: the testing coded for
456 is_tainted() had an off-by-one error in the overenthusiastic direction.
457 Find and fix by Gavan. Although NetBSD is not a supported platform for
458 4.94 this bug could affect other platforms.
460 PP/01 Fix default prime selection to be consistent.
461 One path used ike23 still, instead of exim.dev.20160529.3; now both
462 execution flows will use the same DH primes (currently
463 exim.dev.20160529.3).
465 JH/25 OpenSSL: Fix back-compatibility behaviour surrounding tls_certificates
466 option in smtp transport, to match the documentation. Previously
467 verification was not being done in some cases where it should have been.
469 JH/26 Bug 2646: fix a memory usage issue in ldap lookups. Previously, when more
470 than one server was defined and depending on the platform memory layout
471 details, an internal consistency trap could be hit while walking the list
474 JH/27 Bug 2648: fix the passing of an authenticator public-name through spool
475 files. The value is used by the authresults expansion item. Previously
476 if this was used in a router or transport, a crash could result.
478 JH/28 Fix spurious logging of select error. Some platforms, notably FreeBSD,
479 have a sufficient incidence of EINTR returns from select that an
480 interaction with other operations done by the main daemon loop exposed
481 a bug in the error-handling. This was benign apart from the log
484 JH/29 Bug 2675: add outgoing-interface I= element to deferred "==" log lines,
485 for consistency with delivered "=>" and failed "**" lines. While we're
486 there, handle PRX and TFO.
488 JH/30 Bug 2677: fix matching of long addresses. Since 4.93 a limit of 256 was
489 applied. This resulted, if any header-line rewrite rules were configured,
490 in a panic-log triggerable by sending a message with a long address in
491 a header. Fix by increasing the arbitrary limit to larger than a single
492 (dewrapped) 5322 header line maximum size.
494 JH/31 The ESMTP option name advertised for the SUPPORT_EARLY_PIPE build option
495 is changed from X_PIPE_CONNECT to PIPE_CONNECT. This is in line with
496 RFC 6648 which deprecates X- options in protocols as a general practice.
497 Changeover between the implementations is handled by the mechanisms
500 JH/32 Bug 2599: fix delay of delivery to a local address where there is also
501 a remote which uses callout/hold. Previously the local was queued.
503 JH/33 Fix a taint trap in the ${listextract } expansion when the source data
506 JH/34 Fix the placement of a multiple-message delivery marker in the delivery
507 log line. The asterisk is now consistently appended to the remote IP
508 (and port, if given), and will also be provided on defer and fail log
509 lines. Previously it could be placed on the local IP if that was being
510 logged, and was only provided on delivery lines.
512 JH/35 Bug 2343: Harden exim_tidydb against corrupt wait- files.
514 JH/36 Bug 2687: Fix interpretation of multiple ^ chars in a plaintext
515 authenticator client_send option. Previously the next char, after a pair
516 was collapsed, was taken verbatim (so ^^^foo became ^^foo; ^^^^foo became
517 ^^\x00foo). Fixed to get ^\x00foo and ^^foo respectively to match the
518 documentation. There is still no way to get a leading ^ immediately
519 after a NUL (ie. for the password of a PLAIN method authenticator.
521 JH/37 Enforce the expected size, for fixed-size records read from hints-DB
522 files. For bad sizes read, delete the record and whine to paniclog.
524 JH/38 When logging an AUTH failure, as server, do not include sensitive
525 information. Previously, the credentials would be included if given
526 as part of the AUTH command line and an ACL denied authentication.
528 JH/39 Bug 2691: fix $local_part_data. When the matching list element
529 referred to a file, bad data was returned. This likely also affected
532 JH/40 The gsasl authenticator now supports caching of the salted password
533 generated by the client-side implementation. This required the addition
534 of a new variable: $auth4.
536 JH/41 Fix daemon SIGHUP on FreeBSD. Previously, a named socket for IPC was
537 left undeleted; the attempt to re-create it then failed - resulting in
538 the usual "SIGHUP tp have daemon reload configuration" to not work.
539 This affected any platform not supporting "abstract" Unix-domain
540 sockets (i.e. not Linux).
542 JH/42 Bug 2693: Harden against a peer which reneges on a 452 "too many
543 recipients" response to RCPT in a later response, with a 250. The
544 previous coding assumed this would not happen, and under PIPELINING
545 would result in both lost and duplicate recipients for a message.
547 JH/43 Bug 2694: Fix weighted distribution of work to multiple spamd servers.
548 Previously the weighting was incorrectly applied. Similar fix for socks
549 proxies. Found and fixed by Heiko Schlichting.
551 JH/44 Bug 2701: Fix list-expansion of dns_ipv4_lookup. Previously, it did
552 not handle sub-lists included using the +namedlist syntax. While
553 investigating, the same found for dns_trust_aa, dns_again_means_nonexist,
554 dnssec_require_domains, dnssec_request_domains, srv_fail_domains,
557 JH/45 Use a (new) separate store pool-pair for DKIM verify working data.
558 Previously the permanent pool was used, so the sore could not be freed.
559 This meant a connection with many messages would use continually-growing
562 JH/46 Use an exponentially-increasing block size when malloc'ing store. Do it
563 per-pool so as not to waste too much space. Previously a constant size
564 was used which resulted in O(n^2) behaviour; now we get O(n log n) making
565 DOS attacks harder. The cost is wasted memory use in the larger blocks.
567 JH/47 Use explicit alloc/free for DNS lookup workspace. This permits using the
568 same space repeatedly, and a smaller process footprint.
570 JH/48 Use a less bogus-looking filename for a temporary used for DH-parameters
571 for GnuTLS. Previously the name started "%s" which, while not a bug,
572 looked as if if might be one.
574 JH/49 Bug 2710: when using SOCKS for additional messages after the first (a
575 "continued connection") make the $proxy_* variables available. Previously
576 the information was not passed across the exec() call for subsequent
577 transport executions. This also mean that the log lines for the
578 messages can show the proxy information.
580 JH/50 Bug 2672: QT elements in log lines, unless disabled, now exclude the
581 receive time. With modern systems the difference is significant.
582 The historical behaviour can be restored by disabling (a new) log_selector
583 "queue_time_exclusive".
585 JH/51 Taint-check ACL line. Previously, only filenames (for out-of-line ACL
586 content) were specifically tested for. Now, also cover expansions
587 resulting in ACL names and inline ACL content.
589 JH/52 Fix ${ip6norm:} operator. Previously, any trailing line text was dropped,
590 making it unusable in complex expressions.
592 JH/53 Bug 2743: fix immediate-delivery via named queue. Previously this would
593 fail with a taint-check on the spoolfile name, and leave the message
596 HS/01 Enforce absolute PID file path name.
598 HS/02 Handle SIGINT as we handle SIGTERM: terminate the Exim process.
600 PP/01 Add a too-many-bad-recipients guard to the default config's RCPT ACL.
602 PP/02 Bug 2643: Correct TLS DH constants.
603 A missing NUL termination in our code-generation tool had led to some
604 incorrect Diffie-Hellman constants in the Exim source.
605 Reported by kylon94, code-gen tool fix by Simon Arlott.
607 PP/03 Impose security length checks on various command-line options.
608 Fixes CVE-2020-SPRSS reported by Qualys.
610 PP/04 Fix Linux security issue CVE-2020-SLCWD and guard against PATH_MAX
611 better. Reported by Qualys.
613 PP/05 Fix security issue CVE-2020-PFPSN and guard against cmdline invoker
614 providing a particularly obnoxious sender full name.
617 PP/06 Fix CVE-2020-28016 (PFPZA): Heap out-of-bounds write in parse_fix_phrase()
619 PP/07 Refuse to allocate too little memory, block negative/zero allocations.
622 PP/08 Change default for recipients_max from unlimited to 50,000.
624 PP/09 Fix security issue with too many recipients on a message (to remove a
625 known security problem if someone does set recipients_max to unlimited,
626 or if local additions add to the recipient list).
627 Fixes CVE-2020-RCPTL reported by Qualys.
629 PP/10 Fix security issue in SMTP verb option parsing
630 Fixes CVE-2020-EXOPT reported by Qualys.
632 PP/11 Fix security issue in BDAT state confusion.
633 Ensure we reset known-good where we know we need to not be reading BDAT
634 data, as a general case fix, and move the places where we switch to BDAT
635 mode until after various protocol state checks.
636 Fixes CVE-2020-BDATA reported by Qualys.
638 HS/03 Die on "/../" in msglog file names
640 QS/01 Creation of (database) files in $spool_dir: only uid=0 or the uid of
641 the Exim runtime user are allowed to create files.
643 QS/02 PID file creation/deletion: only possible if uid=0 or uid is the Exim
646 QS/03 When reading the output from interpreted forward files we do not
647 pass the pipe between the parent and the interpreting process to
648 executed child processes (if any).
650 QS/04 Always die if requested from internal logging, even is logging is
653 JH/54 DMARC: recent versions of the OpenDMARC library appear to have broken
654 the API; compilation noo longer completes with DMARC support included.
655 This affects 1.4.1-1 on Fedora 33 (1.3.2-3 is functional); and has
656 been reported on other platforms.
658 JH/55 TLS: as server, reject connections with ALPN indicating non-smtp use.
660 JH/56 Make the majority of info read from config files readonly, for defence-in-
661 depth against exploits. Suggestion by Qualys.
662 Not supported on Solaris 10.
664 JH/57 Fix control=fakreject for a custom message containing tainted data.
665 Previously this resulted in a log complaint, due to a re-expansion present
666 since fakereject was originally introduced.
668 JH/58 GnuTLS: Fix certextract expansion. If a second modifier after a tag
669 modifier was given, a loop resulted.
671 JH/59 DKIM: Fix small-message verification under TLS with chunking. If a
672 pipelined SMTP command followed the BDAT LAST then it would be
673 incorrectly treated as part of the message body, causing a verification
676 JH/60 Bug 2805: Fix logging of domain-literals in Message_ID: headers. They
677 require looser validation rules than those for 821-level addresses,
678 which only permit IP addresses.
684 JH/01 Avoid costly startup code when not strictly needed. This reduces time
685 for some exim process initialisations. It does mean that the logging
686 of TLS configuration problems is only done for the daemon startup.
688 JH/02 Early-pipelining support code is now included unless disabled in Makefile.
690 JH/03 DKIM verification defaults no long accept sha1 hashes, to conform to
691 RFC 8301. They can still be enabled, using the dkim_verify_hashes main
694 JH/04 Support CHUNKING from an smtp transport using a transport_filter, when
695 DKIM signing is being done. Previously a transport_filter would always
696 disable CHUNKING, falling back to traditional DATA.
698 JH/05 Regard command-line recipients as tainted.
700 JH/06 Bug 340: Remove the daemon pid file on exit, when due to SIGTERM.
702 JH/07 Bug 2489: Fix crash in the "pam" expansion condition. It seems that the
703 PAM library frees one of the arguments given to it, despite the
704 documentation. Therefore a plain malloc must be used.
706 JH/08 Bug 2491: Use tainted buffers for the transport smtp context. Previously
707 on-stack buffers were used, resulting in a taint trap when DSN information
708 copied from a received message was written into the buffer.
710 JH/09 Bug 2493: Harden ARC verify against Outlook, whick has been seen to mix
711 the ordering of its ARC headers. This caused a crash.
713 JH/10 Bug 2492: Use tainted memory for retry record when needed. Previously when
714 a new record was being constructed with information from the peer, a trap
717 JH/11 Bug 2494: Unset the default for dmarc_tld_file. Previously a naiive
718 installation would get error messages from DMARC verify, when it hit the
719 nonexistent file indicated by the default. Distros wanting DMARC enabled
720 should both provide the file and set the option.
721 Also enforce no DMARC verification for command-line sourced messages.
723 JH/12 Fix an uninitialised flag in early-pipelining. Previously connections
724 could, depending on the platform, hang at the STARTTLS response.
726 JH/13 Bug 2498: Reset a counter used for ARC verify before handling another
727 message on a connection. Previously if one message had ARC headers and
728 the following one did not, a crash could result when adding an
729 Authentication-Results: header.
731 JH/14 Bug 2500: Rewind some of the common-coding in string handling between the
732 Exim main code and Exim-related utities. The introduction of taint
733 tracking also did many adjustments to string handling. Since then, eximon
734 frequently terminated with an assert failure.
736 JH/15 When PIPELINING, synch after every hundred or so RCPT commands sent and
737 check for 452 responses. This slightly helps the inefficieny of doing
738 a large alias-expansion into a recipient-limited target. The max_rcpt
739 transport option still applies (and at the current default, will override
740 the new feature). The check is done for either cause of synch, and forces
741 a fast-retry of all 452'd recipients using a new MAIL FROM on the same
742 connection. The new facility is not tunable at this time.
744 JH/16 Fix the variables set by the gsasl authenticator. Previously a pointer to
745 library live data was being used, so the results became garbage. Make
746 copies while it is still usable.
748 JH/17 Logging: when the deliver_time selector ise set, include the DT= field
749 on delivery deferred (==) and failed (**) lines (if a delivery was
750 attemtped). Previously it was only on completion (=>) lines.
752 JH/18 Authentication: the gsasl driver not provides the $authN variables in time
753 for the expansion of the server_scram_iter and server_scram_salt options.
755 WB/01 SPF: DNS lookups for the obsolete SPF RR type done by the libspf2 library
756 are now specifically given a NO_DATA response without hitting the system
757 resolver. The library goes on to do the now-standard TXT lookup.
758 Use of dnsdb lookups is not affected.
760 JH/19 Bug 2507: Modules: on handling a dynamic-module (lookups) open failure,
761 only retrieve the errormessage once. Previously two calls to dlerror()
762 were used, and the second one (for mainlog/paniclog) retrieved null
765 JH/20 Taint checking: disallow use of tainted data for
766 - the appendfile transport file and directory options
767 - the pipe transport command
768 - the autoreply transport file, log and once options
769 - file names used by the redirect router (including filter files)
771 - paths used by single-key lookups
772 Previously this was permitted.
774 JH/21 Bug 2501: Fix init call in the heimdal authenticator. Previously it
775 adjusted the size of a major service buffer; this failed because the
776 buffer was in use at the time. Change to a compile-time increase in the
777 buffer size, when this authenticator is compiled into exim.
779 JH/22 Taint-checking: move to safe-mode taint checking on all platforms. The
780 previous fast-mode was untenable in the face of glibs using mmap to
781 support larger malloc requests.
783 PP/01 Update the openssl_options possible values through OpenSSL 1.1.1c.
784 New values supported, if defined on system where compiled:
785 allow_no_dhe_kex, cryptopro_tlsext_bug, enable_middlebox_compat,
786 no_anti_replay, no_encrypt_then_mac, prioritize_chacha, tlsext_padding
788 JH/23 Performance improvement in the initial phase of a two-pass queue run. By
789 running a limited number of proceses in parallel, a benefit is gained. The
790 amount varies with the platform hardware and load. The use of the option
791 queue_run_in_order means we cannot do this, as ordering becomes
794 JH/24 Bug 2524: fix the cyrus_sasl auth driver gssapi usage. A previous fix
795 had introduced a string-copy (for ensuring NUL-termination) which was not
796 appropriate for that case, which can include embedded NUL bytes in the
797 block of data. Investigation showed the copy to actually be needless, the
798 data being length-specified.
800 JH/25 Fix use of concurrent TLS connections under GnuTLS. When a callout was
801 done during a receiving connection, and both used TLS, global info was
802 used rather than per-connection info for tracking the state of data
803 queued for transmission. This could result in a connection hang.
805 JH/26 Fix use of the SIZE parameter on MAIL commands, on continued connections.
806 Previously, when delivering serveral messages down a single connection
807 only the first would provide a SIZE. This was due to the size information
808 not being properly tracked.
810 JH/27 Bug 2530: When operating in a timezone with sub-minute offset, such as
811 TAI (at 37 seconds currently), pretend to be in UTC for time-related
812 expansion and logging. Previously, spurious values such as a future
813 minute could be seen.
815 JH/28 Bug 2533: Fix expansion of ${tr } item. When called in some situations
816 it could crash from a null-deref. This could also affect the
817 ${addresses: } operator and ${readsock } item.
819 JH/29 Bug 2537: Fix $mime_part_count. When a single connection had a non-mime
820 message following a mime one, the variable was not reset.
822 JH/30 When an pipelined-connect fails at the first response, assume incorrect
823 cached capability (perhaps the peer reneged?) and immediately retry in
826 JH/31 Fix spurious detection of timeout while writing to transport filter.
828 JH/32 Bug 2541: Fix segfault on bad cmdline -f (sender) argument. Previously
829 an attempt to copy the string was made before checking it.
831 JH/33 Fix the dsearch lookup to return an untainted result. Previously the
832 taint of the lookup key was maintained; we now regard the presence in the
833 filesystem as sufficient validation.
835 JH/34 Fix the readsocket expansion to not segfault when an empty "options"
836 argument is supplied.
838 JH/35 The dsearch lookup now requires that the directory is an absolute path.
839 Previously this was not checked, and nonempty relative paths made an
840 access under Exim's current working directory.
842 JH/36 Bug 2554: Fix msg:defer event for the hosts_max_try_hardlimit case.
843 Previously no event was raised.
845 JH/37 Bug 2552: Fix the check on spool space during reception to use the SIZE
846 parameter supplied by the sender MAIL FROM command. Previously it was
847 ignored, and only the check_spool_space option value for the required
850 JH/38 Fix $dkim_key_length. This should, after a DKIM verification, present
851 the size of the signing public-key. Previously it was instead giving
852 the size of the signature hash.
854 JH/39 DKIM verification: the RFC 8301 restriction on sizes of RSA keys is now
855 the default. See the (new) dkim_verify_min_keysizes option.
857 JH/40 Fix a memory-handling bug: when a connection carried multiple messages
858 and an ACL use a lookup for checking either the local_part or domain,
859 stale data could be accessed. Ensure that variable references are
860 dropped between messages.
862 JH/41 Bug 2571: Fix SPA authenticator. Running as a server, an offset supplied
863 by the client was not checked as pointing within response data before
864 being used. A malicious client could thus cause an out-of-bounds read and
865 possibly gain authentication. Fix by adding the check.
867 JH/42 Internationalisation: change the default for downconversion in the smtp
868 transport to be "if needed". Previously it was "as previously set" for
869 the message, which usually meant "if needed" for message-submission but
870 "no" for everything else. However, MTAs have been seen using SMTPUTF8
871 even when the envelope addresses did not need it, resulting in forwarding
872 failures to non-supporting MTAs. A downconvert in such cases will be
873 a no-op on the addresses, merely dropping the use of SMTPUTF8 by the
874 transport. The change does mean that addresses needing conversion will
875 be converted when previously a delivery failure would occur.
877 JH/43 Fix possible long line in DSN. Previously when a very long SMTP error
878 response was received it would be used unchecked in a fail-DSN, violating
879 standards on line-length limits. Truncate if needed.
881 HS/01 Remove parameters of the link to www.open-spf.org. The linked form
882 doesn't work. (Additionally add a new main config option to configure the
889 JH/01 OpenSSL: With debug enabled output keying information sufficient, server
890 side, to decode a TLS 1.3 packet capture.
892 JH/02 OpenSSL: Suppress the sending of (stateful) TLS1.3 session tickets.
893 Previously the default library behaviour applied, sending two, each in
896 JH/03 Debug output for ACL now gives the config file name and line number for
899 JH/04 The default received_header_text now uses the RFC 8314 tls cipher clause.
901 JH/05 DKIM: ensure that dkim_domain elements are lowercased before use.
903 JH/06 Fix buggy handling of autoreply bounce_return_size_limit, and a possible
904 buffer overrun for (non-chunking) other transports.
906 JH/07 GnuTLS: Our use of late (post-handshake) certificate verification, under
907 TLS1.3, means that a server rejecting a client certificate is not visible
908 to the client until the first read of encrypted data (typically the
909 response to EHLO). Add detection for that case and treat it as a failed
910 TLS connection attempt, so that the normal retry-in-clear can work (if
911 suitably configured).
913 JB/01 Bug 2375: fix expansions of 822 addresses having comments in local-part
914 and/or domain. Found and fixed by Jason Betts.
916 JH/08 Add hardening against SRV & TLSA lookups the hit CNAMEs (a nonvalid
917 configuration). If a CNAME target was not a wellformed name pattern, a
920 JH/09 Logging: Fix initial listening-on line for multiple ports for an IP when
921 the OS reports them interleaved with other addresses.
923 JH/10 OpenSSL: Fix aggregation of messages. Previously, when PIPELINING was
924 used both for input and for a verify callout, both encrypted, SMTP
925 responses being sent by the server could be lost. This resulted in
926 dropped connections and sometimes bounces generated by a peer sending
929 JH/11 Harden plaintext authenticator against a badly misconfigured client-send
930 string. Previously it was possible to cause undefined behaviour in a
931 library routine (usually a crash). Found by "zerons".
933 JH/12 Bug 2384: fix "-bP smtp_receive_timeout". Previously it returned no
936 JH/13 Bug 2386: Fix builds with Dane under LibreSSL 2.9.0 onward. Some old
937 API was removed, so update to use the newer ones.
939 JH/14 Bug 1891: Close the log file if receiving a non-smtp message, without
940 any timeout set, is taking a long time. Previously we would hang on to a
941 rotated logfile "forever" if the input was arriving with long gaps
942 (a previous attempt to fix addressed lack, for a long time, of initial
945 HS/01 Bug 2390: Use message_id for tempfile creation to avoid races in a
946 shared (NFS) environment. The length of the tempfile name is now
947 4 + 16 ("hdr.$message_exim_id") which might break on file
948 systems which restrict the file name length to lower values.
951 HS/02 Bug 2390: Use message_id for tempfile creation to avoid races in a
952 shared (NFS) environment.
954 HS/03 Bug 2392: exigrep does case sensitive *option* processing (as it
955 did for all versions <4.90). Notably -M, -m, --invert, -I may be
958 JH/15 Use unsigned when creating bitmasks in macros, to avoid build errors
959 on some platforms for bit 31.
961 JH/16 GnuTLS: rework ciphersuite strings under recent library versions. Thanks
962 to changes apparently associated with TLS1.3 handling some of the APIs
963 previously used were either nonfunctional or inappropriate. Strings
964 like TLS1.3:ECDHE_SECP256R1__RSA_PSS_RSAE_SHA256__AES_256_GCM__AEAD:256
965 and TLS1.2:ECDHE_SECP256R1__RSA_SHA256__AES_128_CBC__SHA256:128 replace
966 the previous TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256 .
967 This affects log line X= elements, the $tls_{in,out}_cipher variables,
968 and the use of specific cipher names in the encrypted= ACL condition.
970 JH/17 OpenSSL: the default openssl_options now disables ssl_v3.
972 JH/18 GnuTLS: fix $tls_out_ocsp under hosts_request_ocsp. Previously the
973 verification result was not updated unless hosts_require_ocsp applied.
975 JH/19 Bug 2398: fix listing of a named-queue. Previously, even with the option
976 queue_list_requires_admin set to false, non-admin users were denied the
979 JH/20 Bug 2389: fix server advertising of usable certificates, under GnuTLS in
980 directory-of-certs mode. Previously they were advertised despite the
983 JH/21 The smtp transport option "hosts_noproxy_tls" is now unset by default.
984 A single TCP connection by a client will now hold a TLS connection open
985 for multiple message deliveries, by default. Previously the default was to
988 JH/22 The smtp transport option "hosts_try_dane" now enables all hosts by
989 default. If built with the facility, DANE will be used. The facility
990 SUPPORT_DANE is now enabled in the prototype build Makefile "EDITME".
992 JH/23 The build default is now for TLS to be included; the SUPPORT_TLS define
993 is replaced with DISABLE_TLS. Either USE_GNUTLS or (the new) USE_OPENSSL
994 must be defined and you must still, unless you define DISABLE_TLS, manage
995 the the include-dir and library-file requirements that go with that
996 choice. Non-TLS builds are still supported.
998 JH/24 Fix duplicated logging of peer name/address, on a transport connection-
1001 JH/25 The smtp transport option "hosts_try_fastopen" now enables all hosts by
1002 default. If the platform supports and has the facility enabled, it will
1003 be requested on all coneections.
1005 JH/26 The PIPE_CONNECT facility is promoted from experimental status and is now
1006 controlled by the build-time option SUPPORT_PIPE_CONNECT.
1008 PP/01 Unbreak heimdal_gssapi, broken in 4.92.
1010 JH/27 Bug 2404: Use the main-section configuration option "dsn_from" for
1011 success-DSN messages. Previously the From: header was always the default
1012 one for these; the option was ignored.
1014 JH/28 Fix the timeout on smtp response to apply to the whole response.
1015 Previously it was reset for every read, so a teergrubing peer sending
1016 single bytes within the time limit could extend the connection for a
1017 long time. Credit to Qualsys Security Advisory Team for the discovery.
1019 JH/29 Fix DSN Final-Recipient: field. Previously it was the post-routing
1020 delivery address, which leaked information of the results of local
1021 forwarding. Change to the original envelope recipient address, per
1024 JH/30 Bug 2411: Fix DSN generation when RFC 3461 failure notification is
1025 requested. Previously not bounce was generated and a log entry of
1026 error ignored was made.
1028 JH/31 Avoid re-expansion in ${sort } expansion. (CVE-2019-13917)
1030 JH/32 Introduce a general tainting mechanism for values read from the input
1031 channel, and values derived from them. Refuse to expand any tainted
1032 values, to catch one form of exploit.
1034 JH/33 Bug 2413: Fix dkim_strict option. Previously the expansion result
1035 was unused and the unexpanded text used for the test. Found and
1036 fixed by Ruben Jenster.
1038 JH/34 Fix crash after TLS shutdown. When the TCP/SMTP channel was left open,
1039 an attempt to use a TLS library read routine dereffed a nul pointer,
1042 JH/35 Bug 2409: filter out-of-spec chars from callout response before using
1043 them in our smtp response.
1045 JH/36 Have the general router option retry_use_local_part default to true when
1046 any of the restrictive preconditions are set (to anything). Previously it
1047 was only for check_local user. The change removes one item of manual
1048 configuration which is required for proper retries when a remote router
1049 handles a subset of addresses for a domain.
1051 JH/37 Appendfile: when evaluating quota use (non-quota_size_regex) take the file
1052 link count into consideration.
1054 HS/04 Fix handling of very log lines in -H files. If a -<key> <value> line
1055 caused the extension of big_buffer, the following lines were ignored.
1057 JH/38 Bug 1395: Teach the DNS negative-cache about TTL value from the SOA in
1058 accordance with RFC 2308. Previously there was no expiry, so a longlived
1059 receive process (eg. due to ACL delays) versus a short SOA value could
1062 HS/05 Handle trailing backslash gracefully. (CVE-2019-15846)
1064 JH/39 Promote DMARC support to mainline.
1066 JH/40 Bug 2452: Add a References: header to DSNs.
1068 JH/41 With GnuTLS 3.6.0 (and later) do not attempt to manage Diffie-Hellman
1069 parameters. The relevant library call is documented as "Deprecated: This
1070 function is unnecessary and discouraged on GnuTLS 3.6.0 or later. Since
1071 3.6.0, DH parameters are negotiated following RFC7919."
1073 HS/06 Change the default of dnssec_request_domains to "*"
1075 JH/42 Bug 2545: Fix CHUNKING for all RCPT commands rejected. Previously we
1076 carried on and emitted a BDAT command, even when PIPELINING was not
1079 JH/43 Bug 2465: Fix taint-handling in dsearch lookup. Previously a nontainted
1080 buffer was used for the filename, resulting in a trap when tainted
1081 arguments (eg. $domain) were used.
1083 JH/44 With OpenSSL 1.1.1 (onwards) disable renegotiation for TLS1.2 and below;
1084 recommended to avoid a possible server-load attack. The feature can be
1085 re-enabled via the openssl_options main cofiguration option.
1087 JH/45 local_scan API: documented the current smtp_printf() call. This changed
1088 for version 4.90 - adding a "more data" boolean to the arguments.
1089 Bumped the ABI version number also, this having been missed previously;
1090 release versions 4.90 to 4.92.3 inclusive were effectively broken in
1091 respect of usage of smtp_printf() by either local_scan code or libraries
1092 accessed via the ${dlfunc } expansion item. Both will need coding
1093 adjustment for any calls to smtp_printf() to match the new function
1094 signature; a FALSE value for the new argument is always safe.
1096 JH/46 FreeBSD: fix use of the sendfile() syscall. The shim was not updating
1097 the file-offset (which the Linux syscall does, and exim expects); this
1098 resulted in an indefinite loop.
1100 JH/47 ARC: fix crash in signing, triggered when a configuration error failed
1101 to do ARC verification. The Authentication-Results: header line added
1102 by the configuration then had no ARC item.
1104 JH/48 Bug 2784: fix shutdown=no in the ${readsocket) expansion item. Previously
1105 an incorrect mode was used for reading the result, resulting in it being
1112 JH/01 Remove code calling the customisable local_scan function, unless a new
1113 definition "HAVE_LOCAL_SCAN=yes" is present in the Local/Makefile.
1115 JH/02 Bug 1007: Avoid doing logging from signal-handlers, as that can result in
1116 non-signal-safe functions being used.
1118 JH/03 Bug 2269: When presented with a received message having a stupidly large
1119 number of DKIM-Signature headers, disable DKIM verification to avoid
1120 a resource-consumption attack. The limit is set at twenty.
1122 JH/04 Add variables $arc_domains, $arc_oldest_pass for ARC verify. Fix the
1123 report of oldest_pass in ${authres } in consequence, and separate out
1124 some descriptions of reasons for verification fail.
1126 JH/05 Bug 2273: Cutthrough delivery left a window where the received messsage
1127 files in the spool were present and unlocked. A queue-runner could spot
1128 them, resulting in a duplicate delivery. Fix that by doing the unlock
1129 after the unlink. Investigation by Tim Stewart. Take the opportunity to
1130 add more error-checking on spoolfile handling while that code is being
1133 PP/01 Refuse to open a spool data file (*-D) if it's a symlink.
1134 No known attacks, no CVE, this is defensive hardening.
1136 JH/06 Bug 2275: The MIME ACL unlocked the received message files early, and
1137 a queue-runner could start a delivery while other operations were ongoing.
1138 Cutthrough delivery was a common victim, resulting in duplicate delivery.
1139 Found and investigated by Tim Stewart. Fix by using the open message data
1140 file handle rather than opening another, and not locally closing it (which
1141 releases a lock) for that case, while creating the temporary .eml format
1142 file for the MIME ACL. Also applies to "regex" and "spam" ACL conditions.
1144 JH/07 Bug 177: Make a random-recipient callout success visible in ACL, by setting
1145 $sender_verify_failure/$recipient_verify_failure to "random".
1147 JH/08 When generating a selfsigned cert, use serial number 1 since zero is not
1150 JH/09 Bug 2274: Fix logging of cmdline args when starting in an unlinked cwd.
1151 Previously this would segfault.
1153 JH/10 Fix ARC signing for case when DKIM signing failed. Previously this would
1156 JH/11 Bug 2264: Exim now only follows CNAME chains one step by default. We'd
1157 like zero, since the resolver should be doing this for us, But we need one
1158 as a CNAME but no MX presence gets the CNAME returned; we need to check
1159 that doesn't point to an MX to declare it "no MX returned" rather than
1160 "error, loop". A new main option is added so the older capability of
1161 following some limited number of chain links is maintained.
1163 JH/12 Add client-ip info to non-pass iprev ${authres } lines.
1165 JH/13 For receent Openssl versions (1.1 onward) use modern generic protocol
1166 methods. These should support TLS 1.3; they arrived with TLS 1.3 and the
1167 now-deprecated earlier definitions used only specified the range up to TLS
1168 1.2 (in the older-version library docs).
1170 JH/14 Bug 2284: Fix DKIM signing for body lines starting with a pair of dots.
1172 JH/15 Rework TLS client-side context management. Stop using a global, and
1173 explicitly pass a context around. This enables future use of TLS for
1174 connections to service-daemons (eg. malware scanning) while a client smtp
1175 connection is using TLS; with cutthrough connections this is quite likely.
1177 JH/16 Fix ARC verification to do AS checks in reverse order.
1179 JH/17 Support a "tls" option on the ${readsocket } expansion item.
1181 JH/18 Bug 2287: Fix the protocol name (eg utf8esmtp) for multiple messages
1182 using the SMTPUTF8 option on their MAIL FROM commands, in one connection.
1183 Previously the "utf8" would be re-prepended for every additional message.
1185 JH/19 Reject MAIL FROM commands with SMTPUTF8 when the facility was not advertised.
1186 Previously thery were accepted, resulting in issues when attempting to
1187 forward messages to a non-supporting MTA.
1189 PP/02 Let -n work with printing macros too, not just options.
1191 JH/20 Bug 2296: Fix cutthrough for >1 address redirection. Previously only
1192 one parent address was copied, and bogus data was used at delivery-logging
1193 time. Either a crash (after delivery) or bogus log data could result.
1194 Discovery and analysis by Tim Stewart.
1196 PP/03 Make ${utf8clean:} expansion operator detect incomplete final character.
1197 Previously if the string ended mid-character, we did not insert the
1198 promised '?' replacement.
1200 PP/04 Documentation: current string operators work on bytes, not codepoints.
1202 JH/21 Change as many as possible of the global flags into one-bit bitfields; these
1203 should pack well giving a smaller memory footprint so better caching and
1204 therefore performance. Group the declarations where this can't be done so
1205 that the byte-sized flag variables are not interspersed among pointer
1206 variables, giving a better chance of good packing by the compiler.
1208 JH/22 Bug 1896: Fix the envelope from for DMARC forensic reports to be possibly
1209 non-null, to avoid issues with sites running BATV. Previously reports were
1210 sent with an empty envelope sender so looked like bounces.
1212 JH/23 Bug 2318: Fix the noerror command within filters. It wasn't working.
1213 The ignore_error flag wasn't being returned from the filter subprocess so
1214 was not set for later routers. Investigation and fix by Matthias Kurz.
1216 JH/24 Bug 2310: Raise a msg:fail:internal event for each undelivered recipient,
1217 and a msg:complete for the whole, when a message is manually removed using
1218 -Mrm. Developement by Matthias Kurz, hacked on by JH.
1220 JH/25 Avoid fixed-size buffers for pathnames in DB access. This required using
1221 a "Gnu special" function, asprintf() in the DB utility binary builds; I
1222 hope that is portable enough.
1224 JH/26 Bug 2311: Fix DANE-TA verification under GnuTLS. Previously it was also
1225 requiring a known-CA anchor certificate; make it now rely entirely on the
1226 TLSA as an anchor. Checking the name on the leaf cert against the name
1227 on the A-record for the host is still done for TA (but not for EE mode).
1229 JH/27 Fix logging of proxy address. Previously, a pointless "PRX=[]:0" would be
1230 included in delivery lines for non-proxied connections, when compiled with
1231 SUPPORT_SOCKS and running with proxy logging enabled.
1233 JH/28 Bug 2314: Fire msg:fail:delivery event even when error is being ignored.
1234 Developement by Matthias Kurz, tweaked by JH. While in that bit of code,
1235 move the existing event to fire before the normal logging of message
1236 failure so that custom logging is bracketed by normal logging.
1238 JH/29 Bug 2322: A "fail" command in a non-system filter (file) now fires the
1239 msg:fail:internal event. Developement by Matthias Kurz.
1241 JH/30 Bug 2329: Increase buffer size used for dns lookup from 2k, which was
1242 far too small for todays use of crypto signatures stored there. Go all
1243 the way to the max DNS message size of 64kB, even though this might be
1244 overmuch for IOT constrained device use.
1246 JH/31 Fix a bad use of a copy function, which could be used to pointlessly
1247 copy a string over itself. The library routine is documented as not
1248 supporting overlapping copies, and on MacOS it actually raised a SIGABRT.
1250 JH/32 For main options check_spool_space and check_inode_space, where the
1251 platform supports 64b integers, support more than the previous 2^31 kB
1252 (i.e. more than 2 TB). Accept E, P and T multipliers in addition to
1253 the previous G, M, k.
1255 JH/33 Bug 2338: Fix the cyrus-sasl authenticator to fill in the
1256 $authenticated_fail_id variable on authentication failure. Previously
1259 JH/34 Increase RSA keysize of autogen selfsign cert from 1024 to 2048. RHEL 8.0
1260 OpenSSL didn't want to use such a weak key. Do for GnuTLS also, and for
1261 more-modern GnuTLS move from GNUTLS_SEC_PARAM_LOW to
1262 GNUTLS_SEC_PARAM_MEDIUM.
1264 JH/35 OpenSSL: fail the handshake when SNI processing hits a problem, server
1265 side. Previously we would continue as if no SNI had been received.
1267 JH/36 Harden the handling of string-lists. When a list consisted of a sole
1268 "<" character, which should be a list-separator specification, we walked
1269 off past the nul-terimation.
1271 JH/37 Bug 2341: Send "message delayed" warning MDNs (restricted to external
1272 causes) even when the retry time is not yet met. Previously they were
1273 not, meaning that when (say) an account was over-quota and temp-rejecting,
1274 and multiple senders' messages were queued, only one sender would get
1275 notified on each configured delay_warning cycle.
1277 JH/38 Bug 2351: Log failures to extract envelope addresses from message headers.
1279 JH/39 OpenSSL: clear the error stack after an SSL_accept(). With anon-auth
1280 cipher-suites, an error can be left on the stack even for a succeeding
1281 accept; this results in impossible error messages when a later operation
1284 AM/01 Bug 2359: GnuTLS: repeat lowlevel read and write operations while they
1285 return error codes indicating retry. Under TLS1.3 this becomes required.
1287 JH/40 Fix the feature-cache refresh for EXPERIMENTAL_PIPE_CONNECT. Previously
1288 it only wrote the new authenticators, resulting in a lack of tracking of
1289 peer changes of ESMTP extensions until the next cache flush.
1291 JH/41 Fix the loop reading a message header line to check for integer overflow,
1292 and more-often against header_maxsize. Previously a crafted message could
1293 induce a crash of the recive process; now the message is cleanly rejected.
1295 JH/42 Bug 2366: Fix the behaviour of the dkim_verify_signers option. It had
1296 been totally disabled for all of 4.91. Discovery and fix by "Mad Alex".
1302 GF/01 DEFER rather than ERROR on redis cluster MOVED response.
1303 When redis_servers is set to a list of > 1 element, and the Redis servers
1304 in that list are in cluster configuration, convert the REDIS_REPLY_ERROR
1305 case of MOVED into a DEFER case instead, thus moving the query onto the
1306 next server in the list. For a cluster of N elements, all N servers must
1307 be defined in redis_servers.
1309 GF/02 Catch and remove uninitialized value warning in exiqsumm
1310 Check for existence of @ARGV before looking at $ARGV[0]
1312 JH/01 Replace the store_release() internal interface with store_newblock(),
1313 which internalises the check required to safely use the old one, plus
1314 the allocate and data copy operations duplicated in both (!) of the
1315 extant use locations.
1317 JH/02 Disallow '/' characters in queue names specified for the "queue=" ACL
1318 modifier. This matches the restriction on the commandline.
1320 JH/03 Fix pgsql lookup for multiple result-tuples with a single column.
1321 Previously only the last row was returned.
1323 JH/04 Bug 2217: Tighten up the parsing of DKIM signature headers. Previously
1324 we assumed that tags in the header were well-formed, and parsed the
1325 element content after inspecting only the first char of the tag.
1326 Assumptions at that stage could crash the receive process on malformed
1329 JH/05 Bug 2215: Fix crash associated with dnsdb lookup done from DKIM ACL.
1330 While running the DKIM ACL we operate on the Permanent memory pool so that
1331 variables created with "set" persist to the DATA ACL. Also (at any time)
1332 DNS lookups that fail create cache records using the Permanent pool. But
1333 expansions release any allocations made on the current pool - so a dnsdb
1334 lookup expansion done in the DKIM ACL releases the memory used for the
1335 DNS negative-cache, and bad things result. Solution is to switch to the
1336 Main pool for expansions.
1337 While we're in that code, add checks on the DNS cache during store_reset,
1338 active in the testsuite.
1339 Problem spotted, and debugging aided, by Wolfgang Breyha.
1341 JH/06 Fix issue with continued-connections when the DNS shifts unreliably.
1342 When none of the hosts presented to a transport match an already-open
1343 connection, close it and proceed with the list. Previously we would
1344 queue the message. Spotted by Lena with Yahoo, probably involving
1347 JH/07 Bug 2214: Fix SMTP responses resulting from non-accept result of MIME ACL.
1348 Previously a spurious "250 OK id=" response was appended to the proper
1351 JH/08 The "support for" informational output now, which built with Content
1352 Scanning support, has a line for the malware scanner interfaces compiled
1353 in. Interface can be individually included or not at build time.
1355 JH/09 The "aveserver", "kavdaemon" and "mksd" interfaces are now not included
1356 by the template makefile "src/EDITME". The "STREAM" support for an older
1357 ClamAV interface method is removed.
1359 JH/10 Bug 2223: Fix mysql lookup returns for the no-data case (when the number of
1360 rows affected is given instead).
1362 JH/11 The runtime Berkeley DB library version is now additionally output by
1363 "exim -d -bV". Previously only the compile-time version was shown.
1365 JH/12 Bug 2230: Fix cutthrough routing for nonfirst messages in an initiating
1366 SMTP connection. Previously, when one had more recipients than the
1367 first, an abortive onward connection was made. Move to full support for
1368 multiple onward connections in sequence, handling cutthrough connection
1369 for all multi-message initiating connections.
1371 JH/13 Bug 2229: Fix cutthrough routing for nonstandard port numbers defined by
1372 routers. Previously, a multi-recipient message would fail to match the
1373 onward-connection opened for the first recipient, and cause its closure.
1375 JH/14 Bug 2174: A timeout on connect for a callout was also erroneously seen as
1376 a timeout on read on a GnuTLS initiating connection, resulting in the
1377 initiating connection being dropped. This mattered most when the callout
1378 was marked defer_ok. Fix to keep the two timeout-detection methods
1381 JH/15 Relax results from ACL control request to enable cutthrough, in
1382 unsupported situations, from error to silently (except under debug)
1383 ignoring. This covers use with PRDR, frozen messages, queue-only and
1386 HS/01 Fix Buffer overflow in base64d() (CVE-2018-6789)
1388 JH/16 Fix bug in DKIM verify: a buffer overflow could corrupt the malloc
1389 metadata, resulting in a crash in free().
1391 PP/01 Fix broken Heimdal GSSAPI authenticator integration.
1392 Broken in f2ed27cf5, missing an equals sign for specified-initialisers.
1393 Broken also in d185889f4, with init system revamp.
1395 JH/17 Bug 2113: Fix conversation closedown with the Avast malware scanner.
1396 Previously we abruptly closed the connection after reading a malware-
1397 found indication; now we go on to read the "scan ok" response line,
1400 JH/18 Bug 2239: Enforce non-usability of control=utf8_downconvert in the mail
1401 ACL. Previously, a crash would result.
1403 JH/19 Speed up macro lookups during configuration file read, by skipping non-
1404 macro text after a replacement (previously it was only once per line) and
1405 by skipping builtin macros when searching for an uppercase lead character.
1407 JH/20 DANE support moved from Experimental to mainline. The Makefile control
1408 for the build is renamed.
1410 JH/21 Fix memory leak during multi-message connections using STARTTLS. A buffer
1411 was allocated for every new TLS startup, meaning one per message. Fix
1412 by only allocating once (OpenSSL) or freeing on TLS-close (GnuTLS).
1414 JH/22 Bug 2236: When a DKIM verification result is overridden by ACL, DMARC
1415 reported the original. Fix to report (as far as possible) the ACL
1416 result replacing the original.
1418 JH/23 Fix memory leak during multi-message connections using STARTTLS under
1419 OpenSSL. Certificate information is loaded for every new TLS startup,
1420 and the resources needed to be freed.
1422 JH/24 Bug 2242: Fix exim_dbmbuild to permit directoryless filenames.
1424 JH/25 Fix utf8_downconvert propagation through a redirect router. Previously it
1427 JH/26 Bug 2253: For logging delivery lines under PRDR, append the overall
1428 DATA response info to the (existing) per-recipient response info for
1429 the "C=" log element. It can have useful tracking info from the
1430 destination system. Patch from Simon Arlott.
1432 JH/27 Bug 2251: Fix ldap lookups that return a single attribute having zero-
1433 length value. Previously this would segfault.
1435 HS/02 Support Avast multiline protoocol, this allows passing flags to
1436 newer versions of the scanner.
1438 JH/28 Ensure that variables possibly set during message acceptance are marked
1439 dead before release of memory in the daemon loop. This stops complaints
1440 about them when the debug_store option is enabled. Discovered specifically
1441 for sender_rate_period, but applies to a whole set of variables.
1442 Do the same for the queue-runner and queue-list loops, for variables set
1443 from spool message files. Do the same for the SMTP per-message loop, for
1444 certain variables indirectly set in ACL operations.
1446 JH/29 Bug 2250: Fix a longstanding bug in heavily-pipelined SMTP input (such
1447 as a multi-recipient message from a mailinglist manager). The coding had
1448 an arbitrary cutoff number of characters while checking for more input;
1449 enforced by writing a NUL into the buffer. This corrupted long / fast
1450 input. The problem was exposed more widely when more pipelineing of SMTP
1451 responses was introduced, and one Exim system was feeding another.
1452 The symptom is log complaints of SMTP syntax error (NUL chars) on the
1453 receiving system, and refused recipients seen by the sending system
1454 (propating to people being dropped from mailing lists).
1455 Discovered and pinpointed by David Carter.
1457 JH/30 The (EXPERIMENTAL_DMARC) variable $dmarc_ar_header is withdrawn, being
1458 replaced by the ${authresults } expansion.
1460 JH/31 Bug 2257: Fix pipe transport to not use a socket-only syscall.
1462 HS/03 Set a handler for SIGTERM and call exit(3) if running as PID 1. This
1463 allows proper process termination in container environments.
1465 JH/32 Bug 2258: Fix spool_wireformat in combination with LMTP transport.
1466 Previously the "final dot" had a newline after it; ensure it is CR,LF.
1468 JH/33 SPF: remove support for the "spf" ACL condition outcome values "err_temp"
1469 and "err_perm", deprecated since 4.83 when the RFC-defined words
1470 "temperror" and "permerror" were introduced.
1472 JH/34 Re-introduce enforcement of no cutthrough delivery on transports having
1473 transport-filters or DKIM-signing. The restriction was lost in the
1474 consolidation of verify-callout and delivery SMTP handling.
1475 Extend the restriction to also cover ARC-signing.
1477 JH/35 Cutthrough: for a final-dot response timeout (and nonunderstood responses)
1478 in defer=pass mode supply a 450 to the initiator. Previously the message
1481 PP/02 DANE: add dane_require_tls_ciphers SMTP Transport option; if unset,
1482 tls_require_ciphers is used as before.
1484 HS/03 Malware Avast: Better match the Avast multiline protocol. Add
1485 "pass_unscanned". Only tmpfails from the scanner are written to
1486 the paniclog, as they may require admin intervention (permission
1487 denied, license issues). Other scanner errors (like decompression
1488 bombs) do not cause a paniclog entry.
1490 JH/36 Fix reinitialisation of DKIM logging variable between messages.
1491 Previously it was possible to log spurious information in receive log
1494 JH/37 Bug 2255: Revert the disable of the OpenSSL session caching. This
1495 triggered odd behaviour from Outlook Express clients.
1497 PP/03 Add util/renew-opendmarc-tlds.sh script for safe renewal of public
1500 JH/38 DKIM: accept Ed25519 pubkeys in SubjectPublicKeyInfo-wrapped form,
1501 since the IETF WG has not yet settled on that versus the original
1502 "bare" representation.
1504 JH/39 Fix syslog logging for syslog_timestamp=no and log_selector +millisec.
1505 Previously the millisecond value corrupted the output.
1506 Fix also for syslog_pid=no and log_selector +pid, for which the pid
1507 corrupted the output.
1513 JH/01 Rework error string handling in TLS interface so that the caller in
1514 more cases is responsible for logging. This permits library-sourced
1515 string to be attached to addresses during delivery, and collapses
1516 pairs of long lines into single ones.
1518 PP/01 Allow PKG_CONFIG_PATH to be set in Local/Makefile and use it correctly
1519 during configuration. Wildcards are allowed and expanded.
1521 JH/02 Rework error string handling in DKIM to pass more info back to callers.
1522 This permits better logging.
1524 JH/03 Rework the transport continued-connection mechanism: when TLS is active,
1525 do not close it down and have the child transport start it up again on
1526 the passed-on TCP connection. Instead, proxy the child (and any
1527 subsequent ones) for TLS via a unix-domain socket channel. Logging is
1528 affected: the continued delivery log lines do not have any DNSSEC, TLS
1529 Certificate or OCSP information. TLS cipher information is still logged.
1531 JH/04 Shorten the log line for daemon startup by collapsing adjacent sets of
1532 identical IP addresses on different listening ports. Will also affect
1535 PP/02 Bug 2070: uClibc defines __GLIBC__ without providing glibc headers;
1536 add noisy ifdef guards to special-case this sillyness.
1537 Patch from Bernd Kuhls.
1539 JH/05 Tighten up the checking in isip4 (et al): dotted-quad components larger
1540 than 255 are no longer allowed.
1542 JH/06 Default openssl_options to include +no_ticket, to reduce load on peers.
1543 Disable the session-cache too, which might reduce our load. Since we
1544 currrectly use a new context for every connection, both as server and
1545 client, there is no benefit for these.
1546 GnuTLS appears to not support tickets server-side by default (we don't
1547 call gnutls_session_ticket_enable_server()) but client side is enabled
1548 by default on recent versions (3.1.3 +) unless the PFS priority string
1551 PP/03 Add $SOURCE_DATE_EPOCH support for reproducible builds, per spec at
1552 <https://reproducible-builds.org/specs/source-date-epoch/>.
1554 JH/07 Fix smtp transport use of limited max_rcpt under mua_wrapper. Previously
1555 the check for any unsuccessful recipients did not notice the limit, and
1556 erroneously found still-pending ones.
1558 JH/08 Pipeline CHUNKING command and data together, on kernels that support
1559 MSG_MORE. Only in-clear (not on TLS connections).
1561 JH/09 Avoid using a temporary file during transport using dkim. Unless a
1562 transport-filter is involved we can buffer the headers in memory for
1563 creating the signature, and read the spool data file once for the
1564 signature and again for transmission.
1566 JH/10 Enable use of sendfile in Linux builds as default. It was disabled in
1567 4.77 as the kernel support then wasn't solid, having issues in 64bit
1568 mode. Now, it's been long enough. Add support for FreeBSD also.
1570 JH/11 Bug 2104: Fix continued use of a transport connection with TLS. In the
1571 case where the routing stage had gathered several addresses to send to
1572 a host before calling the transport for the first, we previously failed
1573 to close down TLS in the old transport process before passing the TCP
1574 connection to the new process. The new one sent a STARTTLS command
1575 which naturally failed, giving a failed delivery and bloating the retry
1576 database. Investigation and fix prototype from Wolfgang Breyha.
1578 JH/12 Fix check on SMTP command input synchronisation. Previously there were
1579 false-negatives in the check that the sender had not preempted a response
1580 or prompt from Exim (running as a server), due to that code's lack of
1581 awareness of the SMTP input buffering.
1583 PP/04 Add commandline_checks_require_admin option.
1584 Exim drops privileges sanely, various checks such as -be aren't a
1585 security problem, as long as you trust local users with access to their
1586 own account. When invoked by services which pass untrusted data to
1587 Exim, this might be an issue. Set this option in main configuration
1588 AND make fixes to the calling application, such as using `--` to stop
1591 JH/13 Do pipelining under TLS. Previously, although safe, no advantage was
1592 taken. Now take care to pack both (client) MAIL,RCPT,DATA, and (server)
1593 responses to those, into a single TLS record each way (this usually means
1594 a single packet). As a side issue, smtp_enforce_sync now works on TLS
1597 PP/05 OpenSSL/1.1: use DH_bits() for more accurate DH param sizes. This
1598 affects you only if you're dancing at the edge of the param size limits.
1599 If you are, and this message makes sense to you, then: raise the
1600 configured limit or use OpenSSL 1.1. Nothing we can do for older
1603 JH/14 For the "sock" variant of the malware scanner interface, accept an empty
1604 cmdline element to get the documented default one. Previously it was
1607 JH/15 Fix a crash in the smtp transport caused when two hosts in succession
1608 are unsuable for non-message-specific reasons - eg. connection timeout,
1609 banner-time rejection.
1611 JH/16 Fix logging of delivery remote port, when specified by router, under
1614 PP/06 Repair manualroute's ability to take options in any order, even if one
1615 is the name of a transport.
1618 HS/01 Cleanup, prevent repeated use of -p/-oMr (CVE-2017-1000369)
1620 JH/17 Change the list-building routines interface to use the expanding-string
1621 triplet model, for better allocation and copying behaviour.
1623 JH/18 Prebuild the data-structure for "builtin" macros, for faster startup.
1624 Previously it was constructed the first time a possibly-matching string
1625 was met in the configuration file input during startup; now it is done
1628 JH/19 Bug 2141: Use the full-complex API for Berkeley DB rather than the legacy-
1629 compatible one, to avoid the (poorly documented) possibility of a config
1630 file in the working directory redirecting the DB files, possibly correpting
1631 some existing file. CVE-2017-10140 assigned for BDB.
1633 JH/20 Bug 2147: Do not defer for a verify-with-callout-and-random which is not
1634 cache-hot. Previously, although the result was properly cached, the
1635 initial verify call returned a defer.
1637 JH/21 Bug 2151: Avoid using SIZE on the MAIL for a callout verify, on any but
1638 the main verify for receipient in uncached-mode.
1640 JH/22 Retire historical build files to an "unsupported" subdir. These are
1641 defined as "ones for which we have no current evidence of testing".
1643 JH/23 DKIM: enforce the DNS pubkey record "h" permitted-hashes optional field,
1644 if present. Previously it was ignored.
1646 JH/24 Start using specified-initialisers in C structure init coding. This is
1647 a C99 feature (it's 2017, so now considered safe).
1649 JH/25 Use one-bit bitfields for flags in the "addr" data structure. Previously
1650 if was a fixed-sized field and bitmask ops via macros; it is now more
1653 PP/07 GitHub PR 56: Apply MariaDB build fix.
1654 Patch provided by Jaroslav Å karvada.
1656 PP/08 Bug 2161: Fix regression in sieve quoted-printable handling introduced
1657 during Coverity cleanups [4.87 JH/47]
1658 Diagnosis and fix provided by Michael Fischer v. Mollard.
1660 JH/26 Fix DKIM bug: when the pseudoheader generated for signing was exactly
1661 the right size to place the terminating semicolon on its own folded
1662 line, the header hash was calculated to an incorrect value thanks to
1663 the (relaxed) space the fold became.
1665 HS/02 Fix Bug 2130: large writes from the transport subprocess were chunked
1666 and confused the parent.
1668 JH/27 Fix SOCKS bug: an unitialized pointer was deref'd by the transport process
1669 which could crash as a result. This could lead to undeliverable messages.
1671 JH/28 Logging: "next input sent too soon" now shows where input was truncated
1674 JH/29 Fix queue_run_in_order to ignore the PID portion of the message ID. This
1675 matters on fast-turnover and PID-randomising systems, which were getting
1676 out-of-order delivery.
1678 JH/30 Fix a logging bug on aarch64: an unsafe routine was previously used for
1679 a possibly-overlapping copy. The symptom was that "Remote host closed
1680 connection in response to HELO" was logged instead of the actual 4xx
1683 JH/31 Fix CHUNKING code to properly flush the unwanted chunk after an error.
1684 Previously only that bufferd was discarded, resulting in SYMTP command
1687 JH/32 DKIM: when a message has multiple signatures matching an identity given
1688 in dkim_verify_signers, run the dkim acl once for each. Previously only
1689 one run was done. Bug 2189.
1691 JH/33 Downgrade an unfound-list name (usually a typo in the config file) from
1692 "panic the current process" to "deliberately defer". The panic log is
1693 still written with the problem list name; the mail and reject logs now
1694 get a temp-reject line for the message that was being handled, saying
1695 something like "domains check lookup or other defer". The SMTP 451
1696 message is still "Temporary local problem".
1698 JH/34 Bug 2199: Fix a use-after-free while reading smtp input for header lines.
1699 A crafted sequence of BDAT commands could result in in-use memory beeing
1700 freed. CVE-2017-16943.
1702 HS/03 Bug 2201: Fix checking for leading-dot on a line during headers reading
1703 from SMTP input. Previously it was always done; now only done for DATA
1704 and not BDAT commands. CVE-2017-16944.
1706 JH/35 Bug 2201: Flush received data in BDAT mode after detecting an error fatal
1707 to the message (such as an overlong header line). Previously this was
1708 not done and we did not exit BDAT mode. Followon from the previous item
1709 though a different problem.
1715 JH/01 Bug 1922: Support IDNA2008. This has slightly different conversion rules
1716 than -2003 did; needs libidn2 in addition to libidn.
1718 JH/02 The path option on a pipe transport is now expanded before use.
1720 PP/01 GitHub PR 50: Do not call ldap_start_tls_s on ldapi:// connections.
1721 Patch provided by "Björn", documentation fix added too.
1723 JH/03 Bug 2003: fix Proxy Protocol v2 handling: the address size field was
1724 missing a wire-to-host endian conversion.
1726 JH/04 Bug 2004: fix CHUNKING in non-PIPELINEING mode. Chunk data following
1727 close after a BDAT command line could be taken as a following command,
1728 giving a synch failure. Fix by only checking for synch immediately
1729 before acknowledging the chunk.
1731 PP/02 GitHub PR 52: many spelling fixes, which include fixing parsing of
1732 no_require_dnssec option and creation of _HAVE_TRANSPORT_APPEND_MAILDIR
1733 macro. Patches provided by Josh Soref.
1735 JH/05 Have the EHLO response advertise VRFY, if there is a vrfy ACL defined.
1736 Previously we did not; the RFC seems ambiguous and VRFY is not listed
1737 by IANA as a service extension. However, John Klensin suggests that we
1740 JH/06 Bug 2017: Fix DKIM verification in -bh test mode. The data feed into
1741 the dkim code may be unix-mode line endings rather than smtp wire-format
1742 CRLF, so prepend a CR to any bare LF.
1744 JH/07 Rationalise the coding for callout smtp conversations and transport ones.
1745 As a side-benfit, callouts can now use PIPELINING hence fewer round-trips.
1747 JH/08 Bug 2016: Fix DKIM verification vs. CHUNKING. Any BDAT commands after
1748 the first were themselves being wrongly included in the feed into dkim
1749 processing; with most chunk sizes in use this resulted in an incorrect
1750 body hash calculated value.
1752 JH/09 Bug 2014: permit inclusion of a DKIM-Signature header in a received
1753 DKIM signature block, for verification. Although advised against by
1754 standards it is specifically not ruled illegal.
1756 JH/10 Bug 2025: Fix reception of (quoted) local-parts with embedded spaces.
1758 JH/11 Bug 2029: Fix crash in DKIM verification when a message signature block is
1759 missing a body hash (the bh= tag).
1761 JH/12 Bug 2018: Re-order Proxy Protocol startup versus TLS-on-connect startup.
1762 It seems that HAProxy sends the Proxy Protocol information in clear and
1763 only then does a TLS startup, so do the same.
1765 JH/13 Bug 2027: Avoid attempting to use TCP Fast Open for non-transport client
1766 TCP connections (such as for Spamd) unless the daemon successfully set
1767 Fast Open mode on its listening sockets. This fixes breakage seen on
1768 too-old kernels or those not configured for Fast Open, at the cost of
1769 requiring both directions being enabled for TFO, and TFO never being used
1770 by non-daemon-related Exim processes.
1772 JH/14 Bug 2000: Reject messages recieved with CHUNKING but with malformed line
1773 endings, at least on the first header line. Try to canonify any that get
1774 past that check, despite the cost.
1776 JH/15 Angle-bracket nesting (an error inserted by broken sendmails) levels are
1777 now limited to an arbitrary five deep, while parsing addresses with the
1778 strip_excess_angle_brackets option enabled.
1780 PP/03 Bug 2018: For Proxy Protocol and TLS-on-connect, do not over-read and
1781 instead leave the unprompted TLS handshake in socket buffer for the
1782 TLS library to consume.
1784 PP/04 Bug 2018: Also handle Proxy Protocol v2 safely.
1786 PP/05 FreeBSD compat: handle that Ports no longer create /usr/bin/perl
1788 JH/16 Drop variables when they go out of scope. Memory management drops a whole
1789 region in one operation, for speed, and this leaves assigned pointers
1790 dangling. Add checks run only under the testsuite which checks all
1791 variables at a store-reset and panics on a dangling pointer; add code
1792 explicitly nulling out all the variables discovered. Fixes one known
1793 bug: a transport crash, where a dangling pointer for $sending_ip_address
1794 originally assigned in a verify callout, is re-used.
1796 PP/06 Drop '.' from @INC in various Perl scripts.
1798 PP/07 Switch FreeBSD iconv to always use the base-system libc functions.
1800 PP/08 Reduce a number of compilation warnings under clang; building with
1801 CC=clang CFLAGS+=-Wno-dangling-else -Wno-logical-op-parentheses
1802 should be warning-free.
1804 JH/17 Fix inbound CHUNKING when DKIM disabled at runtime.
1806 HS/01 Fix portability problems introduced by PP/08 for platforms where
1807 realloc(NULL) is not equivalent to malloc() [SunOS et al].
1809 HS/02 Bug 1974: Fix missing line terminator on the last received BDAT
1810 chunk. This allows us to accept broken chunked messages. We need a more
1811 general solution here.
1813 PP/09 Wrote util/chunking_fixqueue_finalnewlines.pl to help recover
1814 already-broken messages in the queue.
1816 JH/18 Bug 2061: Fix ${extract } corrupting an enclosing ${reduce } $value.
1818 JH/19 Fix reference counting bug in routing-generated-address tracking.
1824 JH/01 Use SIZE on MAIL FROM in a cutthrough connection, if the destination
1825 supports it and a size is available (ie. the sending peer gave us one).
1827 JH/02 The obsolete acl condition "demime" is removed (finally, after ten
1828 years of being deprecated). The replacements are the ACLs
1829 acl_smtp_mime and acl_not_smtp_mime.
1831 JH/03 Upgrade security requirements imposed for hosts_try_dane: previously
1832 a downgraded non-dane trust-anchor for the TLS connection (CA-style)
1833 or even an in-clear connection were permitted. Now, if the host lookup
1834 was dnssec and dane was requested then the host is only used if the
1835 TLSA lookup succeeds and is dnssec. Further hosts (eg. lower priority
1836 MXs) will be tried (for hosts_try_dane though not for hosts_require_dane)
1837 if one fails this test.
1838 This means that a poorly-configured remote DNS will make it incommunicado;
1839 but it protects against a DNS-interception attack on it.
1841 JH/04 Bug 1810: make continued-use of an open smtp transport connection
1842 non-noisy when a race steals the message being considered.
1844 JH/05 If main configuration option tls_certificate is unset, generate a
1845 self-signed certificate for inbound TLS connections.
1847 JH/06 Bug 165: hide more cases of password exposure - this time in expansions
1848 in rewrites and routers.
1850 JH/07 Retire gnutls_require_mac et.al. These were nonfunctional since 4.80
1851 and logged a warning sing 4.83; now they are a configuration file error.
1853 JH/08 Bug 1836: Fix crash in VRFY handling when handed an unqualified name
1854 (lacking @domain). Apply the same qualification processing as RCPT.
1856 JH/09 Bug 1804: Avoid writing msglog files when in -bh or -bhc mode.
1858 JH/10 Support ${sha256:} applied to a string (as well as the previous
1861 JH/11 Cutthrough: avoid using the callout hints db on a verify callout when
1862 a cutthrough deliver is pending, as we always want to make a connection.
1863 This also avoids re-routing the message when later placing the cutthrough
1864 connection after a verify cache hit.
1865 Do not update it with the verify result either.
1867 JH/12 Cutthrough: disable when verify option success_on_redirect is used, and
1868 when routing results in more than one destination address.
1870 JH/13 Cutthrough: expand transport dkim_domain option when testing for dkim
1871 signing (which inhibits the cutthrough capability). Previously only
1872 the presence of an option was tested; now an expansion evaluating as
1873 empty is permissible (obviously it should depend only on data available
1874 when the cutthrough connection is made).
1876 JH/14 Fix logging of errors under PIPELINING. Previously the log line giving
1877 the relevant preceding SMTP command did not note the pipelining mode.
1879 JH/15 Fix counting of empty lines in $body_linecount and $message_linecount.
1880 Previously they were not counted.
1882 JH/16 DANE: treat a TLSA lookup response having all non-TLSA RRs, the same
1883 as one having no matching records. Previously we deferred the message
1884 that needed the lookup.
1886 JH/17 Fakereject: previously logged as a normal message arrival "<="; now
1887 distinguished as "(=".
1889 JH/18 Bug 1867: make the fail_defer_domains option on a dnslookup router work
1890 for missing MX records. Previously it only worked for missing A records.
1892 JH/19 Bug 1850: support Radius libraries that return REJECT_RC.
1894 JH/20 Bug 1872: Ensure that acl_smtp_notquit is run when the connection drops
1895 after the data-go-ahead and data-ack. Patch from Jason Betts.
1897 JH/21 Bug 1846: Send DMARC forensic reports for reject and quarantine results,
1898 even for a "none" policy. Patch from Tony Meyer.
1900 JH/22 Fix continued use of a connection for further deliveries. If a port was
1901 specified by a router, it must also match for the delivery to be
1904 JH/23 Bug 1874: fix continued use of a connection for further deliveries.
1905 When one of the recipients of a message was unsuitable for the connection
1906 (has no matching addresses), we lost track of needing to mark it
1907 deferred. As a result mail would be lost.
1909 JH/24 Bug 1832: Log EHLO response on getting conn-close response for HELO.
1911 JH/25 Decoding ACL controls is now done using a binary search; the source code
1912 takes up less space and should be simpler to maintain. Merge the ACL
1913 condition decode tables also, with similar effect.
1915 JH/26 Fix problem with one_time used on a redirect router which returned the
1916 parent address unchanged. A retry would see the parent address marked as
1917 delivered, so not attempt the (identical) child. As a result mail would
1920 JH/27 Fix a possible security hole, wherein a process operating with the Exim
1921 UID can gain a root shell. Credit to http://www.halfdog.net/ for
1922 discovery and writeup. Ubuntu bug 1580454; no bug raised against Exim
1925 JH/28 Enable {spool,log} filesystem space and inode checks as default.
1926 Main config options check_{log,spool}_{inodes,space} are now
1927 100 inodes, 10MB unless set otherwise in the configuration.
1929 JH/29 Fix the connection_reject log selector to apply to the connect ACL.
1930 Previously it only applied to the main-section connection policy
1933 JH/30 Bug 1897: fix callouts connection fallback from TLS to cleartext.
1935 PP/01 Changed default Diffie-Hellman parameters to be Exim-specific, created
1936 by me. Added RFC7919 DH primes as an alternative.
1938 PP/02 Unbreak build via pkg-config with new hash support when crypto headers
1939 are not in the system include path.
1941 JH/31 Fix longstanding bug with aborted TLS server connection handling. Under
1942 GnuTLS, when a session startup failed (eg because the client disconnected)
1943 Exim did stdio operations after fclose. This was exposed by a recent
1944 change which nulled out the file handle after the fclose.
1946 JH/32 Bug 1909: Fix OCSP proof verification for cases where the proof is
1947 signed directly by the cert-signing cert, rather than an intermediate
1948 OCSP-signing cert. This is the model used by LetsEncrypt.
1950 JH/33 Bug 1914: Ensure socket is nonblocking before draining after SMTP QUIT.
1952 HS/01 Fix leak in verify callout under GnuTLS, about 3MB per recipient on
1953 an incoming connection.
1955 HS/02 Bug 1802: Do not half-close the connection after sending a request
1958 HS/03 Use "auto" as the default EC curve parameter. For OpenSSL < 1.0.2
1959 fallback to "prime256v1".
1961 JH/34 SECURITY: Use proper copy of DATA command in error message.
1962 Could leak key material. Remotely exploitable. CVE-2016-9963.
1968 JH/01 Bug 1664: Disable OCSP for GnuTLS library versions at/before 3.3.16
1969 and 3.4.4 - once the server is enabled to respond to an OCSP request
1970 it does even when not requested, resulting in a stapling non-aware
1971 client dropping the TLS connection.
1973 TF/01 Code cleanup: Overhaul the debug_selector and log_selector machinery to
1974 support variable-length bit vectors. No functional change.
1976 TF/02 Improve the consistency of logging incoming and outgoing interfaces.
1977 The I= interface field on outgoing lines is now after the H= remote
1978 host field, same as incoming lines. There is a separate
1979 outgoing_interface log selector which allows you to disable the
1982 JH/02 Bug 728: Close logfiles after a daemon-process "exceptional" log write.
1983 If not running log_selector +smtp_connection the mainlog would be held
1984 open indefinitely after a "too many connections" event, including to a
1985 deleted file after a log rotate. Leave the per net connection logging
1986 leaving it open for efficiency as that will be quickly detected by the
1987 check on the next write.
1989 HS/01 Bug 1671: Fix post transport crash.
1990 Processing the wait-<transport> messages could crash the delivery
1991 process if the message IDs didn't exist for some reason. When
1992 using 'split_spool_directory=yes' the construction of the spool
1993 file name failed already, exposing the same netto behaviour.
1995 JH/03 Bug 425: Capture substrings in $regex1, $regex2 etc from regex &
1996 mime_regex ACL conditions.
1998 JH/04 Bug 1686: When compiled with EXPERIMENTAL_DSN_INFO: Add extra information
1999 to DSN fail messages (bounces): remote IP, remote greeting, remote response
2000 to HELO, local diagnostic string.
2002 JH/05 Downgrade message for a TLS-certificate-based authentication fail from
2003 log line to debug. Even when configured with a tls authenticator many
2004 client connections are expected to not authenticate in this way, so
2005 an authenticate fail is not an error.
2007 HS/02 Add the Exim version string to the process info. This way exiwhat
2008 gives some more detail about the running daemon.
2010 JH/06 Bug 1395: time-limit caching of DNS lookups, to the TTL value. This may
2011 matter for fast-change records such as DNSBLs.
2013 JH/07 Bug 1678: Always record an interface option value, if set, as part of a
2014 retry record, even if constant. There may be multiple transports with
2015 different interface settings and the retry behaviour needs to be kept
2018 JH/08 Bug 1586: exiqgrep now refuses to run if there are unexpected arguments.
2020 JH/09 Bug 1700: ignore space & tab embedded in base64 during decode.
2022 JH/10 Bug 840: fix log_defer_output option of pipe transport
2024 JH/11 Bug 830: use same host for all RCPTS of a message, even under
2025 hosts_randomize. This matters a lot when combined with mua_wrapper.
2027 JH/12 Bug 1706: percent and underbar characters are no longer escaped by the
2028 ${quote_pgsql:<string>} operator.
2030 JH/13 Bug 1708: avoid misaligned access in cached lookup.
2032 JH/14 Change header file name for freeradius-client. Relevant if compiling
2033 with Radius support; from the Gentoo tree and checked under Fedora.
2035 JH/15 Bug 1712: Introduce $prdr_requested flag variable
2037 JH/16 Bug 1714: Permit an empty string as expansion result for transport
2038 option transport_filter, meaning no filtering.
2040 JH/17 Bug 1713: Fix non-PDKIM_DEBUG build. Patch from Jasen Betts.
2042 JH/18 Bug 1709: When built with TLS support, the tls_advertise_hosts option now
2043 defaults to "*" (all hosts). The variable is now available when not built
2044 with TLS, default unset, mainly to enable keeping the testsuite sane.
2045 If a server certificate is not supplied (via tls_certificate) an error is
2046 logged, and clients will find TLS connections fail on startup. Presumably
2047 they will retry in-clear.
2048 Packagers of Exim are strongly encouraged to create a server certificate
2049 at installation time.
2051 HS/03 Add -bP config_file as a synonym for -bP configure_file, for consistency
2052 with the $config_file variable.
2054 JH/19 Two additional event types: msg:rcpt:defer and msg:rcpt:host:defer. Both
2055 in transport context, after the attempt, and per-recipient. The latter type
2056 is per host attempted. The event data is the error message, and the errno
2057 information encodes the lookup type (A vs. MX) used for the (first) host,
2058 and the trailing two digits of the smtp 4xx response.
2060 GF/01 Bug 1715: Fix for race condition in exicyclog, where exim could attempt
2061 to write to mainlog (or rejectlog, paniclog) in the window between file
2062 creation and permissions/ownership being changed. Particularly affects
2063 installations where exicyclog is run as root, rather than exim user;
2064 result is that the running daemon panics and dies.
2066 JH/20 Bug 1701: For MySQL lookups, support MySQL config file option group names.
2068 JH/21 Bug 1720: Add support for priority groups and weighted-random proxy
2069 selection for the EXPERIMENTAL_SOCKS feature, via new per-proxy options
2070 "pri" and "weight". Note that the previous implicit priority given by the
2071 list order is no longer honoured.
2073 JH/22 Bugs 963, 1721: Fix some corner cases in message body canonicalization
2074 for DKIM processing.
2076 JH/23 Move SOCKS5 support from Experimental to mainline, enabled for a build
2077 by defining SUPPORT_SOCKS.
2079 JH/26 Move PROXY support from Experimental to mainline, enabled for a build
2080 by defining SUPPORT_PROXY. Note that the proxy_required_hosts option
2081 is renamed to hosts_proxy, and the proxy_{host,target}_{address,port}.
2082 variables are renamed to proxy_{local,external}_{address,port}.
2084 JH/27 Move Internationalisation support from Experimental to mainline, enabled
2085 for a build by defining SUPPORT_I18N
2087 JH/28 Bug 1745: Fix redis lookups to handle (quoted) spaces embedded in parts
2088 of the query string, and make ${quote_redis:} do that quoting.
2090 JH/29 Move Events support from Experimental to mainline, enabled by default
2091 and removable for a build by defining DISABLE_EVENT.
2093 JH/30 Updated DANE implementation code to current from Viktor Dukhovni.
2095 JH/31 Fix bug with hosts_connection_nolog and named-lists which were wrongly
2096 cached by the daemon.
2098 JH/32 Move Redis support from Experimental to mainline, enabled for a build
2099 by defining LOOKUP_REDIS. The libhiredis library is required.
2101 JH/33 Bug 1748: Permit ACL dnslists= condition in non-smtp ACLs if explicit
2102 keys are given for lookup.
2104 JH/34 Bug 1192: replace the embedded copy of PolarSSL RSA routines in the DKIM
2105 support, by using OpenSSL or GnuTLS library ones. This means DKIM is
2106 only supported when built with TLS support. The PolarSSL SHA routines
2107 are still used when the TLS library is too old for convenient support.
2109 JH/35 Require SINGLE_DH_USE by default in OpenSSL (main config option
2110 openssl_options), for security. OpenSSL forces this from version 1.1.0
2111 server-side so match that on older versions.
2113 JH/36 Bug 1778: longstanding bug in memory use by the ${run } expansion: A fresh
2114 allocation for $value could be released as the expansion processing
2115 concluded, but leaving the global pointer active for it.
2117 JH/37 Bug 1769: Permit a VRFY ACL to override the default 252 response,
2118 and to use the domains and local_parts ACL conditions.
2120 JH/38 Fix cutthrough bug with body lines having a single dot. The dot was
2121 incorrectly not doubled on cutthrough transmission, hence seen as a
2122 body-termination at the receiving system - resulting in truncated mails.
2123 Commonly the sender saw a TCP-level error, and retransmitted the message
2124 via the normal store-and-forward channel. This could result in duplicates
2125 received - but deduplicating mailstores were liable to retain only the
2126 initial truncated version.
2128 JH/39 Bug 1781: Fix use of DKIM private-keys having trailing '=' in the base-64.
2130 JH/40 Fix crash in queryprogram router when compiled with EXPERIMENTAL_SRS.
2132 JH/41 Bug 1792: Fix selection of headers to sign for DKIM: bottom-up. While
2133 we're in there, support oversigning also; bug 1309.
2135 JH/42 Bug 1796: Fix error logged on a malware scanner connection failure.
2137 HS/04 Add support for keep_environment and add_environment options.
2139 JH/43 Tidy coding issues detected by gcc --fsanitize=undefined. Some remain;
2140 either intentional arithmetic overflow during PRNG, or testing config-
2143 JH/44 Bug 1800: The combination of a -bhc commandline option and cutthrough
2144 delivery resulted in actual delivery. Cancel cutthrough before DATA
2147 JH/45 Fix cutthrough, when connection not opened by verify and target hard-
2148 rejects a recipient: pass the reject to the originator.
2150 JH/46 Multiple issues raised by Coverity. Some were obvious or plausible bugs.
2151 Many were false-positives and ignorable, but it's worth fixing the
2154 JH/47 Fix build on HP-UX and older Solaris, which need (un)setenv now also
2155 for the new environment-manipulation done at startup. Move the routines
2156 from being local to tls.c to being global via the os.c file.
2158 JH/48 Bug 1807: Fix ${extract } for the numeric/3-string case. While preparsing
2159 an extract embedded as result-arg for a map, the first arg for extract
2160 is unavailable so we cannot tell if this is a numbered or keyed
2161 extraction. Accept either.
2167 JH/01 Bug 1545: The smtp transport option "retry_include_ip_address" is now
2170 JH/02 The smtp transport option "multi_domain" is now expanded.
2172 JH/03 The smtp transport now requests PRDR by default, if the server offers
2175 JH/04 Certificate name checking on server certificates, when exim is a client,
2176 is now done by default. The transport option tls_verify_cert_hostnames
2177 can be used to disable this per-host. The build option
2178 EXPERIMENTAL_CERTNAMES is withdrawn.
2180 JH/05 The value of the tls_verify_certificates smtp transport and main options
2181 default to the word "system" to access the system default CA bundle.
2182 For GnuTLS, only version 3.0.20 or later.
2184 JH/06 Verification of the server certificate for a TLS connection is now tried
2185 (but not required) by default. The verification status is now logged by
2186 default, for both outbound TLS and client-certificate supplying inbound
2189 JH/07 Changed the default rfc1413 lookup settings to disable calls. Few
2192 JH/08 The EXPERIMENTAL_DSN compile option is no longer needed; all Delivery
2193 Status Notification (bounce) messages are now MIME format per RFC 3464.
2194 Support for RFC 3461 DSN options NOTIFY,ENVID,RET,ORCPT can be advertised
2195 under the control of the dsn_advertise_hosts option, and routers may
2196 have a dsn_lasthop option.
2198 JH/09 A timeout of 2 minutes is now applied to all malware scanner types by
2199 default, modifiable by a malware= option. The list separator for
2200 the options can now be changed in the usual way. Bug 68.
2202 JH/10 The smtp_receive_timeout main option is now expanded before use.
2204 JH/11 The incoming_interface log option now also enables logging of the
2205 local interface on delivery outgoing connections.
2207 JH/12 The cutthrough-routing facility now supports multi-recipient mails,
2208 if the interface and destination host and port all match.
2210 JH/13 Bug 344: The verify = reverse_host_lookup ACL condition now accepts a
2213 JH/14 Bug 1573: The spam= ACL condition now additionally supports Rspamd.
2214 Patch from Andrew Lewis.
2216 JH/15 Bug 670: The spamd_address main option (for the spam= ACL condition)
2217 now supports optional time-restrictions, weighting, and priority
2218 modifiers per server. Patch originally by <rommer@active.by>.
2220 JH/16 The spamd_address main option now supports a mixed list of local
2221 and remote servers. Remote servers can be IPv6 addresses, and
2222 specify a port-range.
2224 JH/17 Bug 68: The spamd_address main option now supports an optional
2225 timeout value per server.
2227 JH/18 Bug 1581: Router and transport options headers_add/remove can
2228 now have the list separator specified.
2230 JH/19 Bug 392: spamd_address, and clamd av_scanner, now support retry
2233 JH/20 Bug 1571: Ensure that $tls_in_peerdn is set, when verification fails
2236 JH/21 Support for the A6 type of dns record is withdrawn.
2238 JH/22 Bug 608: The result of a QUIT or not-QUIT toplevel ACL now matters
2239 rather than the verbs used.
2241 JH/23 Bug 1572: Increase limit on SMTP confirmation message copy size
2242 from 255 to 1024 chars.
2244 JH/24 Verification callouts now attempt to use TLS by default.
2246 HS/01 DNSSEC options (dnssec_require_domains, dnssec_request_domains)
2247 are generic router options now. The defaults didn't change.
2249 JH/25 Bug 466: Add RFC2322 support for MIME attachment filenames.
2250 Original patch from Alexander Shikoff, worked over by JH.
2252 HS/02 Bug 1575: exigrep falls back to autodetection of compressed
2253 files if ZCAT_COMMAND is not executable.
2255 JH/26 Bug 1539: Add timeout/retry options on dnsdb lookups.
2257 JH/27 Bug 286: Support SOA lookup in dnsdb lookups.
2259 JH/28 Bug 1588: Do not use the A lookup following an AAAA for setting the FQDN.
2260 Normally benign, it bites when the pair was led to by a CNAME;
2261 modern usage is to not canonicalize the domain to a CNAME target
2262 (and we were inconsistent anyway for A-only vs AAAA+A).
2264 JH/29 Bug 1632: Removed the word "rejected" from line logged for ACL discards.
2266 JH/30 Check the forward DNS lookup for DNSSEC, in addition to the reverse,
2267 when evaluating $sender_host_dnssec.
2269 JH/31 Check the HELO verification lookup for DNSSEC, adding new
2270 $sender_helo_dnssec variable.
2272 JH/32 Bug 1397: Enable ECDHE on OpenSSL, just the NIST P-256 curve.
2274 JH/33 Bug 1346: Note MAIL cmd seen in -bS batch, to avoid smtp_no_mail log.
2276 JH/34 Bug 1648: Fix a memory leak seen with "mailq" and large queues.
2278 JH/35 Bug 1642: Fix support of $spam_ variables at delivery time. Was
2279 documented as working, but never had. Support all but $spam_report.
2281 JH/36 Bug 1659: Guard checking of input smtp commands again pseudo-command
2282 added for tls authenticator.
2284 HS/03 Add perl_taintmode main config option
2290 TL/01 When running the test suite, the README says that variables such as
2291 no_msglog_check are global and can be placed anywhere in a specific
2292 test's script, however it was observed that placement needed to be near
2293 the beginning for it to behave that way. Changed the runtest perl
2294 script to read through the entire script once to detect and set these
2295 variables, reset to the beginning of the script, and then run through
2296 the script parsing/test process like normal.
2298 TL/02 The BSD's have an arc4random API. One of the functions to induce
2299 adding randomness was arc4random_stir(), but it has been removed in
2300 OpenBSD 5.5. Detect this OpenBSD version and skip calling this
2301 function when detected.
2303 JH/01 Expand the EXPERIMENTAL_TPDA feature. Several different events now
2304 cause callback expansion.
2306 TL/03 Bugzilla 1518: Clarify "condition" processing in routers; that
2307 syntax errors in an expansion can be treated as a string instead of
2308 logging or causing an error, due to the internal use of bool_lax
2309 instead of bool when processing it.
2311 JH/02 Add EXPERIMENTAL_DANE, allowing for using the DNS as trust-anchor for
2312 server certificates when making smtp deliveries.
2314 JH/03 Support secondary-separator specifier for MX, SRV, TLSA lookups.
2316 JH/04 Add ${sort {list}{condition}{extractor}} expansion item.
2318 TL/04 Bugzilla 1216: Add -M (related messages) option to exigrep.
2320 TL/05 GitHub Issue 18: Adjust logic testing for true/false in redis lookups.
2321 Merged patch from Sebastian Wiedenroth.
2323 JH/05 Fix results-pipe from transport process. Several recipients, combined
2324 with certificate use, exposed issues where response data items split
2325 over buffer boundaries were not parsed properly. This eventually
2326 resulted in duplicates being sent. This issue only became common enough
2327 to notice due to the introduction of connection certificate information,
2328 the item size being so much larger. Found and fixed by Wolfgang Breyha.
2330 JH/06 Bug 1533: Fix truncation of items in headers_remove lists. A fixed
2331 size buffer was used, resulting in syntax errors when an expansion
2334 JH/07 Add support for directories of certificates when compiled with a GnuTLS
2335 version 3.3.6 or later.
2337 JH/08 Rename the TPDA experimental facility to Event Actions. The #ifdef
2338 is EXPERIMENTAL_EVENT, the main-configuration and transport options
2339 both become "event_action", the variables become $event_name, $event_data
2340 and $event_defer_errno. There is a new variable $verify_mode, usable in
2341 routers, transports and related events. The tls:cert event is now also
2342 raised for inbound connections, if the main configuration event_action
2345 TL/06 In test suite, disable OCSP for old versions of openssl which contained
2346 early OCSP support, but no stapling (appears to be less than 1.0.0).
2348 JH/09 When compiled with OpenSSL and EXPERIMENTAL_CERTNAMES, the checks on
2349 server certificate names available under the smtp transport option
2350 "tls_verify_cert_hostname" now do not permit multi-component wildcard
2353 JH/10 Time-related extraction expansions from certificates now use the main
2354 option "timezone" setting for output formatting, and are consistent
2355 between OpenSSL and GnuTLS compilations. Bug 1541.
2357 JH/11 Fix a crash in mime ACL when meeting a zero-length, quoted or RFC2047-
2358 encoded parameter in the incoming message. Bug 1558.
2360 JH/12 Bug 1527: Autogrow buffer used in reading spool files. Since they now
2361 include certificate info, eximon was claiming there were spoolfile
2364 JH/13 Bug 1521: Fix ldap lookup for single-attr request, multiple-attr return.
2366 JH/14 Log delivery-related information more consistently, using the sequence
2367 "H=<name> [<ip>]" wherever possible.
2369 TL/07 Bug 1547: Omit RFCs from release. Draft and RFCs have licenses which
2370 are problematic for Debian distribution, omit them from the release
2373 JH/15 Updates and fixes to the EXPERIMENTAL_DSN feature.
2375 JH/16 Fix string representation of time values on 64bit time_t architectures.
2378 JH/17 Fix a null-indirection in certextract expansions when a nondefault
2379 output list separator was used.
2384 TL/01 Bugzilla 1506: Re-add a 'return NULL' to silence complaints from static
2385 checkers that were complaining about end of non-void function with no
2388 JH/01 Bug 1513: Fix parsing of quoted parameter values in MIME headers.
2389 This was a regression introduced in 4.83 by another bugfix.
2391 JH/02 Fix broken compilation when EXPERIMENTAL_DSN is enabled.
2393 TL/02 Bug 1509: Fix exipick for enhanced spoolfile specification used when
2394 EXPERIMENTAL_DSN is enabled. Fix from Wolfgang Breyha.
2400 TF/01 Correctly close the server side of TLS when forking for delivery.
2402 When a message was received over SMTP with TLS, Exim failed to clear up
2403 the incoming connection properly after forking off the child process to
2404 deliver the message. In some situations the subsequent outgoing
2405 delivery connection happened to have the same fd number as the incoming
2406 connection previously had. Exim would try to use TLS and fail, logging
2407 a "Bad file descriptor" error.
2409 TF/02 Portability fix for building lookup modules on Solaris when the xpg4
2410 utilities have not been installed.
2412 JH/01 Fix memory-handling in use of acl as a conditional; avoid free of
2413 temporary space as the ACL may create new global variables.
2415 TL/01 LDAP support uses per connection or global context settings, depending
2416 upon the detected version of the libraries at build time.
2418 TL/02 Experimental Proxy Protocol support: allows a proxied SMTP connection
2419 to extract and use the src ip:port in logging and expansions as if it
2420 were a direct connection from the outside internet. PPv2 support was
2421 updated based on HAProxy spec change in May 2014.
2423 JH/02 Add ${listextract {number}{list}{success}{fail}}.
2425 TL/03 Bugzilla 1433: Fix DMARC SEGV with specific From header contents.
2426 Properly escape header and check for NULL return.
2428 PP/01 Continue incomplete 4.82 PP/19 by fixing docs too: use dns_dnssec_ok
2431 JH/03 Bugzilla 1157: support log_selector smtp_confirmation for lmtp.
2433 TL/04 Add verify = header_names_ascii check to reject email with non-ASCII
2434 characters in header names, implemented as a verify condition.
2435 Contributed by Michael Fischer v. Mollard.
2437 TL/05 Rename SPF condition results err_perm and err_temp to standardized
2438 results permerror and temperror. Previous values are deprecated but
2439 still accepted. In a future release, err_perm and err_temp will be
2440 completely removed, which will be a backward incompatibility if the
2441 ACL tests for either of these two old results. Patch contributed by
2442 user bes-internal on the mailing list.
2444 JH/04 Add ${utf8clean:} operator. Contributed by Alex Rau.
2446 JH/05 Bugzilla 305: Log incoming-TLS details on rejects, subject to log
2447 selectors, in both main and reject logs.
2449 JH/06 Log outbound-TLS and port details, subject to log selectors, for a
2452 JH/07 Add malware type "sock" for talking to simple daemon.
2454 JH/08 Bugzilla 1371: Add tls_{,try_}verify_hosts to smtp transport.
2456 JH/09 Bugzilla 1431: Support (with limitations) headers_add/headers_remove in
2457 routers/transports under cutthrough routing.
2459 JH/10 Bugzilla 1005: ACL "condition =" should accept values which are negative
2460 numbers. Touch up "bool" conditional to keep the same definition.
2462 TL/06 Remove duplicated language in spec file from 4.82 TL/16.
2464 JH/11 Add dnsdb tlsa lookup. From Todd Lyons.
2466 JH/12 Expand items in router/transport headers_add or headers_remove lists
2467 individually rather than the list as a whole. Bug 1452.
2469 Required for reasonable handling of multiple headers_ options when
2470 they may be empty; requires that headers_remove items with embedded
2471 colons must have them doubled (or the list-separator changed).
2473 TL/07 Add new dmarc expansion variable $dmarc_domain_policy to directly
2474 view the policy declared in the DMARC record. Currently, $dmarc_status
2475 is a combined value of both the record presence and the result of the
2478 JH/13 Fix handling of $tls_cipher et.al. in (non-verify) transport. Bug 1455.
2480 JH/14 New options dnssec_request_domains, dnssec_require_domains on the
2481 dnslookup router and the smtp transport (applying to the forward
2484 TL/08 Bugzilla 1453: New LDAP "SERVERS=" option allows admin to override list
2485 of ldap servers used for a specific lookup. Patch provided by Heiko
2488 JH/18 New options dnssec_lax, dnssec_strict on dnsdb lookups.
2489 New variable $lookup_dnssec_authenticated for observability.
2491 TL/09 Bugzilla 609: Add -C option to exiqgrep, specify which exim.conf to use.
2492 Patch submitted by Lars Timman.
2494 JH/19 EXPERIMENTAL_OCSP support under GnuTLS. Bug 1459.
2496 TL/10 Bugzilla 1454: New -oMm option to pass message reference to Exim.
2497 Requires trusted mode and valid format message id, aborts otherwise.
2498 Patch contributed by Heiko Schlichting.
2500 JH/20 New expansion variables tls_(in,out)_(our,peer)cert, and expansion item
2501 certextract with support for various fields. Bug 1358.
2503 JH/21 Observability of OCSP via variables tls_(in,out)_ocsp. Stapling
2504 is requested by default, modifiable by smtp transport option
2507 JH/22 Expansion operators ${md5:string} and ${sha1:string} can now
2508 operate on certificate variables to give certificate fingerprints
2509 Also new ${sha256:cert_variable}.
2511 JH/23 The PRDR feature is moved from being Experimental into the mainline.
2513 TL/11 Bug 1119: fix memory allocation in string_printing2(). Patch from
2514 Christian Aistleitner.
2516 JH/24 The OCSP stapling feature is moved from Experimental into the mainline.
2518 TL/12 Bug 1444: Fix improper \r\n sequence handling when writing spool
2519 file. Patch from Wolfgang Breyha.
2521 JH/25 Expand the coverage of the delivery $host and $host_address to
2522 client authenticators run in verify callout. Bug 1476.
2524 JH/26 Port service names are now accepted for tls_on_connect_ports, to
2525 align with daemon_smtp_ports. Bug 72.
2527 TF/03 Fix udpsend. The ip_connectedsocket() function's socket type
2528 support and error reporting did not work properly.
2530 TL/13 Bug 1495: Exiqgrep check if -C config file specified on cli exists
2531 and is readable. Patch from Andrew Colin Kissa.
2533 TL/14 Enhance documentation of ${run expansion and how it parses the
2534 commandline after expansion, particularly in the case when an
2535 unquoted variable expansion results in an empty value.
2537 JH/27 The TLS SNI feature was broken in 4.82. Fix it.
2539 PP/02 Fix internal collision of T_APL on systems which support RFC3123
2540 by renaming away from it. Addresses GH issue 15, reported by
2543 JH/28 Fix parsing of MIME headers for parameters with quoted semicolons.
2545 TL/15 SECURITY: prevent double expansion in math comparison functions
2546 (can expand unsanitized data). Not remotely exploitable.
2553 PP/01 Add -bI: framework, and -bI:sieve for querying sieve capabilities.
2555 PP/02 Make -n do something, by making it not do something.
2556 When combined with -bP, the name of an option is not output.
2558 PP/03 Added tls_dh_min_bits SMTP transport driver option, only honoured
2561 PP/04 First step towards DNSSEC, provide $sender_host_dnssec for
2562 $sender_host_name and config options to manage this, and basic check
2565 PP/05 DSCP support for outbound connections and control modifier for inbound.
2567 PP/06 Cyrus SASL: set local and remote IP;port properties for driver.
2568 (Only plugin which currently uses this is kerberos4, which nobody should
2569 be using, but we should make it available and other future plugins might
2570 conceivably use it, even though it would break NAT; stuff *should* be
2571 using channel bindings instead).
2573 PP/07 Handle "exim -L <tag>" to indicate to use syslog with tag as the process
2574 name; added for Sendmail compatibility; requires admin caller.
2575 Handle -G as equivalent to "control = suppress_local_fixups" (we used to
2576 just ignore it); requires trusted caller.
2577 Also parse but ignore: -Ac -Am -X<logfile>
2580 TL/01 Bugzilla 1258 - Refactor MAIL FROM optional args processing.
2582 TL/02 Add +smtp_confirmation as a default logging option.
2584 TL/03 Bugzilla 198 - Implement remove_header ACL modifier.
2585 Patch by Magnus Holmgren from 2007-02-20.
2587 TL/04 Bugzilla 1281 - Spec typo.
2588 Bugzilla 1283 - Spec typo.
2589 Bugzilla 1290 - Spec grammar fixes.
2591 TL/05 Bugzilla 1285 - Spec omission, fix docbook errors for spec.txt creation.
2593 TL/06 Add Experimental DMARC support using libopendmarc libraries.
2595 TL/07 Fix an out of order global option causing a segfault. Reported to dev
2596 mailing list by by Dmitry Isaikin.
2598 JH/01 Bugzilla 1201 & 304 - New cutthrough-delivery feature, with TLS support.
2600 JH/02 Support "G" suffix to numbers in ${if comparisons.
2602 PP/08 Handle smtp transport tls_sni option forced-fail for OpenSSL.
2604 NM/01 Bugzilla 1197 - Spec typo
2605 Bugzilla 1196 - Spec examples corrections
2607 JH/03 Add expansion operators ${listnamed:name} and ${listcount:string}
2609 PP/09 Add gnutls_allow_auto_pkcs11 option (was originally called
2610 gnutls_enable_pkcs11, but renamed to more accurately indicate its
2613 PP/10 Let Linux makefile inherit CFLAGS/CFLAGS_DYNAMIC.
2614 Pulled from Debian 30_dontoverridecflags.dpatch by Andreas Metzler.
2616 JH/04 Add expansion item ${acl {name}{arg}...}, expansion condition
2617 "acl {{name}{arg}...}", and optional args on acl condition
2620 JH/05 Permit multiple router/transport headers_add/remove lines.
2622 JH/06 Add dnsdb pseudo-lookup "a+" to do an "aaaa" + "a" combination.
2624 JH/07 Avoid using a waiting database for a single-message-only transport.
2625 Performance patch from Paul Fisher. Bugzilla 1262.
2627 JH/08 Strip leading/trailing newlines from add_header ACL modifier data.
2630 JH/09 Add $headers_added variable, with content from use of ACL modifier
2631 add_header (but not yet added to the message). Bugzilla 199.
2633 JH/10 Add 8bitmime log_selector, for 8bitmime status on the received line.
2634 Pulled from Bugzilla 817 by Wolfgang Breyha.
2636 PP/11 SECURITY: protect DKIM DNS decoding from remote exploit.
2638 (nb: this is the same fix as in Exim 4.80.1)
2640 JH/11 Add A= logging on delivery lines, and a client_set_id option on
2643 JH/12 Add optional authenticated_sender logging to A= and a log_selector
2646 PP/12 Unbreak server_set_id for NTLM/SPA auth, broken by 4.80 PP/29.
2648 PP/13 Dovecot auth: log better reason to rejectlog if Dovecot did not
2649 advertise SMTP AUTH mechanism to us, instead of a generic
2650 protocol violation error. Also, make Exim more robust to bad
2651 data from the Dovecot auth socket.
2653 TF/01 Fix ultimate retry timeouts for intermittently deliverable recipients.
2655 When a queue runner is handling a message, Exim first routes the
2656 recipient addresses, during which it prunes them based on the retry
2657 hints database. After that it attempts to deliver the message to
2658 any remaining recipients. It then updates the hints database using
2661 So if a recipient address works intermittently, it can get repeatedly
2662 deferred at routing time. The retry hints record remains fresh so the
2663 address never reaches the final cutoff time.
2665 This is a fairly common occurrence when a user is bumping up against
2666 their storage quota. Exim had some logic in its local delivery code
2667 to deal with this. However it did not apply to per-recipient defers
2668 in remote deliveries, e.g. over LMTP to a separate IMAP message store.
2670 This change adds a proper retry rule check during routing so that the
2671 final cutoff time is checked against the message's age. We only do
2672 this check if there is an address retry record and there is not a
2673 domain retry record; this implies that previous attempts to handle
2674 the address had the retry_use_local_parts option turned on. We use
2675 this as an approximation for the destination being like a local
2676 delivery, as in LMTP.
2678 I suspect this new check makes the old local delivery cutoff check
2679 redundant, but I have not verified this so I left the code in place.
2681 TF/02 Correct gecos expansion when From: is a prefix of the username.
2683 Test 0254 submits a message to Exim with the header
2687 When I ran the test suite under the user fanf2, Exim expanded
2688 the header to contain my full name, whereas it should have added
2689 a Resent-Sender: header. It erroneously treats any prefix of the
2690 username as equal to the username.
2692 This change corrects that bug.
2694 GF/01 DCC debug and logging tidyup
2695 Error conditions log to paniclog rather than rejectlog.
2696 Debug lines prefixed by "DCC: " to remove any ambiguity.
2698 TF/03 Avoid unnecessary rebuilds of lookup-related code.
2700 PP/14 Fix OCSP reinitialisation in SNI handling for Exim/TLS as server.
2701 Bug spotted by Jeremy Harris; was flawed since initial commit.
2702 Would have resulted in OCSP responses post-SNI triggering an Exim
2703 NULL dereference and crash.
2705 JH/13 Add $router_name and $transport_name variables. Bugzilla 308.
2707 PP/15 Define SIOCGIFCONF_GIVES_ADDR for GNU Hurd.
2708 Bug detection, analysis and fix by Samuel Thibault.
2709 Bugzilla 1331, Debian bug #698092.
2711 SC/01 Update eximstats to watch out for senders sending 'HELO [IpAddr]'
2713 JH/14 SMTP PRDR (http://www.eric-a-hall.com/specs/draft-hall-prdr-00.txt).
2714 Server implementation by Todd Lyons, client by JH.
2715 Only enabled when compiled with EXPERIMENTAL_PRDR. A new
2716 config variable "prdr_enable" controls whether the server
2717 advertises the facility. If the client requests PRDR a new
2718 acl_data_smtp_prdr ACL is called once for each recipient, after
2719 the body content is received and before the acl_smtp_data ACL.
2720 The client is controlled by both of: a hosts_try_prdr option
2721 on the smtp transport, and the server advertisement.
2722 Default client logging of deliveries and rejections involving
2723 PRDR are flagged with the string "PRDR".
2725 PP/16 Fix problems caused by timeouts during quit ACLs trying to double
2726 fclose(). Diagnosis by Todd Lyons.
2728 PP/17 Update configure.default to handle IPv6 localhost better.
2729 Patch by Alain Williams (plus minor tweaks).
2732 PP/18 OpenSSL made graceful with empty tls_verify_certificates setting.
2733 This is now consistent with GnuTLS, and is now documented: the
2734 previous undocumented portable approach to treating the option as
2735 unset was to force an expansion failure. That still works, and
2736 an empty string is now equivalent.
2738 PP/19 Renamed DNSSEC-enabling option to "dns_dnssec_ok", to make it
2739 clearer that Exim is using the DO (DNSSEC OK) EDNS0 resolver flag,
2740 not performing validation itself.
2742 PP/20 Added force_command boolean option to pipe transport.
2743 Patch from Nick Koston, of cPanel Inc.
2745 JH/15 AUTH support on callouts (and hence cutthrough-deliveries).
2748 TF/04 Added udpsend ACL modifier and hexquote expansion operator
2750 PP/21 Fix eximon continuous updating with timestamped log-files.
2751 Broken in a format-string cleanup in 4.80, missed when I repaired the
2752 other false fix of the same issue.
2753 Report and fix from Heiko Schlichting.
2756 PP/22 Guard LDAP TLS usage against Solaris LDAP variant.
2757 Report from Prashanth Katuri.
2759 PP/23 Support safari_ecdhe_ecdsa_bug for openssl_options.
2760 It's SecureTransport, so affects any MacOS clients which use the
2761 system-integrated TLS libraries, including email clients.
2763 PP/24 Fix segfault from trying to fprintf() to a NULL stdio FILE* if
2764 using a MIME ACL for non-SMTP local injection.
2765 Report and assistance in diagnosis by Warren Baker.
2767 TL/08 Adjust exiqgrep to be case-insensitive for sender/receiver.
2769 JH/16 Fix comparisons for 64b. Bugzilla 1385.
2771 TL/09 Add expansion variable $authenticated_fail_id to keep track of
2772 last id that failed so it may be referenced in subsequent ACL's.
2774 TL/10 Bugzilla 1375 - Prevent TLS rebinding in ldap. Patch provided by
2777 TL/11 Bugzilla 1382 - Option ldap_require_cert overrides start_tls
2778 ldap library initialization, allowing self-signed CA's to be
2779 used. Also properly sets require_cert option later in code by
2780 using NULL (global ldap config) instead of ldap handle (per
2781 session). Bug diagnosis and testing by alxgomz.
2783 TL/12 Enhanced documentation in the ratelimit.pl script provided in
2784 the src/util/ subdirectory.
2786 TL/13 Bug 1031 - Imported transport SQL logging patch from Axel Rau
2787 renamed to Transport Post Delivery Action by Jeremy Harris, as
2790 TL/14 Bugzilla 1217 - Redis lookup support has been added. It is only enabled
2791 when Exim is compiled with EXPERIMENTAL_REDIS. A new config variable
2792 redis_servers = needs to be configured which will be used by the redis
2793 lookup. Patch from Warren Baker, of The Packet Hub.
2795 TL/15 Fix exiqsumm summary for corner case. Patch provided by Richard Hall.
2797 TL/16 Bugzilla 1289 - Clarify host/ip processing when have errors looking up a
2798 hostname or reverse DNS when processing a host list. Used suggestions
2799 from multiple comments on this bug.
2801 TL/17 Bugzilla 1057 - Multiple clamd TCP targets patch from Mark Zealey.
2803 TL/18 Had previously added a -CONTINUE option to runtest in the test suite.
2804 Missed a few lines, added it to make the runtest require no keyboard
2807 TL/19 Bugzilla 1402 - Test 533 fails if any part of the path to the test suite
2808 contains upper case chars. Make router use caseful_local_part.
2810 TL/20 Bugzilla 1400 - Add AVOID_GNUTLS_PKCS11 build option. Allows GnuTLS
2811 support when GnuTLS has been built with p11-kit.
2817 PP/01 SECURITY: protect DKIM DNS decoding from remote exploit.
2819 This, or similar/improved, will also be change PP/11 of 4.82.
2825 PP/01 Handle short writes when writing local log-files.
2826 In practice, only affects FreeBSD (8 onwards).
2827 Bugzilla 1053, with thanks to Dmitry Isaikin.
2829 NM/01 Bugzilla 949 - Documentation tweak
2831 NM/02 Bugzilla 1093 - eximstats DATA reject detection regexps
2834 NM/03 Bugzilla 1169 - primary_hostname spelling was incorrect in docs.
2836 PP/02 Implemented gsasl authenticator.
2838 PP/03 Implemented heimdal_gssapi authenticator with "server_keytab" option.
2840 PP/04 Local/Makefile support for (AUTH|LOOKUP)_*_PC=foo to use
2841 `pkg-config foo` for cflags/libs.
2843 PP/05 Swapped $auth1/$auth2 for gsasl GSSAPI mechanism, to be more consistent
2844 with rest of GSASL and with heimdal_gssapi.
2846 PP/06 Local/Makefile support for USE_(GNUTLS|OPENSSL)_PC=foo to use
2847 `pkg-config foo` for cflags/libs for the TLS implementation.
2849 PP/07 New expansion variable $tls_bits; Cyrus SASL server connection
2850 properties get this fed in as external SSF. A number of robustness
2851 and debugging improvements to the cyrus_sasl authenticator.
2853 PP/08 cyrus_sasl server now expands the server_realm option.
2855 PP/09 Bugzilla 1214 - Log authentication information in reject log.
2856 Patch by Jeremy Harris.
2858 PP/10 Added dbmjz lookup type.
2860 PP/11 Let heimdal_gssapi authenticator take a SASL message without an authzid.
2862 PP/12 MAIL args handles TAB as well as SP, for better interop with
2863 non-compliant senders.
2864 Analysis and variant patch by Todd Lyons.
2866 NM/04 Bugzilla 1237 - fix cases where printf format usage not indicated
2867 Bug report from Lars Müller <lars@samba.org> (via SUSE),
2868 Patch from Dirk Mueller <dmueller@suse.com>
2870 PP/13 tls_peerdn now print-escaped for spool files.
2871 Observed some $tls_peerdn in wild which contained \n, which resulted
2872 in spool file corruption.
2874 PP/14 TLS fixes for OpenSSL: support TLS 1.1 & 1.2; new "openssl_options"
2875 values; set SSL_MODE_AUTO_RETRY so that OpenSSL will retry a read
2876 or write after TLS renegotiation, which otherwise led to messages
2879 TK/01 Bugzilla 1239 - fix DKIM verification when signature was not inserted
2880 as a tracking header (ie: a signed header comes before the signature).
2881 Patch from Wolfgang Breyha.
2883 JH/01 Bugzilla 660 - Multi-valued attributes from ldap now parseable as a
2884 comma-sep list; embedded commas doubled.
2886 JH/02 Refactored ACL "verify =" logic to table-driven dispatch.
2888 PP/15 LDAP: Check for errors of TLS initialisation, to give correct
2890 Report and patch from Dmitry Banschikov.
2892 PP/16 Removed "dont_insert_empty_fragments" from "openssl_options".
2893 Removed SSL_clear() after SSL_new() which led to protocol negotiation
2894 failures. We appear to now support TLS1.1+ with Exim.
2896 PP/17 OpenSSL: new expansion var $tls_sni, which if used in tls_certificate
2897 lets Exim select keys and certificates based upon TLS SNI from client.
2898 Also option tls_sni on SMTP Transports. Also clear $tls_bits correctly
2899 before an outbound SMTP session. New log_selector, +tls_sni.
2901 PP/18 Bugzilla 1122 - check localhost_number expansion for failure, avoid
2902 NULL dereference. Report and patch from Alun Jones.
2904 PP/19 DNS resolver init changes for NetBSD compatibility. (Risk of breakage
2905 on less well tested platforms). Obviates NetBSD pkgsrc patch-ac.
2906 Not seeing resolver debug output on NetBSD, but suspect this is a
2907 resolver implementation change.
2909 PP/20 Revert part of NM/04, it broke log_path containing %D expansions.
2910 Left warnings. Added "eximon gdb" invocation mode.
2912 PP/21 Defaulting "accept_8bitmime" to true, not false.
2914 PP/22 Added -bw for inetd wait mode support.
2916 PP/23 Added PCRE_CONFIG=yes support to Makefile for using pcre-config to
2917 locate the relevant includes and libraries. Made this the default.
2919 PP/24 Fixed headers_only on smtp transports (was not sending trailing dot).
2920 Bugzilla 1246, report and most of solution from Tomasz Kusy.
2922 JH/03 ${eval } now uses 64-bit and supports a "g" suffix (like to "k" and "m").
2923 This may cause build issues on older platforms.
2925 PP/25 Revamped GnuTLS support, passing tls_require_ciphers to
2926 gnutls_priority_init, ignoring Exim options gnutls_require_kx,
2927 gnutls_require_mac & gnutls_require_protocols (no longer supported).
2928 Added SNI support via GnuTLS too.
2929 Made ${randint:..} supplier available, if using not-too-old GnuTLS.
2931 PP/26 Added EXPERIMENTAL_OCSP for OpenSSL.
2933 PP/27 Applied dnsdb SPF support patch from Janne Snabb.
2934 Applied second patch from Janne, implementing suggestion to default
2935 multiple-strings-in-record handling to match SPF spec.
2937 JH/04 Added expansion variable $tod_epoch_l for a higher-precision time.
2939 PP/28 Fix DCC dcc_header content corruption (stack memory referenced,
2940 read-only, out of scope).
2941 Patch from Wolfgang Breyha, report from Stuart Northfield.
2943 PP/29 Fix three issues highlighted by clang analyser static analysis.
2944 Only crash-plausible issue would require the Cambridge-specific
2945 iplookup router and a misconfiguration.
2946 Report from Marcin Mirosław.
2948 PP/30 Another attempt to deal with PCRE_PRERELEASE, this one less buggy.
2950 PP/31 %D in printf continues to cause issues (-Wformat=security), so for
2951 now guard some of the printf checks behind WANT_DEEPER_PRINTF_CHECKS.
2952 As part of this, removing so much warning spew let me fix some minor
2953 real issues in debug logging.
2955 PP/32 GnuTLS was always using default tls_require_ciphers, due to a missing
2956 assignment on my part. Fixed.
2958 PP/33 Added tls_dh_max_bits option, defaulting to current hard-coded limit
2959 of NSS, for GnuTLS/NSS interop. Problem root cause diagnosis by
2960 Janne Snabb (who went above and beyond: thank you).
2962 PP/34 Validate tls_require_ciphers on startup, since debugging an invalid
2963 string otherwise requires a connection and a bunch more work and it's
2964 relatively easy to get wrong. Should also expose TLS library linkage
2967 PP/35 Pull in <features.h> on Linux, for some portability edge-cases of
2968 64-bit ${eval} (JH/03).
2970 PP/36 Define _GNU_SOURCE in exim.h; it's needed for some releases of
2971 GNU libc to support some of the 64-bit stuff, should not lead to
2972 conflicts. Defined before os.h is pulled in, so if a given platform
2973 needs to override this, it can.
2975 PP/37 Unbreak Cyrus SASL auth: SSF retrieval was incorrect, Exim thought
2976 protection layer was required, which is not implemented.
2977 Bugzilla 1254, patch from Wolfgang Breyha.
2979 PP/38 Overhaul DH prime handling, supply RFC-specified DH primes as built
2980 into Exim, default to IKE id 23 from RFC 5114 (2048 bit). Make
2981 tls_dhparam take prime identifiers. Also unbreak combination of
2982 OpenSSL+DH_params+TLSSNI.
2984 PP/39 Disable SSLv2 by default in OpenSSL support.
2990 PP/01 Solaris build fix for Oracle's LDAP libraries.
2991 Bugzilla 1109, patch from Stephen Usher.
2993 TF/01 HP/UX build fix: avoid arithmetic on a void pointer.
2995 TK/01 DKIM Verification: Fix relaxed canon for empty headers w/o
2998 TF/02 Fix a couple more cases where we did not log the error message
2999 when unlink() failed. See also change 4.74-TF/03.
3001 TF/03 Make the exiwhat support code safe for signals. Previously Exim might
3002 lock up or crash if it happened to be inside a call to libc when it
3003 got a SIGUSR1 from exiwhat.
3005 The SIGUSR1 handler appends the current process status to the process
3006 log which is later printed by exiwhat. It used to use the general
3007 purpose logging code to do this, but several functions it calls are
3008 not safe for signals.
3010 The new output code in the SIGUSR1 handler is specific to the process
3011 log, and simple enough that it's easy to inspect for signal safety.
3012 Removing some special cases also simplifies the general logging code.
3013 Removing the spurious timestamps from the process log simplifies
3016 TF/04 Improved ratelimit ACL condition.
3018 The /noupdate option has been deprecated in favour of /readonly which
3019 has clearer semantics. The /leaky, /strict, and /readonly update modes
3020 are mutually exclusive. The update mode is no longer included in the
3021 database key; it just determines when the database is updated. (This
3022 means that when you upgrade Exim will forget old rate measurements.)
3024 Exim now checks that the per_* options are used with an update mode that
3025 makes sense for the current ACL. For example, when Exim is processing a
3026 message (e.g. acl_smtp_rcpt or acl_smtp_data, etc.) you can specify
3027 per_mail/leaky or per_mail/strict; otherwise (e.g. in acl_smtp_helo) you
3028 must specify per_mail/readonly. If you omit the update mode it defaults to
3029 /leaky where that makes sense (as before) or /readonly where required.
3031 The /noupdate option is now undocumented but still supported for
3032 backwards compatibility. It is equivalent to /readonly except that in
3033 ACLs where /readonly is required you may specify /leaky/noupdate or
3034 /strict/noupdate which are treated the same as /readonly.
3036 A useful new feature is the /count= option. This is a generalization
3037 of the per_byte option, so that you can measure the throughput of other
3038 aggregate values. For example, the per_byte option is now equivalent
3039 to per_mail/count=${if >{0}{$message_size} {0} {$message_size} }.
3041 The per_rcpt option has been generalized using the /count= mechanism
3042 (though it's more complicated than the per_byte equivalence). When it is
3043 used in acl_smtp_rcpt, the per_rcpt option adds recipients to the
3044 measured rate one at a time; if it is used later (e.g. in acl_smtp_data)
3045 or in a non-SMTP ACL it adds all the recipients in one go. (The latter
3046 /count=$recipients_count behaviour used to work only in non-SMTP ACLs.)
3047 Note that using per_rcpt with a non-readonly update mode in more than
3048 one ACL will cause the recipients to be double-counted. (The per_mail
3049 and per_byte options don't have this problem.)
3051 The handling of very low rates has changed slightly. If the computed rate
3052 is less than the event's count (usually one) then this event is the first
3053 after a long gap. In this case the rate is set to the same as this event's
3054 count, so that the first message of a spam run is counted properly.
3056 The major new feature is a mechanism for counting the rate of unique
3057 events. The new per_addr option counts the number of different
3058 recipients that someone has sent messages to in the last time period. It
3059 behaves like per_rcpt if all the recipient addresses are different, but
3060 duplicate recipient addresses do not increase the measured rate. Like
3061 the /count= option this is a general mechanism, so the per_addr option
3062 is equivalent to per_rcpt/unique=$local_part@$domain. You can, for
3063 example, measure the rate that a client uses different sender addresses
3064 with the options per_mail/unique=$sender_address. There are further
3065 details in the main documentation.
3067 TF/05 Removed obsolete $Cambridge$ CVS revision strings.
3069 TF/06 Removed a few PCRE remnants.
3071 TF/07 Automatically extract Exim's version number from tags in the git
3072 repository when doing development or release builds.
3074 PP/02 Raise smtp_cmd_buffer_size to 16kB.
3075 Bugzilla 879. Patch from Paul Fisher.
3077 PP/03 Implement SSL-on-connect outbound with protocol=smtps on smtp transport.
3078 Heavily based on revision 40f9a89a from Simon Arlott's tree.
3081 PP/04 Use .dylib instead of .so for dynamic library loading on MacOS.
3083 PP/05 Variable $av_failed, true if the AV scanner deferred.
3084 Bugzilla 1078. Patch from John Horne.
3086 PP/06 Stop make process more reliably on build failure.
3087 Bugzilla 1087. Patch from Heiko Schlittermann.
3089 PP/07 Make maildir_use_size_file an _expandable_ boolean.
3090 Bugzilla 1089. Patch from Heiko Schlittermann.
3092 PP/08 Handle ${run} returning more data than OS pipe buffer size.
3093 Bugzilla 1131. Patch from Holger Weiß.
3095 PP/09 Handle IPv6 addresses with SPF.
3096 Bugzilla 860. Patch from Wolfgang Breyha.
3098 PP/10 GnuTLS: support TLS 1.2 & 1.1.
3100 Use gnutls_certificate_verify_peers2() [patch from Andreas Metzler].
3103 PP/11 match_* no longer expand right-hand-side by default.
3104 New compile-time build option, EXPAND_LISTMATCH_RHS.
3105 New expansion conditions, "inlist", "inlisti".
3107 PP/12 fix uninitialised greeting string from PP/03 (smtps client support).
3109 PP/13 shell and compiler warnings fixes for RC1-RC4 changes.
3111 PP/14 fix log_write() format string regression from TF/03.
3112 Bugzilla 1152. Patch from Dmitry Isaikin.
3118 PP/01 The new ldap_require_cert option would segfault if used. Fixed.
3120 PP/02 Harmonised TLS library version reporting; only show if debugging.
3121 Layout now matches that introduced for other libraries in 4.74 PP/03.
3123 PP/03 New openssl_options items: no_sslv2 no_sslv3 no_ticket no_tlsv1
3125 PP/04 New "dns_use_edns0" global option.
3127 PP/05 Don't segfault on misconfiguration of ref:name exim-user as uid.
3130 PP/06 Extra paranoia around buffer usage at the STARTTLS transition.
3131 nb: Exim is not vulnerable to http://www.kb.cert.org/vuls/id/555316
3133 TK/01 Updated PolarSSL code to 0.14.2.
3134 Bugzilla 1097. Patch from Andreas Metzler.
3136 PP/07 Catch divide-by-zero in ${eval:...}.
3137 Fixes bugzilla 1102.
3139 PP/08 Condition negation of bool{}/bool_lax{} did not negate. Fixed.
3142 TK/02 Bugzilla 1106: CVE-2011-1764 - DKIM log line was subject to a
3143 format-string attack -- SECURITY: remote arbitrary code execution.
3145 TK/03 SECURITY - DKIM signature header parsing was double-expanded, second
3146 time unintentionally subject to list matching rules, letting the header
3147 cause arbitrary Exim lookups (of items which can occur in lists, *not*
3148 arbitrary string expansion). This allowed for information disclosure.
3150 PP/09 Fix another SIGFPE (x86) in ${eval:...} expansion, this time related to
3151 INT_MIN/-1 -- value coerced to INT_MAX.
3157 NM/01 Workaround for PCRE version dependency in version reporting
3160 TF/01 Update valgrind.h and memcheck.h to copies from valgrind-3.6.0.
3161 This fixes portability to compilers other than gcc, notably
3162 Solaris CC and HP-UX CC. Fixes Bugzilla 1050.
3164 TF/02 Bugzilla 139: Avoid using the += operator in the modular lookup
3165 makefiles for portability to HP-UX and POSIX correctness.
3167 PP/01 Permit LOOKUP_foo enabling on the make command-line.
3168 Also via indented variable definition in the Makefile.
3169 (Debugging by Oliver Heesakkers).
3171 PP/02 Restore caching of spamd results with expanded spamd_address.
3172 Patch from author of expandable spamd_address patch, Wolfgang Breyha.
3174 PP/03 Build issue: lookups-Makefile now exports LC_ALL=C
3175 Improves build reliability. Fix from: Frank Elsner
3177 NM/02 Fix wide character breakage in the rfc2047 coding
3178 Fixes bug 1064. Patch from Andrey N. Oktyabrski
3180 NM/03 Allow underscore in dnslist lookups
3181 Fixes bug 1026. Patch from Graeme Fowler
3183 PP/04 Bugzilla 230: Support TLS-enabled LDAP (in addition to ldaps).
3184 Code patches from Adam Ciarcinski of NetBSD.
3186 NM/04 Fixed exiqgrep to cope with mailq missing size issue
3189 PP/05 Bugzilla 1083: when lookup expansion defers, escape the output which
3190 is logged, to avoid truncation. Patch from John Horne.
3192 PP/06 Bugzilla 1042: implement freeze_signal on pipe transports.
3193 Patch from Jakob Hirsch.
3195 PP/07 Bugzilla 1061: restrict error messages sent over SMTP to not reveal
3196 SQL string expansion failure details.
3197 Patch from Andrey Oktyabrski.
3199 PP/08 Bugzilla 486: implement %M datestamping in log filenames.
3200 Patch from Simon Arlott.
3202 PP/09 New lookups functionality failed to compile on old gcc which rejects
3203 extern declarations in function scope.
3204 Patch from Oliver Fleischmann
3206 PP/10 Use sig_atomic_t for flags set from signal handlers.
3207 Check getgroups() return and improve debugging.
3208 Fixed developed for diagnosis in bug 927 (which turned out to be
3211 PP/11 Bugzilla 1055: Update $message_linecount for maildir_tag.
3212 Patch from Mark Zealey.
3214 PP/12 Bugzilla 1056: Improved spamd server selection.
3215 Patch from Mark Zealey.
3217 PP/13 Bugzilla 1086: Deal with maildir quota file races.
3218 Based on patch from Heiko Schlittermann.
3220 PP/14 Bugzilla 1019: DKIM multiple signature generation fix.
3221 Patch from Uwe Doering, sign-off by Michael Haardt.
3223 NM/05 Fix to spam.c to accommodate older gcc versions which dislike
3224 variable declaration deep within a block. Bug and patch from
3227 PP/15 lookups-Makefile IRIX compatibility coercion.
3229 PP/16 Make DISABLE_DKIM build knob functional.
3231 NM/06 Bugzilla 968: child_open_uid: restore default SIGPIPE handler
3232 Patch by Simon Arlott
3234 TF/03 Fix valgrind.h portability to C89 compilers that do not support
3235 variable argument macros. Our copy now differs from upstream.
3241 TF/01 Failure to get a lock on a hints database can have serious
3242 consequences so log it to the panic log.
3244 TF/02 Log LMTP confirmation messages in the same way as SMTP,
3245 controlled using the smtp_confirmation log selector.
3247 TF/03 Include the error message when we fail to unlink a spool file.
3249 DW/01 Bugzilla 139: Support dynamically loaded lookups as modules.
3250 With thanks to Steve Haslam, Johannes Berg & Serge Demonchaux
3251 for maintaining out-of-tree patches for some time.
3253 PP/01 Bugzilla 139: Documentation and portability issues.
3254 Avoid GNU Makefile-isms, let Exim continue to build on BSD.
3255 Handle per-OS dynamic-module compilation flags.
3257 PP/02 Let /dev/null have normal permissions.
3258 The 4.73 fixes were a little too stringent and complained about the
3259 permissions on /dev/null. Exempt it from some checks.
3260 Reported by Andreas M. Kirchwitz.
3262 PP/03 Report version information for many libraries, including
3263 Exim version information for dynamically loaded libraries. Created
3264 version.h, now support a version extension string for distributors
3265 who patch heavily. Dynamic module ABI change.
3267 PP/04 CVE-2011-0017 - check return value of setuid/setgid. This is a
3268 privilege escalation vulnerability whereby the Exim run-time user
3269 can cause root to append content of the attacker's choosing to
3272 PP/05 Bugzilla 1041: merged DCC maintainer's fixes for return code.
3275 PP/06 Bugzilla 1071: fix delivery logging with untrusted macros.
3276 If dropping privileges for untrusted macros, we disabled normal logging
3277 on the basis that it would fail; for the Exim run-time user, this is not
3278 the case, and it resulted in successful deliveries going unlogged.
3279 Fixed. Reported by Andreas Metzler.
3285 PP/01 Date: & Message-Id: revert to normally being appended to a message,
3286 only prepend for the Resent-* case. Fixes regression introduced in
3287 Exim 4.70 by NM/22 for Bugzilla 607.
3289 PP/02 Include check_rfc2047_length in configure.default because we're seeing
3290 increasing numbers of administrators be bitten by this.
3292 JJ/01 Added DISABLE_DKIM and comment to src/EDITME
3294 PP/03 Bugzilla 994: added openssl_options main configuration option.
3296 PP/04 Bugzilla 995: provide better SSL diagnostics on failed reads.
3298 PP/05 Bugzilla 834: provide a permit_coredump option for pipe transports.
3300 PP/06 Adjust NTLM authentication to handle SASL Initial Response.
3302 PP/07 If TLS negotiated an anonymous cipher, we could end up with SSL but
3303 without a peer certificate, leading to a segfault because of an
3304 assumption that peers always have certificates. Be a little more
3305 paranoid. Problem reported by Martin Tscholak.
3307 PP/08 Bugzilla 926: switch ClamAV to use the new zINSTREAM API for content
3308 filtering; old API available if built with WITH_OLD_CLAMAV_STREAM=yes
3309 NB: ClamAV planning to remove STREAM in "middle of 2010".
3310 CL also introduces -bmalware, various -d+acl logging additions and
3311 more caution in buffer sizes.
3313 PP/09 Implemented reverse_ip expansion operator.
3315 PP/10 Bugzilla 937: provide a "debug" ACL control.
3317 PP/11 Bugzilla 922: Documentation dusting, patch provided by John Horne.
3319 PP/12 Bugzilla 973: Implement --version.
3321 PP/13 Bugzilla 752: Refuse to build/run if Exim user is root/0.
3323 PP/14 Build without WITH_CONTENT_SCAN. Path from Andreas Metzler.
3325 PP/15 Bugzilla 816: support multiple condition rules on Routers.
3327 PP/16 Add bool_lax{} expansion operator and use that for combining multiple
3328 condition rules, instead of bool{}. Make both bool{} and bool_lax{}
3329 ignore trailing whitespace.
3331 JJ/02 prevent non-panic DKIM error from being sent to paniclog
3333 JJ/03 added tcp_wrappers_daemon_name to allow host entries other than
3336 PP/17 Fix malware regression for cmdline scanner introduced in PP/08.
3337 Notification from Dr Andrew Aitchison.
3339 PP/18 Change ClamAV response parsing to be more robust and to handle ClamAV's
3340 ExtendedDetectionInfo response format.
3341 Notification from John Horne.
3343 PP/19 OpenSSL 1.0.0a compatibility const-ness change, should be backwards
3346 PP/20 Added a CONTRIBUTING file. Fixed the documentation build to use http:
3347 XSL and documented dependency on system catalogs, with examples of how
3350 DW/21 Added Valgrind hooks in store.c to help it capture out-of-bounds store
3353 DW/22 Bugzilla 1044: CVE-2010-4345 - partial fix: restrict default behaviour
3354 of CONFIGURE_OWNER and CONFIGURE_GROUP options to no longer allow a
3355 configuration file which is writeable by the Exim user or group.
3357 DW/23 Bugzilla 1044: CVE-2010-4345 - part two: extend checks for writeability
3358 of configuration files to cover files specified with the -C option if
3359 they are going to be used with root privileges, not just the default
3362 DW/24 Bugzilla 1044: CVE-2010-4345 - part three: remove ALT_CONFIG_ROOT_ONLY
3363 option (effectively making it always true).
3365 DW/25 Add TRUSTED_CONFIG_PREFIX_FILE option to allow alternative configuration
3366 files to be used while preserving root privileges.
3368 DW/26 Set FD_CLOEXEC on SMTP sockets after forking in the daemon, to ensure
3369 that rogue child processes cannot use them.
3371 PP/27 Bugzilla 1047: change the default for system_filter_user to be the Exim
3372 run-time user, instead of root.
3374 PP/28 Add WHITELIST_D_MACROS option to let some macros be overridden by the
3375 Exim run-time user without dropping privileges.
3377 DW/29 Remove use of va_copy() which breaks pre-C99 systems. Duplicate the
3378 result string, instead of calling string_vformat() twice with the same
3381 DW/30 Allow TRUSTED_CONFIG_PREFIX_FILE only for Exim or CONFIGURE_OWNER, not
3382 for other users. Others should always drop root privileges if they use
3383 -C on the command line, even for a whitelisted configure file.
3385 DW/31 Turn TRUSTED_CONFIG_PREFIX_FILE into TRUSTED_CONFIG_FILE. No prefixes.
3387 NM/01 Fixed bug #1002 - Message loss when using multiple deliveries
3393 JJ/01 installed exipick 20100104.1, adding $max_received_linelength,
3394 $data_path, and $header_path variables; fixed documentation bugs and
3397 JJ/02 installed exipick 20100222.0, added --input-dir and --finput to allow
3398 exipick to access non-standard spools, including the "frozen" queue
3401 NM/01 Bugzilla 965: Support mysql stored procedures.
3402 Patch from Alain Williams
3404 NM/02 Bugzilla 961: Spacing fix (syntax error) on Makefile directives for NetBSD
3406 NM/03 Bugzilla 955: Documentation fix for max_rcpts.
3407 Patch from Andreas Metzler
3409 NM/04 Bugzilla 954: Fix for unknown responses from Dovecot authenticator.
3410 Patch from Kirill Miazine
3412 NM/05 Bugzilla 671: Added umask to procmail example.
3414 JJ/03 installed exipick 20100323.0, fixing doc bug
3416 NM/06 Bugzilla 988: CVE-2010-2023 - prevent hardlink attack on sticky mail
3417 directory. Notification and patch from Dan Rosenberg.
3419 TK/01 PDKIM: Upgrade PolarSSL files to upstream version 0.12.1.
3421 TK/02 Improve log output when DKIM signing operation fails.
3423 MH/01 Treat the transport option dkim_domain as a colon separated
3424 list, not as a single string, and sign the message with each element,
3425 omitting multiple occurences of the same signer.
3427 NM/07 Null terminate DKIM strings, Null initialise DKIM variable
3428 Bugzilla 985, 986. Patch by Simon Arlott
3430 NM/08 Bugzilla 967. dnsdb DNS TXT record bug fix (DKIM-related)
3431 Patch by Simon Arlott
3433 PP/01 Bugzilla 989: CVE-2010-2024 - work round race condition on
3434 MBX locking. Notification from Dan Rosenberg.
3440 TK/01 Bugzilla 912: Fix DKIM segfault on empty headers/body.
3442 NM/01 Bugzilla 913: Documentation fix for gnutls_* options.
3444 NM/02 Bugzilla 722: Documentation for randint. Better randomness defaults.
3446 NM/03 Bugzilla 847: Enable DNSDB lookup by default.
3448 NM/04 Bugzilla 915: Flag broken perl installation during build.
3454 TK/01 Added patch by Johannes Berg that expands the main option
3455 "spamd_address" if it starts with a dollar sign.
3457 TK/02 Write list of recipients to X-Envelope-Sender header when building
3458 the mbox-format spool file for content scanning (suggested by Jakob
3461 TK/03 Added patch by Wolfgang Breyha that adds experimental DCC
3462 (http://www.dcc-servers.net/) support via dccifd. Activated by
3463 setting EXPERIMENTAL_DCC=yes in Local/Makefile.
3465 TK/04 Bugzilla 673: Add f-protd malware scanner support. Patch submitted
3466 by Mark Daniel Reidel <mr@df.eu>.
3468 NM/01 Bugzilla 657: Embedded PCRE removed from the exim source tree.
3469 When building exim an external PCRE library is now needed -
3470 PCRE is a system library on the majority of modern systems.
3471 See entry on PCRE_LIBS in EDITME file.
3473 NM/02 Bugzilla 646: Removed unwanted C/R in Dovecot authenticator
3474 conversation. Added nologin parameter to request.
3475 Patch contributed by Kirill Miazine.
3477 TF/01 Do not log submission mode rewrites if they do not change the address.
3479 TF/02 Bugzilla 662: Fix stack corruption before exec() in daemon.c.
3481 NM/03 Bugzilla 602: exicyclog now handles panic log, and creates empty
3482 log files in place. Contributed by Roberto Lima.
3484 NM/04 Bugzilla 667: Close socket used by dovecot authenticator.
3486 TF/03 Bugzilla 615: When checking the local_parts router precondition
3487 after a local_part_suffix or local_part_prefix option, Exim now
3488 does not use the address's named list lookup cache, since this
3489 contains cached lookups for the whole local part.
3491 NM/05 Bugzilla 521: Integrated SPF Best Guess support contributed by
3492 Robert Millan. Documentation is in experimental-spec.txt.
3494 TF/04 Bugzilla 668: Fix parallel build (make -j).
3496 NM/05.2 Bugzilla 437: Prevent Maildir aux files being created with mode 000.
3498 NM/05.3 Bugzilla 598: Improvement to Dovecot authenticator handling.
3499 Patch provided by Jan Srzednicki.
3501 TF/05 Leading white space used to be stripped from $spam_report which
3502 wrecked the formatting. Now it is preserved.
3504 TF/06 Save $spam_score, $spam_bar, and $spam_report in spool files, so
3505 that they are available at delivery time.
3507 TF/07 Fix the way ${extract is skipped in the untaken branch of a conditional.
3509 TF/08 TLS error reporting now respects the incoming_interface and
3510 incoming_port log selectors.
3512 TF/09 Produce a more useful error message if an SMTP transport's hosts
3513 setting expands to an empty string.
3515 NM/06 Bugzilla 744: EXPN did not work under TLS.
3516 Patch provided by Phil Pennock.
3518 NM/07 Bugzilla 769: Extraneous comma in usage fprintf
3519 Patch provided by Richard Godbee.
3521 NM/08 Fixed erroneous documentation references to smtp_notquit_acl to be
3522 acl_smtp_notquit, added index entry.
3524 NM/09 Bugzilla 787: Potential buffer overflow in string_format.
3525 Patch provided by Eugene Bujak.
3527 NM/10 Bugzilla 770: Problem on some platforms modifying the len parameter to
3528 accept(). Patch provided by Maxim Dounin.
3530 NM/11 Bugzilla 749: Preserve old behaviour of blanks comparing equal to zero.
3531 Patch provided by Phil Pennock.
3533 NM/12 Bugzilla 497: Correct behaviour of exiwhat when no config exists.
3535 NM/13 Bugzilla 590: Correct handling of Resent-Date headers.
3536 Patch provided by Brad "anomie" Jorsch.
3538 NM/14 Bugzilla 622: Added timeout setting to transport filter.
3539 Patch provided by Dean Brooks.
3541 TK/05 Add native DKIM support (does not depend on external libraries).
3543 NM/15 Bugzilla 854: Removed code that symlinks to pcre as its no longer useful.
3544 Patch provided by Graeme Fowler.
3546 NM/16 Bugzilla 851: Documentation example syntax fix.
3548 NM/17 Changed NOTICE file to remove references to embedded PCRE.
3550 NM/18 Bugzilla 894: Fix issue with very long lines including comments in
3553 NM/19 Bugzilla 745: TLS version reporting.
3554 Patch provided by Phil Pennock.
3556 NM/20 Bugzilla 167: bool: condition support.
3557 Patch provided by Phil Pennock.
3559 NM/21 Bugzilla 665: gnutls_compat_mode to allow compatibility with broken
3560 clients. Patch provided by Phil Pennock.
3562 NM/22 Bugzilla 607: prepend (not append) Resent-Message-ID and Resent-Date.
3563 Patch provided by Brad "anomie" Jorsch.
3565 NM/23 Bugzilla 687: Fix misparses in eximstats.
3566 Patch provided by Heiko Schlittermann.
3568 NM/24 Bugzilla 688: Fix exiwhat to handle log_selector = +pid.
3569 Patch provided by Heiko Schlittermann.
3571 NM/25 Bugzilla 727: Use transport mode as default mode for maildirsize file.
3572 plus update to original patch.
3574 NM/26 Bugzilla 799: Documentation correction for ratelimit.
3576 NM/27 Bugzilla 802: Improvements to local interface IP addr detection.
3577 Patch provided by David Brownlee.
3579 NM/28 Bugzilla 807: Improvements to LMTP delivery logging.
3581 NM/29 Bugzilla 862, 866, 875: Documentation bugfixes.
3583 NM/30 Bugzilla 888: TLS documentation bugfixes.
3585 NM/31 Bugzilla 896: Dovecot buffer overrun fix.
3587 NM/32 Bugzilla 889: Change all instances of "expr" in shell scripts to "expr --"
3588 Unlike the original bugzilla I have changed all shell scripts in src tree.
3590 NM/33 Bugzilla 898: Transport filter timeout fix.
3591 Patch by Todd Rinaldo.
3593 NM/34 Bugzilla 901: Fix sign/unsigned and UTF mismatches.
3594 Patch by Serge Demonchaux.
3596 NM/35 Bugzilla 39: Base64 decode bug fixes.
3597 Patch by Jakob Hirsch.
3599 NM/36 Bugzilla 909: Correct connect() call in dcc code.
3601 NM/37 Bugzilla 910: Correct issue with relaxed/simple handling.
3603 NM/38 Bugzilla 908: Removed NetBSD3 support as no longer needed.
3605 NM/39 Bugzilla 911: Fixed MakeLinks build script.
3611 TK/01 Add preliminary DKIM support. Currently requires a forked version of
3612 ALT-N's libdkim that I have put here:
3613 http://duncanthrax.net/exim-experimental/
3615 Note to Michael Haardt: I had to rename some vars in sieve.c. They
3616 were called 'true' and it seems that C99 defines that as a reserved
3617 keyword to be used with 'bool' variable types. That means you could
3618 not include C99-style headers which use bools without triggering
3619 build errors in sieve.c.
3621 NM/01 Bugzilla 592: --help option is handled incorrectly if exim is invoked
3622 as mailq or other aliases. Changed the --help handling significantly
3623 to do whats expected. exim_usage() emits usage/help information.
3625 SC/01 Added the -bylocaldomain option to eximstats.
3627 NM/02 Bugzilla 619: Defended against bad data coming back from gethostbyaddr.
3629 NM/03 Bugzilla 613: Documentation fix for acl_not_smtp.
3631 NM/04 Bugzilla 628: PCRE update to 7.4 (work done by John Hall).
3637 PH/01 Another patch from the Sieve maintainer.
3639 PH/02 When an IPv6 address is converted to a string for single-key lookup
3640 in an address list (e.g. for an item such as "net24-dbm;/net/works"),
3641 dots are used instead of colons so that keys in lsearch files need not
3642 contain colons. This was done some time before quoting was made available
3643 in lsearch files. However, iplsearch files do require colons in IPv6 keys
3644 (notated using the quote facility) so as to distinguish them from IPv4
3645 keys. This meant that lookups for IP addresses in host lists did not work
3646 for iplsearch lookups.
3648 This has been fixed by arranging for IPv6 addresses to be expressed with
3649 colons if the lookup type is iplsearch. This is not incompatible, because
3650 previously such lookups could never work.
3652 The situation is now rather anomalous, since one *can* have colons in
3653 ordinary lsearch keys. However, making the change in all cases is
3654 incompatible and would probably break a number of configurations.
3656 TK/01 Change PRVS address formatting scheme to reflect latests BATV draft
3659 MH/01 The "spam" ACL condition code contained a sscanf() call with a %s
3660 conversion specification without a maximum field width, thereby enabling
3661 a rogue spamd server to cause a buffer overflow. While nobody in their
3662 right mind would setup Exim to query an untrusted spamd server, an
3663 attacker that gains access to a server running spamd could potentially
3664 exploit this vulnerability to run arbitrary code as the Exim user.
3666 TK/02 Bugzilla 502: Apply patch to make the SPF-Received: header use
3667 $primary_hostname instead of what libspf2 thinks the hosts name is.
3669 MH/02 The dsearch lookup now uses lstat(2) instead of stat(2) to look for
3670 a directory entry by the name of the lookup key. Previously, if a
3671 symlink pointed to a non-existing file or a file in a directory that
3672 Exim lacked permissions to read, a lookup for a key matching that
3673 symlink would fail. Now it is enough that a matching directory entry
3674 exists, symlink or not. (Bugzilla 503.)
3676 PH/03 The body_linecount and body_zerocount variables are now exported in the
3679 PH/04 Added the $dnslist_matched variable.
3681 PH/05 Unset $tls_cipher and $tls_peerdn before making a connection as a client.
3682 This means they are set thereafter only if the connection becomes
3685 PH/06 Added the client_condition to authenticators so that some can be skipped
3686 by clients under certain conditions.
3688 PH/07 The error message for a badly-placed control=no_multiline_responses left
3689 "_responses" off the end of the name.
3691 PH/08 Added -Mvc to output a copy of a message in RFC 2822 format.
3693 PH/09 Tidied the code for creating ratelimiting keys, creating them explicitly
3694 (without spaces) instead of just copying the configuration text.
3696 PH/10 Added the /noupdate option to the ratelimit ACL condition.
3698 PH/11 Added $max_received_linelength.
3700 PH/12 Added +ignore_defer and +include_defer to host lists.
3702 PH/13 Installed PCRE version 7.2. This needed some changes because of the new
3703 way in which PCRE > 7.0 is built.
3705 PH/14 Implemented queue_only_load_latch.
3707 PH/15 Removed an incorrect (int) cast when reading the value of SIZE in a
3708 MAIL command. The effect was to mangle the value on 64-bit systems.
3710 PH/16 Another patch from the Sieve maintainer.
3712 PH/17 Added the NOTQUIT ACL, based on a patch from Ted Cooper.
3714 PH/18 If a system quota error occurred while trying to create the file for
3715 a maildir delivery, the message "Mailbox is full" was not appended to the
3716 bounce if the delivery eventually timed out. Change 4.67/27 below applied
3717 only to a quota excession during the actual writing of the file.
3719 PH/19 It seems that peer DN values may contain newlines (and other non-printing
3720 characters?) which causes problems in log lines. The DN values are now
3721 passed through string_printing() before being added to log lines.
3723 PH/20 Added the "servers=" facility to MySQL and PostgreSQL lookups. (Oracle
3724 and InterBase are left for another time.)
3726 PH/21 Added message_body_newlines option.
3728 PH/22 Guard against possible overflow in moan_check_errorcopy().
3730 PH/23 POSIX allows open() to be a macro; guard against that.
3732 PH/24 If the recipient of an error message contained an @ in the local part
3733 (suitably quoted, of course), incorrect values were put in $domain and
3734 $local_part during the evaluation of errors_copy.
3740 MH/01 Fix for bug #448, segfault in Dovecot authenticator when interface_address
3741 is unset (happens when testing with -bh and -oMi isn't used). Thanks to
3744 PH/01 Added a new log selector smtp_no_mail, to log SMTP sessions that do not
3745 issue a MAIL command.
3747 PH/02 In an ACL statement such as
3749 deny dnslists = X!=127.0.0.2 : X=127.0.0.2
3751 if a client was not listed at all, or was listed with a value other than
3752 127.0.0.2, in the X list, but was listed with 127.0.0.2 in the Y list,
3753 the condition was not true (as it should be), so access was not denied.
3754 The bug was that the ! inversion was incorrectly passed on to the second
3755 item. This has been fixed.
3757 PH/03 Added additional dnslists conditions == and =& which are different from
3758 = and & when the dns lookup returns more than one IP address.
3760 PH/04 Added gnutls_require_{kx,mac,protocols} to give more control over the
3761 cipher suites used by GnuTLS. These options are ignored by OpenSSL.
3763 PH/05 After discussion on the list, added a compile time option ENABLE_DISABLE_
3764 FSYNC, which compiles an option called disable_fsync that allows for
3765 bypassing fsync(). The documentation is heavily laced with warnings.
3767 SC/01 Updated eximstats to collate all SpamAssassin rejects into one bucket.
3769 PH/06 Some tidies to the infrastructure of the Test Suite that is concerned
3770 with the auxiliary C programs that it uses: (1) Arrange for BIND_8_COMPAT
3771 to be defined when compiling on OSX (Darwin); (2) Tidies to the Makefile,
3772 including adding "make clean"; (3) Added -fPIC when compiling the test
3773 dynamically loaded module, to get rid of a warning.
3775 MH/02 Fix for bug #451, causing paniclog entries to be written if a bounce
3776 message fails, move_frozen_messages = true and ignore_bounce_errors_after
3777 = 0s. The bug is otherwise harmless.
3779 PH/07 There was a bug in the dovecot authenticator such that the value of
3780 $auth1 could be overwritten, and so not correctly preserved, after a
3781 successful authentication. This usually meant that the value preserved by
3782 the server_setid option was incorrect.
3784 PH/08 Added $smtp_count_at_connection_start, deliberately with a long name.
3786 PH/09 Installed PCRE release 7.0.
3788 PH/10 The acl_not_smtp_start ACL was, contrary to the documentation, not being
3789 run for batched SMTP input. It is now run at the start of every message
3790 in the batch. While fixing this I discovered that the process information
3791 (output by running exiwhat) was not always getting set for -bs and -bS
3792 input. This is fixed, and it now also says "batched" for BSMTP.
3794 PH/11 Added control=no_pipelining.
3796 PH/12 Added $sending_ip_address and $sending_port (mostly Magnus Holmgren's
3797 patch, slightly modified), and move the expansion of helo_data till after
3798 the connection is made in the smtp transport (so it can use these
3801 PH/13 Added ${rfc2047d: to decoded RFC 2047 strings.
3803 PH/14 Added log_selector = +pid.
3805 PH/15 Flush SMTP output before delaying, unless control=no_delay_flush is set.
3807 PH/16 Add ${if forany and ${if forall.
3809 PH/17 Added dsn_from option to vary the From: line in DSNs.
3811 PH/18 Flush SMTP output before performing a callout, unless control =
3812 no_callout_flush is set.
3814 PH/19 Change 4.64/PH/36 introduced a bug: when address_retry_include_sender
3815 was true (the default) a successful delivery failed to delete the retry
3816 item, thus causing premature timeout of the address. The bug is now
3819 PH/20 Added hosts_avoid_pipelining to the smtp transport.
3821 PH/21 Long custom messages for fakedefer and fakereject are now split up
3822 into multiline responses in the same way that messages for "deny" and
3823 other ACL rejections are.
3825 PH/22 Applied Jori Hamalainen's speed-up changes and typo fixes to exigrep,
3826 with slight modification.
3828 PH/23 Applied sieve patches from the maintainer "tracking the latest notify
3829 draft, changing the syntax and factoring some duplicate code".
3831 PH/24 When the log selector "outgoing_port" was set, the port was shown as -1
3832 for deliveries of the second and subsequent messages over the same SMTP
3835 PH/25 Applied Magnus Holmgren's patch for ${addresses, ${map, ${filter, and
3836 ${reduce, with only minor "tidies".
3838 SC/02 Applied Daniel Tiefnig's patch to improve the '($parent) =' pattern match.
3840 PH/26 Added a "continue" ACL modifier that does nothing, for the benefit of its
3841 expansion side effects.
3843 PH/27 When a message times out after an over-quota error from an Exim-imposed
3844 quota, the bounce message says "mailbox is full". This message was not
3845 being given when it was a system quota that was exceeded. It now should
3848 MH/03 Made $recipients available in local_scan(). local_scan() already has
3849 better access to the recipient list through recipients_list[], but
3850 $recipients can be useful in postmaster-provided expansion strings.
3852 PH/28 The $smtp_command and $smtp_command_argument variables were not correct
3853 in the case of a MAIL command with additional options following the
3854 address, for example: MAIL FROM:<foo@bar> SIZE=1234. The option settings
3855 were accidentally chopped off.
3857 PH/29 SMTP synchronization checks are implemented when a command is read -
3858 there is a check that no more input is waiting when there shouldn't be
3859 any. However, for some commands, a delay in an ACL can mean that it is
3860 some time before the response is written. In this time, more input might
3861 arrive, invalidly. So now there are extra checks after an ACL has run for
3862 HELO/EHLO and after the predata ACL, and likewise for MAIL and RCPT when
3863 pipelining has not been advertised.
3865 PH/30 MH's patch to allow iscntrl() characters to be list separators.
3867 PH/31 Unlike :fail:, a custom message specified with :defer: was not being
3868 returned in the SMTP response when smtp_return_error_details was false.
3869 This has been fixed.
3871 PH/32 Change the Dovecot authenticator to use read() and write() on the socket
3872 instead of the C I/O that was originally supplied, because problems were
3873 reported on Solaris.
3875 PH/33 Compile failed with OpenSSL 0.9.8e. This was due to a coding error in
3876 Exim which did not show up earlier: it was assuming that a call to
3877 SSL_CTX_set_info_callback() might give an error value. In fact, there is
3878 no error. In previous releases of OpenSSL, SSL_CTX_set_info_callback()
3879 was a macro that became an assignment, so it seemed to work. This has
3880 changed to a proper function call with a void return, hence the compile
3881 error. Exim's code has been fixed.
3883 PH/34 Change HDA_SIZE in oracle.c from 256 to 512. This is needed for 64-bit
3886 PH/35 Applied a patch from the Sieve maintainer which fixes a bug in "notify".
3888 PH/36 Applied John Jetmore's patch to add -v functionality to exigrep.
3890 PH/37 If a message is not accepted after it has had an id assigned (e.g.
3891 because it turns out to be too big or there is a timeout) there is no
3892 "Completed" line in the log. When some messages of this type were
3893 selected by exigrep, they were listed as "not completed". Others were
3894 picked up by some special patterns. I have improved the selection
3895 criteria to be more general.
3897 PH/38 The host_find_failed option in the manualroute router can now be set
3898 to "ignore", to completely ignore a host whose IP address cannot be
3899 found. If all hosts are ignored, the behaviour is controlled by the new
3900 host_all_ignored option.
3902 PH/39 In a list of hosts for manualroute, if one item (either because of multi-
3903 homing or because of multiple MX records with /mx) generated more than
3904 one IP address, and the following item turned out to be the local host,
3905 all the secondary addresses of the first item were incorrectly removed
3906 from the list, along with the local host and any following hosts (which
3907 is what is supposed to happen).
3909 PH/40 When Exim receives a message, it writes the login name, uid, and gid of
3910 whoever called Exim into the -H file. In the case of the daemon it was
3911 behaving confusingly. When first started, it used values for whoever
3912 started the daemon, but after a SIGHUP it used the Exim user (because it
3913 calls itself on a restart). I have changed the code so that it now always
3916 PH/41 (Following a suggestion from Tony Finch) If all the RCPT commands in a
3917 message are rejected with the same error (e.g. no authentication or bad
3918 sender address), and a DATA command is nevertheless sent (as can happen
3919 with PIPELINING or a stupid MUA), the error message that was given to the
3920 RCPT commands is included in the rejection of the DATA command. This is
3921 intended to be helpful for MUAs that show only the final error to their
3924 PH/42 Another patch from the Sieve maintainer.
3926 SC/02 Eximstats - Differentiate between permanent and temporary rejects.
3927 Eximstats - Fixed some broken HTML links and added missing column headers
3929 Eximstats - Fixed Grand Total Summary Domains, Edomains, and Email
3930 columns for Rejects, Temp Rejects, Ham, and Spam rows.
3932 SC/03 Eximstats - V1.58 Fix to get <> and blackhole to show in edomain tables.
3934 PH/43 Yet another patch from the Sieve maintainer.
3936 PH/44 I found a way to check for a TCP/IP connection going away before sending
3937 the response to the final '.' that terminates a message, but only in the
3938 case where the client has not sent further data following the '.'
3939 (unfortunately, this is allowed). However, in many cases there won't be
3940 any further data because there won't be any more messages to send. A call
3941 to select() can be used: if it shows that the input is "ready", there is
3942 either input waiting, or the socket has been closed. An attempt to read
3943 the next input character can distinguish the two cases. Previously, Exim
3944 would have sent an OK response which the client would never have see.
3945 This could lead to message repetition. This fix should cure that, at
3946 least in a lot of common cases.
3948 PH/45 Do not advertise STARTTLS in response to HELP unless it would be
3949 advertised in response to EHLO.
3955 PH/01 Two more bugs that were introduced by 4.64/PH/07, in addition to the one
3956 fixed by 4.65/MH/01 (is this a record?) are fixed:
3958 (i) An empty string was always treated as zero by the numeric comparison
3959 operators. This behaviour has been restored.
3961 (ii) It is documented that the numeric comparison operators always treat
3962 their arguments as decimal numbers. This was broken in that numbers
3963 starting with 0 were being interpreted as octal.
3965 While fixing these problems I realized that there was another issue that
3966 hadn't been noticed. Values of message_size_limit (both the global option
3967 and the transport option) were treated as octal if they started with 0.
3968 The documentation was vague. These values are now always treated as
3969 decimal, and I will make that clear in the documentation.
3975 TK/01 Disable default definition of HAVE_LINUX_SENDFILE. Clashes with
3976 Linux large file support (_FILE_OFFSET_BITS=64) on older glibc
3979 MH/01 Don't check that the operands of numeric comparison operators are
3980 integers when their expansion is in "skipping" mode (fixes bug
3981 introduced by 4.64-PH/07).
3983 PH/01 If a system filter or a router generates more than SHRT_MAX (32767)
3984 child addresses, Exim now panics and dies. Previously, because the count
3985 is held in a short int, deliveries were likely to be lost. As such a
3986 large number of recipients for a single message is ridiculous
3987 (performance will be very, very poor), I have chosen to impose a limit
3988 rather than extend the field.
3994 TK/01 Bugzilla #401. Fix DK spooling code so that it can overwrite a
3995 leftover -K file (the existence of which was triggered by #402).
3996 While we were at it, introduced process PID as part of the -K
3997 filename. This should rule out race conditions when creating
4000 TK/02 Bugzilla #402. Apply patch from Simon Arlott, speeding up DK signing
4001 processing considerably. Previous code took too long for large mails,
4002 triggering a timeout which in turn triggers #401.
4004 TK/03 Introduced HAVE_LINUX_SENDFILE to os.h-Linux. Currently only used
4005 in the DK code in transports.c. sendfile() is not really portable,
4006 hence the _LINUX specificness.
4008 TF/01 In the add_headers option to the mail command in an Exim filter,
4009 there was a bug that Exim would claim a syntax error in any
4010 header after the first one which had an odd number of characters
4013 PH/01 If a server that rejects MAIL FROM:<> was the target of a sender
4014 callout verification, Exim cached a "reject" for the entire domain. This
4015 is correct for most verifications, but it is not correct for a recipient
4016 verification with use_sender or use_postmaster set, because in that case
4017 the callout does not use MAIL FROM:<>. Exim now distinguishes the special
4018 case of MAIL FROM:<> rejection from other early rejections (e.g.
4019 rejection of HELO). When verifying a recipient using a non-null MAIL
4020 address, the cache is ignored if it shows MAIL FROM:<> rejection.
4021 Whatever the result of the callout, the value of the domain cache is
4022 left unchanged (for any other kind of callout, getting as far as trying
4023 RCPT means that the domain itself is ok).
4025 PH/02 Tidied a number of unused variable and signed/unsigned warnings that
4028 PH/03 On Solaris, an unexpectedly close socket (dropped connection) can
4029 manifest itself as EPIPE rather than ECONNECT. When tidying away a
4030 session, the daemon ignores ECONNECT errors and logs others; it now
4031 ignores EPIPE as well.
4033 PH/04 Applied Nico Erfurth's refactoring patch to tidy up mime.c
4034 (quoted-printable decoding).
4036 PH/05 Applied Nico Erfurth's refactoring patch to tidy up spool_mbox.c, and
4037 later the small subsequent patch to fix an introduced bug.
4039 PH/06 Installed the latest Cygwin Makefile from the Cygwin maintainer.
4041 PH/07 There was no check for overflow in expansions such as ${if >{1}{4096M}}.
4043 PH/08 An error is now given if message_size_limit is specified negative.
4045 PH/09 Applied and tidied up Jakob Hirsch's patch for allowing ACL variables
4046 to be given (somewhat) arbitrary names.
4048 JJ/01 exipick 20060919.0, allow for arbitrary acl_ variables introduced
4051 JJ/02 exipick 20060919.0, --show-vars args can now be regular expressions,
4052 miscellaneous code fixes
4054 PH/10 Added the log_reject_target ACL modifier to specify where to log
4057 PH/11 Callouts were setting the name used for EHLO/HELO from $smtp_active_
4058 hostname. This is wrong, because it relates to the incoming message (and
4059 probably the interface on which it is arriving) and not to the outgoing
4060 callout (which could be using a different interface). This has been
4061 changed to use the value of the helo_data option from the smtp transport
4062 instead - this is what is used when a message is actually being sent. If
4063 there is no remote transport (possible with a router that sets up host
4064 addresses), $smtp_active_hostname is used.
4066 PH/12 Installed Andrey Panin's patch to add a dovecot authenticator. Various
4067 tweaks were necessary in order to get it to work (see also 21 below):
4068 (a) The code assumed that strncpy() returns a negative number on buffer
4069 overflow, which isn't the case. Replaced with Exim's string_format()
4071 (b) There were several signed/unsigned issues. I just did the minimum
4072 hacking in of casts. There is scope for a larger refactoring.
4073 (c) The code used strcasecmp() which is not a standard C function.
4074 Replaced with Exim's strcmpic() function.
4075 (d) The code set only $1; it now sets $auth1 as well.
4076 (e) A simple test gave the error "authentication client didn't specify
4077 service in request". It would seem that Dovecot has changed its
4078 interface. Fortunately there's a specification; I followed it and
4079 changed what the client sends and it appears to be working now.
4081 PH/13 Added $message_headers_raw to provide the headers without RFC 2047
4084 PH/14 Corrected misleading output from -bv when -v was also used. Suppose the
4085 address A is aliased to B and C, where B exists and C does not. Without
4086 -v the output is "A verified" because verification stops after a
4087 successful redirection if more than one address is generated. However,
4088 with -v the child addresses are also verified. Exim was outputting "A
4089 failed to verify" and then showing the successful verification for C,
4090 with its parentage. It now outputs "B failed to verify", showing B's
4091 parentage before showing the successful verification of C.
4093 PH/15 Applied Michael Deutschmann's patch to allow DNS black list processing to
4094 look up a TXT record in a specific list after matching in a combined
4097 PH/16 It seems that the options setting for the resolver (RES_DEFNAMES and
4098 RES_DNSRCH) can affect the behaviour of gethostbyname() and friends when
4099 they consult the DNS. I had assumed they would set it the way they
4100 wanted; and indeed my experiments on Linux seem to show that in some
4101 cases they do (I could influence IPv6 lookups but not IPv4 lookups).
4102 To be on the safe side, however, I have now made the interface to
4103 host_find_byname() similar to host_find_bydns(), with an argument
4104 containing the DNS resolver options. The host_find_byname() function now
4105 sets these options at its start, just as host_find_bydns() does. The smtp
4106 transport options dns_qualify_single and dns_search_parents are passed to
4107 host_find_byname() when gethostbyname=TRUE in this transport. Other uses
4108 of host_find_byname() use the default settings of RES_DEFNAMES
4109 (qualify_single) but not RES_DNSRCH (search_parents).
4111 PH/17 Applied (a modified version of) Nico Erfurth's patch to make
4112 spool_read_header() do less string testing, by means of a preliminary
4113 switch on the second character of optional "-foo" lines. (This is
4114 overdue, caused by the large number of possibilities that now exist.
4115 Originally there were few.) While I was there, I also converted the
4116 str(n)cmp tests so they don't re-test the leading "-" and the first
4117 character, in the hope this might squeeze out yet more improvement.
4119 PH/18 Two problems with "group" syntax in header lines when verifying: (1) The
4120 flag allowing group syntax was set by the header_syntax check but not
4121 turned off, possible causing trouble later; (2) The flag was not being
4122 set at all for the header_verify test, causing "group"-style headers to
4123 be rejected. I have now set it in this case, and also caused header_
4124 verify to ignore an empty address taken from a group. While doing this, I
4125 came across some other cases where the code for allowing group syntax
4126 while scanning a header line wasn't quite right (mostly, not resetting
4127 the flag correctly in the right place). These bugs could have caused
4128 trouble for malformed header lines. I hope it is now all correct.
4130 PH/19 The functions {pwcheck,saslauthd}_verify_password() are always called
4131 with the "reply" argument non-NULL. The code, however (which originally
4132 came from elsewhere) had *some* tests for NULL when it wrote to *reply,
4133 but it didn't always do it. This confused somebody who was copying the
4134 code for some other use. I have removed all the tests.
4136 PH/20 It was discovered that the GnuTLS code had support for RSA_EXPORT, a
4137 feature that was used to support insecure browsers during the U.S. crypto
4138 embargo. It requires special client support, and Exim is probably the
4139 only MTA that supported it -- and would never use it because real RSA is
4140 always available. This code has been removed, because it had the bad
4141 effect of slowing Exim down by computing (never used) parameters for the
4142 RSA_EXPORT functionality.
4144 PH/21 On the advice of Timo Sirainen, added a check to the dovecot
4145 authenticator to fail if there's a tab character in the incoming data
4146 (there should never be unless someone is messing about, as it's supposed
4147 to be base64-encoded). Also added, on Timo's advice, the "secured" option
4148 if the connection is using TLS or if the remote IP is the same as the
4149 local IP, and the "valid-client-cert option" if a client certificate has
4152 PH/22 As suggested by Dennis Davis, added a server_condition option to *all*
4153 authenticators. This can be used for authorization after authentication
4154 succeeds. (In the case of plaintext, it servers for both authentication
4157 PH/23 Testing for tls_required and lost_connection in a retry rule didn't work
4158 if any retry times were supplied.
4160 PH/24 Exim crashed if verify=helo was activated during an incoming -bs
4161 connection, where there is no client IP address to check. In this
4162 situation, the verify now always succeeds.
4164 PH/25 Applied John Jetmore's -Mset patch.
4166 PH/26 Added -bem to be like -Mset, but loading a message from a file.
4168 PH/27 In a string expansion for a processed (not raw) header when multiple
4169 headers of the same name were present, leading whitespace was being
4170 removed from all of them, but trailing whitespace was being removed only
4171 from the last one. Now trailing whitespace is removed from each header
4172 before concatenation. Completely empty headers in a concatenation (as
4173 before) are ignored.
4175 PH/28 Fixed bug in backwards-compatibility feature of PH/09 (thanks to John
4176 Jetmore). It would have mis-read ACL variables from pre-4.61 spool files.
4178 PH/29 [Removed. This was a change that I later backed out, and forgot to
4179 correct the ChangeLog entry (that I had efficiently created) before
4180 committing the later change.]
4182 PH/30 Exim was sometimes attempting to deliver messages that had suffered
4183 address errors (4xx response to RCPT) over the same connection as other
4184 messages routed to the same hosts. Such deliveries are always "forced",
4185 so retry times are not inspected. This resulted in far too many retries
4186 for the affected addresses. The effect occurred only when there were more
4187 hosts than the hosts_max_try setting in the smtp transport when it had
4188 the 4xx errors. Those hosts that it had tried were not added to the list
4189 of hosts for which the message was waiting, so if all were tried, there
4190 was no problem. Two fixes have been applied:
4192 (i) If there are any address or message errors in an SMTP delivery, none
4193 of the hosts (tried or untried) are now added to the list of hosts
4194 for which the message is waiting, so the message should not be a
4195 candidate for sending over the same connection that was used for a
4196 successful delivery of some other message. This seems entirely
4197 reasonable: after all the message is NOT "waiting for some host".
4198 This is so "obvious" that I'm not sure why it wasn't done
4199 previously. Hope I haven't missed anything, but it can't do any
4200 harm, as the worst effect is to miss an optimization.
4202 (ii) If, despite (i), such a delivery is accidentally attempted, the
4203 routing retry time is respected, so at least it doesn't keep
4204 hammering the server.
4206 PH/31 Installed Andrew Findlay's patch to close the writing end of the socket
4207 in ${readsocket because some servers need this prod.
4209 PH/32 Added some extra debug output when updating a wait-xxx database.
4211 PH/33 The hint "could be header name not terminated by colon", which has been
4212 given for certain expansion errors for a long time, was not being given
4213 for the ${if def:h_colon_omitted{... case.
4215 PH/34 The spec says: "With one important exception, whenever a domain list is
4216 being scanned, $domain contains the subject domain." There was at least
4217 one case where this was not true.
4219 PH/35 The error "getsockname() failed: connection reset by peer" was being
4220 written to the panic log as well as the main log, but it isn't really
4221 panic-worthy as it just means the connection died rather early on. I have
4222 removed the panic log writing for the ECONNRESET error when getsockname()
4225 PH/36 After a 4xx response to a RCPT error, that address was delayed (in queue
4226 runs only) independently of the message's sender address. This meant
4227 that, if the 4xx error was in fact related to the sender, a different
4228 message to the same recipient with a different sender could confuse
4229 things. In particular, this can happen when sending to a greylisting
4230 server, but other circumstances could also provoke similar problems.
4231 I have changed the default so that the retry time for these errors is now
4232 based a combination of the sender and recipient addresses. This change
4233 can be overridden by setting address_retry_include_sender=false in the
4236 PH/37 For LMTP over TCP/IP (the smtp transport), error responses from the
4237 remote server are returned as part of bounce messages. This was not
4238 happening for LMTP over a pipe (the lmtp transport), but now it is the
4239 same for both kinds of LMTP.
4241 PH/38 Despite being documented as not happening, Exim was rewriting addresses
4242 in header lines that were in fact CNAMEs. This is no longer the case.
4244 PH/39 If -R or -S was given with -q<time>, the effect of -R or -S was ignored,
4245 and queue runs started by the daemon processed all messages. This has
4246 been fixed so that -R and -S can now usefully be given with -q<time>.
4248 PH/40 Import PCRE release 6.7 (fixes some bugs).
4250 PH/41 Add bitwise logical operations to eval (courtesy Brad Jorsch).
4252 PH/42 Give an error if -q is specified more than once.
4254 PH/43 Renamed the variables $interface_address and $interface_port as
4255 $received_ip_address and $received_port, to make it clear that these
4256 values apply to message reception, and not to the outgoing interface when
4257 a message is delivered. (The old names remain recognized, of course.)
4259 PH/44 There was no timeout on the connect() call when using a Unix domain
4260 socket in the ${readsocket expansion. There now is.
4262 PH/45 Applied a modified version of Brad Jorsch's patch to allow "message" to
4263 be meaningful with "accept".
4265 SC/01 Eximstats V1.43
4266 Bug fix for V1.42 with -h0 specified. Spotted by Chris Lear.
4268 SC/02 Eximstats V1.44
4269 Use a glob alias rather than an array ref in the generated
4270 parser. This improves both readability and performance.
4272 SC/03 Eximstats V1.45 (Marco Gaiarin / Steve Campbell)
4273 Collect SpamAssassin and rejection statistics.
4274 Don't display local sender or destination tables unless
4275 there is data to show.
4276 Added average volumes into the top table text output.
4278 SC/04 Eximstats V1.46
4279 Collect data on the number of addresses (recipients)
4280 as well as the number of messages.
4282 SC/05 Eximstats V1.47
4283 Added 'Message too big' to the list of mail rejection
4284 reasons (thanks to Marco Gaiarin).
4286 SC/06 Eximstats V1.48
4287 Mainlog lines which have GMT offsets and are too short to
4288 have a flag are now skipped.
4290 SC/07 Eximstats V1.49 (Alain Williams)
4291 Added the -emptyok flag.
4293 SC/08 Eximstats V1.50
4294 Fixes for obtaining the IP address from reject messages.
4296 JJ/03 exipick.20061117.2, made header handling as similar to exim as possible
4297 (added [br]h_ prefixes, implemented RFC2047 decoding. Fixed
4298 whitespace changes from 4.64-PH/27
4300 JJ/04 exipick.20061117.2, fixed format and added $message_headers_raw to
4303 JJ/05 exipick.20061117.2, bug fixes (error out sooner when invalid criteria
4304 are found, allow negative numbers in numeric criteria)
4306 JJ/06 exipick.20061117.2, added new $message_body_missing variable
4308 JJ/07 exipick.20061117.2, added $received_ip_address and $received_port
4309 to match changes made in 4.64-PH/43
4311 PH/46 Applied Jori Hamalainen's patch to add features to exiqsumm.
4313 PH/47 Put in an explicit test for a DNS lookup of an address record where the
4314 "domain" is actually an IP address, and force a failure. This locks out
4315 those revolvers/nameservers that support "A-for-A" lookups, in
4316 contravention of the specifications.
4318 PH/48 When a host name was looked up from an IP address, and the subsequent
4319 forward lookup of the name timed out, the host name was left in
4320 $sender_host_name, contrary to the specification.
4322 PH/49 Although default lookup types such as lsearch* or cdb*@ have always been
4323 restricted to single-key lookups, Exim was not diagnosing an error if
4324 * or *@ was used with a query-style lookup.
4326 PH/50 Increased the value of DH_BITS in tls-gnu.c from 768 to 1024.
4328 MH/01 local_scan ABI version incremented to 1.1. It should have been updated
4329 long ago, but noone interested enough thought of it. Let's just say that
4330 the "1.1" means that there are some new functions that weren't there at
4331 some point in the past.
4333 PH/51 Error processing for expansion failure of helo_data from an smtp
4334 transport during callout processing was broken.
4336 PH/52 Applied John Jetmore's patch to allow tls-on-connect and STARTTLS to be
4337 tested/used via the -bh/-bhc/-bs options.
4339 PH/53 Added missing "#include <time.h>" to pcre/pcretest.c (this was a PCRE
4340 bug, fixed in subsequent PCRE releases).
4342 PH/54 Applied Robert Bannocks' patch to avoid a problem with references that
4343 arises when using the Solaris LDAP libraries (but not with OpenLDAP).
4345 PH/55 Check for a ridiculously long file name in exim_dbmbuild.
4351 SC/01 Use a glob alias rather than an array ref in eximstats generated
4352 parser. This improves both readability and performance.
4354 SC/02 Collect SpamAssassin and rejection statistics in eximstats.
4355 Don't display local sender or destination tables in eximstats unless
4356 there is data to show.
4357 Added average volumes into the eximstats top table text output.
4359 SC/03 Collect data on the number of addresses (recipients) as well
4360 as the number of messages in eximstats.
4362 TF/01 Correct an error in the documentation for the redirect router. Exim
4363 does (usually) call initgroups() when daemonizing.
4365 TF/02 Call initgroups() when dropping privilege in exim.c, so that Exim runs
4366 with consistent privilege compared to when running as a daemon.
4368 TF/03 Note in the spec that $authenticated_id is not set for local
4369 submissions from trusted users.
4371 TF/04 The ratelimit per_rcpt option now works correctly in acl_not_smtp.
4372 Thanks to Dean Brooks <dean@iglou.com> for the patch.
4374 TF/05 Make it easier to get SMTP authentication and TLS/SSL support working
4375 by adding some example configuration directives to the default
4376 configuration file. A little bit of work is required to uncomment the
4377 directives and define how usernames and passwords are checked, but
4378 there is now a framework to start from.
4380 PH/01 Added #define LDAP_DEPRECATED 1 to ldap.c because some of the "old"
4381 functions that Exim currently uses aren't defined in ldap.h for OpenLDAP
4382 without this. I don't know how relevant this is to other LDAP libraries.
4384 PH/02 Add the verb name to the "unknown ACL verb" error.
4386 PH/03 Magnus Holmgren's patch for filter_prepend_home.
4388 PH/03 Fixed Bugzilla #101: macro definition between ACLs doesn't work.
4390 PH/04 Applied Magnus Holmgren's patch to fix Bugzilla #98: transport's home
4391 directory not expanded when it should be if an expanded home directory
4392 was set for the address (which is overridden by the transport).
4394 PH/05 Applied Alex Kiernan's patch to fix Bugzilla #99: a problem with
4397 PH/06 Added acl_not_smtp_start, based on Johannes Berg's patch, and set the
4398 bit to forbid control=suppress_local_fixups in the acl_not_smtp ACL,
4399 because it is too late at that time, and has no effect.
4401 PH/07 Changed ${quote_pgsql to quote ' as '' instead of \' because of a
4402 security issue with \' (bugzilla #107). I could not use the
4403 PQescapeStringConn() function, because it needs a PGconn value as one of
4406 PH/08 When testing addresses using -bt, indicate those final addresses that
4407 are duplicates that would not cause an additional delivery. At least one
4408 person was confused, thinking that -bt output corresponded to deliveries.
4409 (Suppressing duplicates isn't a good idea as you lose the information
4410 about possibly different redirections that led to the duplicates.)
4412 PH/09 Applied patch from Erik to use select() instead of poll() in spam.c on
4413 systems where poll() doesn't work, in particular OS X.
4415 PH/10 Added more information to debugging output for retry time not reached.
4417 PH/11 Applied patch from Arkadiusz Miskiewicz to apply a timeout to read
4418 operations in malware.c.
4420 PH/12 Applied patch from Magnus Holmgren to include the "h" tag in Domain Keys
4423 PH/13 If write_rejectlog was set false when logging was sent to syslog with
4424 syslog_duplication set false, log lines that would normally be written
4425 both the the main log and to the reject log were not written to syslog at
4428 PH/14 In the default configuration, change the use of "message" in ACL warn
4429 statements to "add_header".
4431 PH/15 Diagnose a filter syntax error for "seen", "unseen", or "noerror" if not
4432 not followed by a command (e.g. "seen endif").
4434 PH/16 Recognize SMTP codes at the start of "message" in ACLs and after :fail:
4435 and :defer: in a redirect router. Add forbid_smtp_code to suppress the
4438 PH/17 Added extra conditions to the default value of delay_warning_condition
4442 { !eq{$h_list-id:$h_list-post:$h_list-subscribe:}{} } \
4443 { match{$h_precedence:}{(?i)bulk|list|junk} } \
4444 { match{$h_auto-submitted:}{(?i)auto-generated|auto-replied} } \
4447 The Auto-Submitted: and various List- headers are standardised, whereas I
4448 don't think Precedence: ever was.
4450 PH/18 Refactored debugging code in route_finduser() to show more information,
4451 in particular, the error code if getpwnam() issues one.
4453 PH/19 Added PQsetClientEncoding(conn, "SQL_ASCII") to the pgsql code module.
4454 This is apparently needed in addition to the PH/07 change above to avoid
4455 any possible encoding problems.
4457 PH/20 Perl can change the locale. Exim was resetting it after a ${perl call,
4458 but not after initializing Perl.
4460 PH/21 Added a call to PQsetNoticeProcessor() to catch pgsql "notices" and
4461 output them only if debugging. By default they are written stderr,
4462 apparently, which is not desirable.
4464 PH/22 Added Alain Williams' LDAP patch to support setting REFERRALS=off on
4467 JJ/01 exipick: added --reverse (and -R synonym), --random, --size, --sort and
4470 JJ/02 exipick: rewrote --help documentation to hopefully make more clear.
4472 PH/23 Made -oMaa and -oMt work with -bh and -bs to pretend the connection is
4473 authenticated or an ident call has been made. Suppress the default
4474 values for $authenticated_id and $authenticated_sender (but permit -oMai
4475 and -oMas) when testing with -bh.
4477 PH/24 Re-jigged the order of the tests in the default configuration so that the
4478 tests for valid domains and recipients precede the DNS black list and CSA
4479 tests, on the grounds that those ones are more expensive.
4481 PH/25 Exim was not testing for a space following SMTP commands such as EHLO
4482 that require one. Thus, EHLORHUBARB was interpreted as a valid command.
4483 This bug exists in every version of Exim that I still have, right back to
4486 PH/26 (n)wildlsearch lookups are documented as being done case-insensitively.
4487 However, an attempt to turn on case-sensitivity in a regex key by
4488 including (?-i) didn't work because the subject string was already
4489 lowercased, and the effects were non-intuitive. It turns out that a
4490 one-line patch can be used to allow (?-i) to work as expected.
4496 TF/01 Fix the add_header change below (4.61 PH/55) which had a bug that (amongst
4497 other effects) broke the use of negated acl sub-conditions.
4499 PH/01 ${readsocket now supports Internet domain sockets (modified John Jetmore
4502 PH/02 When tcp-wrappers is called from Exim, it returns only "deny" or "allow".
4503 "Deny" causes Exim to reject the incoming connection with a 554 error.
4504 Unfortunately, if there is a major crisis, such as a disk failure,
4505 tcp-wrappers gives "deny", whereas what one would like would be some
4506 kind of temporary error. A kludge has been added to help with this.
4507 Before calling hosts_ctl(), errno is set zero. If the result is "deny", a
4508 554 error is used if errno is still zero or contains ENOENT (which occurs
4509 if either of the /etc/hosts.{allow,deny} files is missing). Otherwise, a
4512 PH/03 Add -lutil to the default FreeBSD LIBS setting.
4514 PH/04 Change PH/19 for 4.61 was too wide. It should not be applied to host
4515 errors. Otherwise a message that provokes a temporary error (when other
4516 messages do not) can cause a whole host to time out.
4518 PH/05 Batch deliveries by appendfile and pipe transports did not work when the
4519 addresses were routed directly to files or pipes from a redirect router.
4520 File deliveries just didn't batch; pipe deliveries might have suffered
4523 PH/06 A failure to get a lock for a hints database would erroneously always say
4524 "Failed to get write lock", even when it was really a read lock.
4526 PH/07 The appendfile transport was creating MBX lock files with a fixed mode
4527 of 0600. This has been changed to use the value of the lockfile_mode
4528 option (which defaults to 0600).
4530 PH/08 Applied small patch from the Sieve maintainer.
4532 PH/09 If maildir_quota_directory_regex was set to exclude (say) the .Trash
4533 folder from quota calculations, a direct delivery into this folder messed
4534 up the contents of the maildirsize file. This was because the regex was
4535 used only to exclude .Trash (or whatever) when the size of the mailbox
4536 was calculated. There was no check that a delivery was happening into an
4537 excluded directory. This bug has been fixed by ignoring all quota
4538 processing for deliveries into excluded directories.
4540 PH/10 Added the maildirfolder_create_regex option to appendfile.
4546 PH/01 The code for finding all the local interface addresses on a FreeBSD
4547 system running IPv6 was broken. This may well have applied to all BSD
4548 systems, as well as to others that have similar system calls. The broken
4549 code found IPv4 interfaces correctly, but gave incorrect values for the
4550 IPv6 interfaces. In particular, ::1 was not found. The effect in Exim was
4551 that it would not match correctly against @[] and not recognize the IPv6
4554 PH/02 The ipliteral router was not recognizing addresses of the form user@
4555 [ipv6:....] because it didn't know about the "ipv6:" prefix.
4557 PH/03 Added disable_ipv6.
4559 PH/04 Changed $reply_address to use the raw form of the headers instead of the
4560 decoded form, because it is most often used to construct To: headers
4561 lines in autoreplies, and the decoded form may well be syntactically
4562 invalid. However, $reply_address has leading white space removed, and all
4563 newlines turned into spaces so that the autoreply transport does not
4566 PH/05 If group was specified without a user on a router, and no group or user
4567 was specified on a transport, the group from the router was ignored.
4569 PH/06 Increased the number of ACL variables to 20 of each type, and arranged
4570 for visible compile-time settings that can be used to change these
4571 numbers, for those that want even more. Backwards compatibility with old
4572 spool files has been maintained. However, going back to a previous Exim
4573 release will lost any variables that are in spool files.
4575 PH/07 Two small changes when running in the test harness: increase delay when
4576 passing a TCP/IP connection to a new process, in case the original
4577 process has to generate a bounce, and remove special handling of
4578 127.0.0.2 (sic), which is no longer necessary.
4580 PH/08 Changed debug output of dbfn_open() flags from numbers to names, so as to
4581 be the same on different OS.
4583 PH/09 Moved a debug statement in filter processing to avoid a race problem when
4586 JJ/01 exipick: fixed bug where -b (brief) output option showed "Vars:"
4587 whether --show-vars was specified or not
4589 JJ/02 exipick: Added support for new ACL variable spool format introduced
4592 PH/10 Fixed another bug related to PH/04 above: if an incoming message had a
4593 syntactically invalid From: or Reply-to: line, and a filter used this to
4594 generate an autoreply, and therefore failed to obtain an address for the
4595 autoreply, Exim could try to deliver to a non-existent relative file
4596 name, causing unrelated and misleading errors. What now happens is that
4597 it logs this as a hard delivery error, but does not attempt to create a
4600 PH/11 The exinext utility has a -C option for testing purposes, but although
4601 the given file was scanned by exinext itself; it wasn't being passed on
4602 when Exim was called.
4604 PH/12 In the smtp transport, treat an explicit ECONNRESET error the same as
4605 an end-of-file indication when reading a command response.
4607 PH/13 Domain literals for IPv6 were not recognized unless IPv6 support was
4608 compiled. In many other places in Exim, IPv6 addresses are always
4609 recognized, so I have changed this. It also means that IPv4 domain
4610 literals of the form [IPV4:n.n.n.n] are now always recognized.
4612 PH/14 When a uid/gid is specified for the queryprogram router, it cannot be
4613 used if the router is not running as root, for example, when verifying at
4614 ACL time, or when using -bh. The debugging output from this situation was
4615 non-existent - all you got was a failure to exec. I have made two
4618 (a) Failures to set uid/gid, the current directory, or a process leader
4619 in a subprocess such as that created by queryprogram now generate
4620 suitable debugging output when -d is set.
4622 (b) The queryprogram router detects when it is not running as root,
4623 outputs suitable debugging information if -d is set, and then runs
4624 the subprocess without attempting to change uid/gid.
4626 PH/15 Minor change to Makefile for building test_host (undocumented testing
4629 PH/16 As discussed on the list in Nov/Dec: Exim no longer looks at the
4630 additional section of a DNS packet that returns MX or SRV records.
4631 Instead, it always explicitly searches for A/AAAA records. This avoids
4632 major problems that occur when a DNS server includes only records of one
4633 type (A or AAAA) in an MX/SRV packet. A byproduct of this change has
4634 fixed another bug: if SRV records were looked up and the corresponding
4635 address records were *not* found in the additional section, the port
4636 values from the SRV records were lost.
4638 PH/17 If a delivery to a pipe, file, or autoreply was deferred, Exim was not
4639 using the correct key (the original address) when searching the retry
4640 rules in order to find which one to use for generating the retry hint.
4642 PH/18 If quota_warn_message contains a From: header, Exim now refrains from
4643 adding the default one. Similarly, if it contains a Reply-To: header, the
4644 errors_reply_to option, if set, is not used.
4646 PH/19 When calculating a retry time, Exim used to measure the "time since
4647 failure" by looking at the "first failed" field in the retry record. Now
4648 it does not use this if it is later than than the arrival time of the
4649 message. Instead it uses the arrival time. This makes for better
4650 behaviour in cases where some deliveries succeed, thus re-setting the
4651 "first failed" field. An example is a quota failure for a huge message
4652 when small messages continue to be delivered. Without this change, the
4653 "time since failure" will always be short, possible causing more frequent
4654 delivery attempts for the huge message than are intended.
4655 [Note: This change was subsequently modified - see PH/04 for 4.62.]
4657 PH/20 Added $auth1, $auth2, $auth3 to contain authentication data (as well as
4658 $1, $2, $3) because the numerical variables can be reset during some
4659 expansion items (e.g. "match"), thereby losing the authentication data.
4661 PH/21 Make -bV show the size of off_t variables so that the test suite can
4662 decide whether to run tests for quotas > 2G.
4664 PH/22 Test the values given for quota, quota_filecount, quota_warn_threshold,
4665 mailbox_size, and mailbox_filecount in the appendfile transport. If a
4666 filecount value is greater than 2G or if a quota value is greater than 2G
4667 on a system where the size of off_t is not greater than 4, a panic error
4670 PH/23 When a malformed item such as 1.2.3/24 appears in a host list, it can
4671 never match. The debug and -bh output now contains an explicit error
4672 message indicating a malformed IPv4 address or mask.
4674 PH/24 An host item such as 1.2.3.4/abc was being treated as the IP address
4675 1.2.3.4 without a mask. Now it is not recognized as an IP address, and
4676 PH/23 above applies.
4678 PH/25 Do not write to syslog when running in the test harness. The only
4679 occasion when this arises is a failure to open the main or panic logs
4680 (for which there is an explicit test).
4682 PH/26 Added the /no_tell option to "control=freeze".
4684 PH/27 If a host name lookup failed very early in a connection, for example, if
4685 the IP address matched host_lookup and the reverse lookup yielded a name
4686 that did not have a forward lookup, an error message of the form "no IP
4687 address found for host xxx.xxx.xxx (during SMTP connection from NULL)"
4688 could be logged. Now it outputs the IP address instead of "NULL".
4690 PH/28 An enabling patch from MH: add new function child_open_exim2() which
4691 allows the sender and the authenticated sender to be set when
4692 submitting a message from within Exim. Since child_open_exim() is
4693 documented for local_scan(), the new function should be too.
4695 PH/29 In GnuTLS, a forced expansion failure for tls_privatekey was not being
4696 ignored. In both GnuTLS and OpenSSL, an expansion of tls_privatekey that
4697 results in an empty string is now treated as unset.
4699 PH/30 Fix eximon buffer overflow bug (Bugzilla #73).
4701 PH/31 Added sender_verify_fail logging option.
4703 PH/32 In November 2003, the code in Exim that added an empty Bcc: header when
4704 needed by RFC 822 but not by RFC 2822 was commented out. I have now
4705 tidied the source and removed it altogether.
4707 PH/33 When a queue run was abandoned because the load average was too high, a
4708 log line was always written; now it is written only if the queue_run log
4709 selector is set. In addition, the log line for abandonment now contains
4710 information about the queue run such as the pid. This is always present
4711 in "start" and "stop" lines but was omitted from the "abandon" line.
4713 PH/34 Omit spaces between a header name and the colon in the error message that
4714 is given when verify = headers_syntax fails (if there are lots of them,
4715 the message gets confusing).
4717 PH/35 Change the default for dns_check_names_pattern to allow slashes within
4718 names, as there are now some PTR records that contain slashes. This check
4719 is only to protect against broken name servers that fall over on strange
4720 characters, so the fact that it applies to all lookups doesn't matter.
4722 PH/36 Now that the new test suite is complete, we can remove some of the
4723 special code in Exim that was needed for the old test suite. For example,
4724 sorting DNS records because real resolvers return them in an arbitrary
4725 order. The new test suite's fake resolver always returns records in the
4728 PH/37 When running in the test harness, use -odi for submitted messages (e.g.
4729 bounces) except when queue_only is set, to avoid logging races between
4730 the different processes.
4732 PH/38 Panic-die if .include specifies a non-absolute path.
4734 PH/39 A tweak to the "H" retry rule from its user.
4736 JJ/03 exipick: Removed parentheses from 'next' and 'last' calls that specified
4737 a label. They prevented compilation on older perls.
4739 JJ/04 exipick: Refactored code to prevent implicit split to @_ which caused
4740 a warning to be raised on newish perls.
4742 JJ/05 exipick: Fixed bug where -bpc always showed a count of all messages
4743 on queue. Changes to match documented behaviour of showing count of
4744 messages matching specified criteria.
4746 PH/40 Changed the default ident timeout from 30s to 5s.
4748 PH/41 Added support for the use of login_cap features, on those BSD systems
4749 that have them, for controlling the resources used by pipe deliveries.
4751 PH/42 The content-scanning code uses fopen() to create files in which to put
4752 message data. Previously it was not paying any attention to the mode of
4753 the files. Exim runs with umask(0) because the rest of the code creates
4754 files with open(), and sets the required mode explicitly. Thus, these
4755 files were ending up world-writeable. This was not a big issue, because,
4756 being within the spool directory, they were not world-accessible. I have
4757 created a function called modefopen, which takes an additional mode
4758 argument. It sets umask(777), creates the file, chmods it to the required
4759 mode, then resets the umask. All the relevant calls to fopen() in the
4760 content scanning code have been changed to use this function.
4762 PH/43 If retry_interval_max is set greater than 24 hours, it is quietly reset
4763 to 24 hours. This avoids potential overflow problems when processing G
4764 and H retry rules. I suspect nobody ever tinkers with this value.
4766 PH/44 Added STRIP_COMMAND=/usr/bin/strip to the FreeBSD Makefile.
4768 PH/45 When the plaintext authenticator is running as a client, the server's
4769 challenges are checked to ensure they are valid base64 strings. By
4770 default, the authentication attempt is cancelled if an invalid string is
4771 received. Setting client_ignore_invalid_base64 true ignores these errors.
4772 The decoded challenge strings are now placed in $auth1, $auth2, etc. as
4773 they are received. Thus, the responses can be made to depend on the
4774 challenges. If an invalid string is ignored, an empty string is placed in
4777 PH/46 Messages that are created by the autoreply transport now contains a
4778 References: header, in accordance with RFCs 2822 and 3834.
4780 PH/47 Added authenticated_sender_force to the smtp transport.
4782 PH/48 The ${prvs expansion was broken on systems where time_t was long long.
4784 PH/49 Installed latest patch from the Sieve maintainer.
4786 PH/50 When an Exim quota was set without a file count quota, and mailbox_size
4787 was also set, the appendfile transport was unnecessarily scanning a
4788 directory of message files (e.g. for maildir delivery) to find the count
4789 of files (along with the size), even though it did not need this
4790 information. It now does the scan only if it needs to find either the
4791 size of the count of files.
4793 PH/51 Added ${time_eval: to convert Exim time strings into seconds.
4795 PH/52 Two bugs concerned with error handling when the smtp transport is
4798 (i) Exim was not creating retry information for temporary errors given
4799 for individual recipients after the DATA command when the smtp transport
4800 was used in LMTP mode. This meant that they could be retried too
4801 frequently, and not timed out correctly.
4803 (ii) Exim was setting the flag that allows error details to be returned
4804 for LMTP errors on RCPT commands, but not for LMTP errors for individual
4805 recipients that were returned after the DATA command.
4807 PH/53 This is related to PH/52, but is more general: for any failing address,
4808 when detailed error information was permitted to be returned to the
4809 sender, but the error was temporary, then after the final timeout, only
4810 "retry timeout exceeded" was returned. Now it returns the full error as
4811 well as "retry timeout exceeded".
4813 PH/54 Added control=allow_auth_unadvertised, as it seems there are clients that
4814 do this, and (what is worse) MTAs that accept it.
4816 PH/55 Added the add_header modified to ACLs. The use of "message" with "warn"
4817 will now be deprecated.
4819 PH/56 New os.c-cygwin from the Cygwin maintainer.
4821 JJ/06 exipick: added --unsorted option to allow unsorted output in all output
4822 formats (previously only available in exim formats via -bpr, -bpru,
4823 and -bpra. Now also available in native and exiqgrep formats)
4825 JJ/07 exipick: added --freeze and --thaw options to allow faster interaction
4826 with very large, slow to parse queues
4828 JJ/08 exipick: added ! as generic prefix to negate any criteria format
4830 JJ/09 exipick: miscellaneous performance enhancements (~24% improvements)
4832 PH/57 Tidies in SMTP dialogue display in debug output: (i) It was not showing
4833 responses to authentication challenges, though it was showing the
4834 challenges; (ii) I've removed the CR characters from the debug output for
4837 PH/58 Allow for the insertion of a newline as well as a space when a string
4838 is turned into more than one encoded-word during RFC 2047 encoding. The
4839 Sieve code now uses this.
4841 PH/59 Added the following errors that can be detected in retry rules: mail_4xx,
4842 data_4xx, lost_connection, tls_required.
4844 PH/60 When a VRFY deferred or FAILED, the log message rather than the user
4845 message was being sent as an SMTP response.
4847 PH/61 Add -l and -k options to exicyclog.
4849 PH/62 When verifying, if an address was redirected to one new address, so that
4850 verification continued, and the new address failed or deferred after
4851 having set something in $address_data, the value of $address_data was not
4852 passed back to the ACL. This was different to the case when no
4853 redirection occurred. The value is now passed back in both cases.
4855 PH/63 Changed the macro HAVE_LOGIN_CAP (see PH/41 for this release above) to
4856 HAVE_SETCLASSRESOURCES because there are different APIs in use that all
4857 use login_cap.h, so on its own it isn't the distinguishing feature. The
4858 new name refers directly to the setclassresources() function.
4860 PH/65 Added configuration files for NetBSD3.
4862 PH/66 Updated OS/Makefile-HP-UX for gcc 4.1.0 with HP-UX 11.
4864 PH/67 Fixed minor infelicity in the sorting of addresses to ensure that IPv6
4865 is preferred over IPv4.
4867 PH/68 The bounce_return_message and bounce_return_body options were not being
4868 honoured for bounces generated during the reception of non-SMTP messages.
4869 In particular, this applied to messages rejected by the ACL. This bug has
4870 been fixed. However, if bounce_return_message is true and bounce_return_
4871 body is false, the headers that are returned for a non-SMTP message
4872 include only those that have been read before the error was detected.
4873 (In the case of an ACL rejection, they have all been read.)
4875 PH/69 The HTML version of the specification is now built in a directory called
4876 spec_html instead of spec.html, because the latter looks like a path with
4877 a MIME-type, and this confuses some software.
4879 PH/70 Catch two compiler warnings in sieve.c.
4881 PH/71 Fixed an obscure and subtle bug (thanks Alexander & Matthias). The
4882 function verify_get_ident() calls ip_connect() to connect a socket, but
4883 if the "connect()" function timed out, ip_connect() used to close the
4884 socket. However, verify_get_ident() also closes the socket later, and in
4885 between Exim writes to the log, which may get opened at this point. When
4886 the socket was closed in ip_connect(), the log could get the same file
4887 descriptor number as the socket. This naturally causes chaos. The fix is
4888 not to close the socket in ip_connect(); the socket should be closed by
4889 the function that creates it. There was only one place in the code where
4890 this was missing, in the iplookup router, which I don't think anybody now
4891 uses, but I've fixed it anyway.
4893 PH/72 Make dns_again_means_nonexist apply to lookups using gethostbyname() as
4894 well as to direct DNS lookups. Otherwise the handling of names in host
4895 lists is inconsistent and therefore confusing.
4901 PH/01 Two changes to the default runtime configuration:
4903 (1) Move the checks for relay_from_hosts and authenticated clients from
4904 after to before the (commented out) DNS black list checks.
4906 (2) Add control=submission to the relay_from_hosts and authenticated
4907 clients checks, on the grounds that messages accepted by these
4908 statements are most likely to be submissions.
4910 PH/02 Several tidies to the handling of ${prvs and ${prvscheck:
4912 (1) Generate an error if the third argument for the ${prvs expansion is
4915 (2) Treat a missing third argument of ${prvscheck as if it were an empty
4918 (3) Reset the variables that are obtained from the first argument of
4919 ${prvscheck and used in the second argument before leaving the code,
4920 because their memory is reclaimed, so using them afterwards may do
4923 (4) Tidy up the code for expanding the arguments of ${prvscheck one by
4924 one (it's much easier than Tom thought :-).
4926 (5) Because of (4), we can now allow for the use of $prvscheck_result
4927 inside the third argument.
4929 PH/03 For some reason, the default setting of PATH when running a command from
4930 a pipe transport was just "/usr/bin". I have changed it to
4933 PH/04 SUPPORT_TRANSLATE_IP_ADDRESS and MOVE_FROZEN_MESSAGES did not cause
4934 anything to be listed in the output from -bV.
4936 PH/05 When a filter generated an autoreply, the entire To: header line was
4937 quoted in the delivery log line, like this:
4939 => >A.N.Other <ano@some.domain> <original@ddress> ...
4941 This has been changed so that it extracts the operative address. There
4942 may be more than one such address. If so, they are comma-separated, like
4945 => >ano@some.domain,ona@other.domain <original@ddress> ...
4947 PH/06 When a client host used a correct literal IP address in a HELO or EHLO
4948 command, (for example, EHLO [1.2.3.4]) and the client's IP address was
4949 not being looked up in the rDNS to get a host name, Exim was showing the
4950 IP address twice in Received: lines, even though the IP addresses were
4951 identical. For example:
4953 Received: from [1.2.3.4] (helo=[1.2.3.4])
4955 However, if the real host name was known, it was omitting the HELO data
4956 if it matched the actual IP address. This has been tidied up so that it
4957 doesn't show the same IP address twice.
4959 PH/07 When both +timestamp and +memory debugging was on, the value given by
4960 $tod_xxx expansions could be wrong, because the tod_stamp() function was
4961 called by the debug printing, thereby overwriting the timestamp buffer.
4962 Debugging no longer uses the tod_stamp() function when +timestamp is set.
4964 PH/08 When the original message was included in an autoreply transport, it
4965 always said "this is a copy of the message, including all the headers",
4966 even if body_only or headers_only was set. It now gives an appropriate
4969 PH/09 Applied a patch from the Sieve maintainer which:
4971 o fixes some comments
4972 o adds the (disabled) notify extension core
4973 o adds some debug output for the result of if/elsif tests
4974 o points to the current vacation draft in the documentation
4975 and documents the missing references header update
4979 o fixes a bug in processing the envelope test (when testing
4980 multiple envelope elements, the last element determined the
4983 PH/10 Exim was violating RFC 3834 ("Recommendations for Automatic Responses to
4984 Electronic Mail") by including:
4986 Auto-submitted: auto-generated
4988 in the messages that it generates (bounce messages and others, such as
4989 warnings). In the case of bounce messages for non-SMTP messages, there was
4990 also a typo: it was using "Auto_submitted" (underscore instead of
4991 hyphen). Since every message generated by Exim is necessarily in response
4992 to another message, thes have all been changed to:
4994 Auto-Submitted: auto-replied
4996 in accordance with these statements in the RFC:
4998 The auto-replied keyword:
5000 - SHOULD be used on messages sent in direct response to another
5001 message by an automatic process,
5003 - MUST NOT be used on manually-generated messages,
5005 - MAY be used on Delivery Status Notifications (DSNs) and Message
5006 Disposition Notifications (MDNs),
5008 - MUST NOT be used on messages generated by automatic or periodic
5009 processes, except for messages which are automatic responses to
5012 PH/11 Added "${if def:sender_address {(envelope-from <$sender_address>)\n\t}}"
5013 to the default Received: header definition.
5015 PH/12 Added log selector acl_warn_skipped (default on).
5017 PH/13 After a successful wildlsearch lookup, discard the values of numeric
5018 variables because (a) they are in the wrong storage pool and (b) even if
5019 they were copied, it wouldn't work properly because of the caching.
5021 PH/14 Add check_rfc2047_length to disable enforcement of RFC 2047 length
5022 checking when decoding. Apparently there are clients that generate
5023 overlong encoded strings. Why am I not surprised?
5025 PH/15 If the first argument of "${if match_address" was not empty, but did not
5026 contain an "@" character, Exim crashed. Now it writes a panic log message
5027 and treats the condition as false.
5029 PH/16 In autoreply, treat an empty string for "once" the same as unset.
5031 PH/17 A further patch from the Sieve maintainer: "Introduce the new Sieve
5032 extension "envelope-auth". The code is finished and in agreement with
5033 other implementations, but there is no documentation so far and in fact,
5034 nobody wrote the draft yet. This extension is currently #undef'ed, thus
5035 not changing the active code.
5037 Print executed "if" and "elsif" statements when debugging is used. This
5038 helps a great deal to understand what a filter does.
5040 Document more things not specified clearly in RFC3028. I had all this
5041 sorted out, when out of a sudden new issues came to my mind. Oops."
5043 PH/18 Exim was not recognizing the "net-" search type prefix in match_ip lists
5046 PH/19 Exim expands the IPv6 address given to -bh to its full non-abbreviated
5047 canonical form (as documented). However, after a host name lookup from
5048 the IP address, check_host() was doing a simple string comparison with
5049 addresses acquired from the DNS when checking that the found name did
5050 have the original IP as one of its addresses. Since any found IPv6
5051 addresses are likely to be in abbreviated form, the comparison could
5052 fail. Luckily, there already exists a function for doing the comparison
5053 by converting both addresses to binary, so now that is used instead of
5054 the text comparison.
5056 PH/20 There was another similar case to PH/19, when a complete host name was
5057 given in a host list; looking up its IP address could give an abbreviated
5058 form, whereas the current host's name might or might not be abbreviated.
5059 The same fix has been applied.
5065 PH/01 The ${base62: operator adjusted itself to base 36 when BASE_62 was
5066 set to 36 (for Darwin and Cygwin), but the ${base62d: operator did not.
5069 PH/02 Two minor problems detected in Cygwin: the os.{c,h} files had lost */ on
5070 the CVS lines, and there was a missing #if HAVE_IPV6 in host.c.
5072 PH/03 Typo: missing ".o" in src/pcre/Makefile.
5074 PH/04 Tighten up "personal" tests: Instead of testing for any "List-"
5075 header line, restrict the check to what is listed in RFCs 2369 and 2929.
5076 Also, for "Auto-Submitted", treat anything other than "no" as
5077 non-personal, in accordance with RFC 3834. (Previously it treated
5078 anything starting "auto-" as non-personal.)
5080 TF/01 The control=submission/name=... option had a problem with syntax
5081 errors if the name included a slash character. The /name= option
5082 now slurps the rest of the string, so it can include any characters
5083 but it must come last in the list of options (after /sender_retain
5086 PH/05 Some modifications to the interface to the fake nameserver for the new
5094 TK/01 Added the "success_on_redirect" address verification option. See
5095 NewStuff for rationale and an example.
5097 PH/01 Added support for SQLite, basic code supplied by David Woodhouse.
5099 PH/02 Patch to exigrep to allow it to work on syslog lines.
5101 PH/03 When creating an mbox file for a virus/spam scan, use fseek() instead of
5102 fread() to skip over the body file's header line, because in Cygwin the
5103 header line is locked and is inaccessible.
5105 PH/04 Added $message_exim_id, ultimately to replace $message_id (they will both
5106 co-exist for some time) to make it clear that it is the Exim ID that is
5107 referenced, not the Message-ID: header line.
5109 PH/05 Replaced all Tom's calls to snprintf() with calls to the internal
5110 string_format() function, because snprintf() does not exist on all
5113 PH/06 The use of forbid_filter_existstest now also locks out the use of the
5114 ${stat: expansion item.
5116 PH/07 Changed "SMTP protocol violation: synchronization error" into "SMTP
5117 protocol synchronization error", to keep the pedants happy.
5119 PH/08 Arrange for USE_INET_NTOA_FIX to be set in config.h for AIX systems as
5120 well as for IRIX systems, when gcc is being used. See the host.c source
5123 PH/09 Installed latest Cygwin configuration files from the Cygwin maintainer.
5125 PH/10 Named domain lists were not working if used in a queue_smtp_domains
5128 PH/11 Added support for the IGNOREQUOTA extension to LMTP, both to the lmtp
5129 transport and to the smtp transport in LMTP mode.
5131 TK/02 Remove one case of BASE64 error detection FTTB (undocumented anyway).
5133 PH/12 There was a missing call to search_tidyup() before the fork() in rda.c to
5134 run a filter in a subprocess. This could lead to confusion in subsequent
5135 lookups in the parent process. There should also be a search_tidyup() at
5136 the end of the subprocess.
5138 PH/13 Previously, if "verify = helo" was set in an ACL, the condition was true
5139 only if the host matched helo_try_verify_hosts, which caused the
5140 verification to occur when the EHLO/HELO command was issued. The ACL just
5141 tested the remembered result. Now, if a previous verification attempt has
5142 not happened, "verify = helo" does it there and then.
5144 JJ/01 exipick: added $message_exim_id variable (see 4.53-PH/04)
5146 TK/03 Fix log output including CR from clamd.
5148 PH/14 A reference to $reply_address when Reply-to: was empty and From: did not
5149 exist provoked a memory error which could cause a segfault.
5151 PH/15 Installed PCRE 6.2
5153 PH/17 Defined BIND_8_COMPAT in the Darwin os.h file.
5155 PH/18 Reversed 4.52/PH/17 because the HP-UX user found it wasn't the cause
5156 of the problem. Specifically, suggested +O2 rather than +O1 for the
5159 PH/19 Added sqlite_lock_timeout option (David Woodhouse's patch).
5161 PH/20 If a delivery was routed to a non-standard port by means of an SRV
5162 record, the port was not correctly logged when the outgoing_port log
5163 selector was set (it logged the transort's default port).
5165 PH/21 Added support for host-specific ports to manualroute, queryprogram,
5166 fallback_hosts, and "hosts" in the smtp transport.
5168 PH/22 If the log selector "outgoing_port" is set, the port is now also given on
5169 host errors such as "Connection refused".
5171 PH/23 Applied a patch to fix problems with exim-4.52 while doing radius
5172 authentication with radiusclient 0.4.9:
5174 - Error returned from rc_read_config was caught wrongly
5175 - Username/password not passed on to radius server due to wrong length.
5177 The presumption is that some radiusclient API changes for 4.51/PH/17
5178 were not taken care of correctly. The code is still untested by me (my
5179 Linux distribution still has 0.3.2 of radiusclient), but it was
5180 contributed by a Radius user.
5182 PH/24 When doing a callout, the value of $domain wasn't set correctly when
5183 expanding the "port" option of the smtp transport.
5185 TK/04 MIME ACL: Fix buffer underrun that occurs when EOF condition is met
5186 while reading a MIME header. Thanks to Tom Hughes for a patch.
5188 PH/24 Include config.h inside local_scan.h so that configuration settings are
5191 PH/25 Make $smtp_command_argument available after all SMTP commands. This means
5192 that in an ACL for RCPT (for example), you can examine exactly what was
5195 PH/26 Exim was recognizing IPv6 addresses of the form [IPv6:....] in EHLO
5196 commands, but it was not correctly comparing the address with the actual
5197 client host address. Thus, it would show the EHLO address in Received:
5198 header lines when this was not necessary.
5200 PH/27 Added the % operator to ${eval:}.
5202 PH/28 Exim tries to create and chdir to its spool directory when it starts;
5203 it should be ignoring failures (because with -C, for example, it has lost
5204 privilege). It wasn't ignoring creation failures other than "already
5207 PH/29 Added "crypteq" to the list of supported features that Exim outputs when
5210 PH/30 Fixed (presumably very longstanding) bug in exim_dbmbuild: if it failed
5211 because an input line was too long, either on its own, or by virtue of
5212 too many continuations, the temporary file was not being removed, and the
5213 return code was incorrect.
5215 PH/31 Missing "BOOL" in function definition in filtertest.c.
5217 PH/32 Applied Sieve patches from the maintainer.
5219 TK/05 Domainkeys: Accomodate for a minor API change in libdomainkeys 0.67.
5221 PH/33 Added "verify = not_blind".
5223 PH/34 There are settings for CHOWN_COMMAND and MV_COMMAND that can be used in
5224 Local/Makefile (with some defaults set). These are used in built scripts
5225 such as exicyclog, but they have never been used in the exim_install
5226 script (though there are many overriding facilities there). I have
5227 arranged that the exim_install script now takes note of these two
5230 PH/35 Installed configuration files for Dragonfly.
5232 PH/36 When a locally submitted message by a trusted user did not contain a
5233 From: header, and the sender address was obtained from -f or from an SMTP
5234 MAIL command, and the trusted user did not use -F to supply a sender
5235 name, $originator_name was incorrectly used when constructing a From:
5236 header. Furthermore, $originator_name was used for submission mode
5237 messages from external hosts without From: headers in a similar way,
5238 which is clearly wrong.
5240 PH/37 Added control=suppress_local_fixups.
5242 PH/38 When log_selector = +received_sender was set, and the addition of the
5243 sender made the log line's construction buffer exactly full, or one byte
5244 less than full, an overflow happened when the terminating "\n" was
5247 PH/39 Added a new log selector, "unknown_in_list", which provokes a log entry
5248 when the result of a list match is failure because a DNS lookup failed.
5250 PH/40 RM_COMMAND is now used in the building process.
5252 PH/41 Added a "distclean" target to the top-level Makefile; it deletes all
5253 the "build-* directories that it finds.
5255 PH/42 (But a TF fix): In a domain list, Exim incorrectly matched @[] if the IP
5256 address in a domain literal was a prefix of an interface address.
5258 PH/43 (Again a TF fix): In the dnslookup router, do not apply widen_domains
5259 when verifying a sender address, unless rewrite_headers is false.
5261 PH/44 Wrote a long comment about why errors_to addresses are verified as
5262 recipients, not senders.
5264 TF/01 Add missing LIBS=-lm to OS/Makefile-OpenBSD which was overlooked when
5265 the ratelimit ACL was added.
5267 PH/45 Added $smtp_command for the full command (cf $smtp_command_argument).
5269 PH/46 Added extra information about PostgreSQL errors to the error string.
5271 PH/47 Added an interface to a fake DNS resolver for use by the new test suite,
5272 avoiding the need to install special zones in a real server. This is
5273 backwards compatible; if it can't find the fake resolver, it drops back.
5274 Thus, both old and new test suites can be run.
5276 TF/02 Added util/ratelimit.pl
5278 TF/03 Minor fix to the ratelimit code to improve its behaviour in case the
5279 clock is set back in time.
5281 TF/04 Fix the ratelimit support in exim_fixdb. Patch provided by Brian
5282 Candler <B.Candler@pobox.com>.
5284 TF/05 The fix for PH/43 was not completely correct; widen_domains is always
5285 OK for addresses that are the result of redirections.
5287 PH/48 A number of further additions for the benefit of the new test suite,
5288 including a fake gethostbyname() that interfaces to the fake DNS resolver
5291 TF/06 The fix for widen_domains has also been applied to qualify_single and
5292 search_parents which are the other dnslookup options that can cause
5295 PH/49 Michael Haardt's randomized retrying, but as a separate retry parameter
5298 PH/50 Make never_users, trusted_users, admin_groups, trusted_groups expandable.
5300 TF/07 Exim produced the error message "an SRV record indicated no SMTP
5301 service" if it encountered an MX record with an empty target hostname.
5302 The message is now "an MX or SRV record indicated no SMTP service".
5304 TF/08 Change PH/13 introduced the possibility that verify=helo may defer,
5305 if the DNS of the sending site is misconfigured. This is quite a
5306 common situation. This change restores the behaviour of treating a
5307 helo verification defer as a failure.
5309 PH/51 If self=fail was set on a router, the bounce message did not include the
5310 actual error message.
5316 TF/01 Added support for Client SMTP Authorization. See NewStuff for details.
5318 PH/01 When a transport filter timed out in a pipe delivery, and the pipe
5319 command itself ended in error, the underlying message about the transport
5320 filter timeout was being overwritten with the pipe command error. Now the
5321 underlying error message should be appended to the second error message.
5323 TK/01 Fix poll() being unavailable on Mac OSX 10.2.
5325 PH/02 Reduce the amount of output that "make" produces by default. Full output
5326 can still be requested.
5328 PH/03 The warning log line about a condition test deferring for a "warn" verb
5329 was being output only once per connection, rather than after each
5330 occurrence (because it was using the same function as for successful
5331 "warn" verbs). This seems wrong, so I have changed it.
5333 TF/02 Two buglets in acl.c which caused Exim to read a few bytes of memory that
5334 it should not have, which might have caused a crash in the right
5335 circumstances, but probably never did.
5337 PH/04 Installed a modified version of Tony Finch's patch to make submission
5338 mode fix the return path as well as the Sender: header line, and to
5339 add a /name= option so that you can make the user's friendly name appear
5342 TF/03 Added the control = fakedefer ACL modifier.
5344 TF/04 Added the ratelimit ACL condition. See NewStuff for details. Thanks to
5345 Mark Lowes for thorough testing.
5347 TK/02 Rewrote SPF support to work with libspf2 versions >1.2.0.
5349 TK/03 Merged latest SRS patch from Miles Wilton.
5351 PH/05 There's a shambles in IRIX6 - it defines EX_OK in unistd.h which conflicts
5352 with the definition in sysexits.h (which is #included earlier).
5353 Fortunately, Exim does not actually use EX_OK. The code used to try to
5354 preserve the sysexits.h value, by assuming that macro definitions were
5355 scanned for macro replacements. I have been disabused of this notion,
5356 so now the code just undefines EX_OK before #including unistd.h.
5358 PH/06 There is a timeout for writing blocks of data, set by, e.g. data_timeout
5359 in the smtp transport. When a block could not be written in a single
5360 write() function, the timeout was being re-applied to each part-write.
5361 This seems wrong - if the receiver was accepting one byte at a time it
5362 would take for ever. The timeout is now adjusted when this happens. It
5363 doesn't have to be particularly precise.
5365 TK/04 Added simple SPF lookup method in EXPERIMENTAL_SPF. See NewStuff for
5366 details. Thanks to Chris Webb <chris@arachsys.com> for the patch!
5368 PH/07 Added "fullpostmaster" verify option, which does a check to <postmaster>
5369 without a domain if the check to <postmaster@domain> fails.
5371 SC/01 Eximstats: added -xls and the ability to specify output files
5372 (patch written by Frank Heydlauf).
5374 SC/02 Eximstats: use FileHandles for outputting results.
5376 SC/03 Eximstats: allow any combination of xls, txt, and html output.
5378 SC/04 Eximstats: fixed display of large numbers with -nvr option
5380 SC/05 Eximstats: fixed merging of reports with empty tables.
5382 SC/06 Eximstats: added the -include_original_destination flag
5384 SC/07 Eximstats: removed tabs and trailing whitespace.
5386 TK/05 Malware: Improve on aveserver error handling. Patch from Alex Miller.
5388 TK/06 MBOX spool code: Add real "From " MBOX separator line
5389 so the .eml file is really in mbox format (even though
5390 most programs do not really care). Patch from Alex Miller.
5392 TK/07 MBOX spool code: Add X-Envelope-From: and X-Envelope-To: headers.
5393 The latter is generated from $received_to and is only set if the
5394 message has one envelope recipient. SA can use these headers,
5395 obviously out-of-the-box. Patch from Alex Miller.
5397 PH/08 The ${def test on a variable was returning false if the variable's
5398 value was "0", contrary to what the specification has always said!
5399 The result should be true unless the variable is empty.
5401 PH/09 The syntax error of a character other than { following "${if
5402 def:variable_name" (after optional whitespace) was not being diagnosed.
5403 An expansion such as ${if def:sender_ident:{xxx}{yyy}} in which an
5404 accidental colon was present, for example, could give incorrect results.
5406 PH/10 Tidied the code in a number of places where the st_size field of a stat()
5407 result is used (not including appendfile, where other changes are about
5410 PH/11 Upgraded appendfile so that quotas larger than 2G are now supported.
5411 This involved changing a lot of size variables from int to off_t. It
5412 should work with maildirs and everything.
5414 TK/08 Apply fix provided by Michael Haardt to prevent deadlock in case of
5415 spamd dying while we are connected to it.
5417 TF/05 Fixed a ${extract error message typo reported by Jeremy Harris
5420 PH/12 Applied Alex Kiernan's patch for the API change for the error callback
5421 function for BDB 4.3.
5423 PH/13 Changed auto_thaw such that it does not apply to bounce messages.
5425 PH/14 Imported PCRE 6.0; this was more than just a trivial operation because
5426 the sources for PCRE have been re-arranged and more files are now
5429 PH/15 The code I had for printing potentially long long variables in PH/11
5430 above was not the best (it lost precision). The length of off_t variables
5431 is now inspected at build time, and an appropriate printing format (%ld
5432 or %lld) is chosen and #defined by OFF_T_FMT. We also define LONGLONG_T
5433 to be "long long int" or "long int". This is needed for the internal
5434 formatting function string_vformat().
5436 PH/16 Applied Matthew Newton's patch to exicyclog: "If log_file_path is set in
5437 the configuration file to be ":syslog", then the script "guesses" where
5438 the logs files are, rather than using the compiled in default. In our
5439 case the guess is not the same as the compiled default, so the script
5440 suddenly stopped working when I started to use syslog. The patch checks
5441 to see if log_file_path is "". If so, it attempts to read it from exim
5442 with no configuration file to get the compiled in version, before it
5443 falls back to the previous guessing code."
5445 TK/09 Added "prvs" and "prvscheck" expansion items. These help a lot with
5446 implementing BATV in an Exim configuration. See NewStuff for the gory
5449 PH/17 Applied Michael Haardt's patch for HP-UX, affecting only the os.h and
5450 Makefile that are specific to HP-UX.
5452 PH/18 If the "use_postmaster" option was set for a recipient callout together
5453 with the "random" option, the postmaster address was used as the MAIL
5454 FROM address for the random test, but not for the subsequent recipient
5455 test. It is now used for both.
5457 PH/19 Applied Michael Haardt's patch to update Sieve to RFC3028bis. "The
5458 patch removes a few documentation additions to RFC 3028, because the
5459 latest draft now contains them. It adds the new en;ascii-case comparator
5460 and a new error check for 8bit text in MIME parts. Comparator and
5461 require names are now matched exactly. I enabled the subaddress
5462 extension, but it is not well tested yet (read: it works for me)."
5464 PH/20 Added macros for time_t as for off_t (see PH/15 above) and used them to
5465 rework some of the code of TK/09 above to avoid the hardwired use of
5466 "%lld" and "long long". Replaced the call to snprintf() with a call to
5469 PH/21 Added some other messages to those in 4.51/PH/42, namely "All relevant MX
5470 records point to non-existent hosts", "retry timeout exceeded", and
5471 "retry time not reached for any host after a long failure period".
5473 PH/22 Fixed some oversights/typos causing bugs when Exim is compiled with
5474 experimental DomainKeys support:
5476 (1) The filter variables $n0-$n9 and $sn0-$sn9 were broken.
5477 (2) On an error such as an illegally used "control", the wrong name for
5478 the control was given.
5480 These problems did NOT occur unless DomainKeys support was compiled.
5482 PH/23 Added daemon_startup_retries and daemon_startup_sleep.
5484 PH/24 Added ${if match_ip condition.
5486 PH/25 Put debug statements on either side of calls to EXIM_DBOPEN() for hints
5487 databases so that it will be absolutely obvious if a crash occurs in the
5488 DB library. This is a regular occurrence (often caused by mis-matched
5491 PH/26 Insert a lot of missing (void) casts for functions such as chown(),
5492 chmod(), fcntl(), sscanf(), and other functions from stdio.h. These were
5493 picked up on a user's system that detects such things. There doesn't seem
5494 to be a gcc warning option for this - only an attribute that has to be
5495 put on the function's prototype. It seems that in Fedora Core 4 they have
5496 set this on a number of new functions. No doubt there will be more in due
5499 PH/27 If a dnslookup or manualroute router is set with verify=only, it need not
5500 specify a transport. However, if an address that was verified by such a
5501 router was the subject of a callout, Exim crashed because it tried to
5502 read the rcpt_include_affixes from the non-existent transport. Now it
5503 just assumes that the setting of that option is false. This bug was
5504 introduced by 4.51/PH/31.
5506 PH/28 Changed -d+all to exclude +memory, because that information is very
5507 rarely of interest, but it makes the output a lot bigger. People tend to
5508 do -d+all out of habit.
5510 PH/29 Removed support for the Linux-libc5 build, as it is obsolete and the
5511 code in os-type was giving problems when libc.so lives in lib64, like on
5514 PH/30 Exim's DNS code uses the original T_xxx names for DNS record times. These
5515 aren't the modern standard, and it seems that some systems' include files
5516 don't always have them. Exim was already checking for some of the newer
5517 ones like T_AAAA, and defining it itself. I've added checks for all the
5518 record types that Exim uses.
5520 PH/31 When using GnuTLS, if the parameters cache file did not exist, Exim was
5521 not automatically generating a new one, as it is supposed to. This
5522 prevented TLS from working. If the file did exist, but contained invalid
5523 data, a new version was generated, as expected. It was only the case of a
5524 non-existent file that was broken.
5526 TK/10 Domainkeys: Fix a bug in verification that caused a crash in conjunction
5527 with a change in libdomainkeys > 0.64.
5529 TK/11 Domainkeys: Change the logic how the "testing" policy flag is retrieved
5530 from DNS. If the selector record carries the flag, it now has
5531 precedence over the domain-wide flag.
5533 TK/12 Cleared some compiler warnings related to SPF, SRS and DK code.
5535 PH/32 In mua_wrapper mode, if an smtp transport configuration error (such as
5536 the use of a port name that isn't defined in /etc/services) occurred, the
5537 message was deferred as in a normal delivery, and thus remained on the
5538 spool, instead of being failed because of the mua_wrapper setting. This
5539 is now fixed, and I tidied up some of the mua_wrapper messages at the
5542 SC/08 Eximstats: whilst parsing the mainlog(s), store information about
5543 the messages in a hash of arrays rather than using individual hashes.
5544 This is a bit cleaner and results in dramatic memory savings, albeit
5545 at a slight CPU cost.
5547 SC/09 Eximstats: added the -show_rt<list> and the -show_dt<list> flags
5548 as requested by Marc Sherman.
5550 SC/10 Eximstats: added histograms for user specified patterns as requested
5553 SC/11 Eximstats: v1.43 - bugfix for pattern histograms with -h0 specified.
5555 PH/33 Patch from the Cygwin maintainer to add "b" to all occurences of
5556 fopen() in the content-scanning modules that did not already have it.
5562 TK/01 Added Yahoo DomainKeys support via libdomainkeys. See
5563 doc/experimental-spec.txt for details. (http://domainkeys.sf.net)
5565 TK/02 Fix ACL "control" statement not being available in MIME ACL.
5567 TK/03 Fix ACL "regex" condition not being available in MIME ACL.
5569 PH/01 Installed a patch from the Sieve maintainer that allows -bf to be used
5570 to test Sieve filters that use "vacation".
5572 PH/02 Installed a slightly modified version of Nikos Mavrogiannopoulos' patch
5573 that changes the way the GnuTLS parameters are stored in the cache file.
5574 The new format can be generated externally. For backward compatibility,
5575 if the data in the cache doesn't make sense, Exim assumes it has read an
5576 old-format file, and it generates new data and writes a new file. This
5577 means that you can't go back to an older release without removing the
5580 PH/03 A redirect router that has both "unseen" and "one_time" set does not
5581 work if there are any delivery delays because "one_time" forces the
5582 parent to be marked "delivered", so its unseen clone is never tried
5583 again. For this reason, Exim now forbids the simultaneous setting of
5586 PH/04 Change 4.11/85 fixed an obscure bug concerned with addresses that are
5587 redirected to themselves ("homonym" addresses). Read the long ChangeLog
5588 entry if you want to know the details. The fix, however, neglected to
5589 consider the case when local delivery batching is involved. The test for
5590 "previously delivered" was not happening when checking to see if an
5591 address could be batched with a previous (undelivered) one; under
5592 certain circumstances this could lead to multiple deliveries to the same
5595 PH/05 Renamed the macro SOCKLEN_T as EXIM_SOCKLEN_T because AIX uses SOCKLEN_T
5596 in its include files, and this causes problems building Exim.
5598 PH/06 A number of "verify =" ACL conditions have no options (e.g. verify =
5599 header_syntax) but Exim was just ignoring anything given after a slash.
5600 In particular, this caused confusion with an attempt to use "verify =
5601 reverse_host_lookup/defer_ok". An error is now given when options are
5602 supplied for verify items that do not have them. (Maybe reverse_host_
5603 lookup should have a defer_ok option, but that's a different point.)
5605 PH/07 Increase the size of the buffer for incoming SMTP commands from 512 (as
5606 defined by RFC 821) to 2048, because there were problems with some AUTH
5607 commands, and RFC 1869 says the size should be increased for extended
5608 SMTP commands that take arguments.
5610 PH/08 Added ${dlfunc dynamically loaded function for expansion (code from Tony
5613 PH/09 Previously, an attempt to use ${perl when it wasn't compiled gave an
5614 "unknown" error; now it says that the functionality isn't in the binary.
5616 PH/10 Added a nasty fudge to try to recognize and flatten LDAP passwords in
5617 an address' error message when a string expansion fails (syntax or
5618 whatever). Otherwise the password may appear in the log. Following change
5619 PH/42 below, there is no longer a chance of it appearing in a bounce
5622 PH/11 Installed exipick version 20050225.0 from John Jetmore.
5624 PH/12 If the last host in a fallback_hosts list was multihomed, only the first
5625 of its addresses was ever tried. (Bugzilla bug #2.)
5627 PH/13 If "headers_add" in a transport didn't end in a newline, Exim printed
5628 the result incorrectly in the debug output. (It correctly added a newline
5629 to what was transported.)
5631 TF/01 Added $received_time.
5633 PH/14 Modified the default configuration to add an acl_smtp_data ACL, with
5634 commented out examples of how to interface to a virus scanner and to
5635 SpamAssassin. Also added commented examples of av_scanner and
5636 spamd_address settings.
5638 PH/15 Further to TK/02 and TK/03 above, tidied up the tables of what conditions
5639 and controls are allowed in which ACLs. There were a couple of minor
5640 errors. Some of the entries in the conditions table (which is a table of
5641 where they are NOT allowed) were getting very unwieldy; rewrote them as a
5642 negation of where the condition IS allowed.
5644 PH/16 Installed updated OS/os.c-cygwin from the Cygwin maintainer.
5646 PH/17 The API for radiusclient changed at release 0.4.0. Unfortunately, the
5647 header file does not have a version number, so I've had to invent a new
5648 value for RADIUS_LIB_TYPE, namely "RADIUSCLIENTNEW" to request the new
5649 API. The code is untested by me (my Linux distribution still has 0.3.2 of
5650 radiusclient), but it was contributed by a Radius user.
5652 PH/18 Installed Lars Mainka's patch for the support of CRL collections in
5653 files or directories, for OpenSSL.
5655 PH/19 When an Exim process that is running as root has to create an Exim log
5656 file, it does so in a subprocess that runs as exim:exim so as to get the
5657 ownership right at creation (otherwise, other Exim processes might see
5658 the file with the wrong ownership). There was no test for failure of this
5659 fork() call, which would lead to the process getting stuck as it waited
5660 for a non-existent subprocess. Forks do occasionally fail when resources
5661 run out. I reviewed all the other calls to fork(); they all seem to check
5664 PH/20 When checking for unexpected SMTP input at connect time (before writing
5665 the banner), Exim was not dealing correctly with a non-positive return
5666 from the read() function. If the client had disconnected by this time,
5667 the result was a log entry for a synchronization error with an empty
5668 string after "input=" when read() returned zero. If read() returned -1
5669 (an event I could not check), uninitialized data bytes were printed.
5670 There were reports of junk text (parts of files, etc) appearing after
5673 PH/21 Added acl_not_smtp_mime to allow for MIME scanning for non-SMTP messages.
5675 PH/22 Added support for macro redefinition, and (re)definition in between
5676 driver and ACL definitions.
5678 PH/23 The cyrus_sasl authenticator was expanding server_hostname, but then
5679 forgetting to use the resulting value; it was using the unexpanded value.
5681 PH/24 The cyrus_sasl authenticator was advertising mechanisms for which it
5682 hadn't been configured. The fix is from Juergen Kreileder, who
5683 understands it better than I do:
5685 "Here's what I see happening with three configured cyrus_sasl
5686 authenticators configured (plain, login, cram-md5):
5688 On startup auth_cyrus_sasl_init() gets called for each of these.
5689 This means three calls to sasl_listmech() without a specified mech_list.
5690 => SASL tests which mechs of all available mechs actually work
5691 => three warnings about OTP not working
5692 => the returned list contains: plain, login, cram-md5, digest-md5, ...
5694 With the patch, sasl_listmech() also gets called three times. But now
5695 SASL's mech_list option is set to the server_mech specified in the the
5696 authenticator. Or in other words, the answer from sasl_listmech()
5697 gets limited to just the mech you're testing for (which is different
5699 => the return list contains just 'plain' or 'login', 'cram-md5' or
5700 nothing depending on the value of ob->server_mech.
5702 I've just tested the patch: Authentication still works fine,
5703 unavailable mechs specified in the exim configuration are still
5704 caught, and the auth.log warnings about OTP are gone."
5706 PH/25 When debugging is enabled, the contents of the command line are added
5707 to the debugging output, even when log_selector=+arguments is not
5710 PH/26 Change scripts/os-type so that when "uname -s" returns just "GNU", the
5711 answer is "GNU", and only if the return is "GNU/something" is the answer
5714 PH/27 $acl_verify_message is now set immediately after the failure of a
5715 verification in an ACL, and so is available in subsequent modifiers. In
5716 particular, the message can be preserved by coding like this:
5718 warn !verify = sender
5719 set acl_m0 = $acl_verify_message
5721 Previously, $acl_verify_message was set only while expanding "message"
5722 and "log_message" when a very denied access.
5724 PH/28 Modified OS/os.c-Linux with
5726 -#ifndef OS_LOAD_AVERAGE
5727 +#if !defined(OS_LOAD_AVERAGE) && defined(__linux__)
5729 to make Exim compile on kfreebsd-gnu. (I'm totally confused about the
5730 nomenclature these days.)
5732 PH/29 Installed patch from the Sieve maintainer that adds the options
5733 sieve_useraddress and sieve_subaddress to the redirect router.
5735 PH/30 In these circumstances:
5736 . Two addresses routed to the same list of hosts;
5737 . First host does not offer TLS;
5738 . First host accepts first address;
5739 . First host gives temporary error to second address;
5740 . Second host offers TLS and a TLS session is established;
5741 . Second host accepts second address.
5742 Exim incorrectly logged both deliveries with the TLS parameters (cipher
5743 and peerdn, if requested) that were in fact used only for the second
5746 PH/31 When doing a callout as part of verifying an address, Exim was not paying
5747 attention to any local part prefix or suffix that was matched by the
5748 router that accepted the address. It now behaves in the same way as it
5749 does for delivery: the affixes are removed from the local part unless
5750 rcpt_include_affixes is set on the transport.
5752 PH/32 Add the sender address, as F=<...>, to the log line when logging a
5753 timeout during the DATA phase of an incoming message.
5755 PH/33 Sieve envelope tests were broken for match types other than :is. I have
5756 applied a patch sanctioned by the Sieve maintainer.
5758 PH/34 Change 4.50/80 broke Exim in that it could no longer handle cases where
5759 the uid or gid is negative. A case of a negative gid caused this to be
5760 noticed. The fix allows for either to be negative.
5762 PH/35 ACL_WHERE_MIME is now declared unconditionally, to avoid too much code
5763 clutter, but the tables that are indexed by ACL_WHERE_xxx values had been
5766 PH/36 The change PH/12 above was broken. Fixed it.
5768 PH/37 Exim used to check for duplicate addresses in the middle of routing, on
5769 the grounds that routing the same address twice would always produce the
5770 same answer. This might have been true once, but it is certainly no
5771 longer true now. Routing a child address may depend on the previous
5772 routing that produced that child. Some complicated redirection strategies
5773 went wrong when messages had multiple recipients, and made Exim's
5774 behaviour dependent on the order in which the addresses were given.
5776 I have moved the duplicate checking until after the routing is complete.
5777 Exim scans the addresses that are assigned to local and remote
5778 transports, and removes any duplicates. This means that more work will be
5779 done, as duplicates will always all be routed, but duplicates are
5780 presumably rare, so I don't expect this is of any significance.
5782 For deliveries to pipes, files, and autoreplies, the duplicate checking
5783 still happens during the routing process, since they are not going to be
5786 PH/38 Installed a patch from Ian Freislich, with the agreement of Tom Kistner.
5787 It corrects a timeout issue with spamd. This is Ian's comment: "The
5788 background is that sometimes spamd either never reads data from a
5789 connection it has accepted, or it never writes response data. The exiscan
5790 spam.[ch] uses a 3600 second timeout on spamd socket reads, further, it
5791 blindly assumes that writes won't block so it may never time out."
5793 PH/39 Allow G after quota size as well as K and M.
5795 PH/40 The value set for $authenticated_id in an authenticator may not contain
5796 binary zeroes or newlines because the value is written to log lines and
5797 to spool files. There was no check on this. Now the value is run through
5798 the string_printing() function so that such characters are converted to
5799 printable escape sequences.
5801 PH/41 $message_linecount is a new variable that contains the total number of
5802 lines in the message. Compare $body_linecount, which is the count for the
5805 PH/42 Exim no longer gives details of delivery errors for specific addresses in
5806 bounce and delay warning messages, except in certain special cases, which
5809 (a) An SMTP error message from a remote host;
5810 (b) A message specified in a :fail: redirection;
5811 (c) A message specified in a "fail" command in a system filter;
5812 (d) A message specified in a FAIL return from the queryprogram router;
5813 (e) A message specified by the cannot_route_message router option.
5815 In these cases only, Exim does include the error details in bounce and
5816 warning messages. There are also a few cases where bland messages such
5817 as "unrouteable address" or "local delivery error" are given.
5819 PH/43 $value is now also set for the "else" part of a ${run expansion.
5821 PH/44 Applied patch from the Sieve maintainer: "The vacation draft is still
5822 being worked on, but at least Exim now implements the latest version to
5825 PH/45 In a pipe transport, although a timeout while waiting for the pipe
5826 process to complete was treated as a delivery failure, a timeout while
5827 writing the message to the pipe was logged, but erroneously treated as a
5828 successful delivery. Such timeouts include transport filter timeouts. For
5829 consistency with the overall process timeout, these timeouts are now
5830 treated as errors, giving rise to delivery failures by default. However,
5831 there is now a new Boolean option for the pipe transport called
5832 timeout_defer, which, if set TRUE, converts the failures into defers for
5833 both kinds of timeout. A transport filter timeout is now identified in
5836 PH/46 The "scripts/Configure-config.h" script calls "make" at one point. On
5837 systems where "make" and "gmake" are different, calling "gmake" at top
5838 level broke things. I've arranged for the value of $(MAKE) to be passed
5839 from the Makefile to this script so that it can call the same version of
5843 A note about Exim versions 4.44 and 4.50
5844 ----------------------------------------
5846 Exim 4.50 was meant to be the next release after 4.43. It contains a lot of
5847 changes of various kinds. As a consequence, a big documentation update was
5848 needed. This delayed the release for rather longer than seemed good, especially
5849 in the light of a couple of (minor) security issues. Therefore, the changes
5850 that fixed bugs were backported into 4.43, to create a 4.44 maintenance
5851 release. So 4.44 and 4.50 are in effect two different branches that both start
5854 I have left the 4.50 change log unchanged; it contains all the changes since
5855 4.43. The change log for 4.44 is below; many of its items are identical to
5856 those for 4.50. This seems to be the most sensible way to preserve the
5857 historical information.
5863 1. Minor wording change to the doc/README.SIEVE file.
5865 2. Change 4.43/35 introduced a bug: if quota_filecount was set, the
5866 computation of the current number of files was incorrect.
5868 3. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
5869 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
5872 4. Give more explanation in the error message when the command for a transport
5873 filter fails to execute.
5875 5. There are several places where Exim runs a non-Exim command in a
5876 subprocess. The SIGUSR1 signal should be disabled for these processes. This
5877 was being done only for the command run by the queryprogram router. It is
5878 now done for all such subprocesses. The other cases are: ${run, transport
5879 filters, and the commands run by the lmtp and pipe transports.
5881 6. Added CONFIGURE_GROUP build-time option.
5883 7. Some older OS have a limit of 256 on the maximum number of file
5884 descriptors. Exim was using setrlimit() to set 1000 as a large value
5885 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
5886 systems. I've change it so that if it can't get 1000, it tries for 256.
5888 8. "control=submission" was allowed, but had no effect, in a DATA ACL. This
5889 was an oversight, and furthermore, ever since the addition of extra
5890 controls (e.g. 4.43/32), the checks on when to allow different forms of
5891 "control" were broken. There should now be diagnostics for all cases when a
5892 control that does not make sense is encountered.
5894 9. Added the /retain_sender option to "control=submission".
5896 10. $recipients is now available in the predata ACL (oversight).
5898 11. Tidy the search cache before the fork to do a delivery from a message
5899 received from the command line. Otherwise the child will trigger a lookup
5900 failure and thereby defer the delivery if it tries to use (for example) a
5901 cached ldap connection that the parent has called unbind on.
5903 12. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
5904 of $address_data from the recipient verification was clobbered by the
5905 sender verification.
5907 13. The value of address_data from a sender verification is now available in
5908 $sender_address_data in subsequent conditions in the ACL statement.
5910 14. Added forbid_sieve_filter and forbid_exim_filter to the redirect router.
5912 15. Added a new option "connect=<time>" to callout options, to set a different
5915 16. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
5916 was its contents. (It was OK if the option was not defined at all.)
5918 17. A "Completed" log line is now written for messages that are removed from
5919 the spool by the -Mrm option.
5921 18. New variables $sender_verify_failure and $recipient_verify_failure contain
5922 information about exactly what failed.
5924 19. Added -dd to debug only the daemon process.
5926 20. Incorporated Michael Haardt's patch to ldap.c for improving the way it
5927 handles timeouts, both on the server side and network timeouts. Renamed the
5928 CONNECT parameter as NETTIMEOUT (but kept the old name for compatibility).
5930 21. The rare case of EHLO->STARTTLS->HELO was setting the protocol to "smtp".
5931 It is now set to "smtps".
5933 22. $host_address is now set to the target address during the checking of
5934 ignore_target_hosts.
5936 23. When checking ignore_target_hosts for an ipliteral router, no host name was
5937 being passed; this would have caused $sender_host_name to have been used if
5938 matching the list had actually called for a host name (not very likely,
5939 since this list is usually IP addresses). A host name is now passed as
5942 24. Changed the calls that set up the SIGCHLD handler in the daemon to use the
5943 code that specifies a non-restarting handler (typically sigaction() in
5944 modern systems) in an attempt to fix a rare and obscure crash bug.
5946 25. Narrowed the window for a race in the daemon that could cause it to ignore
5947 SIGCHLD signals. This is not a major problem, because they are used only to
5948 wake it up if nothing else does.
5950 26. A malformed maildirsize file could cause Exim to calculate negative values
5951 for the mailbox size or file count. Odd effects could occur as a result.
5952 The maildirsize information is now recalculated if the size or filecount
5955 27. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
5956 support for a long time. Removed HAVE_SYS_VFS_H.
5958 28. Installed the latest version of exipick from John Jetmore.
5960 29. In an address list, if the pattern was not a regular expression, an empty
5961 subject address (from a bounce message) matched only if the pattern was an
5962 empty string. Non-empty patterns were not even tested. This was the wrong
5963 because it is perfectly reasonable to use an empty address as part of a
5964 database query. An empty address is now tested by patterns that are
5965 lookups. However, all the other forms of pattern expect the subject to
5966 contain a local part and a domain, and therefore, for them, an empty
5967 address still always fails if the pattern is not itself empty.
5969 30. Exim went into a mad DNS loop when attempting to do a callout where the
5970 host was specified on an smtp transport, and looking it up yielded more
5971 than one IP address.
5973 31. Re-factored the code for checking spool and log partition space into a
5974 function that finds that data and another that does the check. The former
5975 is then used to implement four new variables: $spool_space, $log_space,
5976 $spool_inodes, and $log_inodes.
5978 32. The RFC2047 encoding function was originally intended for short strings
5979 such as real names; it was not keeping to the 75-character limit for
5980 encoded words that the RFC imposes. It now respects the limit, and
5981 generates multiple encoded words if necessary. To be on the safe side, I
5982 have increased the buffer size for the ${rfc2047: expansion operator from
5985 33. It is now permitted to omit both strings after an "if" condition; if the
5986 condition is true, the result is "true". As before, when the second string
5987 is omitted, a false condition yields an empty string. This makes it less
5988 cumbersome to write custom ACL and router conditions.
5990 34. Failure to deliver a bounce message always caused it to be frozen, even if
5991 there was an errors_to setting on the router. The errors_to setting is now
5994 35. If an IPv6 address is given for -bh or -bhc, it is now converted to the
5995 canonical form (fully expanded) before being placed in
5996 $sender_host_address.
5998 36. The table in the code that translates DNS record types into text (T_A to
5999 "A" for instance) was missing entries for NS and CNAME. It is just possible
6000 that this could have caused confusion if both these types were looked up
6001 for the same domain, because the text type is used as part of Exim's
6002 per-process caching. But the chance of anyone hitting this buglet seems
6005 37. The dnsdb lookup has been extended in a number of ways.
6007 (1) There is a new type, "zns", which walks up the domain tree until it
6008 finds some nameserver records. It should be used with care.
6010 (2) There is a new type, "mxh", which is like "mx" except that it returns
6011 just the host names, not the priorities.
6013 (3) It is now possible to give a list of domains (or IP addresses) to be
6014 looked up. The behaviour when one of the lookups defers can be
6015 controlled by a keyword.
6017 (4) It is now possible to specify the separator character for use when
6018 multiple records are returned.
6020 38. The dnslists ACL condition has been extended: it is now possible to supply
6021 a list of IP addresses and/or domains to be looked up in a particular DNS
6024 39. Added log_selector=+queue_time_overall.
6026 40. When running the queue in the test harness, wait just a tad after forking a
6027 delivery process, to get repeatability of debugging output.
6029 41. Include certificate and key file names in error message when GnuTLS fails
6030 to set them up, because the GnuTLS error message doesn't include the name
6031 of the failing file when there is a problem reading it.
6033 42. Allow both -bf and -bF in the same test run.
6035 43. Did the same fix as 41 above for OpenSSL, which had the same infelicity.
6037 44. The "Exiscan patch" is now merged into the mainline Exim source.
6039 45. Sometimes the final signoff response after QUIT could fail to get
6040 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
6041 < 0 before doing a fflush(). This bug looks as though it goes back to the
6042 introduction of TLS in release 3.20, but "sometimes" must have been rare
6043 because the tests only now provoked it.
6045 46. Reset the locale to "C" after calling embedded Perl, in case it was changed
6046 (this can affect the format of dates).
6048 47. exim_tidydb, when checking for the continued existence of a message for
6049 which it has found a message-specific retry record, was not finding
6050 messages that were in split spool directories. Consequently, it was
6051 deleting retry records that should have stayed in existence.
6053 48. Steve fixed some bugs in eximstats.
6055 49. The SPA authentication driver was not abandoning authentication and moving
6056 on to the next authenticator when an expansion was forced to fail,
6057 contradicting the general specification for all authenticators. Instead it
6058 was generating a temporary error. It now behaves as specified.
6060 50. The default ordering of permitted cipher suites for GnuTLS was pessimal
6061 (the order specifies the preference for clients). The order is now AES256,
6062 AES128, 3DES, ARCFOUR128.
6064 51. Small patch to Sieve code - explicitly set From: when generating an
6067 52. Exim crashed if a remote delivery caused a very long error message to be
6068 recorded - for instance if somebody sent an entire SpamAssassin report back
6069 as a large number of 550 error lines. This bug was coincidentally fixed by
6070 increasing the size of one of Exim's internal buffers (big_buffer) that
6071 happened as part of the Exiscan merge. However, to be on the safe side, I
6072 have made the code more robust (and fixed the comments that describe what
6075 53. Now that there can be additional text after "Completed" in log lines (if
6076 the queue_time_overall log selector is set), a one-byte patch to exigrep
6077 was needed to allow it to recognize "Completed" as not the last thing in
6080 54. The LDAP lookup was not handling a return of LDAP_RES_SEARCH_REFERENCE. A
6081 patch that reportedly fixes this has been added. I am not expert enough to
6082 create a test for it. This is what the patch creator wrote:
6084 "I found a little strange behaviour of ldap code when working with
6085 Windows 2003 AD Domain, where users was placed in more than one
6086 Organization Units. When I tried to give exim partial DN, the exit code
6087 of ldap_search was unknown to exim because of LDAP_RES_SEARCH_REFERENCE.
6088 But simultaneously result of request was absolutely normal ldap result,
6089 so I produce this patch..."
6091 Later: it seems that not all versions of LDAP support LDAP_RES_SEARCH_
6092 REFERENCE, so I have modified the code to exclude the patch when that macro
6095 55. Some experimental protocols are using DNS PTR records for new purposes. The
6096 keys for these records are domain names, not reversed IP addresses. The
6097 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
6098 leaves it alone. Component reversal etc. now happens only for IP addresses.
6101 56. Improve error message when ldap_search() fails in OpenLDAP or Solaris LDAP.
6103 57. Double the size of the debug message buffer (to 2048) so that more of very
6104 long debug lines gets shown.
6106 58. The exicyclog utility now does better if the number of log files to keep
6107 exceeds 99. In this case, it numbers them 001, 002 ... instead of 01, 02...
6109 59. Two changes related to the smtp_active_hostname option:
6111 (1) $smtp_active_hostname is now available as a variable.
6112 (2) The default for smtp_banner uses $smtp_active_hostname instead
6113 of $primary_hostname.
6115 60. The host_aton() function is supposed to be passed a string that is known
6116 to be a valid IP address. However, in the case of IPv6 addresses, it was
6117 not checking this. This is a hostage to fortune. Exim now panics and dies
6118 if the condition is not met. A case was found where this could be provoked
6119 from a dnsdb PTR lookup with an IPv6 address that had more than 8
6120 components; fortuitously, this particular loophole had already been fixed
6121 by change 4.50/55 above.
6123 If there are any other similar loopholes, the new check in host_aton()
6124 itself should stop them being exploited. The report I received stated that
6125 data on the command line could provoke the exploit when Exim was running as
6126 exim, but did not say which command line option was involved. All I could
6127 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
6128 running as the user.
6131 61. There was a buffer overflow vulnerability in the SPA authentication code
6132 (which came originally from the Samba project). I have added a test to the
6133 spa_base64_to_bits() function which I hope fixes it.
6136 62. Configuration update for GNU/Hurd and variations. Updated Makefile-GNU and
6137 os.h-GNU, and added configuration files for GNUkFreeBSD and GNUkNetBSD.
6139 63. The daemon start-up calls getloadavg() while still root for those OS that
6140 need the first call to be done as root, but it missed one case: when
6141 deliver_queue_load_max is set with deliver_drop_privilege. This is
6142 necessary for the benefit of the queue runner, because there is no re-exec
6143 when deliver_drop_privilege is set.
6145 64. A call to exiwhat cut short delays set up by "delay" modifiers in ACLs.
6146 This has been fixed.
6148 65. Caching of lookup data for "hosts =" ACL conditions, when a named host list
6149 was in use, was not putting the data itself into the right store pool;
6150 consequently, it could be overwritten for a subsequent message in the same
6151 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
6154 66. Added hosts_max_try_hardlimit to the smtp transport, default 50.
6156 67. The string_is_ip_address() function returns 0, 4, or 6, for "no an IP
6157 address", "IPv4 address", and "IPv6 address", respectively. Some calls of
6158 the function were treating the return as a boolean value, which happened to
6159 work because 0=false and not-0=true, but is not correct code.
6161 68. The host_aton() function was not handling scoped IPv6 addresses (those
6162 with, for example, "%eth0" on the end) correctly.
6164 69. Fixed some compiler warnings in acl.c for the bitmaps specified with
6165 negated items (that is, ~something) in unsigned ints. Some compilers
6166 apparently mutter when there is no cast.
6168 70. If an address verification called from an ACL failed, and did not produce a
6169 user-specific message (i.e. there was only a "system" message), nothing was
6170 put in $acl_verify_message. In this situation, it now puts the system
6173 71. Change 4.23/11 added synchronization checking at the start of an SMTP
6174 session; change 4.31/43 added the unwanted input to the log line - except
6175 that it did not do this in the start of session case. It now does.
6177 72. After a timeout in a callout SMTP session, Exim still sent a QUIT command.
6178 This is wrong and can cause the other end to generate a synchronization
6179 error if it is another Exim or anything else that does the synchronization
6180 check. A QUIT command is no longer sent after a timeout.
6182 73. $host_lookup_deferred has been added, to make it easier to detect DEFERs
6183 during host lookups.
6185 74. The defer_ok option of callout verification was not working if it was used
6186 when verifying addresses in header lines, that is, for this case:
6188 verify = header_sender/callout=defer_ok
6190 75. A backgrounded daemon closed stdin/stdout/stderr on entry; this meant that
6191 those file descriptors could be used for SMTP connections. If anything
6192 wrote to stderr (the example that came up was "warn" in embedded Perl), it
6193 could be sent to the SMTP client, causing chaos. The daemon now opens
6194 stdin, stdout, and stderr to /dev/null when it puts itself into the
6197 76. Arrange for output from Perl's "warn" command to be written to Exim's main
6198 log by default. The user can override this with suitable Perl magic.
6200 77. The use of log_message on a "discard" ACL verb, which is supposed to add to
6201 the log message when discard triggers, was not working for the DATA ACL or
6202 for the non-SMTP ACL.
6204 78. Error message wording change in sieve.c.
6206 79. If smtp_accept_max_per_host was set, the number of connections could be
6207 restricted to fewer than expected, because the daemon was trying to set up
6208 a new connection before checking whether the processes handling previous
6209 connections had finished. The check for completed processes is now done
6210 earlier. On busy systems, this bug wouldn't be noticed because something
6211 else would have woken the daemon, and it would have reaped the completed
6214 80. If a message was submitted locally by a user whose login name contained one
6215 or more spaces (ugh!), the spool file that Exim wrote was not re-readable.
6216 It caused a spool format error. I have fixed the spool reading code. A
6217 related problem was that the "from" clause in the Received: line became
6218 illegal because of the space(s). It is now covered by ${quote_local_part.
6220 81. Included the latest eximstats from Steve (adds average sizes to HTML Top
6223 82. Updated OS/Makefile-AIX as per message from Mike Meredith.
6225 83. Patch from Sieve maintainer to fix unterminated string problem in
6226 "vacation" handling.
6228 84. Some minor changes to the Linux configuration files to help with other
6229 OS variants using glibc.
6231 85. One more patch for Sieve to update vacation handling to latest spec.
6234 ----------------------------------------------------
6235 See the note above about the 4.44 and 4.50 releases.
6236 ----------------------------------------------------
6242 1. Change 4.43/35 introduced a bug that caused file counts to be
6243 incorrectly computed when quota_filecount was set in an appendfile
6246 2. Closing a stable door: arrange to panic-die if setitimer() ever fails. The
6247 bug fixed in 4.43/37 would have been diagnosed quickly if this had been in
6250 3. Give more explanation in the error message when the command for a transport
6251 filter fails to execute.
6253 4. There are several places where Exim runs a non-Exim command in a
6254 subprocess. The SIGUSR1 signal should be disabled for these processes. This
6255 was being done only for the command run by the queryprogram router. It is
6256 now done for all such subprocesses. The other cases are: ${run, transport
6257 filters, and the commands run by the lmtp and pipe transports.
6259 5. Some older OS have a limit of 256 on the maximum number of file
6260 descriptors. Exim was using setrlimit() to set 1000 as a large value
6261 unlikely to be exceeded. Change 4.43/17 caused a lot of logging on these
6262 systems. I've change it so that if it can't get 1000, it tries for 256.
6264 6. "control=submission" was allowed, but had no effect, in a DATA ACL. This
6265 was an oversight, and furthermore, ever since the addition of extra
6266 controls (e.g. 4.43/32), the checks on when to allow different forms of
6267 "control" were broken. There should now be diagnostics for all cases when a
6268 control that does not make sense is encountered.
6270 7. $recipients is now available in the predata ACL (oversight).
6272 8. Tidy the search cache before the fork to do a delivery from a message
6273 received from the command line. Otherwise the child will trigger a lookup
6274 failure and thereby defer the delivery if it tries to use (for example) a
6275 cached ldap connection that the parent has called unbind on.
6277 9. If verify=recipient was followed by verify=sender in a RCPT ACL, the value
6278 of $address_data from the recipient verification was clobbered by the
6279 sender verification.
6281 10. If FIXED_NEVER_USERS was defined, but empty, Exim was assuming the uid 0
6282 was its contents. (It was OK if the option was not defined at all.)
6284 11. A "Completed" log line is now written for messages that are removed from
6285 the spool by the -Mrm option.
6287 12. $host_address is now set to the target address during the checking of
6288 ignore_target_hosts.
6290 13. When checking ignore_target_hosts for an ipliteral router, no host name was
6291 being passed; this would have caused $sender_host_name to have been used if
6292 matching the list had actually called for a host name (not very likely,
6293 since this list is usually IP addresses). A host name is now passed as
6296 14. Changed the calls that set up the SIGCHLD handler in the daemon to use the
6297 code that specifies a non-restarting handler (typically sigaction() in
6298 modern systems) in an attempt to fix a rare and obscure crash bug.
6300 15. Narrowed the window for a race in the daemon that could cause it to ignore
6301 SIGCHLD signals. This is not a major problem, because they are used only to
6302 wake it up if nothing else does.
6304 16. A malformed maildirsize file could cause Exim to calculate negative values
6305 for the mailbox size or file count. Odd effects could occur as a result.
6306 The maildirsize information is now recalculated if the size or filecount
6309 17. Added HAVE_SYS_STATVFS_H to the os.h file for Linux, as it has had this
6310 support for a long time. Removed HAVE_SYS_VFS_H.
6312 18. Updated exipick to current release from John Jetmore.
6314 19. Allow an empty sender to be matched against a lookup in an address list.
6315 Previously the only cases considered were a regular expression, or an
6318 20. Exim went into a mad DNS lookup loop when doing a callout where the
6319 host was specified on the transport, if the DNS lookup yielded more than
6322 21. The RFC2047 encoding function was originally intended for short strings
6323 such as real names; it was not keeping to the 75-character limit for
6324 encoded words that the RFC imposes. It now respects the limit, and
6325 generates multiple encoded words if necessary. To be on the safe side, I
6326 have increased the buffer size for the ${rfc2047: expansion operator from
6329 22. Failure to deliver a bounce message always caused it to be frozen, even if
6330 there was an errors_to setting on the router. The errors_to setting is now
6333 23. If an IPv6 address is given for -bh or -bhc, it is now converted to the
6334 canonical form (fully expanded) before being placed in
6335 $sender_host_address.
6337 24. Updated eximstats to version 1.33
6339 25. Include certificate and key file names in error message when GnuTLS fails
6340 to set them up, because the GnuTLS error message doesn't include the name
6341 of the failing file when there is a problem reading it.
6343 26. Expand error message when OpenSSL has problems setting up cert/key files.
6346 27. Reset the locale to "C" after calling embedded Perl, in case it was changed
6347 (this can affect the format of dates).
6349 28. exim_tidydb, when checking for the continued existence of a message for
6350 which it has found a message-specific retry record, was not finding
6351 messages that were in split spool directories. Consequently, it was
6352 deleting retry records that should have stayed in existence.
6354 29. eximstats updated to version 1.35
6355 1.34 - allow eximstats to parse syslog lines as well as mainlog lines
6356 1.35 - bugfix such that pie charts by volume are generated correctly
6358 30. The SPA authentication driver was not abandoning authentication and moving
6359 on to the next authenticator when an expansion was forced to fail,
6360 contradicting the general specification for all authenticators. Instead it
6361 was generating a temporary error. It now behaves as specified.
6363 31. The default ordering of permitted cipher suites for GnuTLS was pessimal
6364 (the order specifies the preference for clients). The order is now AES256,
6365 AES128, 3DES, ARCFOUR128.
6367 31. Small patch to Sieve code - explicitly set From: when generating an
6370 32. Exim crashed if a remote delivery caused a very long error message to be
6371 recorded - for instance if somebody sent an entire SpamAssassin report back
6372 as a large number of 550 error lines. This bug was coincidentally fixed by
6373 increasing the size of one of Exim's internal buffers (big_buffer) that
6374 happened as part of the Exiscan merge. However, to be on the safe side, I
6375 have made the code more robust (and fixed the comments that describe what
6378 33. Some experimental protocols are using DNS PTR records for new purposes. The
6379 keys for these records are domain names, not reversed IP addresses. The
6380 dnsdb PTR lookup now tests whether its key is an IP address. If not, it
6381 leaves it alone. Component reversal etc. now happens only for IP addresses.
6384 34. The host_aton() function is supposed to be passed a string that is known
6385 to be a valid IP address. However, in the case of IPv6 addresses, it was
6386 not checking this. This is a hostage to fortune. Exim now panics and dies
6387 if the condition is not met. A case was found where this could be provoked
6388 from a dnsdb PTR lookup with an IPv6 address that had more than 8
6389 components; fortuitously, this particular loophole had already been fixed
6390 by change 4.50/55 or 4.44/33 above.
6392 If there are any other similar loopholes, the new check in host_aton()
6393 itself should stop them being exploited. The report I received stated that
6394 data on the command line could provoke the exploit when Exim was running as
6395 exim, but did not say which command line option was involved. All I could
6396 find was the use of -be with a bad dnsdb PTR lookup, and in that case it is
6397 running as the user.
6400 35. There was a buffer overflow vulnerability in the SPA authentication code
6401 (which came originally from the Samba project). I have added a test to the
6402 spa_base64_to_bits() function which I hope fixes it.
6405 36. The daemon start-up calls getloadavg() while still root for those OS that
6406 need the first call to be done as root, but it missed one case: when
6407 deliver_queue_load_max is set with deliver_drop_privilege. This is
6408 necessary for the benefit of the queue runner, because there is no re-exec
6409 when deliver_drop_privilege is set.
6411 37. Caching of lookup data for "hosts =" ACL conditions, when a named host list
6412 was in use, was not putting the data itself into the right store pool;
6413 consequently, it could be overwritten for a subsequent message in the same
6414 SMTP connection. (Fix 4.40/11 dealt with the non-cache case, but overlooked
6417 38. Sometimes the final signoff response after QUIT could fail to get
6418 transmitted in the non-TLS case. Testing !tls_active instead of tls_active
6419 < 0 before doing a fflush(). This bug looks as though it goes back to the
6420 introduction of TLS in release 3.20, but "sometimes" must have been rare
6421 because the tests only now provoked it.
6427 1. Fixed a longstanding but relatively impotent bug: a long time ago, before
6428 PIPELINING, the function smtp_write_command() used to return TRUE or FALSE.
6429 Now it returns an integer. A number of calls were still expecting a T/F
6430 return. Fortuitously, in all cases, the tests worked in OK situations,
6431 which is the norm. However, things would have gone wrong on any write
6432 failures on the smtp file descriptor. This function is used when sending
6433 messages over SMTP and also when doing verify callouts.
6435 2. When Exim is called to do synchronous delivery of a locally submitted
6436 message (the -odf or -odi options), it no longer closes stderr before doing
6439 3. Implemented the mua_wrapper option.
6441 4. Implemented mx_fail_domains and srv_fail_domains for the dnslookup router.
6443 5. Implemented the functions header_remove(), header_testname(),
6444 header_add_at_position(), and receive_remove_recipient(), and exported them
6447 6. If an ACL "warn" statement specified the addition of headers, Exim already
6448 inserted X-ACL-Warn: at the start if there was no header name. However, it
6449 was not making this test for the second and subsequent header lines if
6450 there were newlines in the string. This meant that an invalid header could
6451 be inserted if Exim was badly configured.
6453 7. Allow an ACL "warn" statement to add header lines at the start or after all
6454 the Received: headers, as well as at the end.
6456 8. Added the rcpt_4xx retry error code.
6458 9. Added postmaster_mailfrom=xxx to callout verification option.
6460 10. Added mailfrom=xxxx to the callout verification option, for verify=
6463 11. ${substr_1_:xxxx} and ${substr__3:xxxx} are now diagnosed as syntax errors
6464 (they previously behaved as ${substr_1_0:xxxx} and ${substr:_0_3:xxxx}).
6466 12. Inserted some casts to stop certain compilers warning when using pointer
6467 differences as field lengths or precisions in printf-type calls (mostly
6468 affecting debugging statements).
6470 13. Added optional readline() support for -be (dynamically loaded).
6472 14. Obscure bug fix: if a message error (e.g. 4xx to MAIL) happened within the
6473 same clock tick as a message's arrival, so that its received time was the
6474 same as the "first fail" time on the retry record, and that message
6475 remained on the queue past the ultimate address timeout, every queue runner
6476 would try a delivery (because it was past the ultimate address timeout) but
6477 after another failure, the ultimate address timeout, which should have then
6478 bounced the address, did not kick in. This was a "< instead of <=" error;
6479 in most cases the first failure would have been in the next clock tick
6480 after the received time, and all would be well.
6482 15. The special items beginning with @ in domain lists (e.g. @mx_any) were not
6483 being recognized when the domain list was tested by the match_domain
6484 condition in an expansion string.
6486 16. Added the ${str2b64: operator.
6488 17. Exim was always calling setrlimit() to set a large limit for the number of
6489 processes, without checking whether the existing limit was already
6490 adequate. (It did check for the limit on file descriptors.) Furthermore,
6491 errors from getrlimit() and setrlimit() were being ignored. Now they are
6492 logged to the main and panic logs, but Exim does carry on, to try to do its
6493 job under whatever limits there are.
6495 18. Imported PCRE 5.0.
6497 19. Trivial typo in log message " temporarily refused connection" (the leading
6500 20. If the log selector return_path_on_delivery was set and an address was
6501 redirected to /dev/null, the delivery process crashed because it assumed
6502 that a return path would always be set for a "successful" delivery. In this
6503 case, the whole delivery is bypassed as an optimization, and therefore no
6506 21. Internal re-arrangement: the function for sending a challenge and reading
6507 a response while authentication was assuming a zero-terminated challenge
6508 string. It's now changed to take a pointer and a length, to allow for
6509 binary data in such strings.
6511 22. Added the cyrus_sasl authenticator (code supplied by MBM).
6513 23. Exim was not respecting finduser_retries when seeking the login of the
6514 uid under which it was called; it was always trying 10 times. (The default
6515 setting of finduser_retries is zero.) Also, it was sleeping after the final
6516 failure, which is pointless.
6518 24. Implemented tls_on_connect_ports.
6520 25. Implemented acl_smtp_predata.
6522 26. If the domain in control=submission is set empty, Exim assumes that the
6523 authenticated id is a complete email address when it generates From: or
6524 Sender: header lines.
6526 27. Added "#define SOCKLEN_T int" to OS/os.h-SCO and OS/os.h-SCO_SV. Also added
6527 definitions to OS/Makefile-SCO and OS/Makefile-SCO_SV that put basename,
6528 chown and chgrp in /bin and hostname in /usr/bin.
6530 28. Exim was keeping the "process log" file open after each use, just as it
6531 does for the main log. This opens the possibility of it remaining open for
6532 long periods when the USR1 signal hits a daemon. Occasional processlog
6533 errors were reported, that could have been caused by this. Anyway, it seems
6534 much more sensible not to leave this file open at all, so that is what now
6537 29. The long-running daemon process does not normally write to the log once it
6538 has entered its main loop, and it closes the log before doing so. This is
6539 so that log files can straightforwardly be renamed and moved. However,
6540 there are a couple of unusual error situations where the daemon does write
6541 log entries, and I had neglected to close the log afterwards.
6543 30. The text of an SMTP error response that was received during a remote
6544 delivery was being truncated at 512 bytes. This is too short for some of
6545 the long messages that one sometimes sees. I've increased the limit to
6548 31. It is now possible to make retry rules that apply only when a message has a
6549 specific sender, in particular, an empty sender.
6551 32. Added "control = enforce_sync" and "control = no_enforce_sync". This makes
6552 it possible to be selective about when SMTP synchronization is enforced.
6554 33. Added "control = caseful_local_part" and "control = "caselower_local_part".
6556 32. Implemented hosts_connection_nolog.
6558 33. Added an ACL for QUIT.
6560 34. Setting "delay_warning=" to disable warnings was not working; it gave a
6563 35. Added mailbox_size and mailbox_filecount to appendfile.
6565 36. Added control = no_multiline_responses to ACLs.
6567 37. There was a bug in the logic of the code that waits for the clock to tick
6568 in the case where the clock went backwards by a substantial amount such
6569 that the microsecond fraction of "now" was more than the microsecond
6570 fraction of "then" (but the whole seconds number was less).
6572 38. Added support for the libradius Radius client library this is found on
6573 FreeBSD (previously only the radiusclient library was supported).
6579 1. When certain lookups returned multiple values in the form name=value, the
6580 quoting of the values was not always being done properly. Specifically:
6581 (a) If the value started with a double quote, but contained no whitespace,
6583 (b) If the value contained whitespace other than a space character (i.e.
6584 tabs or newlines or carriage returns) it was not quoted.
6585 This fix has been applied to the mysql and pgsql lookups by writing a
6586 separate quoting function and calling it from the lookup code. The fix
6587 should probably also be applied to nisplus, ibase and oracle lookups, but
6588 since I cannot test any of those, I have not disturbed their existing code.
6590 2. A hit in the callout cache for a specific address caused a log line with no
6591 reason for rejecting RCPT. Now it says "Previous (cached) callout
6592 verification failure".
6594 3. There was an off-by-one bug in the queryprogram router. An over-long
6595 return line was truncated at 256 instead of 255 characters, thereby
6596 overflowing its buffer with the terminating zero. As well as fixing this, I
6597 have increased the buffer size to 1024 (and made a note to document this).
6599 4. If an interrupt, such as the USR1 signal that is send by exiwhat, arrives
6600 when Exim is waiting for an SMTP response from a remote server, Exim
6601 restarts its select() call on the socket, thereby resetting its timeout.
6602 This is not a problem when such interrupts are rare. Somebody set up a cron
6603 job to run exiwhat every 2 minutes, which is less than the normal select()
6604 timeout (5 or 10 minutes). This meant that the select() timeout never
6605 kicked in because it was always reset. I have fixed this by comparing the
6606 time when an interrupt arrives with the time at the start of the first call
6607 to select(). If more time than the timeout has elapsed, the interrupt is
6608 treated as a timeout.
6610 5. Some internal re-factoring in preparation for the addition of Sieve
6611 extensions (by MH). In particular, the "personal" test is moved to a
6612 separate function, and given an option for scanning Cc: and Bcc: (which is
6613 not set for Exim filters).
6615 6. When Exim created an email address using the login of the caller as the
6616 local part (e.g. when creating a From: or Sender: header line), it was not
6617 quoting the local part when it contained special characters such as @.
6619 7. Installed new OpenBSD configuration files.
6621 8. Reworded some messages for syntax errors in "and" and "or" conditions to
6622 try to make them clearer.
6624 9. Callout options, other than the timeout value, were being ignored when
6625 verifying sender addresses in header lines. For example, when using
6627 verify = header_sender/callout=no_cache
6629 the cache was (incorrectly) being used.
6631 10. Added a missing instance of ${EXE} to the exim_install script; this affects
6632 only the Cygwin environment.
6634 11. When return_path_on_delivery was set as a log selector, if different remote
6635 addresses in the same message used different return paths and parallel
6636 remote delivery occurred, the wrong values would sometimes be logged.
6637 (Whenever a remote delivery process finished, the return path value from
6638 the most recently started remote delivery process was logged.)
6640 12. RFC 3848 specifies standard names for the "with" phrase in Received: header
6641 lines when AUTH and/or TLS are in use. This is the "received protocol"
6642 field. Exim used to use "asmtp" for authenticated SMTP, without any
6643 indication (in the protocol name) for TLS use. Now it follows the RFC and
6644 uses "esmtpa" if the connection is authenticated, "esmtps" if it is
6645 encrypted, and "esmtpsa" if it is both encrypted and authenticated. These
6646 names appear in log lines as well as in Received: header lines.
6648 13. Installed MH's patches for Sieve to add the "copy" and "vacation"
6649 extensions, and comparison tests, and to fix some bugs.
6651 14. Changes to the "personal" filter test:
6653 (1) The test was buggy in that it was just doing the equivalent of
6654 "contains" tests on header lines. For example, if a user's address was
6655 anne@some.where, the "personal" test would incorrectly be true for
6657 To: susanne@some.where
6659 This test is now done by extracting each address from the header in turn,
6660 and checking the entire address. Other tests that are part of "personal"
6661 are now done using regular expressions (for example, to check local parts
6662 of addresses in From: header lines).
6664 (2) The list of non-personal local parts in From: addresses has been
6665 extended to include "listserv", "majordomo", "*-request", and "owner-*",
6666 taken from the Sieve specification recommendations.
6668 (3) If the message contains any header line starting with "List-" it is
6669 treated as non-personal.
6671 (4) The test for "circular" in the Subject: header line has been removed
6672 because it now seems ill-conceived.
6674 15. Minor typos in src/EDITME comments corrected.
6676 16. Installed latest exipick from John Jetmore.
6678 17. If headers_add on a router specified a text string that was too long for
6679 string_sprintf() - that is, longer than 8192 bytes - Exim panicked. The use
6680 of string_sprintf() is now avoided.
6682 18. $message_body_size was not set (it was always zero) when running the DATA
6683 ACL and the local_scan() function.
6685 19. For the "mail" command in an Exim filter, no default was being set for
6686 the once_repeat time, causing a random time value to be used if "once" was
6687 specified. (If the value happened to be <= 0, no repeat happened.) The
6688 default is now 0s, meaning "never repeat". The "vacation" command was OK
6689 (its default is 7d). It's somewhat surprising nobody ever noticed this bug
6690 (I found it when inspecting the code).
6692 20. There is now an overall timeout for performing a callout verification. It
6693 defaults to 4 times the callout timeout, which applies to individual SMTP
6694 commands during the callout. The overall timeout applies when there is more
6695 than one host that can be tried. The timeout is checked before trying the
6696 next host. This prevents very long delays if there are a large number of
6697 hosts and all are timing out (e.g. when the network connections are timing
6698 out). The value of the overall timeout can be changed by specifying an
6699 additional sub-option for "callout", called "maxwait". For example:
6701 verify = sender/callout=5s,maxwait=20s
6703 21. Add O_APPEND to the open() call for maildirsize files (Exim already seeks
6704 to the end before writing, but this should make it even safer).
6706 22. Exim was forgetting that it had advertised PIPELINING for the second and
6707 subsequent messages on an SMTP connection. It was also not resetting its
6708 memory on STARTTLS and an internal HELO.
6710 23. When Exim logs an SMTP synchronization error within a session, it now
6711 records whether PIPELINING has been advertised or not.
6713 24. Added 3 instances of "(long int)" casts to time_t variables that were being
6714 formatted using %ld, because on OpenBSD (and perhaps others), time_t is int
6715 rather than long int.
6717 25. Installed the latest Cygwin configuration files from the Cygwin maintainer.
6719 26. Added the never_mail option to autoreply.
6725 1. A reorganization of the code in order to implement 4.40/8 caused a daemon
6726 crash if the getsockname() call failed; this can happen if a connection is
6727 closed very soon after it is established. The problem was simply in the
6728 order in which certain operations were done, causing Exim to try to write
6729 to the SMTP stream before it had set up the file descriptor. The bug has
6730 been fixed by making things happen in the correct order.
6736 1. If "drop" was used in a DATA ACL, the SMTP output buffer was not flushed
6737 before the connection was closed, thus losing the rejection response.
6739 2. Commented out the definition of SOCKLEN_T in os.h-SunOS5. It is needed for
6740 some early Solaris releases, but causes trouble in current releases where
6741 socklen_t is defined.
6743 3. When std{in,out,err} are closed, re-open them to /dev/null so that they
6746 4. Minor refactoring of os.c-Linux to avoid compiler warning when IPv6 is not
6749 5. Refactoring in expand.c to improve memory usage. Pre-allocate a block so
6750 that releasing the top of it at the end releases what was used for sub-
6751 expansions (unless the block got too big). However, discard this block if
6752 the first thing is a variable or header, so that we can use its block when
6753 it is dynamic (useful for very large $message_headers, for example).
6755 6. Lookups now cache *every* query, not just the most recent. A new, separate
6756 store pool is used for this. It can be recovered when all lookup caches are
6757 flushed. Lookups now release memory at the end of their result strings.
6758 This has involved some general refactoring of the lookup sources.
6760 7. Some code has been added to the store_xxx() functions to reduce the amount
6761 of flapping under certain conditions.
6763 8. log_incoming_interface used to affect only the <= reception log lines. Now
6764 it causes the local interface and port to be added to several more SMTP log
6765 lines, for example "SMTP connection from", and rejection lines.
6767 9. The Sieve author supplied some patches for the doc/README.SIEVE file.
6769 10. Added a conditional definition of _BSD_SOCKLEN_T to os.h-Darwin.
6771 11. If $host_data was set by virtue of a hosts lookup in an ACL, its value
6772 could be overwritten at the end of the current message (or the start of a
6773 new message if it was set in a HELO ACL). The value is now preserved for
6774 the duration of the SMTP connection.
6776 12. If a transport had a headers_rewrite setting, and a matching header line
6777 contained an unqualified address, that address was qualified, even if it
6778 did not match any rewriting rules. The underlying bug was that the values
6779 of the flags that permit the existence of unqualified sender and recipient
6780 addresses in header lines (set by {sender,recipient}_unqualified_hosts for
6781 non-local messages, and by -bnq for local messages) were not being
6782 preserved with the message after it was received.
6784 13. When Exim was logging an SMTP synchronization error, it could sometimes log
6785 "next input=" as part of the text comprising the host identity instead of
6786 the correct text. The code was using the same buffer for two different
6787 strings. However, depending on which order the printing function evaluated
6788 its arguments, the bug did not always show up. Under Linux, for example, my
6789 test suite worked just fine.
6791 14. Exigrep contained a use of Perl's "our" scoping after change 4.31/70. This
6792 doesn't work with some older versions of Perl. It has been changed to "my",
6793 which in any case is probably the better facility to use.
6795 15. A really picky compiler found some instances of statements for creating
6796 error messages that either had too many or two few arguments for the format
6799 16. The size of the buffer for calls to the DNS resolver has been increased
6800 from 1024 to 2048. A larger buffer is needed when performing PTR lookups
6801 for addresses that have a lot of PTR records. This alleviates a problem; it
6802 does not fully solve it.
6804 17. A dnsdb lookup for PTR records that receives more data than will fit in the
6805 buffer now truncates the list and logs the incident, which is the same
6806 action as happens when Exim is looking up a host name and its aliases.
6807 Previously in this situation something unpredictable would happen;
6808 sometimes it was "internal error: store_reset failed".
6810 18. If a server dropped the connection unexpectedly when an Exim client was
6811 using GnuTLS and trying to read a response, the client delivery process
6812 crashed while trying to generate an error log message.
6814 19. If a "warn" verb in an ACL added multiple headers to a message in a single
6815 string, for example:
6817 warn message = H1: something\nH2: something
6819 the text was added as a single header line from Exim's point of view
6820 though it ended up OK in the delivered message. However, searching for the
6821 second and subsequent header lines using $h_h2: did not work. This has been
6822 fixed. Similarly, if a system filter added multiple headers in this way,
6823 the routers could not see them.
6825 20. Expanded the error message when iplsearch is called with an invalid key to
6826 suggest using net-iplsearch in a host list.
6828 21. When running tests using -bh, any delays imposed by "delay" modifiers in
6829 ACLs are no longer actually imposed (and a message to that effect is
6832 22. If a "gecos" field in a passwd entry contained escaped characters, in
6833 particular, if it contained a \" sequence, Exim got it wrong when building
6834 a From: or a Sender: header from that name. A second bug also caused
6835 incorrect handling when an unquoted " was present following a character
6836 that needed quoting.
6838 23. "{crypt}" as a password encryption mechanism for a "crypteq" expansion item
6839 was not being matched caselessly.
6841 24. Arranged for all hyphens in the exim.8 source to be escaped with
6844 25. Change 16 of 4.32, which reversed 71 or 4.31 didn't quite do the job
6845 properly. Recipient callout cache records were still being keyed to include
6846 the sender, even when use_sender was set false. This led to far more
6847 callouts that were necessary. The sender is no longer included in the key
6848 when use_sender is false.
6850 26. Added "control = submission" modifier to ACLs.
6852 27. Added the ${base62d: operator to decode base 62 numbers.
6854 28. dnsdb lookups can now access SRV records.
6856 29. CONFIGURE_OWNER can be set at build time to define an alternative owner for
6857 the configuration file.
6859 30. The debug message "delivering xxxxxx-xxxxxx-xx" is now output in verbose
6860 (-v) mode. This makes the output for a verbose queue run more intelligible.
6862 31. Added a use_postmaster feature to recipient callouts.
6864 32. Added the $body_zerocount variable, containing the number of binary zero
6865 bytes in the message body.
6867 33. The time of last modification of the "new" subdirectory is now used as the
6868 "mailbox time last read" when there is a quota error for a maildir
6871 34. Added string comparison operators lt, lti, le, lei, gt, gti, ge, gei.
6873 35. Added +ignore_unknown as a special item in host lists.
6875 36. Code for decoding IPv6 addresses in host lists is now included, even if
6876 IPv6 support is not being compiled. This fixes a bug in which an IPv6
6877 address was recognized as an IP address, but was then not correctly decoded
6878 into binary, causing unexpected and incorrect effects when compared with
6885 1. Very minor rewording of debugging text in manualroute to say "list of
6886 hosts" instead of "hostlist".
6888 2. If verify=header_syntax was set, and a header line with an unqualified
6889 address (no domain) and a large number of spaces between the end of the
6890 name and the colon was received, the reception process suffered a buffer
6891 overflow, and (when I tested it) crashed. This was caused by some obsolete
6892 code that should have been removed. The fix is to remove it!
6894 3. When running in the test harness, delay a bit after writing a bounce
6895 message to get a bit more predictability in the log output.
6897 4. Added a call to search_tidyup() just before forking a reception process. In
6898 theory, someone could use a lookup in the expansion of smtp_accept_max_
6899 per_host which, without the tidyup, could leave open a database connection.
6901 5. Added the variables $recipient_data and $sender_data which get set from a
6902 lookup success in an ACL "recipients" or "senders" condition, or a router
6903 "senders" option, similar to $domain_data and $local_part_data.
6905 6. Moved the writing of debug_print from before to after the "senders" test
6908 7. Change 4.31/66 (moving the time when the Received: is generated) caused
6909 problems for message scanning, either using a data ACL, or using
6910 local_scan() because the Received: header was not generated till after they
6911 were called (in order to set the time as the time of reception completion).
6912 I have revised the way this works. The header is now generated after the
6913 body is received, but before the ACL or local_scan() are called. After they
6914 are run, the timestamp in the header is updated.
6920 1. Change 4.24/6 introduced a bug because the SIGALRM handler was disabled
6921 before starting a queue runner without re-exec. This happened only when
6922 deliver_drop_privilege was set or when the Exim user was set to root. The
6923 effect of the bug was that timeouts during subsequent deliveries caused
6924 crashes instead of being properly handled. The handler is now left at its
6925 default (and expected) setting.
6927 2. The other case in which a daemon avoids a re-exec is to deliver an incoming
6928 message, again when deliver_drop_privilege is set or Exim is run as root.
6929 The bug described in (1) was not present in this case, but the tidying up
6930 of the other signals was missing. I have made the two cases consistent.
6932 3. The ignore_target_hosts setting on a manualroute router was being ignored
6933 for hosts that were looked up using the /MX notation.
6935 4. Added /ignore=<ip list> feature to @mx_any, @mx_primary, and @mx_secondary
6938 5. Change 4.31/55 was buggy, and broke when there was a rewriting rule that
6939 operated on the sender address. After changing the $sender_address to <>
6940 for the sender address verify, Exim was re-instated it as the original
6941 (before rewriting) address, but remembering that it had rewritten it, so it
6942 wasn't rewriting it again. This bug also had the effect of breaking the
6943 sender address verification caching when the sender address was rewritten.
6945 6. The ignore_target_hosts option was being ignored by the ipliteral router.
6946 This has been changed so that if the ip literal address matches
6947 ignore_target_hosts, the router declines.
6949 7. Added expansion conditions match_domain, match_address, and match_local_
6950 part (NOT match_host).
6952 8. The placeholder for the Received: header didn't have a length field set.
6954 9. Added code to Exim itself and to exim_lock to test for a specific race
6955 condition that could lead to file corruption when using MBX delivery. The
6956 issue is with the lockfile that is created in /tmp. If this file is removed
6957 after a process has opened it but before that process has acquired a lock,
6958 there is the potential for a second process to recreate the file and also
6959 acquire a lock. This could lead to two Exim processes writing to the file
6960 at the same time. The added code performs the same test as UW imapd; it
6961 checks after acquiring the lock that its file descriptor still refers to
6962 the same named file.
6964 10. The buffer for building added header lines was of fixed size, 8192 bytes.
6965 It is now parameterized by HEADER_ADD_BUFFER_SIZE and this can be adjusted
6968 11. Added the smtp_active_hostname option. If used, this will typically be made
6969 to depend on the incoming interface address. Because $interface_address is
6970 not set up until the daemon has forked a reception process, error responses
6971 that can happen earlier (such as "too many connections") no longer contain
6974 12. If an expansion in a condition on a "warn" statement fails because a lookup
6975 defers, the "warn" statement is abandoned, and the next ACL statement is
6976 processed. Previously this caused the whole ACL to be aborted.
6978 13. Added the iplsearch lookup type.
6980 14. Added ident_timeout as a log selector.
6982 15. Added tls_certificate_verified as a log selector.
6984 16. Added a global option tls_require_ciphers (compare the smtp transport
6985 option of the same name). This controls incoming TLS connections.
6987 17. I finally figured out how to make tls_require_ciphers do a similar thing
6988 in GNUtls to what it does in OpenSSL, that is, set up an appropriate list
6989 before starting the TLS session.
6991 18. Tabs are now shown as \t in -bP output.
6993 19. If the log selector return_path_on_delivery was set, Exim crashed when
6994 bouncing a message because it had too many Received: header lines.
6996 20. If two routers both had headers_remove settings, and the first one included
6997 a superfluous trailing colon, the final name in the first list and the
6998 first name in the second list were incorrectly joined into one item (with a
6999 colon in the middle).
7005 1. Added -C and -D options to the exinext utility, mainly to make it easier
7006 to include in the automated testing, but these could be helpful when
7007 multiple configurations are in use.
7009 2. The exinext utility was not formatting the output nicely when there was
7010 an alternate port involved in the retry record key, nor when there was a
7011 message id as well (for retries that were specific to a specific message
7012 and a specific host). It was also confused by IPv6 addresses, because of
7013 the additional colons they contain. I have fixed the IPv4 problem, and
7014 patched it up to do a reasonable job for IPv6.
7016 3. When there is an error after a MAIL, RCPT, or DATA SMTP command during
7017 delivery, the log line now contains "pipelined" if PIPELINING was used.
7019 4. An SMTP transport process used to panic and die if the bind() call to set
7020 an explicit outgoing interface failed. This has been changed; it is now
7021 treated in the same way as a connect() failure.
7023 5. A reference to $sender_host_name in the part of a conditional expansion
7024 that was being skipped was still causing a DNS lookup. This no longer
7027 6. The def: expansion condition was not recognizing references to header lines
7028 that used bh_ and bheader_.
7030 7. Added the _cache feature to named lists.
7032 8. The code for checking quota_filecount in the appendfile transport was
7033 allowing one more file than it should have been.
7035 9. For compatibility with Sendmail, the command line option
7043 and sets the incoming protocol and host name (for trusted callers). The
7044 host name and its colon can be omitted when only the protocol is to be set.
7045 Note the Exim already has two private options, -pd and -ps, that refer to
7046 embedded Perl. It is therefore impossible to set a protocol value of "d" or
7047 "s", but I don't think that's a major issue.
7049 10. A number of refactoring changes to the code, none of which should affect
7052 (a) The number of logging options was getting close to filling up the
7053 32-bit word that was used as a bit map. I have split them into two classes:
7054 those that are passed in the argument to log_write(), and those that are
7055 only ever tested independently outside of that function. These are now in
7056 separate 32-bit words, so there is plenty of room for expansion again.
7057 There is no change in the user interface or the logging behaviour.
7059 (b) When building, for example, log lines, the code previously used a
7060 macro that called string_cat() twice, in order to add two strings. This is
7061 not really sufficiently general. Furthermore, there was one instance where
7062 it was actually wrong because one of the argument was used twice, and in
7063 one call a function was used. (As it happened, calling the function twice
7064 did not affect the overall behaviour.) The macro has been replaced by a
7065 function that can join an arbitrary number of extra strings onto a growing
7068 (c) The code for expansion conditions now uses a table and a binary chop
7069 instead of a serial search (which was left over from when there were very
7070 few conditions). Also, it now recognizes conditions like "pam" even when
7071 the relevant support is not compiled in: a suitably worded error message is
7072 given if an attempt is made to use such a condition.
7074 11. Added ${time_interval:xxxxx}.
7076 12. A bug was causing one of the ddress fields not to be passed back correctly
7077 from remote delivery subprocesses. The field in question was not being
7078 subsequently used, so this caused to problems in practice.
7080 13. Added new log selectors queue_time and deliver_time.
7082 14. Might have fixed a bug in maildirsizefile handling that threw up
7083 "unexpected character" debug warnings, and recalculated the data
7084 unnecessarily. In any case, I expanded the warning message to give more
7087 15. Added the message "Restricted characters in address" to the statements in
7088 the default ACL that block characters like @ and % in local parts.
7090 16. Change 71 for release 4.31 proved to be much less benign that I imagined.
7091 Three changes have been made:
7093 (a) There was a serious bug; a negative response to MAIL caused the whole
7094 recipient domain to be cached as invalid, thereby blocking all messages
7095 to all local parts at the same domain, from all senders. This bug has
7096 been fixed. The domain is no longer cached after a negative response to
7097 MAIL if the sender used is not empty.
7099 (b) The default behaviour of using MAIL FROM:<> for recipient callouts has
7102 (c) A new callout option, "use_sender" has been added for people who want
7103 the modified behaviour.
7109 1. Removed "EXTRALIBS=-lwrap" from OS/Makefile-Unixware7 on the advice of
7112 2. Removed "LIBS = -lresolv" from OS/Makefile-Darwin as it is not needed, and
7113 indeed breaks things for older releases.
7115 3. Added additional logging to the case where there is a problem reading data
7116 from a filter that is running in a subprocess using a pipe, in order to
7117 try to track down a specific problem.
7119 4. Testing facility fudge: when running in the test harness and attempting
7120 to connect to 10.x.x.x (expecting a connection timeout) I'm now sometimes
7121 getting "No route to host". Convert this to a timeout.
7123 5. Define ICONV_ARG2_TYPE as "char **" for Unixware7 to avoid compiler
7126 6. Some OS don't have socklen_t but use size_t instead. This affects the
7127 fifth argument of getsockopt() amongst other things. This is now
7128 configurable by a macro called SOCKLEN_T which defaults to socklen_t, but
7129 can be set for individual OS. I have set it for SunOS5, OSF1, and
7130 Unixware7. Current versions of SunOS5 (aka Solaris) do have socklen_t, but
7131 some earlier ones do not.
7133 7. Change 4.30/15 was not doing the test caselessly.
7135 8. The standard form for an IPv6 address literal was being rejected by address
7136 parsing in, for example, MAIL and RCPT commands. An example of this kind of
7137 address is [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
7138 this, as well as the form without the "IPv6" on the front (but only when
7139 address literals are enabled, of course).
7141 9. Added some casts to avoid compiler warnings in OS/os.c-Linux.
7143 10. Exim crashed if a message with an empty sender address specified by -f
7144 encountered a router with an errors_to setting. This could be provoked only
7145 by a command such as
7149 where an empty string was supplied; "<>" did not hit this bug.
7151 11. Installed PCRE release 4.5.
7153 12. If EHLO/HELO was rejected by an ACL, the value of $sender_helo_name
7154 remained set. It is now erased.
7156 13. exiqgrep wasn't working on MacOS X because it didn't correctly compute
7157 times from message ids (which are base 36 rather than the normal 62).
7159 14. "Expected" SMTP protocol errors that can arise when PIPELINING is in use
7160 were being counted as actual protocol errors, and logged if the log
7161 selector +smtp_protocol_error was set. One cannot be perfect in this test,
7162 but now, if PIPELINING has been advertised, RCPT following a rejected MAIL,
7163 and DATA following a set of rejected RCPTs do not count as protocol errors.
7164 In other words, Exim assumes they were pipelined, though this may not
7165 actually be the case. Of course, in all cases the client gets an
7166 appropriate error code.
7168 15. If a lookup fails in an ACL condition, a message about the failure may
7169 be available; it is used if testing the ACL cannot continue, because most
7170 such messages specify what the cause of the deferral is. However, some
7171 messages (e.g. "MYSQL: no data found") do not cause a defer. There was bug
7172 that caused an old message to be retained and used if a later statement
7173 caused a defer, replacing the real cause of the deferral.
7175 16. If an IP address had so many PTR records that the DNS lookup buffer
7176 was not large enough to hold them, Exim could crash while trying to process
7177 the truncated data. It now detects and logs this case.
7179 17. Further to 4.21/58, another change has been made: if (and only if) the
7180 first line of a message (the first header line) ends with CRLF, a bare LF
7181 in a subsequent header line has a space inserted after it, so as not to
7182 terminate the header.
7184 18. Refactoring: tidied an ugly bit of code in appendfile that copied data
7185 unnecessarily, used atoi() instead of strtol(), and didn't check the
7186 termination when getting file sizes from file names by regex.
7188 19. Completely re-implemented the support for maildirsize files, in the light
7189 of a number of problems with the previous contributed implementation
7190 (4.30/29). In particular:
7192 . If the quota is zero, the maildirsize file is maintained, but no quota is
7195 . If the maildir directory does not exist, it is created before any attempt
7196 to write a maildirsize file.
7198 . The quota value in the file is just a cache; if the quota is changed in
7199 the transport, the new value overrides.
7201 . A regular expression is available for excluding directories from the
7204 20. The autoreply transport checks the characters in options that define the
7205 message's headers; it allows continued headers, but it was checking with
7206 isspace() after an embedded newline instead of explicitly looking for a
7209 21. If all the "regular" hosts to which an address was routed had passed their
7210 expiry times, and had not reached their retry times, the address was
7211 bounced, even if fallback hosts were defined. Now Exim should go on to try
7214 22. Increased buffer sizes in the callout code from 1024 to 4096 to match the
7215 equivalent code in the SMTP transport. Some hosts send humungous responses
7216 to HELO/EHLO, more than 1024 it seems.
7218 23. Refactoring: code in filter.c used (void *) for "any old type" but this
7219 gives compiler warnings in some environments. I've now done it "properly",
7222 24. The replacement for inet_ntoa() that is used with gcc on IRIX systems
7223 (because of problems with the built-in one) was declared to return uschar *
7224 instead of char *, causing compiler failure.
7226 25. Fixed a file descriptor leak when processing alias/forward files.
7228 26. Fixed a minor format string issue in dbfn.c.
7230 27. Typo in exim.c: ("dmbnz" for "dbmnz").
7232 28. If a filter file refered to $h_xxx or $message_headers, and the headers
7233 contained RFC 2047 "words", Exim's memory could, under certain conditions,
7236 29. When a sender address is verified, it is cached, to save repeating the test
7237 when there is more than one recipient in a message. However, when the
7238 verification involves a callout, it is possible for different callout
7239 options to be set for different recipients. It is too complicated to keep
7240 track of this in the cache, so now Exim always runs a verification when a
7241 callout is required, relying on the callout cache for the optimization.
7242 The overhead is duplication of the address routing, but this should not be
7245 30. Fixed a bug in callout caching. If a RCPT command caused the sender address
7246 to be verified with callout=postmaster, and the main callout worked but the
7247 postmaster check failed, the verification correctly failed. However, if a
7248 subsequent RCPT command asked for sender verification *without* the
7249 postmaster check, incorrect caching caused this verification also to fail,
7252 31. Exim caches DNS lookup failures so as to avoid multiple timeouts; however,
7253 it was not caching the DNS options (qualify_single, search_parents) that
7254 were used when the lookup failed. A subsequent lookup with different
7255 options therefore always gave the same answer, though there were cases
7256 where it should not have. (Example: a "domains = !$mx_any" option on a
7257 dnslookup router: the "domains" option is always processed without any
7258 widening, but the router might have qualify_single set.) Now Exim uses the
7259 cached value only when the same options are set.
7261 32. Added John Jetmore's "exipick" utility to the distribution.
7263 33. GnuTLS: When an attempt to start a TLS session fails for any reason other
7264 than a timeout (e.g. a certificate is required, and is not provided), an
7265 Exim server now closes the connection immediately. Previously it waited for
7266 the client to close - but if the client is SSL, it seems that they each
7267 wait for each other, leading to a delay before one of them times out.
7269 34: GnuTLS: Updated the code to use the new GnuTLS 1.0.0 API. I have not
7270 maintained 0.8.x compatibility because I don't think many are using it, and
7271 it is clearly obsolete.
7273 35. Added TLS support for CRLs: a tls_crl global option and one for the smtp
7276 36. OpenSSL: $tls_certificate_verified was being set to 1 even if the
7277 client certificate was expired. A simple patch fixes this, though I don't
7278 understand the full logic of why the verify callback is called multiple
7281 37. OpenSSL: a patch from Robert Roselius: "Enable client-bug workaround.
7282 Versions of OpenSSL as of 0.9.6d include a 'CBC countermeasure' feature,
7283 which causes problems with some clients (such as the Certicom SSL Plus
7284 library used by Eudora). This option, SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS,
7285 disables the coutermeasure allowing Eudora to connect."
7287 38. Exim was not checking that a write() to a log file succeeded. This could
7288 lead to Bad Things if a log got too big, in particular if it hit a file
7289 size limit. Exim now panics and dies if it cannot write to a log file, just
7290 as it does if it cannot open a log file.
7292 39. Modified OS/Makefile-Linux so that it now contains
7294 CFLAGS=-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE
7296 The two -D definitions ensure that Exim is compiled with large file
7297 support, which makes it possible to handle log files that are bigger than
7300 40. Fixed a subtle caching bug: if (in an ACL or a set of routers, for
7301 instance) a domain was checked against a named list that involved a lookup,
7302 causing $domain_data to be set, then another domain was checked against the
7303 same list, then the first domain was re-checked, the value of $domain_data
7304 after the final check could be wrong. In particular, if the second check
7305 failed, it could be set empty. This bug probably also applied to
7308 41. The strip_trailing_dot option was not being applied to the address given
7309 with the -f command-line option.
7311 42. The code for reading a message's header from the spool was incrementing
7312 $received_count, but never initializing it. This meant that the value was
7313 incorrect (doubled) while delivering a message in the same process in which
7314 it was received. In the most common configuration of Exim, this never
7315 happens - a fresh exec is done - but it can happen when
7316 deliver_drop_privilege is set.
7318 43. When Exim logs an SMTP synchronization error - client data sent too soon -
7319 it now includes up to 150 characters of the unexpected data in the log
7322 44. The exim_dbmbuild utility uses fixed size buffers for reading input lines
7323 and building data strings. The size of both of these buffers was 10 000
7324 bytes - far larger than anybody would *ever* want, thought I. Needless to
7325 say, somebody hit the limit. I have increased the maximum line length to
7326 20 000 and the maximum data length of concatenated lines to 100 000. I have
7327 also fixed two bugs, because there was no checking on these buffers. Tsk,
7328 tsk. Now exim_dbmbuild gives a message and exits with an error code if a
7329 buffer is too small.
7331 45. The exim_dbmbuild utility did not support quoted keys, as Exim does in
7332 lsearch lookups. Now it does.
7334 46. When parsing a route_list item in a manualroute router, a fixed-length
7335 buffer was used for the list of hosts. I made this 1024 bytes long,
7336 thinking that nobody would ever have a list of hosts that long. Wrong.
7337 Somebody had a whole pile of complicated expansion conditions, and the
7338 string was silently truncated, leading to an expansion error. It turns out
7339 that it is easier to change to an unlimited length (owing to other changes
7340 that have happened since this code was originally written) than to build
7341 structure for giving a limitation error. The length of the item that
7342 expands into the list of hosts is now unlimited.
7344 47. The lsearch lookup could not handle data where the length of text line was
7345 more than 4095 characters. Such lines were truncated, leading to shortened
7346 data being returned. It should now handle lines of any length.
7348 48. Minor wording revision: "cannot test xxx in yyy ACL" becomes "cannot test
7349 xxx condition in yyy ACL" (e.g. "cannot test domains condition in DATA
7352 49. Cosmetic tidy to scripts like exicyclog that are generated by globally
7353 replacing strings such as BIN_DIRECTORY in a source file: the replacement
7354 no longer happens in comment lines. A list of replacements is now placed
7355 at the head of all of the source files, except those whose only change is
7356 to replace PERL_COMMAND in the very first #! line.
7358 50. Replaced the slow insertion sort in queue.c, for sorting the list of
7359 messages on the queue, with a bottom-up merge sort, using code contributed
7360 by Michael Haardt. This should make operations like -bp somewhat faster on
7361 large queues. It won't affect queue runners, except when queue_run_in_order
7364 51. Installed eximstats 1.31 in the distribution.
7366 52. Added support for SRV lookups to the dnslookup router.
7368 53. If an ACL referred to $message_body or $message_body_end, the value was not
7369 reset for any messages that followed in the same SMTP session.
7371 54. The store-handling optimization for building very long strings was not
7372 differentiating between the different store pools. I don't think this
7373 actually made any difference in practice, but I've tidied it.
7375 55. While running the routers to verify a sender address, $sender_address
7376 was still set to the sender address. This is wrong, because when routing to
7377 send a bounce to the sender, it would be empty. Therefore, I have changed
7378 it so that, while verifying a sender address, $sender_address is set to <>.
7379 (There is no change to what happens when verifying a recipient address.)
7381 56. After finding MX (or SRV) records, Exim was doing a DNS lookup for the
7382 target A or AAAA records (if not already returned) without resetting the
7383 qualify_single or search_parents options of the DNS resolver. These are
7384 inappropriate in this case because the targets of MX and SRV records must
7385 be FQDNs. A broken DNS record could cause trouble if it happened to have a
7386 target that, when qualified, matched something in the local domain. These
7387 two options are now turned off when doing these lookups.
7389 57. It seems that at least some releases of Reiserfs (which does not have the
7390 concept of a fixed number of inodes) returns zero and not -1 for the
7391 number of available inodes. This interacted badly with check_spool_inodes,
7392 which assumed that -1 was the "no such thing" setting. What I have done is
7393 to check that the total number of inodes is greater than zero before doing
7394 the test of how many are available.
7396 58. When a "warn" ACL statement has a log_message modifier, the message is
7397 remembered, and not repeated. This is to avoid a lot of repetition when a
7398 message has many recipients that cause the same warning to be written.
7399 However, Exim was preserving the list of already written lines for an
7400 entire SMTP session, which doesn't seem right. The memory is now reset if a
7401 new message is started.
7403 59. The "rewrite" debugging flag was not showing the result of rewriting in the
7404 debugging output unless log_rewrite was also set.
7406 60. Avoid a compiler warning on 64-bit systems in dsearch.c by avoiding the use
7407 of (int)(handle) when we know that handle contains (void *)(-1).
7409 61. The Exim daemon panic-logs an error return when it closes the incoming
7410 connection. However "connection reset by peer" seems to be common, and
7411 isn't really an error worthy of noting specially, so that particular error
7414 62. When Exim is trying to find all the local interfaces, it used to panic and
7415 die if the ioctl to get the interface flags failed. However, it seems that
7416 on at least one OS (Solaris 9) it is possible to have an interface that is
7417 included in the list of interfaces, but for which you get a failure error
7418 for this call. This happens when the interface is not "plumbed" into a
7419 protocol (i.e. neither IPv4 nor IPv6). I've changed the code so that a
7420 failure of the "get flags" call assumes that the interface is down.
7422 63. Added a ${eval10: operator, which assumes all numbers are decimal. This
7423 makes life easier for people who are doing arithmetic on fields extracted
7424 from dates, where you often get leading zeros that should not be
7425 interpreted as octal.
7427 64. Added qualify_domain to the redirect router, to override the global
7430 65. If a pathologically long header line contained very many addresses (the
7431 report of this problem mentioned 10 000) and each of them was rewritten,
7432 Exim could use up a very large amount of memory. (It kept on making new
7433 copies of the header line as it rewrote, and never released the old ones.)
7434 At the expense of a bit more processing, the header rewriting function has
7435 been changed so that it no longer eats memory in this way.
7437 66. The generation of the Received: header has been moved from the time that a
7438 message starts to be received, to the time that it finishes. The timestamp
7439 in the Received: header should now be very close to that of the <= log
7440 line. There are two side-effects of this change:
7442 (a) If a message is rejected by a DATA or non-SMTP ACL or local_scan(), the
7443 logged header lines no longer include the local Received: line, because
7444 it has not yet been created. The same applies to a copy of the message
7445 that is returned to a non-SMTP sender when a message is rejected.
7447 (b) When a filter file is tested using -bf, no additional Received: header
7448 is added to the test message. After some thought, I decided that this
7451 This change does not affect the value of $received_for. It is still set
7452 after address rewriting, but before local_scan() is called.
7454 67. Installed the latest Cygwin-specific files from the Cygwin maintainer.
7456 68. GnuTLS: If an empty file is specified for tls_verify_certificates, GnuTLS
7457 gave an unhelpful panic error message, and a defer error. I have managed to
7458 change this behaviour so that it now rejects any supplied certificate,
7459 which seems right, as the list of acceptable certificates is empty.
7461 69. OpenSSL: If an empty file is specified for tls_verify_certificates, OpenSSL
7462 gave an unhelpful defer error. I have not managed to make this reject any
7463 supplied certificates, but the error message it gives is "no certificate
7464 supplied", which is not helpful.
7466 70. exigrep's output now also includes lines that are not associated with any
7467 message, but which match the given pattern. Implemented by a patch from
7468 Martin Sluka, which also tidied up the Perl a bit.
7470 71. Recipient callout verification, like sender verification, was using <> in
7471 the MAIL FROM command. This isn't really the right thing, since the actual
7472 sender may affect whether the remote host accepts the recipient or not. I
7473 have changed it to use the actual sender in the callout; this means that
7474 the cache record is now keyed on a recipient/sender pair, not just the
7475 recipient address. There doesn't seem to be a real danger of callout loops,
7476 since a callout by the remote host to check the sender would use <>.
7477 [SEE ABOVE: changed after hitting problems.]
7479 72. Exim treats illegal SMTP error codes that do not begin with 4 or 5 as
7480 temporary errors. However, in the case of such a code being given after
7481 the end of a data transmission (i.e. after ".") Exim was failing to write
7482 a retry record for the message. (Yes, there was some broken host that was
7483 actually sending 8xx at this point.)
7485 73. An unknown lookup type in a host list could cause Exim to panic-die when
7486 the list was checked. (An example that provoked this was putting <; in the
7487 middle of a list instead of at the start.) If this happened during a DATA
7488 ACL check, a -D file could be left lying around. This kind of configuration
7489 error no longer causes Exim to die; instead it causes a defer error. The
7490 incident is still logged to the main and panic logs.
7492 74. Buglet left over from Exim 3 conversion. The message "too many messages
7493 in one connection" was written to the rejectlog but not the mainlog, except
7494 when address rewriting (yes!) was being logged.
7496 75. Added write_rejectlog option.
7498 76. When a system filter was run not as root (that is, when system_filter_user
7499 was set), the values of the $n variables were not being returned to the
7500 main process; thus, they were not subsequently available in the $sn
7503 77. Added +return_path_on_delivery log selector.
7505 78. A connection timeout was being treated differently from recipients deferred
7506 when testing hosts_max_try with a message that was older than the host's
7507 retry timeout. (The host should not be counted, thus allowing all hosts to
7508 be tried at least once before bouncing.) This may have been the cause of an
7509 occasionally reported bug whereby a message would remain on the queue
7510 longer than the retry timeout, but would be bounced if a delivery was
7511 forced. I say "may" because I never totally pinned down the problem;
7512 setting up timeout/retry tests is difficult. See also the next item.
7514 79. The ultimate address timeout was not being applied to errors that involved
7515 a combination of host plus message (for example, a timeout on a MAIL
7516 command). When an address resolved to a number of possible hosts, and they
7517 were not all tried for each delivery (e.g. because of hosts_max_try), a
7518 message could remain on the queue longer than the retry timeout.
7520 80. Sieve bug: "stop" inside "elsif" was broken. Applied a patch from Michael
7523 81. Fixed an obscure SMTP outgoing bug which required at least the following
7524 conditions: (a) there was another message waiting for the same server;
7525 (b) the server returned 5xx to all RCPT commands in the first message so
7526 that the message was not completed; (c) the server dropped the connection
7527 or gave a negative response to the RSET that Exim sends to abort the
7528 transaction. The observed case was a dropped connection after DATA that had
7529 been sent in pipelining mode. That is, the server had advertised PIPELINING
7530 but was not implementing it correctly. The effect of the bug was incorrect
7531 behaviour, such as trying another host, and this could lead to a crash.
7537 1. The 3rd arguments to getsockname(), getpeername(), and accept() in exim.c
7538 and daemon.c were passed as pointers to ints; they should have been
7539 pointers to socklen_t variables (which are typically unsigned ints).
7541 2. Some signed/unsigned type warnings in the os.c file for Linux have been
7544 3. Fixed a really odd bug that affected only the testing scheme; patching a
7545 certain fixed string in the binary changed the value of another string that
7546 happened to be identical to the end of the original first string.
7548 4. When gethostbyname() (or equivalent) is passed an IP address as a "host
7549 name", it returns that address as the IP address. On some operating
7550 systems (e.g. Solaris), it also passes back the IP address string as the
7551 "host name". However, on others (e.g. Linux), it passes back an empty
7552 string. Exim wasn't checking for this, and was changing the host name to an
7553 empty string, assuming it had been canonicalized.
7555 5. Although rare, it is permitted to have more than one PTR record for a given
7556 IP address. I thought that gethostbyaddr() or getipnodebyaddr() always gave
7557 all the names associated with an address, because they do in Solaris.
7558 However, it seems that they do not in Linux for data that comes from the
7559 DNS. If an address in /etc/hosts has multiple names, they _are_ all given.
7560 I found this out when I moved to a new Linux workstation and tried to run
7561 the Exim test suite.
7563 To get round this problem I have changed the code so that it now does its
7564 own call to the DNS to look up PTR records when searching for a host name.
7565 If nothing can be found in the DNS, it tries gethostbyaddr(), so that
7566 addresses that are only in /etc/hosts are still found.
7568 This behaviour is, however, controlled by an option called host_lookup_
7569 order, which defaults to "bydns:byaddr". If people want to use the other
7570 order, or indeed, just use one or the other means of lookup, they can
7571 specify it in this variable.
7573 6. If a PTR record yields an empty name, Exim treats it as non-existent. In
7574 some operating systems, this comes back from gethostbyaddr() as an empty
7575 string, and this is what Exim used to test for. However, it seems that in
7576 other systems, "." is yielded. Exim now tests for this case too.
7578 7. The values of check_spool_space and check_log_space are now held internally
7579 as a number of kilobytes instead of an absolute number of bytes. If a
7580 numbers is specified without 'K' or 'M', it is rounded up to the nearest
7581 kilobyte. This means that much larger values can be stored.
7583 8. Exim monitor: an attempt to get the action menu when not actually pointing
7584 at a message produces an empty menu entitled "No message selected". This
7585 works on Solaris (OpenWindows). However, XFree86 does not like a menu with
7586 no entries in it ("Shell widget menu has zero width and/or height"). So I
7587 have added a single, blank menu entry in this case.
7589 9. Added ${quote_local_part.
7591 10. MIME decoding is now applied to the contents of Subject: header lines when
7594 11. Now that a reference to $sender_host_address automatically causes a reverse
7595 lookup to occur if necessary (4.13/18), there is no need to arrange for a
7596 host lookup before query-style lookups in lists that might use this
7597 variable. This has therefore been abolished, and the "net-" prefix is no
7598 longer necessary for query-style lookups.
7600 12. The Makefile for SCO_SV contained a setting of LDFLAGS. This appears to
7601 have been a typo for LFLAGS, so it has been changed.
7603 13. The install script calls Exim with "-C /dev/null" in order to find the
7604 version number. If ALT_CONFIG_PREFIX was set, this caused an error message
7605 to be output. However, since Exim outputs its version number before the
7606 error, it didn't break the script. It just looked ugly. I fixed this by
7607 always allowing "-C /dev/null" if the caller is root.
7609 14. Ignore overlarge ACL variable number when reading spool file - insurance
7610 against a later release with more variables having written the file.
7612 15. The standard form for an IPv6 address literal was being rejected by EHLO.
7613 Example: [IPv6:2002:c1ed:8229:10:202:2dff:fe07:a42a]. Exim now accepts
7614 this, as well as the form without the "IPv6" on the front.
7616 16. Added CHOWN_COMMAND=/usr/sbin/chown and LIBS=-lresolv to the
7617 OS/Makefile-Darwin file.
7619 17. Fixed typo in lookups/ldap.c: D_LOOKUP should be D_lookup. This applied
7620 only to LDAP libraries that do not have LDAP_OPT_DEREF.
7622 18. After change 4.21/52, "%ld" was used to format the contents of the $inode
7623 variable. However, some OS use ints for inodes. I've added cast to long int
7624 to get rid of the compiler warning.
7626 19. I had forgotten to lock out "/../" in configuration file names when
7627 ALT_CONFIG_PREFIX was set.
7629 20. Routers used for verification do not need to specify transports. However,
7630 if such a router generated a host list, and callout was configured, Exim
7631 crashed, because it could not find a port number from the (non-existent)
7632 transport. It now assumes port 25 in this circumstance.
7634 21. Added the -t option to exigrep.
7636 22. If LOOKUP_LSEARCH is defined, all three linear search methods (lsearch,
7637 wildlsearch, nwildlsearch) are compiled. LOOKUP_WILDLSEARCH and LOOKUP_
7638 NWILDLSEARCH are now obsolete, but retained for compatibility. If either of
7639 them is set, LOOKUP_LSEARCH is forced.
7641 23. "exim -bV" now outputs a list of lookups that are included in the binary.
7643 24. Added sender and host information to the "rejected by local_scan()" log
7644 line; previously there was no indication of these.
7646 25. Added .include_if_exists.
7648 26. Change 3.952/11 added an explicit directory sync on top of a file sync for
7649 Linux. It turns out that not all file systems support this. Apparently some
7650 versions of NFS do not. (It's rare to put Exim's spool on NFS, but people
7651 do it.) To cope with this, the error EINVAL, which means that sync-ing is
7652 not supported on the file descriptor, is now ignored when Exim is trying to
7653 sync a directory. This applies only to Linux.
7655 27. Added -DBIND_8_COMPAT to the CLFAGS setting for Darwin.
7657 28. In Darwin (MacOS X), the PAM headers are in /usr/include/pam and not in
7658 /usr/include/security. There's now a flag in OS/os.h-Darwin to cope with
7661 29. Added support for maildirsize files from supplied patch (modified a bit).
7663 30. The use of :fail: followed by an empty string could lead Exim to respond to
7664 sender verification failures with (e.g.):
7666 550 Verification failed for <xxx>
7667 550 Sender verify failed
7669 where the first response line was missing the '-' that indicates it is not
7670 the final line of the response.
7672 31. The loop for finding the name of the user that called Exim had a hardwired
7673 limit of 10; it now uses the value of finduser_retries, which is used for
7674 all other user lookups.
7676 32. Added $received_count variable, available in data and not_smtp ACLs, and at
7679 33. Exim was neglecting to zero errno before one call of strtol() when
7680 expanding a string and expecting an integer value. On some systems this
7681 resulted in spurious "integer overflow" errors. Also, it was casting the
7682 result into an int without checking.
7684 34. Testing for a connection timeout using "timeout_connect" in the retry rules
7685 did not work. The code looks as if it has *never* worked, though it appears
7686 to have been documented since at least release 1.62. I have made it work.
7688 35. The "timeout_DNS" error in retry rules, also documented since at least
7689 1.62, also never worked. As it isn't clear exactly what this means, and
7690 clearly it isn't a major issue, I have abolished the feature by treating it
7691 as "timeout", and writing a warning to the main and panic logs.
7693 36. The display of retry rules for -brt wasn't always showing the error code
7696 37. Added new error conditions to retry rules: timeout_A, timeout_MX,
7697 timeout_connect_A, timeout_connect_MX.
7699 38. Rewriting the envelope sender at SMTP time did not allow it to be rewritten
7700 to the empty sender.
7702 39. The daemon was not analysing the content of -oX till after it had closed
7703 stderr and disconnected from the controlling terminal. This meant that any
7704 syntax errors were only noted on the panic log, and the return code from
7705 the command was 0. By re-arranging the code a little, I've made the
7706 decoding happen first, so such errors now appear on stderr, and the return
7707 code is 1. However, the actual setting up of the sockets still happens in
7708 the disconnected process, so errors there are still only recorded on the
7711 40. A daemon listener on a wildcard IPv6 socket that also accepts IPv4
7712 connections (as happens on some IP stacks) was logged at start up time as
7713 just listening for IPv6. It now logs "IPv6 with IPv4". This differentiates
7714 it from "IPv6 and IPv4", which means that two separate sockets are being
7717 41. The debug output for gethostbyname2() or getipnodebyname() failures now
7718 says whether AF_INET or AF_INET6 was passed as an argument.
7720 42. Exiwhat output was messed up when time zones were included in log
7723 43. Exiwhat now gives more information about the daemon's listening ports,
7724 and whether -tls-on-connect was used.
7726 44. The "port" option of the smtp transport is now expanded.
7728 45. A "message" modifier in a "warn" statement in a non-message ACL was being
7729 silently ignored. Now an error message is written to the main and panic
7732 46. There's a new ACL modifier called "logwrite" which writes to a log file
7733 as soon as it is encountered.
7735 47. Added $local_user_uid and $local_user_gid at routing time.
7737 48. Exim crashed when trying to verify a sender address that was being
7740 49. Exim was recognizing only a space character after ".include". It now also
7741 recognizes a tab character.
7743 50. Fixed several bugs in the Perl script that creates the exim.8 man page by
7744 extracting the relevant information from the specification. The man page no
7745 longer contains scrambled data for the -d option, and I've added a section
7746 at the front about calling Exim under different names.
7748 51. Added "extra_headers" argument to the "mail" command in filter files.
7750 52. Redirecting mail to an unqualified address in a Sieve filter caused Exim to
7753 53. Installed eximstats 1.29.
7755 54. Added transport_filter_timeout as a generic transport option.
7757 55. Exim no longer adds an empty Bcc: header to messages that have no To: or
7758 Cc: header lines. This was required by RFC 822, but it not required by RFC
7761 56. Exim used to add From:, Date:, and Message-Id: header lines to any
7762 incoming messages that did not have them. Now it does so only if the
7763 message originates locally, that is, if there is no associated remote host
7764 address. When Resent- header lines are present, this applies to the Resent-
7765 lines rather than the non-Resent- lines.
7767 57. Drop incoming SMTP connection after too many syntax or protocol errors. The
7768 limit is controlled by smtp_max_synprot_errors, defaulting to 3.
7770 58. Messages for configuration errors now include the name of the main
7771 configuration file - useful now that there may be more than one file in a
7772 list (.included file names were always shown).
7774 59. Change 4.21/82 (run initgroups() when starting the daemon) causes problems
7775 for those rare installations that do not start the daemon as root or run it
7776 setuid root. I've cut out the call to initgroups() if the daemon is not
7779 60. The Exim user and group can now be bound into the binary as text strings
7780 that are looked up at the start of Exim's processing.
7782 61. Applied a small patch for the Interbase code, supplied by Ard Biesheuvel.
7784 62. Added $mailstore_basename variable.
7786 63. Installed patch to sieve.c from Michael Haardt.
7788 64. When Exim failed to open the panic log after failing to open the main log,
7789 the original message it was trying to log was written to stderr and debug
7790 output, but if they were not available (the usual case in production), it
7791 was lost. Now it is written to syslog before the two lines that record the
7792 failures to open the logs.
7794 65. Users' Exim filters run in subprocesses under the user's uid. It is
7795 possible for a "deliver" command or an alias in a "personal" command to
7796 provoke an address rewrite. If logging of address rewriting is configured,
7797 this fails because the process is not running as root or exim. There may be
7798 a better way of dealing with this, but for the moment (because 4.30 needs
7799 to be released), I have disabled address rewrite logging when running a
7800 filter in a non-root, non-exim process.
7806 1. The buildconfig auxiliary program wasn't quoting the value set for
7807 HEADERS_CHARSET. This caused a compilation error complaining that 'ISO' was
7808 not defined. This bug was masked in 4.22 by the effect that was fixed in
7811 2. Some messages that were rejected after a message id was allocated were
7812 shown as "incomplete" by exigrep. It no longer does this for messages that
7813 are rejected by local_scan() or the DATA or non-SMTP ACLs.
7815 3. If a Message-ID: header used a domain literal in the ID, and Exim did not
7816 have allow_domain_literals set, the ID did not get logged in the <= line.
7817 Domain literals are now always recognized in Message-ID: header lines.
7819 4. The first argument for a ${extract expansion item is the key name or field
7820 number. Leading and trailing spaces in this item were not being ignored,
7821 causing some misleading effects.
7823 5. When deliver_drop_privilege was set, single queue runner processes started
7824 manually (i.e. by the command "exim -q") or by the daemon (which uses the
7825 same command in the process it spins off) were not dropping privilege.
7827 6. When the daemon running as "exim" started a queue runner, it always
7828 re-executed Exim in the spun-off process. This is a waste of effort when
7829 deliver_drop_privilege is set. The new process now just calls the
7830 queue-runner function directly.
7836 1. Typo in the src/EDITME file: it referred to HEADERS_DECODE_TO instead of
7839 2. Change 4.21/73 introduced a bug. The pid file path set by -oP was being
7840 ignored. Though the use of -oP was forcing the writing of a pid file, it
7841 was always written to the default place.
7843 3. If the message "no IP address found for host xxxx" is generated during
7844 incoming verification, it is now followed by identification of the incoming
7845 connection (so you can more easily find what provoked it).
7847 4. Bug fix for Sieve filters: "stop" inside a block was not working properly.
7849 5. Added some features to "harden" Exim a bit more against certain attacks:
7851 (a) There is now a build-time option called FIXED_NEVER_USERS that can
7852 be put in Local/Makefile. This is like the never_users runtime option,
7853 but it cannot be overridden. The default setting is "root".
7855 (b) If ALT_CONFIG_PREFIX is defined in Local/Makefile, it specifies a
7856 prefix string with which any file named in a -C command line option
7859 (c) If ALT_CONFIG_ROOT_ONLY is defined in Local/Makefile, root privilege
7860 is retained for -C and -D only if the caller of Exim is root. Without
7861 it, the exim user may also use -C and -D and retain privilege.
7863 (d) If DISABLE_D_OPTION is defined in Local/Makefile, the use of the -D
7864 command line option is disabled.
7866 6. Macro names set by the -D option must start with an upper case letter, just
7867 like macro names defined in the configuration file.
7869 7. Added "dereference=" facility to LDAP.
7871 8. Two instances of the typo "uknown" in the source files are fixed.
7873 9. If a PERL_COMMAND setting in Local/Makefile was not at the start of a line,
7874 the Configure-Makefile script screwed up while processing it.
7876 10. Incorporated PCRE 4.4.
7878 11. The SMTP synchronization check was not operating right at the start of an
7879 SMTP session. For example, it could not catch a HELO sent before the client
7880 waited for the greeting. There is now a check for outstanding input at the
7881 point when the greeting is written. Because of the duplex, asynchronous
7882 nature of TCP/IP, it cannot be perfect - the incorrect input may be on its
7883 way, but not yet received, when the check is performed.
7885 12. Added tcp_nodelay to make it possible to turn of the setting of TCP_NODELAY
7886 on TCP/IP sockets, because this apparently causes some broken clients to
7889 13. Installed revised OS/Makefile-CYGWIN and OS/os.c-cygwin (the .h file was
7890 unchanged) from the Cygwin maintainer.
7892 14. The code for -bV that shows what is in the binary showed "mbx" when maildir
7893 was supported instead of testing for mbx. Effectively a typo.
7895 15. The spa authenticator server code was not checking that the input it
7896 received was valid base64.
7898 16. The debug output line for the "set" modifier in ACLs was not showing the
7899 name of the variable that was being set.
7901 17. Code tidy: the variable type "vtype_string" was never used. Removed it.
7903 18. Previously, a reference to $sender_host_name did not cause a DNS reverse
7904 lookup on its own. Something else was needed to trigger the lookup. For
7905 example, a match in host_lookup or the need for a host name in a host list.
7906 Now, if $sender_host_name is referenced and the host name has not yet been
7907 looked up, a lookup is performed. If the lookup fails, the variable remains
7908 empty, and $host_lookup_failed is set to "1".
7910 19. Added "eqi" as a case-independent comparison operator.
7912 20. The saslauthd authentication condition could segfault if neither service
7913 nor realm was specified.
7915 21. If an overflowing value such as "2048M" was set for message_size_limit, the
7916 error message that was logged was misleading, and incoming SMTP
7917 connections were dropped. The message is now more accurate, and temporary
7918 errors are given to SMTP connections.
7920 22. In some error situations (such as 21 above) Exim rejects all SMTP commands
7921 (except RSET) with a 421 error, until QUIT is received. However, it was
7922 failing to send a response to QUIT.
7924 23. The HELO ACL was being run before the code for helo_try_verify_hosts,
7925 which made it impossible to use "verify = helo" in the HELO ACL. The HELO
7926 ACL is now run after the helo_try_verify_hosts code.
7928 24. "{MD5}" and "{SHA1}" are now recognized as equivalent to "{md5"} and
7929 "{sha1}" in the "crypteq" expansion condition (in fact the comparison is
7930 case-independent, so other case variants are also recognized). Apparently
7931 some systems use these upper case variants.
7933 25. If more than two messages were waiting for the same host, and a transport
7934 filter was specified for the transport, Exim sent two messages over the
7935 same TCP/IP connection, and then failed with "socket operation on non-
7936 socket" when it tried to send the third.
7938 26. Added Exim::debug_write and Exim::log_write for embedded Perl use.
7940 27. The extern definition of crypt16() in expand.c was not being excluded when
7941 the OS had its own crypt16() function.
7943 28. Added bounce_return_body as a new option, and bounce_return_size_limit
7944 as a preferred synonym for return_size_limit, both as an option and as an
7947 29. Added LIBS=-liconv to OS/Makefile-OSF1.
7949 30. Changed the default configuration ACL to relax the local part checking rule
7950 for addresses that are not in any local domains. For these addresses,
7951 slashes and pipe symbols are allowed within local parts, but the sequence
7952 /../ is explicitly forbidden.
7954 31. SPA server authentication was not clearing the challenge buffer before
7957 32. log_message in a "warn" ACL statement was writing to the reject log as
7958 well as to the main log, which contradicts the documentation and doesn't
7959 seem right (because no rejection is happening). So I have stopped it.
7961 33. Added Ard Biesheuvel's lookup code for accessing an Interbase database.
7962 However, I am unable to do any testing of this.
7964 34. Fixed an infelicity in the appendfile transport. When checking directories
7965 for a mailbox, to see if any needed to be created, it was accidentally
7966 using path names with one or more superfluous leading slashes; tracing
7967 would show up entries such as stat("///home/ph10", 0xFFBEEA48).
7969 35. If log_message is set on a "discard" verb in a MAIL or RCPT ACL, its
7970 contents are added to the log line that is written for every discarded
7971 recipient. (Previously a log_message setting was ignored.)
7973 36. The ${quote: operator now quotes the string if it is empty.
7975 37. The install script runs exim in order to find its version number. If for
7976 some reason other than non-existence or emptiness, which it checks, it
7977 could not run './exim', it was installing it with an empty version number,
7978 i.e. as "exim-". This error state is now caught, and the installation is
7981 38. An argument was missing from the function that creates an error message
7982 when Exim fails to connect to the socket for saslauthd authentication.
7983 This could cause Exim to crash, or give a corrupted message.
7985 39. Added isip, isip4, and isip6 to ${if conditions.
7987 40. The ACL variables $acl_xx are now saved with the message, and can be
7988 accessed later in routers, transports, and filters.
7990 41. The new lookup type nwildlsearch is like wildlsearch, except that the key
7991 strings in the file are not string-expanded.
7993 42. If a MAIL command specified a SIZE value that was too large to fit into an
7994 int variable, the check against message_size_limit failed. Such values are
7995 now forced to INT_MAX, which is around 2Gb for a 32-bit variable. Maybe one
7996 day this will have to be increased, but I don't think I want to be around
7997 when emails are that large.
8004 1. Removed HAVE_ICONV=yes from OS/Makefile-FreeBSD, since it seems that
8005 iconv() is not standard in FreeBSD.
8007 2. Change 4.21/17 was buggy and could cause stack overwriting on a system with
8008 IPv6 enabled. The observed symptom was a segmentation fault on return from
8009 the function os_common_find_running_interfaces() in src/os.c.
8011 3. In the check_special_case() function in daemon.c I had used "errno" as an
8012 argument name, which causes warnings on some systems. This was basically a
8013 typo, since it was named "eno" in the comments!
8015 4. The code that waits for the clock to tick (at a resolution of some fraction
8016 of a second) so as to ensure message-id uniqueness was always waiting for
8017 at least one whole tick, when it could have waited for less. [This is
8018 almost certainly not relevant at current processor speeds, where it is
8019 unlikely to ever wait at all. But we try to future-proof.]
8021 5. The function that sleeps for a time interval that includes fractions of a
8022 second contained a race. It did not block SIGALRM between setting the
8023 timer, and suspending (a couple of lines later). If the interval was short
8024 and the sigsuspend() was delayed until after it had expired, the suspension
8025 never ended. On busy systems this could lead to processes getting stuck for
8028 6. Some uncommon configurations may cause a lookup to happen in a queue runner
8029 process, before it forks any delivery processes. The open lookup caching
8030 mechanism meant that the open file or database connection was passed into
8031 the delivery process. The problem was that delivery processes always tidy
8032 up cached lookup data. This could cause a problem for the next delivery
8033 process started by the queue runner, because the external queue runner
8034 process does not know about the closure. So the next delivery process
8035 still has data in the lookup cache. In the case of a file lookup, there was
8036 no problem because closing a file descriptor in a subprocess doesn't affect
8037 the parent. However, if the lookup was caching a connection to a database,
8038 the connection was closed, and the second delivery process was likely to
8039 see errors such as "PGSQL: query failed: server closed the connection
8040 unexpectedly". The problem has been fixed by closing all cached lookups
8041 in a queue runner before running a delivery process.
8043 7. Compiler warning on Linux for the second argument of iconv(), which doesn't
8044 seem to have the "const" qualifier which it has on other OS. I've
8047 8. Change 4.21/2 was too strict. It is only if there are two authenticators
8048 *of the same type* (client or server) with the same public name that an
8049 error should be diagnosed.
8051 9. When Exim looked up a host name for an IP address, but failed to find the
8052 original IP address when looking up the host name (a safety check), it
8053 output the message "<ip address> does not match any IP for NULL", which was
8054 confusing, to say the least. The bug was that the host name should have
8055 appeared instead of "NULL".
8057 10. Since release 3.03, if Exim is called by a uid other than root or the Exim
8058 user that is built into the binary, and the -C or -D options is used, root
8059 privilege is dropped before the configuration file is read. In addition,
8060 logging is switched to stderr instead of the normal log files. If the
8061 configuration then re-defines the Exim user, the unprivileged environment
8062 is probably not what is expected, so Exim logs a panic warning message (but
8065 However, if deliver_drop_privilege is set, the unprivileged state may well
8066 be exactly what is intended, so the warning has been cut out in that case,
8067 and Exim is allowed to try to write to its normal log files.
8073 1. smtp_return_error_details was not giving details for temporary sender
8074 or receiver verification errors.
8076 2. Diagnose a configuration error if two authenticators have the same public
8079 3. Exim used not to create the message log file for a message until the first
8080 delivery attempt. This could be confusing when incoming messages were held
8081 for policy or load reasons. The message log file is now created at the time
8082 the message is received, and an initial "Received" line is written to it.
8084 4. The automatically generated man page for command line options had a minor
8085 bug that caused no ill effects; however, a more serious problem was that
8086 the procedure for building the man page automatically didn't always
8087 operate. Consequently, release 4.20 contains an out-of-date version. This
8088 shouldn't happen again.
8090 5. When building Exim with embedded Perl support, the script that builds the
8091 Makefile was calling 'perl' to find its compile-time parameters, ignoring
8092 any setting of PERL_COMMAND in Local/Makefile. This is now fixed.
8094 6. The freeze_tell option was not being used for messages that were frozen on
8095 arrival, either by an ACL or by local_scan().
8097 7. Added the smtp_incomplete_transaction log selector.
8099 8. After STARTTLS, Exim was not forgetting that it had advertised AUTH, so it
8100 was accepting AUTH without a new EHLO.
8102 9. Added tls_remember_esmtp to cope with YAEB. This allows AUTH and other
8103 ESMTP extensions after STARTTLS without a new EHLO, in contravention of the
8106 10. Logging of TCP/IP connections (when configured) now happens in the main
8107 daemon process instead of the child process, so that the TCP/IP connection
8108 count is more accurate (but it can never be perfect).
8110 11. The use of "drop" in a nested ACL was not being handled correctly in the
8111 outer ACL. Now, if condition failure induced by the nested "drop" causes
8112 the outer ACL verb to deny access ("accept" or "discard" after "endpass",
8113 or "require"), the connection is dropped.
8115 12. Similarly, "discard" in a nested ACL wasn't being handled. A nested ACL
8116 that yield "discard" can now be used with an "accept" or a "discard" verb,
8117 but an error is generated for any others (because I can't see a useful way
8118 to define what should happen).
8120 13. When an ACL is read dynamically from a file (or anywhere else), the lines
8121 are now processed in the same way as lines in the Exim configuration file.
8122 In particular, continuation lines are supported.
8124 14. Added the "dnslists = a.b.c!=n.n.n.n" feature.
8126 15. Added -ti meaning -t -i.
8128 16. Check for letters, digits, hyphens, and dots in the names of dnslist
8129 domains, and warn by logging if others are found.
8131 17. At least on BSD, alignment is not guaranteed for the array of ifreq's
8132 returned from GIFCONF when Exim is trying to find the list of interfaces on
8133 a host. The code in os.c has been modified to copy each ifreq to an aligned
8134 structure in all cases.
8136 Also, in some cases, the returned ifreq's were being copied to a 'struct
8137 ifreq' on the stack, which was subsequently passed to host_ntoa(). That
8138 means the last couple of bytes of an IPv6 address could be chopped if the
8139 ifreq contained only a normal sockaddr (14 bytes storage).
8141 18. Named domain lists were not supported in the hosts_treat_as_local option.
8142 An entry such as +xxxx was not recognized, and was treated as a literal
8145 19. Ensure that header lines added by a DATA ACL are included in the reject log
8146 if the ACL subsequently rejects the message.
8148 20. Upgrade the cramtest.pl utility script to use Digest::MD5 instead of just
8149 MD5 (which is deprecated).
8151 21. When testing a filter file using -bf, Exim was writing a message when it
8152 took the sender from a "From " line in the message, but it was not doing so
8153 when it took $return_path from a Return-Path: header line. It now does.
8155 22. If the contents of a "message" modifier for a "warn" ACL verb do not begin
8156 with a valid header line field name (a series of printing characters
8157 terminated by a colon, Exim now inserts X-ACL-Warn: at the beginning.
8159 23. Changed "disc" in the source to "disk" to conform to the documentation and
8160 the book and for uniformity.
8162 24. Ignore Sendmail's -Ooption=value command line item.
8164 25. When execve() failed while trying to run a command in a pipe transport,
8165 Exim was returning EX_UNAVAILABLE (69) from the subprocess. However, this
8166 could be confused with a return value of 69 from the command itself. This
8167 has been changed to 127, the value the shell returns if it is asked to run
8168 a non-existent command. The wording for the related log line suggests a
8169 non-existent command as the problem.
8171 26. If received_header_text expands to an empty string, do not add a Received:
8172 header line to the message. (Well, it adds a token one on the spool, but
8173 marks it "old" so that it doesn't get used or transmitted.)
8175 27. Installed eximstats 1.28 (addition of -nt option).
8177 28. There was no check for failure on the call to getsockname() in the daemon
8178 code. This can fail if there is a shortage of resources on the system, with
8179 ENOMEM, for example. A temporary error is now given on failure.
8181 29. Contrary to the C standard, it seems that in some environments, the
8182 equivalent of setlocale(LC_ALL, "C") is not obeyed at the start of a C
8183 program. Exim now does this explicitly; it affects the formatting of
8184 timestamps using strftime().
8186 30. If exiqsumm was given junk data, it threw up some uninitialized variable
8187 complaints. I've now initialized all the variables, to avoid this.
8189 32. Header lines added by a system filter were not being "seen" during
8190 transport-time rewrites.
8192 33. The info_callback() function passed to OpenSSL is set up with type void
8193 (*)(SSL *, int, int), as described somewhere. However, when calling the
8194 function (actually a macro) that sets it up, the type void(*)() is
8195 expected. I've put in a cast to prevent warnings from picky compilers.
8197 34. If a DNS black list lookup found a CNAME record, but there were no A
8198 records associated with the domain it pointed at, Exim crashed.
8200 35. If a DNS black list lookup returned more than one A record, Exim ignored
8201 all but the first. It now scans all returned addresses if a particular IP
8202 value is being sought. In this situation, the contents of the
8203 $dnslist_value variable are a list of all the addresses, separated by a
8206 36. Tightened up the rules for host name lookups using reverse DNS. Exim used
8207 to accept a host name and all its aliases if the forward lookup for any of
8208 them yielded the IP address of the incoming connection. Now it accepts only
8209 those names whose forward lookup yields the correct IP address. Any other
8210 names are discarded. This closes a loophole whereby a rogue DNS
8211 administrator could create reverse DNS records to break through a
8212 wildcarded host restriction in an ACL.
8214 37. If a user filter or a system filter that ran in a subprocess used any of
8215 the numerical variables ($1, $2 etc), or $thisaddress, in a pipe command,
8216 the wrong values were passed to the pipe command ($thisaddress had the
8217 value of $0, $0 had the value of $1, etc). This bug was introduced by
8218 change 4.11/101, and not discovered because I wrote an inadequate test. :-(
8220 38. Improved the line breaking for long SMTP error messages from ACLs.
8221 Previously, if there was no break point between 40 and 75 characters, Exim
8222 left the rest of the message alone. Two changes have been made: (a) I've
8223 reduced the minimum length to 35 characters; (b) if it can't find a break
8224 point between 35 and 75 characters, it looks ahead and uses the first one
8225 that it finds. This may give the occasional overlong line, but at least the
8226 remaining text gets split now.
8228 39. Change 82 of 4.11 was unimaginative. It assumed the limit on the number of
8229 file descriptors might be low, and that setting 1000 would always raise it.
8230 It turns out that in some environments, the limit is already over 1000 and
8231 that lowering it causes trouble. So now Exim takes care not to decrease it.
8233 40. When delivering a message, the value of $return_path is set to $sender_
8234 address at the start of routing (routers may change the value). By an
8235 oversight, this default was not being set up when an address was tested by
8236 -bt or -bv, which affected the outcome if any router or filter referred to
8239 41. The idea of the "warn" ACL verb is that it adds a header or writes to the
8240 log only when "message" or "log_message" are set. However, if one of the
8241 conditions was an address verification, or a call to a nested ACL, the
8242 messages generated by the underlying test were being passed through. This
8243 no longer happens. The underlying message is available in $acl_verify_
8244 message for both "message" and "log_message" expansions, so it can be
8245 passed through if needed.
8247 42. Added RFC 2047 interpretation of header lines for $h_ expansions, with a
8248 new expansion $bh_ to give the encoded byte string without charset
8249 translation. Translation happens only if iconv() is available; HAVE_ICONV
8250 indicates this at build time. HEADERS_CHARSET gives the charset to
8251 translate to; headers_charset can change it in the configuration, and
8252 "headers charset" can change it in an individual filter file.
8254 43. Now that we have a default RFC 2047 charset (see above), the code in Exim
8255 that creates RFC 2047 encoded "words" labels them as that charset instead
8256 of always using iso-8859-1. The cases are (i) the explicit ${rfc2047:
8257 expansion operator; (ii) when Exim creates a From: line for a local
8258 message; (iii) when a header line is rewritten to include a "phrase" part.
8260 44. Nasty bug in exiqsumm: the regex to skip already-delivered addresses was
8261 buggy, causing it to skip the first lines of messages whose message ID
8262 ended in 'D'. This would not have bitten before Exim release 4.14, because
8263 message IDs were unlikely to end in 'D' before then. The effect was to have
8264 incorrect size information for certain domains.
8266 45. #include "config.h" was missing at the start of the crypt16.c module. This
8267 caused trouble on Tru64 (aka OSF1) systems, because HAVE_CRYPT16 was not
8270 46. If there was a timeout during a "random" callout check, Exim treated it as
8271 a failure of the random address, and carried on sending RSET and the real
8272 address. If the delay was just some slowness somewhere, the response to the
8273 original RCPT would be taken as a response to RSET and so on, causing
8274 mayhem of various kinds.
8276 47. Change 50 for 4.20 was a heap of junk. I don't know what I was thinking
8277 when I implemented it. It didn't allow for the fact that some option values
8278 may legitimately be negative (e.g. size_addition), and it didn't even do
8279 the right test for positive values.
8281 48. Domain names in DNS records are case-independent. Exim always looks them up
8282 in lower case. Some resolvers return domain names in exactly the case they
8283 appear in the zone file, that is, they may contain uppercase letters. Not
8284 all resolvers do this - some return always lower case. Exim was treating a
8285 change of case by a resolver as a change of domain, similar to a widening
8286 of a domain abbreviation. This triggered its re-routing code and so it was
8287 trying to route what was effectively the same domain again. This normally
8288 caused routing to fail (because the router wouldn't handle the domain
8289 twice). Now Exim checks for this case specially, and just changes the
8290 casing of the domain that it ultimately uses when it transmits the message
8293 49. Added Sieve (RFC 3028) support, courtesy of Michael Haardt's contributed
8296 50. If a filter generated a file delivery with a non-absolute name (possible if
8297 no home directory exists for the router), the forbid_file option was not
8300 51. Added '&' feature to dnslists, to provide bit mask matching in addition to
8301 the existing equality matching.
8303 52. Exim was using ints instead of ino_t variables in some places where it was
8304 dealing with inode numbers.
8306 53. If TMPDIR is defined in Local/Makefile (default in src/EDITME is
8307 TMPDIR="/tmp"), Exim checks for the presence of an environment variable
8308 called TMPDIR, and if it finds it is different, it changes its value.
8310 54. The smtp_printf() function is now made available to local_scan() so
8311 additional output lines can be written before returning. There is also an
8312 smtp_fflush() function to enable the detection of a dropped connection.
8313 The variables smtp_input and smtp_batched_input are exported to
8316 55. Changed the default runtime configuration: the message "Unknown user"
8317 has been removed from the ACL, and instead placed on the localuser router,
8318 using the cannot_route_message feature. This means that any verification
8319 failures that generate their own messages won't get overridden. Similarly,
8320 the "Unrouteable address" message that was in the ACL for unverifiable
8321 relay addresses has also been removed.
8323 56. Added hosts_avoid_esmtp to the smtp transport.
8325 57. The exicyclog script was not checking for the esoteric option
8326 CONFIGURE_FILE_USE_EUID in the Local/Makefile. It now does this, but it
8327 will work only if exicyclog is run under the appropriate euid.
8329 58. Following a discussion on the list, the rules by which Exim recognises line
8330 endings on incoming messages have been changed. The -dropcr and drop_cr
8331 options are now no-ops, retained only for backwards compatibility. The
8332 following line terminators are recognized: LF CRLF CR. However, special
8333 processing applies to CR:
8335 (i) The sequence CR . CR does *not* terminate an incoming SMTP message,
8336 nor a local message in the state where . is a terminator.
8338 (ii) If a bare CR is encountered in a header line, an extra space is added
8339 after the line terminator so as not to end the header. The reasoning
8340 behind this is that bare CRs in header lines are most likely either
8341 to be mistakes, or people trying to play silly games.
8343 59. The size of a message, as listed by "-bp" or in the Exim monitor window,
8344 was being incorrectly given as 18 bytes larger than it should have been.
8345 This is a VOB (very old bug).
8347 60. This may never have affected anything current, but just in case it has:
8348 When the local host is found other than at the start of a list of hosts,
8349 the local host, those with the same MX, and any that follow, are discarded.
8350 When the list in question was part of a longer list of hosts, the following
8351 hosts (not currently being processed) were also being discarded. This no
8352 longer happens. I'm not sure if this situation could ever has previously
8355 61. Added the "/MX" feature to lists of hosts in the manualroute and query
8358 62. Whenever Exim generates a new message, it now adds an Auto-Submitted:
8359 header. This is something that is recommended in a new Internet Draft, and
8360 is something that is documented as being done by Sendmail. There are two
8361 possible values. For messages generated by the autoreply transport, Exim
8364 Auto-Submitted: auto-replied
8366 whereas for all other generated messages (e.g. bounces) it adds
8368 Auto-Submitted: auto-generated
8370 63. The "personal" condition in filters now includes a test for the
8371 Auto-Submitted: header. If it contains the string "auto-" the message it
8372 not considered personal.
8374 64. Added rcpt_include_affixes as a generic transport option.
8376 65. Added queue_only_override (default true).
8378 66. Added the syslog_duplication option.
8380 67. If what should have been the first header line of a message consisted of
8381 a space followed by a colon, Exim was mis-interpreting it as a header line.
8382 It isn't of course - it is syntactically invalid and should therefore be
8383 treated as the start of the message body. The misbehaviour could have
8384 caused a number of strange effects, including loss of data in subsequent
8385 header lines, and spool format errors.
8387 68. Formerly, the AUTH parameter on a MAIL command was trusted only if the
8388 client host had authenticated. This control can now be exercised by an ACL
8389 for more flexibility.
8391 69. By default, callouts do not happen when testing with -bh. There is now a
8392 variant, -bhc, which does actually run the callout code, including
8393 consulting and updating the callout cache.
8395 70. Added support for saslauthd authentication, courtesy of Alexander
8398 71. If statvfs() failed on the spool or log directories while checking their
8399 size for availability, Exim confusingly gave the error "space shortage".
8400 Furthermore, in debugging mode it crashed with a floating point exception.
8401 These checks are done if check_{spool,log}_{space,inodes} are set, and when
8402 an SMTP message arrives with SIZE= on the MAIL command. As this is a really
8403 serious problem, Exim now writes to the main and panic logs when this
8404 happens, with details of the failure. It then refuses to accept the
8405 incoming message, giving the message "spool directory problem" or "log
8406 directory problem" with a 421 code for SMTP messages.
8408 72. When Exim is about to re-exec itself, it ensures that the file descriptors
8409 0, 1, and 2 exist, because some OS complain for execs without them (see
8410 ChangeLog 4.05/30). If necessary, Exim opens /dev/null to use for these
8411 descriptors. However, the code omitted to check that the open succeeded,
8412 causing mysterious errors if for some reason the permissions on /dev/null
8413 got screwed. Now Exim writes a message to the main and panic logs, and
8414 bombs out if it can't open /dev/null.
8416 73. Re-vamped the way daemon_smtp_port, local_interfaces, and -oX work and
8417 interact so that it is all more flexible. It is supposed to remain
8418 backwards compatible. Also added extra_local_interfaces.
8420 74. Invalid data sent to a SPA (NTLM) server authenticator could cause the code
8421 to bomb out with an assertion failure - to the client this appears as a
8422 connection drop. This problem occurs in the part of the code that was taken
8423 from the Samba project. Fortunately, the assertion is in a very simple
8424 function, so I have fixed this by reproducing the function inline in the
8425 one place where it is called, and arranging for authentication to fail
8426 instead of killing the process with assert().
8428 75. The SPA client code was not working when the server requested OEM rather
8429 than Unicode encoding.
8431 76. Added code to make require_files with a specific uid setting more usable in
8432 the case where statting the file as root fails - usually a non-root-mounted
8433 NFS file system. When this happens and the failure is EACCES, Exim now
8434 forks a subprocess and does the per-uid checking as the relevant uid.
8436 77. Added process_log_path.
8438 78. If log_file_path was not explicitly set, a setting of check_log_space or
8439 check_log_inodes was ignored.
8441 79. If a space check for the spool or log partitions fails, the incident is now
8442 logged. Of course, in the latter case the data may get lost...
8444 80. Added the %p formatting code to string_format() so that it can be used to
8445 print addresses in debug_print(). Adjusted all the address printing in the
8446 debugging in store.c to use %p rather than %d.
8448 81. There was a concern that a line of code in smtp_in.c could overflow a
8449 buffer if a HELO/EHLO command was given followed by 500 or so spaces. As
8450 initially expressed, the concern was not well-founded, because trailing
8451 spaces are removed early. However, if the trailing spaces were followed by
8452 a NULL, they did not get removed, so the overflow was possible. Two fixes
8455 (a) I re-wrote the offending code in a cleaner fashion.
8456 (b) If an incoming SMTP command contains a NULL character, it is rejected
8459 82. When Exim changes uid/gid to the Exim user at daemon start time, it now
8460 runs initgroups(), so that if the Exim user is in any additional groups,
8461 they will be used during message reception.
8467 The change log for 4.20 and earlier releases has been archived.