1 Fetchmail Release Notes
2 =======================
4 This file is in Unicode charset with UTF-8 encoding.
5 All dates are in Universal Time unless otherwise noted.
7 (The `lines' figures total .c, .h, .l, and .y files under version control.
8 Abbreviations in parentheses are the maintainers who committed the respective
9 change. MA = Matthias Andree, ESR = Eric S. Raymond, RF = Rob Funk.)
11 # ADVANCE WARNING OF FEATURES TO BE REMOVED OR CHANGED IN FUTURE VERSIONS
12 (There are no plans to remove features from a 6.4.X release, but they may be
13 removed from a 6.5.0 or newer release.)
14 * Future fetchmail releases may require compilers and operating systems
15 that adhere to standards issued 2011 or later.
16 (Currently, C89 and Single Unix Specification V2 should suffice.)
17 * Future fetchmail releases may tighten up security and lean towards
18 it a bit more by, for instance, implementing recommendations from
19 RFC-7817 or RFC-8314. This may, for instance, require that TLS v1.1
21 * The MX and host alias DNS lookups that fetchmail performs in multidrop mode
22 are based on assumptions that are rarely met in practice, somewhat defective,
23 deprecated and may be removed from a future fetchmail version.
24 They have never supported IPv6 (including IPv6-mapped IPv4).
25 Non-DNS based alias keywords such as "aka" will remain in fetchmail.
26 * The monitor and interface options may be removed from a future fetchmail
27 version as they are not reasonably portable across operating systems.
28 * POP2 is obsolete, support will be removed from a future fetchmail version.
29 * IMAP2 and IMAP4 (not IMAP4r1) are obsolete, support may be removed from a
30 future fetchmail version.
31 * RPOP is obsolete, support will be removed from a future fetchmail release.
32 * The multidrop To/Cc guessing code along with the fragile duplicate suppressor
33 is deprecated and may be removed from a future release.
34 * The "envelope Received" option may be removed from a future release, because
35 the Received header was never meant to be machine-readable, the format varies
36 widely, and various other differences in behavior make parsing Received an
37 unreliable undertaking. The envelope option as such will remain though, in
38 order to support Delivered-To, X-Envelope-To, X-Original-To and similar.
39 See also <http://home.pages.de/~mandree/mail/multidrop>.
40 * The --enable-fallback (fall back to MDA if MTA unavailable) will be removed
41 from a future fetchmail release, because it makes fetchmail's behavior
42 inconsistent and confusing.
43 * The "protocol auto" default inside fetchmail may be removed from a future
44 fetchmail release. Explicit configuration of the protocol is recommended.
45 * Kerberos IV support may be removed from a future fetchmail release.
46 * Kerberos 5 support may be removed from a future fetchmail release.
47 * The --principal option may be removed from a future fetchmail release.
48 * SIGHUP wakeup support may be removed from a future fetchmail release and
49 cause fetchmail to terminate - it was broken for many years.
50 * Support for operating systems that are not sufficiently POSIX compliant may be
51 removed or operation on such systems may be suboptimal for future releases.
52 This means that fetchmail may only continue to work on C99 and POSIX 2001
54 * The maintainer may migrate fetchmail to C++ with STL or C#, and impose further
55 requirements (dependencies), such as Boost or other class libraries.
56 * The softbounce option default will change to "false" in the next release.
57 * The --bsmtp - mode of operation may be removed in a future release.
58 * Given that OpenSSL is severely underdocumented, and needs license exceptions,
59 fetchmail may switch to a different SSL library.
60 * SSLv3 support may be removed from a future fetchmail release. It has been
61 obsolete for many years and found insecure. Use TLS.
62 * Fetchmailconf is deprecated and will be removed from a future release.
63 * Fetchmail does not guarantee compatibility with EOL OpenSSL versions. Support
64 for end-of-life OpenSSL versions may be removed even from patchlevel releases.
66 --------------------------------------------------------------------------------
67 fetchmail-6.4.6 (released 2020-05-29, 27596 LoC):
69 ## TRANSLATION UPDATE, with thanks to the translator:
70 * eo: Felipe Castro [Esperanto]
72 --------------------------------------------------------------------------------
74 fetchmail-6.4.5 (released 2020-05-07, 27596 LoC):
77 * fetchmail 6.4.0 and 6.4.1 changed the resolution of the home directory
78 in a way that requires SUSv4 semantics of realpath(), which leads to
79 'Cannot find absolute path for... directory' error messages followed by aborts
80 on systems where realpath() follows strict SUSv2 semantics and returns
81 EINVAL if the 2nd argument is NULL.
83 On such systems, for instance, Solaris 10, fetchmail requires PATH_MAX to be
84 defined, and will then work again. Regression reported by David Hough.
86 On systems that neither provide auto-allocation semantics for realpath(),
87 nor PATH_MAX, fetchmail will print this error and abort. Such systems
88 are unsupported, see README.
91 * Add a test program fm_realpath, and a t.realpath script, neither to be
92 installed. These will test resolution of the current working directory.
94 ## TRANSLATION UPDATES in reverse alphabetical order of language codes,
95 ## with my thanks to the translators:
96 * zh_CN: Boyuan Yang [Chinese (simplified)]
97 * sv: Göran Uddeborg [Swedish]
98 * sq: Besnik Bleta [Albanian]
99 * pl: Jakub Bogusz [Polish]
100 * ja: Takeshi Hamasaki [Japanese]
101 * fr: Frédéric Marchal [French]
102 * cs: Petr Pisar [Czech]
104 # KNOWN BUGS AND WORKAROUNDS
105 (This section floats upwards through the NEWS file so it stays with the
106 current release information)
107 * Fetchmail does not handle messages without Message-ID header well
108 (See sourceforge.net bug #780933)
109 * Fetchmail currently uses 31-bit signed integers in several places
110 where unsigned and/or wider types should have been used, for instance,
111 for mailbox sizes, and misreports sizes of 2 GibiB and beyond.
112 Fixing this requires C89 compatibility to be relinquished.
113 * BSMTP is mostly untested and errors can cause corrupt output.
114 * Fetchmail does not track pending deletes across crashes.
115 * The command line interface is sometimes a bit stubborn, for instance,
116 fetchmail -s doesn't work with a daemon running.
117 * Linux systems may return duplicates of an IP address in some circumstances if
118 no or no global IPv6 addresses are configured.
119 (No workaround. Ubuntu Bug#582585, Novell Bug#606980.)
120 * Kerberos 5 may be broken, particularly on Heimdal, and provide bogus error
121 messages. This will not be fixed, because the maintainer has no Kerberos 5
122 server to test against. Use GSSAPI.
124 --------------------------------------------------------------------------------
126 fetchmail-6.4.4 (released 2020-04-26, 27530 LoC):
128 ## UPDATED TRANSLATIONS - WITH THANKS TO THE TRANSLATOR:
129 * ja: Takeshi Hamasaki [Japanese]
131 --------------------------------------------------------------------------------
133 fetchmail-6.4.3 (released 2020-04-05, 27530 LoC):
136 * Plug memory leaks when parts of the configuration (defaults, rcfile, command
137 line) override one another.
138 * fetchmail terminated the placeholder command string too late and included
139 garbage from the heap at the end of the string. Workaround: don't use place-
140 holders %h or %p in the --plugin string. Bug added in 6.4.0 when merging
141 Gitlab merge request !5 in order to fix an input buffer overrun.
142 Faulty commit 418cda65f752e367fa663fd13884a45fcbc39ddd.
143 Reported by Stefan Thurner, Gitlab issue #16.
144 * Fetchmail now checks for errors when trying to read the .idfile,
146 * Fetchmail's error messages that reports that the defaults entry isn't the
147 first was made more precise. It could be misleading if there was a poll or
148 skip statement before the defaults.
151 * Fetchmail documentation was updated to require OpenSSL 1.1.1.
152 OpenSSL 1.0.2 reached End Of Life status at the end of the year 2019.
153 Fetchmail will tolerate, but warn about, 1.0.2 for now on the assumption that
154 distributors backport security fixes as the need arises.
155 Fetchmail will also warn if another SSL library that is API-compatible
156 with OpenSSL lacks TLS v1.3 support.
157 * If the trust anchor is missing, fetchmail refers the user to README.SSL.
160 * The AC_DECLS(getenv) check was removed, its only user was broken and not
161 accounting for that AC_DECLS always defines HAVE_DECL_... to 0 or 1, so
162 fetchmail never declared a missing getenv() symbol (it was testing with
163 #ifdef). Remove the backup declaration. getenv is mandated by SUSv2 anyways.
165 ## UPDATED TRANSLATIONS - WITH THANKS TO THE TRANSLATORS:
166 * sq: Besnik Bleta [Albanian]
167 * zh_CN: Boyuan Yang [Chinese (simplified)]
168 * pl: Jakub Bogusz [Polish]
169 * cs: Petr Pisar [Czech]
170 * fr: Frédéric Marchal [French]
171 * sv: Göran Uddeborg [Swedish]
172 * eo: Felipe Castro [Esperanto]
174 --------------------------------------------------------------------------------
176 fetchmail-6.4.2 (released 2020-02-14, 27473 LoC):
179 * fetchmailconf now supports Python 3 and currently requires the "future"
180 package, see https://pypi.org/project/future/.
181 * fetchmailconf: The minimum supported version is now Python 2.7.13, but it is
182 recommended to use at least 2.7.16 (due to its massive SSL updates).
183 Older Python versions may check SSL certificates not strictly enough,
184 which may cause fetchmail to complain later, if the certificate verify fails.
185 * fetchmailconf now autoprobes SSL-wrapped connections (ports 993 and 995 for
186 IMAP and POP3) as well and by preference.
187 * fetchmailconf now defaults newly created users to "ssl" if either of the
188 existing users sets ssl, or if the server has freshly been probed and
189 found supporting ssl.
190 There is a caveat: adding a user to an existing server without probing it
191 again may skip adding ssl. (This does not prevent STARTTLS.)
194 * Fix three bugs in fetchmail.man (one unterminated string to .IP macro, one
195 line that ran into a .PP macro, .TH date format), and remove one .br request
196 from inside the table, which is unsupported by FreeBSD 12's mandoc(1)
197 formatter. FreeBSD Bug#241032, reported by Helge Oldach.
198 * Further man page fixes and additions by Chris Mayo and Gregor Zattler.
199 * When evaluating the need for STARTTLS in non-default configurations (SSL
200 certificate validation turned off), fetchmail would only consider --sslproto
201 tls1 as requiring STARTTLS, now all non-empty protocol versions do.
202 * fetchmailconf now properly writes "no sslcertck" if sslcertck is disabled.
203 * fetchmailconf now catches and reports OS errors (including DNS errors) when
204 autoprobing. Reported as Gitlab issue #12 by Sergey Alirzaev.
205 * fetchmailconf received a host of other bugfixes, see the Git commit log.
208 * Make t.smoke more robust and use temporary directory as FETCHMAILHOME, to make
209 sure that the home directory resolves for the user running the test suite
210 even if the environment isn't perfect. Reported by Konstantin Belousov,
211 analysed by Corey Halpin, FreeBSD Bug#240914.
213 ## UPDATED TRANSLATION - THANKS TO:
214 * zh_CN: Boyuan Yang [Chinese (simplified)]
216 --------------------------------------------------------------------------------
218 fetchmail 6.4.1 (released 2019-09-28, 27473 LoC):
221 * The bug fix Debian Bug#941129 was incomplete and caused
222 + a regression in the default file locations, so that fetchmail was no longer
223 able to find its configuration files in some situations.
224 Reported by Cy Schubert, Christian Ebert.
225 + a regression under _FORTIFY_SOURCE where PATH_MAX > minimal _POSIX_PATH_MAX.
227 --------------------------------------------------------------------------------
229 fetchmail 6.4.0 (released 2019-09-27, 27429 LoC):
231 # NOTE THAT FETCHMAIL IS NO LONGER PUBLISHED THROUGH IBIBLIO.
232 * They have stopped accepting submissions and consider themselves an archive.
234 ## SECURITY FIXES THAT AFFECT BEHAVIOUR AND MAY REQUIRE RECONFIGURATION
235 * Fetchmail no longer supports SSLv2.
237 * Fetchmail no longer attempts to negotiate SSLv3 by default,
238 even with --sslproto ssl23. Fetchmail can now use SSLv3, or TLSv1.1 or a newer
239 TLS version, with STLS/STARTTLS (it would previously force TLSv1.0 with
240 STARTTLS). If the OpenSSL version used at build and run-time supports these
241 versions, --sslproto ssl3 and --sslproto ssl3+ can be used to re-enable SSLv3.
242 Doing so is discouraged because the SSLv3 protocol is broken.
244 Along the lines suggested - as patch - by Kurt Roeckx, Debian Bug #768843.
246 While this change is supposed to be compatible with common configurations,
247 users may have to and are advised to change all explicit --sslproto ssl2
248 (change to newer protocols required), --sslproto ssl3, --sslproto tls1 to
249 --sslproto auto, so that they can benefit from TLSv1.1 and TLSv1.2 where
250 supported by the server.
252 The --sslproto option now understands the values auto, ssl3+, tls1+, tls1.1,
253 tls1.1+, tls1.2, tls1.2+, tls1.3, tls1.3+ (case insensitively), see CHANGES
256 * Fetchmail defaults to --sslcertck behaviour. A new option --nosslcertck to
257 override this has been added, but may be removed in future fetchmail versions
258 in favour of another configuration option that makes the insecurity in using
262 * Fetchmail prevents buffer overruns in GSSAPI authentication with user names
263 beyond c. 6000 characters in length. Reported by Greg Hudson.
265 ## CHANGED REQUIREMENTS
266 * fetchmail 6.4.0 is written in C99 and requires a SUSv3 (Single Unix
267 Specification v3, a superset of POSIX.1-2001 aka. IEEE Std 1003.1-2001 with
268 XSI extension) compliant system. For now, a C89 compiler should also work
269 if the system is SUSv3 compliant.
271 In particular, older fetchmail versions had workaround for several functions
272 standardized in the Single Unix Specification v3, these have been removed.
274 The trio/ library has been removed from the distribution.
277 * fetchmail 6.3.X is unsupported.
278 * fetchmail now configures OpenSSL support by default.
279 * fetchmail now requires OpenSSL v1.0.2 or newer.
280 * Fetchmail now supports --sslproto auto and --sslproto tls1+ (same as ssl23).
281 * --sslproto tls1.1+, tls1.2+, and tls1.3+ are now supported for
282 auto-negotiation with a minimum specified TLS protocol version, and --sslproto
283 tls1.1, --sslproto tls1.2 and --sslproto tls1.3 to force the specified TLS
284 protocol version. Note that tls1.3 requires OpenSSL v1.1.1 or newer.
285 * Fetchmail now detects if the server hangs up prematurely during SSL_connect()
286 and reports this condition as such, and not just as SSL connection failure.
287 (OpenSSL 1.0.2 reported incompatible with pop3.live.com by Jerry Seibert).
288 * A foreground fetchmail can now accept a few more options while another copy is
289 running in the background.
290 * fetchmail now handles POP3 --keep UID lists more efficiently, by using Rainer
291 Weikusat's P-Tree implementation. This reduces the complexity for handling
292 a large UIDL from O(n^2) to O(n log n) and becomes noticably faster with
293 thousands of kept messages.
294 (IMAP does not currently track UIDs and is unaffected.)
295 At the same time, the UIDL emulation code for deficient servers has been
296 removed. It never worked really well. Servers that do not implement the
297 optional UIDL command only work with --fetchall option set, which in itself is
298 incompatible with the --keep option (it would cause message duplication).
299 * fetchmail, when setting up TLS connections, now uses SSL_set_tlsext_host_name()
300 to set up the SNI (Server Name Indication). Some servers (for instance
301 googlemail) require SNI when using newer SSL protocols.
302 * Fetchmail now sets the expected hostname through OpenSSL 1.0.2's new
303 X509_VERIFY_PARAM_set1_host() function to enable OpenSSL's native certificate
304 verification features.
305 * fetchmail will drop the connection when fetching with IMAP and receiving an
306 unexpected untagged "* BYE" response, to work around certain faulty servers.
307 * The FETCHMAIL_POP3_FORCE_RETR environment variable is now documented,
308 it forces fetchmail, when talking POP3, to always use the RETR command,
309 even if it would otherwise use the TOP command.
310 * Fetchmail's configure stage will try to query pkg-config or pkgconf for libssl
311 and libcrypto, in case other system use .pc files to document specific library
312 dependencies. (contributed by Fabrice Fontaine, GitLab merge request !14.)
313 * The gethostbyname() API calls and compatibility functions have been removed.
314 * These translations are shipped but not installed by default because
315 they have less than 500 translated messages out of 714: el fi gl pt_BR sk tr
316 -> Greek, Finnish, Galician, Brazilian Portuguese, Slovak, Turkish.
317 * Fetchmail now refuses delivery if the MDA option contains single-quoted
321 * Fix a typo in the FAQ. Submitted by David Lawyer, Debian Bug#706776.
322 * Do not translate header tags such as "Subject:". Reported by Gonzalo Pérez de
323 Olaguer Córdoba, Debian Bug#744907.
324 * Convert most links from berlios.de to sourceforge.net.
325 * Report error to stderr, and exit, if --idle is combined with multiple
327 * Point to --idle from GENERAL OPERATION to clarify --idle and multiple
328 mailboxes do not mix. In response to Jeremy Chadwick's trouble 2014-11-19,
329 fetchmail-users mailing list.
330 * Fix SSL-enabled build on systems that do not declare SSLv3_client_method(),
331 or that #define OPENSSL_NO_SSL3 inside #include <openssl/ssl.h>
332 Related to Debian Bug#775255. Fixes Debian Bug #804604.
333 * Version report lists -SSLv3 on SSL-enabled no-ssl3 builds.
334 * Fetchmail no longer adds a NUL byte to the username in GSSAPI authentication.
335 This was reported to break Kerberos-based authentication with Microsoft
336 Exchange 2013 by Greg Hudson.
337 * Set umask properly before writing the .fetchids file, to avoid failing the
338 security check on the next run. Reported by Fabian Raab,
339 Fixes Debian Bug#831611.
340 * When forwarding by LMTP, also check antispam response code when collecting
341 the responses after the CR LF . CR LF sequence at the end of the DATA phase.
342 (Contributed by Evil.2000, GitLab merge request !12.)
343 * fetchmail will not try other protocols after a socket error. This avoids
344 mismatches of how different prococols see messages as "seen" and re-fetches
345 of known mail. (Fix contributed by Lauri Nurmi, GitLab Merge Request !10.)
346 * fetchmail no longer reports "System error during SSL_connect(): Success."
347 Fixes Debian Bug#928916, reported by Paul Kimoto.
348 * fetchmailconf would ignore Edit or Delete actions on the first (topmost)
349 item in a list (no matter if server list, user list, ...).
350 * The mimedecode feature now properly detects multipart/mixed-type matches, so
351 that quoted-printable-encoded multipart messages can get decoded.
352 (Regression in 5.0.0 on 1999-03-27, as a side effect of a PGP-mimedecode fix
353 attributed to Henrik Storner.)
354 * FETCHMAILHOME can now safely be a relative path, which will be qualified
355 through realpath(). Previously, it had to be absolute in daemon mode.
356 Reported by Alex Andreotti, Debian Bug#941129.
358 ## UPDATED TRANSLATIONS - THANKS TO:
359 * CS: Petr Pisar <petr.pisar@atlas.cz> [Czech]
360 * EO: Felipe Castro <fefcas@gmail.com> [Esperanto]
361 * FR: Frédéric Marchal <fmarchal@perso.be> [French]
362 * JP: Takeshi Hamasaki <hmatrjp@users.sourceforge.jp> [Japanese]
363 * PL: Jakub Bogusz <qboosh@pld-linux.org> [Polish]
364 * SV: Göran Uddeborg <goeran@uddeborg.se> [Swedish]
366 --------------------------------------------------------------------------------
368 fetchmail-6.3.26 (released 2013-04-23, 26180 LoC):
370 # CRITICAL BUG FIX for setups using "mimedecode":
371 * The mimedecode feature failed to ship the last line of the body if it was
372 encoded as quoted-printable and had a MIME soft line break in the very last
373 line. Reported by Lars Hecking in June 2011.
375 Bug introduced on 1998-03-20 when the mimedecode support was added by ESR
376 before release 4.4.1 through code contributed by Henrik Storner.
377 Workaround for older releases: do not use mimedecode feature.
379 Earlier versions of this NEWS file claimed this bug fixed in fetchmail-6.3.23,
382 Fixes Launchpad Bug#1171818.
385 fetchmail-6.3.25 (released 2013-03-18, 26149 LoC):
388 * Fix a memory leak in out-of-memory error condition while handling plugins.
389 Report and patch by John Beck (found with Parfait static code analyzer).
390 * Fix a NULL pointer dereference in out-of-memory error condition while handling
392 Report and patch by John Beck (found with Parfait static code analyzer).
395 * Improved reporting when SSL/TLS X.509 certificate validation has failed,
396 working around a not-so-recent swapping of two OpenSSL error codes, and
397 a practical impossibility to distinguish broken certification chains from
398 missing trust anchors (root certificates).
399 * OpenSSL decoded errors are now reported through report(), rather than dumped
400 to stderr, so that they should show up in logfiles and/or syslog.
401 * The fetchmail manual page no longer claims that MD5 were the default OpenSSL
402 hash format (for use with --sslfingerprint). Reported by Jakob Wilk,
403 PARTIAL fix for Debian Bug#700266.
404 * The fetchmail manual page now refers the user to --softbounce from the
405 SMTP/ESMTP ERROR HANDLING section. Reported by Anton Shterenlikht.
408 * Older systems that provide the older RFC-2553 implementation of getaddrinfo,
409 rather than the current RFC-3493, and systems that do not provide this
410 getaddrinfo() interface at all and thus use the replacement functions from
411 libesmtp/getaddrinfo.?, might return EAI_NODATA when a host is registered in
412 DNS as MX or similar, but without A or AAAA records. Handle this situation
413 when checking for multidrop aliases and treat EAI_NODATA the same as
414 EAI_NONAME, i. e. name cannot be resolved.
416 The proper fix, however, is to upgrade the operating system.
418 # TRANSLATION UPDATES
419 [cs] Czech, by Petr Pisar
420 [da] Danish, by Joe Hansen
422 [eo] Esperanto, by Sian Mountbatten and Felipe Castro
423 [fr] French, by Frédéric Marchal
424 [ja] Japanese, by Takeshi Hamasaki
425 [pl] Polish, by Jakub Bogusz
426 [sv] Swedish, by Göran Uddeborg
427 [vi] Vietnamese, by Trần Ngọc Quân
430 fetchmail-6.3.24 (released 2012-12-23, 26108 LoC):
432 # CRITICAL AND REGRESSION FIXES
433 * Plug a memory leak in OpenSSL's certificate verification callback.
434 This would affect fetchmail configurations running with SSL in daemon mode
435 more than one-shot runs.
436 Reported by Erik Thiele, and pinned by Dominik Heeg,
437 fixes Debian Bug #688015.
438 This bug was introduced into fetchmail 6.3.0 (committed 2005-10-29)
439 when support for subjectAltName was added through a patch by Roland
440 Stigge, submitted as Debian Bug#201113.
442 * The --logfile option now works again outside daemon mode, reported by Heinz
443 Diehl. The documentation that I had been reading was inconsistent with the
444 code, and only parts of the manual page claimed that --logfile was only
445 effective in daemon mode.
448 fetchmail-6.3.23 (released 2012-12-10, 26106 LoC):
451 * Fix compilation with OpenSSL implementations before 0.9.8m that lack
452 SSL_CTX_clear_options. Patch by Earl Chew.
453 Note that the use of older OpenSSL versions with fetchmail is unsupported and
457 * Fix combination of --plugin and -f -. Patch by Alexander Zangerl,
458 to fix Debian Bug#671294.
459 * Clean up logfile vs. syslog handling, and in case logfile overrides
460 syslog, send a message to the latter stating where logging goes.
463 * The build process can now be made a bit more silent and concise through
464 ./configure --enable-silent-rules, or by adding "V=0" to the make command.
467 * Make Maillennium POP3 workarounds less specific, to encompass
468 Maillennium POP3/UNIBOX (Maillennium V05.00c++). Reported by Eddie
469 via fetchmail-users mailing list, 2012-10-13.
471 # TRANSLATION UPDATES
472 [cs] Czech, by Petr Pisar
473 [da] Danish, by Joe Hansen
475 [fr] French, Frédéric Marchal
476 [ja] Japanese, Takeshi Hamasaki
477 [pl] Polish, by Jakub Bogusz
478 [sv] Swedish, by Göran Uddeborg
479 [vi] Vietnamese, Trần Ngọc Quân
482 fetchmail-6.3.22 (released 2012-08-29, 26077 LoC):
486 NTLM: fetchmail mistook an error message that the server sent in response to
487 an NTLM request for protocol exchange, tried to decode it, and crashed while
488 reading from a bad memory location.
489 Also, with a carefully crafted NTLM challenge packet sent from the server, it
490 would be possible that fetchmail conveyed confidential data not meant for the
491 server through the NTLM response packet.
492 Fix: Detect base64 decoding errors, validate the NTLM challenge, and abort
493 NTLM authentication in case of error.
494 See fetchmail-SA-2012-02.txt for further details.
495 Reported by J. Porter Clark.
498 SSL/TLS (wrapped and STARTTLS): fetchmail used to disable a countermeasure
499 against a certain kind of attack against cipher block chaining initialization
500 vectors (SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS).
501 Whether this creates an exploitable situation, depends on the server and the
503 As a precaution, fetchmail 6.3.22 enables the countermeasure, by clearing
504 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS.
506 NOTE that this can cause connections to certain non-conforming servers to
507 fail, in which case you can set the environment variable
508 FETCHMAIL_DISABLE_CBC_IV_COUNTERMEASURE to any non-empty value when starting
509 fetchmail to re-instate the compatibility option at the expense of security.
511 Reported by Apple Product Security.
513 For technical details, refer to <http://www.openssl.org/~bodo/tls-cbc.txt>.
514 See fetchmail-SA-2012-01.txt for further details.
517 * The Server certificate: message in verbose mode now appears on stdout like the
518 remainder of the output. Reported by Henry Jensen, to fix Debian Bug #639807.
520 * The GSSAPI-related autoconf code now matches gssapi.c better, and uses
521 a different check to look for GSS_C_NT_HOSTBASED_SERVICE.
522 This fixes the GSSAPI-enabled build on NetBSD 6 Beta.
525 * On systems where SSLv2_client_method isn't defined in OpenSSL (such as
526 newer Debian, and Ubuntu starting with 11.10 oneiric ocelot), don't
527 reference it (to fix the build) and if configured, print a run-time error
528 that the OS does not support SSLv2. Fixes Debian Bug #622054,
529 but note that that bug report has a more thorough patch that does away with
532 * The security and errata notices fetchmail-{EN,SA}-20??-??.txt are now
533 under the more relaxed CC BY-ND 3.0 license (the noncommercial clause
534 was dropped). The Creative Commons address was updated.
536 * The Python-related Makefile.am parts were simplified to avoid an automake
537 1.11.X bug around noinst_PYTHON, Automake Bug #10995.
539 * Configuring fetchmail without SSL now triggers a configure warning,
540 and asks the user to consider running configure --with-ssl.
543 * Some servers, notably Zimbra, return A1234 987 FETCH () in response to
544 a header request, in the face of message corruption. fetchmail now treats
545 these as temporary errors. Report and Patch by Mikulas Patocka, Red Hat.
547 * Some servers, notably Microsoft Exchange, return "A0009 OK FETCH completed."
548 without any header in response to a header request for meeting reminder
549 messages (with a "meeting.ics" attachment). fetchmail now treats these as
550 transient errors. Report by John Connett, Patch by Sunil Shetye.
552 # TRANSLATION UPDATES
553 * [cs] Czech, by Petr Pisar
555 * [fr] French, by Frédéric Marchal
556 * [ja] Japanese, by Takeshi Hamasaki
557 * [pl] Polish, by Jakub Bogusz
558 * [sv] Swedish, by Göran Uddeborg --- NEW TRANSLATION - Thank you!
559 * [vi] Vietnamese, by Trần Ngọc Quân
562 fetchmail-6.3.21 (released 2011-08-21, 26011 LoC):
565 * The IMAP client no longer inserts NUL bytes into the last line of a message
566 when it is not closed with a LF or CRLF sequence. Reported by Antoine Levitt.
567 As a side effect of the fix, and in order to avoid a full rewrite, fetchmail
568 will now CRLF-terminate the last line fetched through IMAP, even if it is
569 originally not terminated by LF or CRLF. This bears no relevance if your
570 messages end up in mbox, but adds line termination for storages (like Maildir)
571 that do not require that the last line be LF- or CRLF-terminated.
574 * There is a patch against fetchnews's source, contrib/rawlog.patch, that can
575 log (and hexdump non-printing characters) raw socket data to a file. It proved
576 useful to debug Antoine's bug described above.
579 fetchmail-6.3.20 (released 2011-06-06, 26005 LoC):
583 STARTTLS: Fetchmail runs the IMAP STARTTLS or POP3 STLS negotiation with the
584 set timeout (default five minutes) now. This was reported missing, with
585 observed fetchmail freezes beyond a week, by Thomas Jarosch.
586 SSL-wrapped connections were unaffected by this timeout, so users of older
587 versions can force ssl-wrapped connections -- if supported by the server --
588 with the --ssl command line or ssl rcfile option.
589 See fetchmail-SA-2011-01.txt for further details.
592 * IMAP: Do not search for UNSEEN messages in ranges. Usually, there are very few
593 new messages and most of the range searches result in nothing. Instead, split
594 the long response to make the IMAP driver think that there are multiple lines
595 of response. (Sunil Shetye)
596 * Do not print "skipping message" for old messages even in verbose mode. If
597 there are too many old messages, the logs just get filled without any real
598 activity. (Sunil Shetye) (suggested by Yunfan Jiang)
599 * Build: fetchmail now always uses its own MD5 implementation rather than trying
600 to find a system library with matched header. The library and header variants
601 found on systems are too diverse, and the code size saving is not worth any
602 more wasted user or programmer time.
605 * Call strlen() only once when removing CRLF from a line. (Sunil Shetye)
606 * fetchmail sets Internet domain sockets to "keepalive" mode now. Note that
607 there is no portable way to configure actual timeouts for this mode, and some
608 systems only support a system-wide timeout setting. fetchmail does not
609 attempt to tune the time spans of keepalive mode.
611 # TRANSLATION UPDATES
612 [cs] Chech (Petr Pisar)
613 [nl] Dutch (Erwin Poeze)
614 [fr] French (Frédéric Marchal)
615 [de] German (Matthias Andree)
616 [ja] Japanese (Takeshi Hamasaki)
617 [pl] Polish (Jakub Bogusz)
618 [sk] Slovak (Marcel Telka)
621 fetchmail-6.3.19 (released 2010-12-10, 25945 LoC):
623 # ERRATUM NOTICE ISSUED
624 * fetchmail 6.3.18 contains several bug fixes that were considered sufficiently
625 grave to warrant the issue of an erratum notice, fetchmail-EN-2010-03.txt.
628 * When specifying multiple local multidrop lists, do not lose wildcard flag.
629 (Affects "user foo is bar baz * is joe here")
630 * In multidrop configurations, an asterisk can now appear anywhere in the list
631 of local users, not just at the end.
632 * In multidrop mode, header parsing is now more verbose in -vv mode, so that it
633 becomes possible to see which header is used.
634 * Make --antispam work from command line (these used to work in rcfiles).
635 Reported by Kees Bakker, BerliOS Bug #17599. (Sunil Shetye)
636 * Smoke test XHTML 1.1 validation, and if it fails, skip validating HTML
637 documents. Skip validating Mailbox-Names-UTF7.html. Several systems have
638 broken XHTML 1.1 DTD installations that jeopardize the build.
639 Reported by Mihail Nechkin against FreeBSD port.
640 Workaround for 6.3.18: build in a separate directory, i. e:
641 mkdir build && cd build && ../configure --options-go-here
642 * Send a NOOP only after a failed STARTTLS in IMAP. (Sunil Shetye)
643 * Demote GSSAPI verbose/debug syslog to INFO severity. Requested by Carlos E. R.
644 and Derek Simkowiak via the fetchmail-users@ mailing list.
645 * Do STARTTLS/STLS negotiation in IMAP/POP3 if it is mandatory even if the
646 server capabilities do not show support for upgradation to TLS.
647 To use this, configure --sslproto tls1. (Sunil Shetye)
648 * IMAP: Understand empty strings as FETCH response, seen on Yahoo. Reported by
649 Yasin Malli to fetchmail-users@ 2010-12-10.
650 Note that fetchmail continues to expect literals as FETCH response for now.
653 * The manual page now links to IANA for GSSAPI service names.
655 # TRANSLATION UPDATES
656 [cs] Czech (Petr Pisar)
657 [fr] French (Frédéric Marchal)
659 [it] Italian (Vincenzo Campanella)
660 [pl] Polish (Jakub Bogusz)
663 fetchmail-6.3.18 (released 2010-10-09, 25936 LoC):
665 # SECURITY IMPROVEMENTS TO DEFANG X.509 CERTIFICATE ABUSE
666 * Fetchmail now only accepts wildcard certificate common names and subject
667 alternative names if they start with "*.". Previous versions would accept
668 wildcards even if no period followed immediately.
669 * Fetchmail now disallows wildcards in certificates to match domain literals
670 (such as 10.9.8.7), or wildcards in domain literals ("*.168.23.23").
671 The test is overly picky and triggers if the pattern (after skipping the
672 initial wildcard "*") or domain consists solely of digits and dots, and thus
673 matches more than needed.
674 * Fetchmail now disallows wildcarding top-level domains.
676 # CRITICAL BUG FIXES AND REGRESSION FIXES
677 * Fetchmail 6.3.15, 6.3.16, and 6.3.17 would pick up libmd5 to obtain MD5*
678 functions, as an effect of an undocumented Solaris MD5 fix.
679 This caused all MD5-related functions to malfunction if, for instance,
680 libmd5.so was installed on other operating systems as part of libwww on
681 machines where long isn't 32-bits, i. e. usually on 64-bit computers.
682 Fixes Gentoo Bug #319283, reported, including libwww hint, by Karl Hakimian.
683 Side effect: fetchmail will now use -lmd on Solaris rather than -lmd5.
684 * Fetchmail 6.3.17 warned about insecure SSL/TLS connections even if a matching
685 --sslfingerprint was specified. This is an omission from an SSL usability
686 change made in 6.3.17.
687 Fixes Debian Bug#580796 reported by Roland Stigge.
688 * Fetchmail will now apply timeouts to the authentication stage.
689 This stage encompasses STARTTLS/STLS negotiation in IMAP/POP3.
690 Reported missing by Thomas Jarosch.
691 * Fetchmail now cancels GSSAPI authentication properly when encountering GSS
692 errors, such as no or unsuitable credentials.
693 It now sends an asterisk on a line by its own, as required in SASL.
694 This fixes protocol synchronization issues that cause Authentication
695 failures, often observed with kerberized MS Exchange servers.
696 Fixes Debian Bug #568455 reported by Patrick Rynhart, and Alan Murrell, to the
697 fetchmail-users list. Fix verified by Thomas Voigtmann and Patrick Rynhart.
700 * Fetchmail will no longer print connection attempts and errors for one host
701 in "silent" and "normal" logging modes, unless all connections fail. This
702 should reduce irritation around refused-connection logging if services are
703 only on an IPv4 socket if the host also supports IPv6. Often observed as
704 connections refused to ::1/25 when the subsequent connection to 127.0.0.1/25
705 then - silently - succeeds. Fetchmail, unless in verbose mode, will collect
706 all connect errors and only report them if all of them fail.
707 * Fetchmail will not try GSSAPI authentication automatically, unless it has GSS
708 credentials. However, if GSSAPI authentication is requested explicitly,
709 fetchmail will always try it.
710 * Fetchmail now parses response to "FETCH n:m RFC822.SIZE" and "FETCH n
711 RFC822.HEADER" in a more flexible manner. (Sunil Shetye)
712 * The manual page clearly states that --principal is for Kerberos 4 only, not
713 for Kerberos 5 or GSSAPI. Found by Thomas Voigtmann.
716 * When encountering incorrect headers, fetchmail will refer to the bad-header
717 option in the manpage.
718 Fixes BerliOS Bug #17272, change suggested by Björn Voigt.
719 * Fetchmail now decodes and reports GSSAPI status codes upon errors.
720 * Fetchmail now autoprobes NTLM also for POP3.
721 * The Fetchmail FAQ has a new item #R15 on authentication failures.
724 * The common NTLM authentication code was factored out from pop3.c and imap.c.
726 # TRANSLATION UPDATES
727 [zh_CN] Chinese/simplified (Ji Zheng-Yu)
728 [cs] Czech (Petr Pisar)
729 [nl] Dutch (Erwin Poeze)
730 [fr] French (Frédéric Marchal)
732 [it] Italian (Vincenzo Campanella)
733 [ja] Japanese (Takeshi Hamasaki)
734 [pl] Polish (Jakub Bogusz)
735 [sk] Slovak (Marcel Telka)
738 fetchmail-6.3.17 (released 2010-05-06, 25767 LoC):
741 * CVE-2010-1167: Fetchmail before release 6.3.17 did not properly sanitize
742 external input (mail headers and UID). When a multi-character locale (such as
743 UTF-8) was in use, this could cause memory exhaustion and thus a denial of
744 service, because fetchmail's report.c functions assumed that non-success of
745 [v]snprintf was due to insufficient buffer size allocation. It would then
746 repeatedly reallocate a larger buffer and fail formatting again.
747 See fetchmail-SA-2010-02.txt.
750 * Fetchmail now supports a --sslcertfile <file> option to specify a "CA bundle"
751 file (a file that contains trusted CA certificates). Since these bundled CA
752 files do not require c_rehash to be run, they are easier to use and immune to
753 OpenSSL library updates that affect the hash function.
754 * Fetchmail now supports a FETCHMAIL_INCLUDE_DEFAULT_X509_CA_CERTS
755 environment variable to force loading the default SSL CA certificate
756 locations even if --sslcertfile or --sslcertpath is used.
757 If neither option is in effect, fetchmail loads the default locations.
760 * Fix string handling in rcfile scanner, which caused fetchmail to misparse a
761 run control file in certain circumstances. Fixes BerliOS bug #14257.
762 Patch by Michael Banack. This fixes a regression introduced before 6.3.0.
765 * Plug memory leak when using a "defaults" entry in the run control file.
766 * Do not print SSL certificate mismatches unless verbose or --sslcertck is
768 * Do not lose "set invisible" in fetchmailconf. (Michael Barnack)
771 * Usability: SSL certificate chains are fully printed in -v -v mode, and there
772 are now helpful pointers to --sslcertpath and c_rehash for "unable to get
773 local issuer certificate" and self-signed certificates -- these usually hint
774 to missing root signing CAs in the certs directory.
775 * Several fixes for compiler (GCC, Intel C++, CLang) and autotools warnings
776 * Memory allocation failures will now cause abnormal program abort (SIGABRT),
777 no longer an exit with unspecified code.
778 * Print a warning if certificate verification failed and the user did not
782 * Fix table of global option to read "set softbounce" where there used to be a
783 2nd copy of "set spambounce". Patch by Michael Banack, BerliOS Bug #17067.
784 * In the --sslcertpath description, mention that OpenSSL upgrade (and a 0.9.X
785 to 1.0.0 upgrade in particular) may require running c_rehash.
787 # TRANSLATION UPDATES
788 [zh_CN] Chinese/simplified (Ji Zheng-Yu)
789 [cs] Czech (Petr Pisar)
790 [nl] Dutch (Erwin Poeze)
791 [fr] French (Frédéric Marchal)
793 [id] Indonesian (Andhika Padmawan)
794 [it] Italian (Vincenzo Campanella)
795 [ja] Japanese (Takeshi Hamasaki)
796 [pl] Polish (Jakub Bogusz)
797 [sk] Slovak (Marcel Telka)
798 [vi] Vietnamese (Clytie Siddall)
801 fetchmail-6.3.16 (released 2010-04-06, 25574 LoC):
804 * Fix --interface option, broken in 6.3.15. Reported by Vladmimir Stavrinov.
805 Fixes Debian Bug #576717.
808 * Call OpenSSL_add_all_algorithms(). This is needed to support non-mandatory
809 and non-standard algorithms in certificates.
810 Sjoerd Simons, to fix Debian Bug #576430.
811 OpenSSL 0.9.8* does not load - for instance - the SHA256 digest by default.
812 Reported as OpenSSL RT#2224.
815 fetchmail-6.3.15 (released 2010-03-28, 25572 LoC):
818 * Fetchmail now supports a bad-header command line or rcfile option that takes
819 exactly one argument, accept or reject (default). This specifies how messages
820 with bad headers retrieved from the current server are to be treated.
823 * In the rcfile, recognize "local" as abbreviation for "localdomains", as
824 documented. The short form has not ever worked since this feature was added in
825 January 1997. Reported by Frédéric Marchal.
826 * Do not close stdout when using mda and "bsmtp -" at the same time.
827 * Log operating system errors when BSMTP writes fail.
828 * Fix verbose mode progress formatting regression from 6.3.10; SMTP trace lines
829 were no longer on a line of their own. Reported by Melchior Franz.
830 * Check seteuid() return value and abort running MDA if switch fails.
831 * Set global flags in a consistent manner. Make --nosoftbounce and
832 --nobounce work from command line (these used to work in rcfiles).
833 Reported and fix confirmed working by N.J. Mann. (Sunil Shetye)
834 * Properly import h_errno declarations, even on systems where h_errno isn't a
835 macro. (Adds ./configure check, fixes Cygwin dllimport warnings.)
838 * The repository has been converted and moved from the Subversion (SVN) format
839 kindly hosted by Graham Wilson over the past years to Git format hosted on
840 Gitorious.org. My deepest thanks to Graham Wilson for this service that
841 kept us going when BerliOS's Subversion service was faulty in its early days.
842 * This opportunity was used to convert BRANCH_6-2 and BRANCH_1-9-9 to
843 GnuPG-signed tags, as a sign that these are now closed.
844 * The outdated SVN trunk is now called "oldtrunk" in Git just to save the work
845 for future reference. All development in the past few years was on BRANCH_6-3.
846 * master was branched from BRANCH_6-3. BRANCH_6-3 is now obsolete (and in fact
847 was also converted to a tag to record where the conversion from SVN to Git
849 * "make check" now skips HTML validation if xmllint or XHTML DTD are missing.
852 * Web site and documentation were adjusted to reflect the SVN->Git move.
853 * The fetchmail manual page is now much clearer on the user id switching
854 (seteuid) when using --mda while running as the super user.
856 # TRANSLATION UPDATES, by language name
857 * [zh_CN] Chinese (Simplified), by Ji Zheng-Yu
858 * [cs] Czech, by Petr Pisar
859 * [nl] Dutch, by Erwin Poeze
860 * [fr] French, by Frédéric Marchal
862 * [id] Indonesian, by Andhika Padmawan
863 * [it] Italian, by Vincenzo Campanella
864 * [ja] Japanese, by Takeshi Hamasaki
865 * [pl] Polish, by Jakub Bogusz
866 * [vi] Vietnamese, by Clytie Siddall
869 fetchmail 6.3.14 (released 2010-02-05, 25487 LoC):
872 * CVE-2010-0562: SSL/TLS certificate information is now also reported properly
873 on computers that consider the "char" type signed. Fixes malloc() buffer
874 overrun. Workaround for older versions: do not use verbose mode.
875 See fetchmail-SA-2010-01.txt for details, including a minimal patch.
878 * The IMAP client no longer skips messages from several IMAP servers including
879 Dovecot if fetchmail's "idle" is in use. Causes were that fetchmail (a)
880 ignored some untagged responses when it should not (b) relied on EXISTS
881 messages in response to EXPUNGE, which aren't mandated by RFC-3501 (the IMAP
882 standard) and aren't sent by Dovecot either.
883 Fix by Sunil Shetye (the fix also consolidates IMAP response handling,
884 improving overall robustness of the IMAP client), bug report and testing by
885 Matt Doran, with further hints from Timo Sirainen.
886 * The SMTP client now recovers from errors (such as servers dropping the
887 connection after errors) when sending an RSET command.
888 Fix by Sunil Shetye. Report by James Moe.
889 * The IMAP client now uses "SEARCH UNSEEN" rather than "SEARCH UNSEEN NOT
890 DELETED" again on IMAP2, to fix a regression in fetchmail 6.2.5 reported by
891 Will Stringer in June 2004. (Sunil Shetye)
892 * The IMAP client now uses "SEARCH UNSEEN UNDELETED" on IMAP4 and IMAP4r1
893 servers (Sunil Shetye).
894 * Workaround: The IMAP client now falls back to "FETCH n:m FLAGS" if the server
895 does not support "SEARCH". (Sunil Shetye)
896 * The IMAP client now requests message numbers in batches of 1,000 to avoid
897 problems if there are more than 1860 unseen messages. (Sunil Shetye)
898 Note that this wasn't security relevant because fetchmail would only read up
899 to the maximum buffer size and leave the remainder of the string unread, going
900 out of synch afterwards.
901 * Stricter validation of IMAP responses containing byte or message counts.
904 * Only include gssapi.h if we're not including gssapi/gssapi.h, to fix a FreeBSD
905 compiler warning about gssapi.h being obsolete.
908 * The README.SSL document was revised for grammar, spelling, and clarity.
909 Courtesy of Robert Mullin.
911 # TRANSLATION UPDATES
912 * [it] Italian, by Vincenzo Campanella
916 fetchmail 6.3.13 (released 2009-10-30, 25333 LoC):
919 * The multiline SMTP error fix in release 6.3.12 caused fetchmail to lose
920 message codes 400..599 and treat all of these as temporary error. This would
921 cause messages to be left on the server even if softbounce was turned off.
922 Reported by Thomas Jarosch.
924 # TRANSLATION UPDATES
925 * [cs] Czech, by Petr Pisar
926 * [zh_CN] Chinese (simplified), by Ji ZhengYu
927 * [nl] Dutch, by Erwin Poeze
928 * [id] Indonesian, by Andhika Padmawan
929 * [ja] Japanese, by Takeshi Hamasaki
930 * [pl] Polish, by Jakub Bogusz
931 * [es] Spanish (Castilian), by Franciso Molinero
932 * [vi] Vietnamese, by Clytie Siddall
935 fetchmail 6.3.12 (released 2009-10-05):
938 * The CVE-2009-2666 fix in fetchmail release 6.3.11 caused a free() of
939 unallocated memory on SSL connections, which caused crashes or program aborts
940 on some systems (depending on how initialization and free() of unallocated
941 memory is handled in compiler and libc).
942 Workaround for older versions: run in verbose mode.
943 Patch courtesy of Thomas Heinz, fixes Gentoo Bug #280760.
944 This regression affected only the 6.3.11 release, but not the patch that was
945 part of the security announcement fetchmail-SA-2009-01.
948 * Fix error reporting for GSSAPI on Heimdal (h5l) Kerberos.
949 * Look for MD5_Init in libcrypto rather than libssl, fixes Gentoo Kerberos
950 builds; fixes upstream parts of Gentoo Bugs #231400 and #185652, and fixes
952 * Report multiline SMTP errors properly, reported by Earl Chew; fixes Debian Bug
953 #529899, reported by Akihiro Terasaki.
954 Note: This fix introduced a regression, fixed in 6.3.13.
955 * Replace control characters in SMTP replies by '?'.
956 * Fetchmailconf: Fix descriptions for smtpaddress and smtpname options;
957 smtpaddress is for RCPT TO, not MAIL FROM. Found by Gerard Seibert.
959 # TRANSLATION UPDATES AND ADDITIONS (ordered by language name):
960 * [ca] Catalan (Ernest Adrogué Calveras)
961 * [zh_CN] Chinese/Simplified (Ji ZhengYu)
962 * [cs] Czech (Petr Pisar)
963 * [ja] Japanese (Takeshi Hamasaki)
964 * [pl] Polish (Jakub Bogusz)
965 * [es] Spanish/Castilian (Francisco Molinero)
966 * [vi] Vietnamese (Clytie Siddall)
969 fetchmail 6.3.11 (released 2009-08-06):
972 * CVE-2009-2666: SSL NUL prefix impersonation attack through NULs in a
973 part of a X.509 certificate's CommonName and subjectAltName fields. These
974 fields use opaque strings with a separate length field, so that the NUL
975 character isn't a special character inside the certificate. Fetchmail, being
976 written in the C language, used to treat these strings as C strings
977 nonetheless, so that the domain comparison would end at the first embedded NUL
978 character, rather than at the real end of the string.
979 Fetchmail will now abort certificate verification as failed if NULs are
980 encountered inside either of these fields regardless of their position, and
981 drop the connection even if --sslcertck is not used, because NUL is not a
982 valid character in legitimate DNS names.
983 See fetchmail-SA-2009-01.txt for details, including a minimal patch.
986 * Remove the spurious message "message delimiter found while scanning headers".
987 RFC-5322 syntax states that the delimiter is part of the body, and the body is
989 * Convert all non-printable characters in certificate Subject/Issuer
990 Common Name or Subject Alternative Name fields to ANSI-C hex escapes (\xnn,
991 where nn are hex digits).
992 Note that this change introduces a regression, fixed in 6.3.12.
993 See the 6.3.12 documentation above for details and a workaround.
995 # TRANSLATION UPDATES AND ADDITIONS (ordered by language name):
996 * [zh_CN] Chinese/Simplified (Ji ZhengYu)
997 * [es] Spanish/Castilian (Francisco Molinero)
1000 fetchmail 6.3.10 (released 2009-07-02):
1002 # INCOMPATIBLE BUGFIXES AND CHANGES
1003 * Fetchmail no longer drops permanently undelivered messages by default, to
1004 match historic documentation. It does this by adding a new "softbounce"
1006 Fixes Debian Bug#471283, demotes Debian Bug#494418 to wishlist.
1007 * There is a new "softbounce" global option that prevents the deletion of
1008 messages that have not been forwarded. It defaults to "true" for fetchmail
1009 6.3.X in order to match historic documentation. This may change its default
1010 in the next major release.
1013 * Fix misuse of canonical autoconf target as _TARGET when it should have been
1014 _HOST. Report and patch courtesy of Diego E. "Flameeyes" Pettenò.
1015 Details: http://blog.flameeyes.eu/2009/01/01/the-canonical-target
1016 * Do not lose PS_MAXFETCH (13) exit status when hitting maxpoll. Reported by
1017 Michelle Konzack, Debian Bug#508667.
1018 * Do not overlap source and destination fields in snprintf() in interface.c.
1019 Courtesy of Nico Golde, Debian.
1020 * When a pre- or post-connect command fails, now report the exit status or
1021 termination signal properly through sys/wait.h macros.
1022 * When acquiring a body, understand NIL ("no such data item"), as returned by
1023 some MS Exchange versions. Fixes BerliOS Bug #11980 by KB Sriram.
1024 * Make progress tickers (-v/--showdots) consistent, and update documentation
1025 accordingly ("." for each 1024 octets read, "#" for a header written, and "*"
1026 for each body line written.)
1027 The conditions under which these had been printed were inconsistent,
1028 illogical, and documentation hadn't matched real behaviour for long.
1029 * For NTLM authentication, use dynamically allocated buffers.
1030 Fixes Debian Bug#449179, reported by Stepan Golosunov.
1031 * Non-delivery notice ("bounce mail") now mentions the original reason again,
1032 before the address list. This fixes a regression introduced in 6.3.0.
1033 * Several compiler warnings were fixed.
1034 * The minimum recommended SMTP (RFC-5321) timeouts are enforced to leave
1035 sufficient time for the listener to respond. Some synchronous listeners,
1036 particularly when used with spam filtering and other policy enforcement
1037 services, take extended amounts of time to process messages after the sender,
1038 recipient, or data block and EOM line. This can cause fetchmail to not wait
1039 long enough for the "250 Ok" and make fetchmail believe the message wasn't
1040 properly delivered when in fact it was; fetchmail would then retry the
1041 download next time and never make progress.
1042 Fixes Berlios Bug #10972, reported by Viktor Binzberger.
1043 * The ESMTP/LMTP client will now apply an application-specific timeout while
1044 waiting for the EHLO/LHLO response, rather than wait for the server or TCP
1046 * Treat 530 errors as temporary, so as not to delete messages on configuration
1047 errors. Partially taken from Petr Cerny's patch in Novell Bugzilla #246829.
1048 The 501 part of said patch was not added, as the maintainer is not convinced
1049 501 is a temporary condition, and softbounce takes care of this anyways.
1052 * Make the comparison of the SSL fingerprints case insensitive, to
1053 ease its use. Suggested by Daniel Richard G.
1054 * Proper precedence ordering for the syslog and logfile options. If the logfile
1055 option is effective (i. e. we're in daemon mode and nodetach isn't used),
1056 reset the syslog option. If logfile is ineffective (we're not in daemon mode,
1057 or nodetach is set), syslog takes precedence.
1058 * The sleeping at/awakened at messages appear in logfiles and syslog only if
1059 verbose mode is enabled. On the console, they will still appear without
1060 verbose mode. Fixes Debian Bug#282259.
1061 * fetchmail only requests IPv6 addresses via name service if at least one is
1062 configured on the local host, likewise for IPv4. (AI_ADDRCONFIG flag to
1063 getaddrinfo()) Extended version of Redhat's patch.
1064 * If the server name contains "yahoo.com", offers the "ID" capability, and we're
1065 polling via IMAP, send an ID ("guid" "1") transaction first, ignoring its
1066 result. This appears needed to be able to log into Yahoo's Zimbra servers, but
1067 there are open issues (such as being only able to download one message and
1068 server certificate mismatches).
1070 # CHANGES TO CONTRIB
1071 * Fix bashism in contrib/fetchsetup. Fixes Debian Bug#530081.
1074 * Some parts of the the manual page were revised for clarity, accuracy, and
1075 updated recommendations (particularly SSL/TLS) and formatting conventions from
1077 * The README and README.SSL documents were updated.
1078 * A document, README.SSL-SERVER, was added to describe server-side requirements
1079 for proper SSL and/or TLS service offerings. These are not specific to
1081 * Documentation on how to make "NOMAIL" (exit code 1) not treated an error has
1082 been added to the EXIT CODES section of the manpage and to the FAQ as item C8.
1083 The suggested solution uses a tiny POSIX shell script fragment.
1084 Fixes Debian Bug #530749, filed by Reuben Thomas.
1086 # TRANSLATION UPDATES AND ADDITIONS (ordered by language name):
1087 * [cs] Czech (Petr Pisar)
1088 * [en_GB] English/British
1090 * [id] Indonesian (Andhika Padmawan)
1091 * [it] Italian (Vincenzo Campanella)
1092 * [ja] Japanese (Takeshi Hamasaki)
1093 * [pl] Polish (Jakub Bogusz)
1094 * [ru] Russian (Pavel Maryanov), fixing Debian Bug #531925
1095 * [es] Spanish/Castilian (Francisco Molinero)
1096 * [zh_CN] Chinese/Simplified (Ji ZhengYu)
1099 fetchmail 6.3.9 (released 2008-11-16):
1101 # SECURITY AND CRITICAL BUG FIXES:
1102 * CVE-2007-4565: Denial of service: When fetchmail tries to inject a warning
1103 message it created itself, and the message is refused by the SMTP listener,
1104 fetchmail dereferences a NULL pointer and crashes. Report & fix by Earl Chew.
1105 Note while this is theoretically a remote denial of service attack vector,
1106 fetchmail by default talks SMTP to the localhost, so the overall risk is
1108 This bug was apparently introduced on 1998-11-27 when the bouncemail facility
1109 was modularized. The bug then made its appearance in fetchmail release 4.6.8.
1110 See also fetchmail-SA-2007-02.txt.
1111 * CVE-2008-2711: Denial of service: When fetchmail logs data blobs
1112 (for instance, a To: header in -v -v verbose mode) in excess of 2048
1113 bytes, it will crash, because it hands an uninitialized argument
1114 pointer (not the format string though) to vsnprintf and reads a
1115 random memory location (it calls va_arg() too often without
1116 resetting it with va_start()). Based on a patch (BerliOS patch #2492)
1117 by Petr Uzel, fixes Novell Bug #354291.
1118 Note 6.3.9-rc1 did not completely fix this issue, so it was redrawn a few
1119 hours after its release.
1120 See also fetchmail-SA-2008-01.txt.
1121 * When expunging, mark the right messages as seen to avoid message loss in "keep
1122 flush" configurations. Workaround for previous versions: "expunge 0".
1123 Report and patch by Alexander Cherepanov - thanks a lot, Berlios Bug #11797,
1124 "imap_mark_seen doesn't consider expunged messages".
1125 * SSL fix: close memory leak when SSL connection fails; fetchmail used to forget
1126 calling SSL_free() on the SSL context, leaking in excess of 500 kB RAM on a
1127 x86_64 system per failed SSL connection attempt.
1128 Bug reported and patch provided by Seiichi Ikarashi, Fujitsu.
1131 * The configure script will additionally check for 'dn_skipname', to fix build
1132 failures with µClibc. The new check still recognizes the resolver libraries on
1133 Ubuntu 7.04, openSUSE 10.2, Solaris 8, NetBSD 4.0_BETA2 and FreeBSD 6.2.
1134 Fixes Gentoo bug #134187.
1135 NOTE: this is a bit of a hack, since we twist the HAVE_RES_SEARCH result, but
1136 res_search() and dn_skipname() are only used together and scheduled for
1137 removal in future versions, so this is probably fine.
1138 * No longer complain about invalid sslproto "" when POP3 CAPA probe fails.
1139 Fixes Debian Bug#421446 (Holger Leskien), Novell Bug #247233 (Jon Nelson),
1141 Thanks to Matthias Strauß for a configuration to reproduce the issue.
1142 * Allow .fetchmailrc and .fetchids to be symlinks, as the manpage does not
1143 document they aren't allowed - fixes Debian Bug #452907 (Roger Leigh).
1144 TOCTOU race persists.
1145 * fetchmailconf quotes mailbox (folder) names when writing the configuration.
1146 Fixes BerliOS Bug #13207 (reported + fix suggested by Terry Brown).
1147 * Only print "Deleting fetchids file" if there actually is one.
1148 Fixes Debian Bug#374514, reported by Dan Jacobson.
1149 * SSL fix: check and report if SSL_set_fd fails.
1152 * autoconf 2.60 is now required to build fetchmail; it uses
1153 AC_USE_SYSTEM_EXTENSIONS to replace AC_AIX, AC_MINIX, and the like.
1154 * Removed dead FETCHMAIL_DEBUG code from fetchmail.h that was disabled by
1155 default with no switches in configure to enable it. However, the macro would
1156 have been prone to a symlink attack. Found by Nico Golde.
1157 * Removed dead FORCE_STUFFING code from socket.c that was disabled by default
1158 with no switches in configure to enable it.
1159 * Include the typedef for int16 in the #ifndef _AIX in smbencrypt.c (Peter
1161 * Correct check for u_int32_t in configure.ac (seems to be typedef'ed in
1162 namser.h on some platforms.) (Peter O'Gorman)
1163 * In configure.ac change all CPFLAGS to CPPFLAGS, CEFLAGS to CFLAGS and LDEFLAGS
1164 to LDFLAGS otherwise the results of some tests (additional -L and -I flags) do
1165 not get used for later tests causing incorrect configure results. Makefile.am
1166 was also changed to reflect this. (Peter O'Gorman)
1167 * m4/gethostbyname_r.m4 does AC_TRY_COMPILE, which unfortunately can pass even
1168 if there is no gethostbyname_r. Changed to AC_TRY_LINK. (Peter O'Gorman)
1169 * Revise getnameinfo check to ensure NULL is defined and the result is properly
1170 evaluated, to avoid bogus results on for instance FreeBSD and redefinitions of
1171 NI_* at compile time. (Matthias Andree).
1172 * __attribute__ ((unused)) is a gccism, removed from libesmtp/gethostbyname.c.
1174 * In KAME/getnameinfo.c it's best to use the correct argument to inet_ntoa.
1176 * In verbose mode, log if --check mode is enabled.
1177 * Add sslcommonname option (rcfile and commandline) as a way to work around
1178 misconfigured upstream SSL servers that use the wrong certificate name. It
1179 specifies which CommonName fetchmail expects and logs. (Daniel Richard G.)
1180 * Changed CRLF to LF line endings in contrib/delete-later (reporter: Petr Uzel)
1181 * SSL change: enable all workarounds with SSL_CTX_set_options(ctx,SSL_OP_ALL)
1182 * All translations have been re-enabled, in an attempt to rekindle translator or
1186 * Add fetchmail-SA-2007-02.txt and fetchmail-SA-2008-01.txt.
1187 * Re-add two lines to the manual page that had accidentally become comments
1188 to nroff. One was part of the --sslproto documentation, and one in the
1189 "Awakening the background daemon" section.
1190 * The manual page no longer asserts that .fetchids were for exclusive POP3 use,
1191 since it is planned to use the file with IMAP4 later.
1192 * Add grammar fixes from Dan Jacobson to fetchmail.man. Debian Bug #461642.
1193 * The manual page now mentions that user descriptions need to come before user
1194 options. Reported by Francensco Pontortì, to fix Debian Bug #467010.
1195 * The manual page no longer hints that multi-user declarations per server were
1196 only useful in daemon mode running as root, to avoid hinting people to doing
1198 * Several manual page rcfile examples now include "ssl".
1199 * The manual page hints that option arguments beginning with numbers can be
1201 * The manual page now mentions that the --logfile must already exist before
1203 * The FAQ now recommends (#I9) not to use Google Mail for their disregard to the
1204 protocols they claim to support.
1205 * Documentation and program output now /consistently/ claim that the rcfile must
1206 not have more than 0700 (u=rwx,g=,o=) permissions, but fetchmail will still
1207 silently accept additional g=x permissions for compatibility with previous
1208 6.2.X and 6.3.X versions.
1209 Inconsistency (program 0710, manpage 0600) reported by Petr Uzel.
1210 * The --logfile documentation is now clearer about requiring detached daemon
1213 # TRANSLATION UPDATES AND ADDITIONS (ordered by language name):
1214 * [sq] Albanian (Besnik Bleta)
1215 * [zh_CN] Chinese, simplified (Ji Zheng-Yu)
1216 * [cs] Czech (Petr Pisar)
1217 * [da] Danish (Byrial Ole Jensen) - outdated, but newer than in 6.3.8
1218 * [nl] Dutch (Tony Vroon, Benno Schulenberg)
1219 * [en_GB] English, British
1220 * [fi] Finnish (Lauri Nurmi)
1222 * [id] Indonesian (Andhika Padmawan)
1223 * [ja] Japanese (Takeshi Hamasaki)
1224 * [pl] Polish (Jakub Bogusz)
1225 * [ru] Russian (Pavel Maryanov)
1226 * [es] Spanish (Javier Fernández-Sanguino Peña, Matthias Andree)
1227 * [tr] Turkish (Engin Gündüz) - outdated, but newer than in 6.3.8
1228 * [vi] Vietnamese (Clytie Siddall)
1232 fetchmail 6.3.8 (released 2007-04-06):
1234 # SECURITY STRENGTHENING:
1235 * Make the APOP challenge parser more distrustful and have it reject challenges
1236 that do not conform to RFC-822 msg-id format, in the hope to make mounting
1237 man-in-the-middle attacks (MITM) against APOP a bit more difficult.
1238 (CVE-2007-1558, reported by Gaëtan Leurent, published 2007-04-02 on Bugtraq)
1240 APOP is claimed insecure by Gaëtan Leurent for MITM scenarios for typical
1241 setups: based on MD5 collisions, it is purportedly possible to recover the
1242 first three characters of the shared secret (password), which would then make
1243 recovery of the shared secret a matter of hours or minutes; this would then
1244 enable the attacker to impersonate the client vis-à-vis the server.
1246 For further details, check
1247 * Gaëtan Leurent, "Message Freedom in MD4 and MD5 Collisions: Application
1248 to APOP", Fast Software Encryption 2007, Luxembourg. (Proceedings to appear in
1249 Springer's Lecture Notes on Computer Science.)
1250 * The mailing list discussion thread at
1251 <http://lists.berlios.de/pipermail/fetchmail-devel/2007-March/000887.html>
1254 * Fix pluralization of oversized-message warning mails.
1255 * Fix manual page: --sslcheck -> --sslcertck, and do not set trailing
1256 "recommended:" in bold. Fixes Debian Bug #413059, reported by Rafal Czlonka.
1257 * Repoll immediately if a protocol error happens during the authentication
1258 attempt after a failed opportunistic TLS upgrade.
1259 Fixes comment #9 in Gentoo Bug #163782, reported by Takuto Matsuu.
1260 * Fix rendering of the "24 - 26, 28, 29" paragraph in the exit codes section.
1261 Reported by Nico Golde.
1262 * If SOCKS support was compiled in, add 'socks' to the feature_options Python
1263 list emitted in --configdump. Reported by Rob MacGregor.
1264 * Do not crash with a null pointer dereference when opening the BSMTP file
1265 fails. Improve error checking and reporting. Reported by Reto Schüttel,
1266 Debian Bug#416625. Fix based on a patch by Nico Golde.
1267 * Make BSMTP output actually work, it would persistently fail with SOCKET error
1268 after writing the first header. Bug independently found and reported in
1269 excellent detail by Reto Schüttel, Debian Bug#416812.
1272 * Add fetchmail-SA-2007-01.txt
1273 * Extend --mda documentation, discourage use of qmail-inject.
1274 Based on a patch by Rob MacGregor.
1275 * Document SOCKS configuration facility (SOCKS_CONF environment variable).
1276 Thanks to Jochen Hayek, Michael Shuldman and Rob MacGregor.
1277 * Use envelope option in multidrop example. Patch by Rob MacGregor.
1278 * Document expected Received: line format when parsing for envelope addressees.
1279 * Stripped option documentation from sample.rcfile, since this is bound to go
1280 out of synch with the manual page, which is the only reference on options.
1281 * Mention that --limit default is 0 bytes, which is special for "no limit".
1282 * Corrected Robert M. Funk's name that I misspelled. My sincere apologies
1286 * Add delete-later and delete-later.README, a script and documentation for
1287 a MySQL/Tcl-based client-side "delete-after" feature.
1288 Kindly donated by Yoo GmbH, Großvoigtsberg, Germany (Carsten Ralle).
1291 fetchmail 6.3.7 (released 2007-02-18):
1293 # FIXES FOR REGRESSIONS IN 6.3.6
1294 * Fix KPOP. Patch by Miloslav Trmac.
1295 * Fix repoll when server disconnects after opportunistic TLS failed for POP3.
1296 Berlios Bug #10133 = Gentoo Bug #163782 reported by Andrej Kacian.
1298 # TRANSLATION UPDATES
1299 * Japanese (Takeshi Hamasaki), Polish (Jakub Bogusz)
1302 * Consider getaddrinfo() on Darwin 9 (Mac OS X 10.5 "Leopard") thread-safe.
1303 Reported by Uli Zappe.
1306 fetchmail 6.3.6 (released 2007-01-04):
1309 * CVE-2006-5867, fetchmail-SA-2006-02.txt:
1310 Password disclosure vulnerability fixed. This has several aspects:
1312 - Fetchmail now implies sslproto 'tls1' if the sslfingerprint or sslcertck
1313 options are used and the ssl option is not used, in order to be sure that
1314 fetchmail gets a certificate from the mail server.
1316 - Fetchmail breaks the connection if the TLS negotiation (or verification, if
1317 requested) fails with sslproto 'tls1', sslfingerprint or sslcheck enabled.
1319 - POP3 connections now use STLS reliably. They used to ignore STLS altogether
1320 for serveral values of the "auth" option, when fetchmail forget to probe
1321 server capabilities - see fetchmail-SA-2006-02.txt for details.
1323 - POP3 connections will no longer fall back USER/PASS authentication if
1324 strong challenge-response authenticators such as CRAM-MD5 are configured
1325 but the server does not advertise these in its CAPA response.
1327 - POP2 is obsolete and does not support STLS or anything beyond password-based
1328 authentication. The attempt to use STLS or strong authenticators now causes
1331 Configurations using both ssl and sslcertck however have been semi-safe in
1332 that they would send the password in the clear. The USER/PASS fallback
1333 problem however applies to these too, so that the password was only safe on
1334 trustworthy servers.
1336 * CVE-2006-5974, fetchmail-SA-2006-03.txt:
1337 Repairs a regression in 6.3.5 that crashes fetchmail when a message with
1338 invalid headers is found while fetchmail's mda option is in use. BerliOS bugs
1339 #9364, #9412, #9449. Stack backtrace provided by Neil Hoggarth - thanks.
1341 # REGRESSION FIXES (recently introduced bugs)
1342 * Repair --logfile, broken in 6.3.5. BerliOS Bug #9059,
1343 reported by Brian Harring.
1344 * Repair --user, broken in 6.3.5 (as a side effect of the authenticate external
1345 patch): using SSL certificate/key authentication overrode the --user option.
1346 Now the latter takes precedence, and only defaults to the certificate's common
1347 name. Debian Bug #400950, reported by Jorgen Schaefer <forcer@debian.org>.
1349 # BUG FIXES (long-standing bugs):
1350 * RPOP: used to log the password locally rather than an asterisk as the other
1351 protocols do. The password is now shrouded in the local logs.
1352 * POP3: Probes capabilities now when Kerberos V5 is enabled, so that we can
1353 actually detect if the server supports it.
1354 * Robustness: If a stale lockfile cannot be deleted, truncate it so that
1355 fetchmail doesn't later believe itself to be running if the PID is recycled
1356 by a non-fetchmail process.
1357 * DNS: Detect /etc/resolv.conf changes: On systems that have res_search(),
1358 assume we also have res_init() and call it (suggested by Ulrich Drepper,
1359 glibc bug #3675) in order to make libc or libresolv reread the resolver
1360 configuration at the beginning of a poll cycle. This is important when
1361 fetchmail is in daemon mode and /etc/resolv.conf is changed later by dhcpcd,
1362 dhclient, pppd, openvpn or other ip-up/ipchange scripts. Should fix Debian
1363 Bug#389270, Bug#391698.
1364 * Robustness: Fix crash on systems that do not provide strdup(), the crash
1365 happens only in out-of-memory conditions when fetchmail cannot proceed
1366 anyways. Patch by Andreas Krennmair.
1367 * Robustness: When HOME and FETCHMAILHOME are unset, be sure to copy user
1368 database information, so it is not trashed later. Patch by Jim Correia.
1371 * Workaround: Improve handling of IMAP IDLE, some servers do not reset their
1372 time counters after sending information asynchronously. Patch by Sunil
1373 Shetye, after report from Andrew Baumann.
1374 * Usability: When requesting Kerberos or GSSAPI, complain and exit with syntax
1375 error if any of these requested features has not been compiled in. This is
1376 to fail early and with precise error message. Reported by Isaac Wilcox.
1377 * --version will now add +KRB4 or +KRB5 if Kerberos v4 or v5, respectively, have
1378 been compiled in. Reported missing by Isaac Wilcox.
1381 * New en_GB (British English) translation by David Lodge.
1382 * Update Japanese (Takeshi Hamasaki), Polish (Jakub Bogusz), Russian (Pavel
1383 Maryanov) and Vietnamese (Clytie Siddall) translations.
1384 ! Note that not all these translations are complete -- this isn't the
1385 translators' fault though, but due to delays at the BerliOS hosting site and
1386 the translation project handlers. You may see a few untranslated messages.
1389 * Dropped exit status 15 from manual page, it's not used by fetchmail.
1390 Reported by Isaac Wilcox.
1391 * Documented exit codes 24 - 29 as internal.
1393 fetchmail 6.3.5 (released 2006-10-09):
1396 * For protocols such as IMAP that are not delimited by "." lines, truncate the
1397 input buffer when the message has been completely read, to avoid taking
1398 trailing garbage into the message if the terminal CRLF is missing. Fixes
1399 Debian Bug#312415. (Patch suggested by Mike Jones, Manchester Univ.).
1400 * When using NTLM authentication, use regular IMAP response code handler after
1401 completing NTLM handshake, for robustness and consistency.
1402 (Taken from the NetBSD portable packages collection, patch-ac.)
1403 * Support Kerberos installations where krb5.h and perhaps roken.h are in
1404 .../include/krb5. Taken from NetBSD portable packages collection patch-ae.
1405 * On NetBSD, link against -lroken -lcom_err if --with-kerberos is enabled.
1406 * Drop #include <com_err.h> from Kerberos 5 header file, fixes compile error on
1408 * Fix des_pcbc_encrypt compile warnings in kerberos.c line 246.
1409 * If krb5-config provides gssapi library information, use that rather than
1411 * Improve --with-gssapi auto detection for /usr-based GSSAPI installs.
1412 * Fix --with-gssapi builds for NetBSD 3.0.
1413 * Improve KAME/getnameinfo.c portability to Linux libc5 systems.
1414 Based on a patch by Dan Fandrich.
1415 * Provide INET6 to KAME/getnameinfo.c (only useful on IPv6-enabled systems that
1416 lack getnameinfo, and there only visible in some Received: headers).
1417 Found by Dan Fandrich.
1418 * POP3: some UID flags may not be set properly on UIDL lists. (Sunil Shetye)
1419 * Make IMAP4 IDLE work on servers that do not update RECENT counts.
1420 Reported by Lars Tewes.
1421 * IMAP4 patch by Sunil Shetye:
1422 - do not depend on server updating RECENT counts at all
1423 - also enter IDLE loop when messages are present on the server.
1424 * Fix --flush description in the manual page, fetchmail does not mark messages
1425 seen unless it has successfully delivered them. Suggested by Frederic Marchal.
1426 * Fetchmail no longer attempts to stat the "-" file in daemon mode -- this is a
1427 special name to read the RC file from stdin, and cannot always be re-read
1428 anyways. BerliOS bug #7858.
1429 * When looking up ports for a service, the lookup succeeds and the returned
1430 address family isn't IPv4 or IPv6, properly free the allocated memory from the
1431 service lookup. Found by Uli Zappe.
1432 * When looking up ports for a service, only look up TCP ports.
1433 * Avoid compiling empty files, to avoid diagnostics from strict compilers.
1434 * If the lockfile ends before the process ID, treat it as stale and unlink it.
1435 Reported by Justin Pryzby, Debian Bug #376603.
1436 * SIGHUP wake-up behavior was broken since 5.9.13's Cygwin changes, in that for
1437 non-root users, SIGHUP would abort the first poll and subsequently interfere
1438 with new polls, and SIGHUP would be ignored for root users. SIGHUP now matches
1439 documented behavior. SIGUSR1 has always been a wakeup signal for both root
1440 (undocumented) and non-root users. See also the deprecation warning above.
1441 * Track getaddrinfo() results to properly free them after timeouts and make sure
1442 that getaddrinfo() isn't interrupted by a timeout (which breaks on MacOS X),
1443 reported by Uli Zappe. This should fix Debian Bug#294547 and Bug#377135.
1444 * --logfile is now handled more carefully, errors opening the logfile are
1445 now reported to the TTY where fetchmail was started from.
1446 * fetchmail now complains and aborts when it cannot properly daemonize itself.
1447 * fix compilation on systems that don't know struct addrinfo (Solaris 2.6).
1448 * ignore SIGPIPE signals and rely on functions to return EPIPE instead. This is
1449 necessary because the former longjmp() from the signal handler is unsafe and
1450 makes the whole fetchmail behavior undefined after the event.
1451 * Avoid crash in env.c/host_fqdn if we cannot canonicalize our own hostname.
1452 Reported by Alexander Holler.
1453 * SSL fix by Miloslav Trmac (Red Hat): free the SSL contexts after the
1454 connection, to avoid from growing SSL certpaths without bounds, avoid using
1455 SSL contexts for unrelated connections, and to fix Red Hat Bug #206346.
1458 * Rename all fetchmail-internal lock_* functions to fm_lock_*. Obsoletes
1459 NetBSD portable packages collection patch-ah, patch-ai and patch-aj.
1460 * Configure prints a warning (but proceeds) if Kerberos IV support is enabled.
1461 * In verbose mode, log every IP fetchmail tries to connect to, to avoid
1462 misleading the user. Suppress EAFNOSUPPORT errors from socket() call, too.
1463 Fixes Debian Bug #361825, reported by Daniel Baur.
1464 * In idle mode, fetchmail complains about the fetchall option.
1465 * When a connection fails, log not only the IP address, but also host and
1466 service name and the port number. Log the latter when trying to connect in
1468 * Keep syslog output at one line per message (this works if no errors occur).
1469 * Fetchmail in verbose mode now logs if it opportunistically upgrades a POP3
1470 or IMAP connection to TLS security with STLS/STARTTLS.
1471 * fetchmail now supports foo@example.org=bar user mappings for multidrop boxes.
1472 * switch setjmp/longjmp to sigsetjmp/siglongjmp
1473 * IMAP now supports the EXTERNAL authentication method, courtesy of
1474 Götz 'nimrill' Babin-Ebell, BerliOS patch #1095 with minor changes.
1475 Note that this change causes --sslcert to override --user.
1476 * The sslproto keywords are now case insensitive, courtesy of
1477 Götz 'nimrill' Babin-Ebell, BerliOS patch #1095.
1478 * When going to sleep, log for how long. Suggested by Claudia Ludwig.
1479 * When the server name cannot be canonicalized, log the gai_strerror value.
1481 # TRANSLATION UPDATES:
1482 * Catalan/ca (Ernest Adrogué Calveras), Japanese/ja (Takeshi Hamasaki) - also
1483 made gettext 0.15 ready, Polish/pl (Jakub Bogusz), Russian/ru (Pavel
1484 Maryanov), Spanish/es (Héctor García Álvarez), Vietnamese/vi (Clytie Siddall)
1486 # CONTRIBUTED SCRIPTS:
1487 * PopDel.py was revised by Joshua Crawford to display the From: address and
1488 list every email, even if it has no Subject: header; and not delete the wrong
1489 message in the presence of mail without Subject: headers.
1491 fetchmail 6.3.4 (released 2006-04-14):
1494 * configure: detect res_* functions properly with newer glibc ABIs.
1495 Patch by Miloslav Trmac.
1496 * tracepolls: add folder information if available. Reported by Terry Brown.
1497 * lexer: add %option noyywrap to avoid link errors about missing yywrap().
1498 * a few more type fixes for report/snprintf, patch by Miloslav Trmac.
1499 * bouncing: fetchmail would still send "General SMTP/ESMTP error." bounces
1500 in spite of "no bouncemail" configuration.
1501 * SSL/TLS: if, for a certain server, an sslfingerprint is specified and
1502 sslcertck is NOT set, suppress printing SSL certificate mismatch errors.
1503 (Reported by Hannes Erven.)
1504 * SSL/TLS: always print if the sslfingerprint mismatches, even in silent
1505 mode. (This is for consistency with certificate verification errors.)
1507 # TRANSLATION UPDATES:
1508 * German/de (Matthias Andree), French/fr (Matthias Andree), Spanish/es (Héctor
1509 García), Polish/pl (Jakub Bogusz), Japanese/ja (Takeshi Hamasaki)
1510 * New Vietnamese/vi translation (Clytie Siddall).
1511 * Updated French descriptions for the .spec file (Stéphane Schildknecht,
1512 Luc Pionchon, Matthias Andree).
1515 * pidfile: there is a new command-line (--pidfile PATH) and global option for
1516 the rcfile (set pidfile [=] "/path/to/pidfile") option to allow overriding
1517 the default location of the PID file.
1518 Requested by Héctor García, Debian maintainer.
1519 * specgen.sh: Converted to UTF-8 to support translated texts better.
1521 fetchmail 6.3.3 (released 2006-03-30):
1524 * SEGFAULT: Do not attempt to overwrite the netrc password if none has been
1525 specified. This fixes a segmentation fault bug introduced into 6.3.2.
1526 Fixes BerliOS bug #6234. BerliOS patch #804 by Craig Leres.
1527 The patch, as accepted into fetchmail, was available separately from
1528 <http://download.berlios.de/fetchmail/patch-6.3.2.1-fix-netrc-SIGSEGV.diff>
1529 * SEGFAULT: Work around C libraries that return a NULL in getaddrinfo()'s
1530 ai_canonname record, to avoid a segfault. Affects for instance FreeBSD 4.10,
1531 4.11 and 5.3 when dotted quads are given as server names.
1532 Analysis and fix by Vladimir Olegovich Ravodin (Владимир Олегович Раводин).
1533 * IMAP: fix hangs in NOOP-based IDLE emulation. Reported by Casper Gripenberg
1534 and Brendan Lynch, fix by Sunil Shetye (his patch was merged) and Brendan Lynch.
1535 * IMAP: Handle other clients concurrently accessing IMAP mailboxes better.
1536 Fetchmail quits the poll if the EXPUNGE count does not match expectations, and
1537 servers not updating RECENT counts after EXPUNGE are handled in a better way.
1538 (Patch by Sunil Shetye.)
1539 * IMAP: Stop sending EXPUNGE after NOOP-idling (patch by Sunil Shetye).
1540 * POP3: fetchmail can now use UIDL in fetchall keep mode, to avoid re-fetching
1541 the same messages again when the fetchall keyword is removed. Patch by
1542 Sunil Shetye. For details, please see
1543 <http://lists.berlios.de/pipermail/fetchmail-users/2006-March/000308.html>
1544 * LMTP: fix bug in LMTP port validation (patch by Miloslav Trmac).
1545 * SDPS: fetchmail no longer replaces the local user ID for an empty envelope
1546 sender when using the proprietary SDPS extension for POP3.
1547 Fixes Debian Bug#353575, reported by Roger Lynn.
1548 * SDPS: Warn and disable SDPS if POP3 is disabled to avoid compilation errors.
1549 * fetchmail no longer prints empty lines in verbose mode when using syslog.
1550 * fetchmail no longer prints UID lists in verbose mode when using syslog.
1551 * ./configure --quiet is now quieter (no SSL and fallback-related output).
1552 * Miloslav Trmac's patch (with minor changes) to fix char * sign consistency,
1553 unused arguments and variables.
1554 * More signedness, unused argument/variable and other warning fixes.
1557 * --idle can now be specified on the command line, too.
1558 * --fetchall is now supported on the command-line.
1559 * POP3: Lower default fastuidl span to 4 (i. e. every 4th run fetches the
1560 whole UIDL list), patch by Sunil Shetye.
1563 * "ssl" is a user option rather than a server option. Patch by Nico Golde.
1564 Fixes Debian Bug#354661, reported by Keith Hellman.
1565 * The manual page now suggests "--" before the addresses in the sendmail MDA
1566 example, for safety.
1567 * The FAQ item X9, Domino IMAP omits Content-Transfer-Encoding header, was
1568 added. Information provided by Anthony Kim on the fetchmail-friends list
1570 * Credit Chris Boyle with the NOOP emulation code for IDLE in fetchmail 6.2.4.
1571 Eric forgot to credit Chris, thanks to Sunil Shetye for providing these links:
1572 http://lists.ccil.org/pipermail/fetchmail-friends/2003-July/007705.html
1573 http://lists.ccil.org/pipermail/fetchmail-friends/2003-July/007713.html
1574 * Added a section about RETR vs. TOP to the manual page.
1575 * Changed section/subsection levels in some areas.
1577 fetchmail 6.3.2 (released 2006-01-22):
1579 Unless otherwise noted, changes to this release were made by Matthias Andree.
1581 # SECURITY FIX IN THIS RELEASE
1582 * CVE-2006-0321: Fix segfault or bus error after bouncing a message. This bug
1583 was introduced into 6.3.0 when removing alloca(); it caused fetchmail to free
1584 random memory. Reported by Nathaniel W. Turner, Debian Bug#348747.
1585 See fetchmail-SA-2006-01.txt
1587 # INCOMPATIBLE CHANGE:
1588 * Automatically disable the POP3 TOP command if the greeting string contains
1589 "Maillennium POP3/PROXY server", which is used by comcast and known to
1590 truncate messages after 80 kByte. Fall back to RETR, and complain if we had
1591 used TOP otherwise (the warning is printed only once per server in daemon
1592 mode). Suggested by Ed Wilts.
1593 *Note* that this means messages are marked read on these servers, which is a
1594 deviation from how 6.3.1 behaved, but we have no alternative, comcast haven't
1595 fixed this bug in years. Preventing the loss of the remainder of the message
1596 justifies this incompatible fix.
1597 * fetchmail, since 6.3.0, requires write permission to the directory holding the
1598 idfile. See the amendment in the 6.3.0 MAJOR INCOMPATIBLE CHANGES section
1599 below for details. The manual page was updated.
1601 # CHANGES RELEVANT TO PACKAGERS:
1602 * The outdated BUGS document was removed from the distribution.
1603 * Added fetchmail-SA-2006-01.txt to the distribution.
1606 * SMTP/LMTP cleanup to fix these two bugs:
1607 - switch back to SMTP after having tried LMTP hosts (multiple smtphost hosts)
1608 - switch back to LMTP after sending a bounce.
1609 The patch removes the global state variable that was the root of this problem.
1610 Patch by Sunil Shetye. (MA)
1611 * Don't complain about fetchall keep in --configdump mode. Bug introduced in
1613 * fetchmailconf.py: Fix novice help for Poll interval and fetchall.
1614 Reported by Justin Pryzby, Debian Bug #344978.
1615 * Some verbose output disappeared in debug mode. Adding further -v options would
1616 alternate between verbose and debug mode. debug mode now comprises all verbose
1617 output, and adding more -v options does not switch back from debug to verbose
1619 * fetchmail.man: Fix accented characters in Héctor García's name. Merged from
1620 downstream debian/patches/01_man_page.dpatch.
1621 * Add missing --help text for "--sslcertck" option.
1622 * fetchmailconf.py: Accept --help and --version.
1623 * fetchmail --version now prints the copyright notice.
1624 * don't complain about READ-ONLY IMAP folders in --fetchall --keep mode.
1625 Reported Alexander Zangerl, Debian Bug#348964.
1626 * the RPM .spec file now generates a -debuginfo package on newer RPM versions.
1628 fetchmail 6.3.1 (released 2005-12-19):
1630 # SECURITY FIX IN THIS RELEASE
1631 * CVE-2005-4348 Fix segmentation fault (null pointer dereference) in
1632 multidrop mode with headerless email. See fetchmail-SA-2005-03.txt.
1633 Reported by Daniel Drake, patch by Sunil Shetye. (MA)
1635 # OTHER BUG FIXES, DOCUMENTATION AND TRANSLATION UPDATES
1636 * Fix broken default port in POP2. Patch by Stanislav Brabec, SUSE [CZ]. (MA)
1637 * Fix manual page, some lines starting with ' were escaped by \&.
1638 Reported by Simon Barner. (MA)
1639 * Ship with gettext-0.14.3 again, as 6.2.9-rc10 did. Found by Sunil Shetye. (MA)
1640 * Actually set default SSL certificate path if --sslcertpath is unset.
1641 Reported by Heino Tiedemann and Rob MacGregor. (MA)
1642 * Remove bogus Netscape IMAP4rev1 Service >= 3.6 warning about BODY[TEXT]
1643 that we are not using. Patch by Sunil Shetye. (MA)
1644 * Plug potential memory and socket leak when polling multiple folders or when
1645 the upstream sends bogus message sizes. Patch by Sunil Shetye. (MA)
1646 * Update Catalan translation, by Ernest Adrogué Calveras. (MA)
1647 * Fix segfault (null pointer dereference) on some operating systems with
1648 fetchmail's obsolete DNS MX/host alias lookups in multidrop mode.
1649 Patch by Dr.-Ing. Andreas Haakh. (MA)
1650 * Close SMTP sockets early, to reduce resource usage, trigger earlier delivery
1651 with some MTAs and avoid SIGPIPE (SIG 13) when the SMTP listener gets bored
1652 and drops the connection after timeout. Patch by Sunil Shetye. (MA)
1653 * Don't treat hitting a fetch limit as error. Patch by Sunil Shetye. (MA)
1654 * Fix negative "messages left on server" on idle/repoll with fetchlimit.
1655 Patch by Sunil Shetye. (MA)
1656 * Properly track logout stage. Patch by Sunil Shetye. (MA)
1657 * Preserve error conditions across postconnect script. Sunil Shetye. (MA)
1658 * Do not trash destination domain if multiple messages are forwarded into the
1659 same SMTP/LMTP connection. Reported by Joachim Feise, Berlios Bug #5849. (MA)
1660 * Manual page: Add "-md5" to "openssl x509" example in --sslfingerprint
1661 documentation, since OpenSSL 0.9.8 changed the default to SHA1.
1662 Suggested by Jason White. (MA)
1663 * Cope with servers that return UID information in response to non-UID
1664 RFC822.{SIZE|HEADER} requests. Reported by Jason White.
1665 Patch suggestion by by Sunil Shetye, simplified by MA.
1667 fetchmail 6.3.0 (released 2005-11-30):
1669 # SECURITY FIXES IN THIS RELEASE
1670 * CVE-2005-2335: The POP3 UIDL code doesn't sufficiently validate/truncate the
1671 input length, so a (malicious or compromised) server that sends UIDs longer
1672 than 128 bytes can corrupt fetchmail's stack and crash fetchmail.
1673 This vulnerability is remotely exploitable to inject code run in a
1674 root shell. Edward J. Shornock, Ludwig Nussel. fetchmail-SA-2005-01.txt
1675 * CVE-2005-3088: fetchmailconf now changes the output file to mode 0600 BEFORE
1676 writing to it, so there is no window where passwords could be read by the
1677 world. Matthias Andree. fetchmail-SA-2005-02.txt
1679 # MAJOR INCOMPATIBLE CHANGES
1680 * Remove support for --netsec/-T options, the required inet6_apps library is no
1682 http://www.inner.net/pub/ipv6/ states, as of 2005-07-03: "/pub/ipv6
1683 Our IPv6 software is now long defunct. Please find a more modern source."
1684 I haven't been able to find a more modern source. Matthias Andree
1685 * Operating systems that do not conform to the Single Unix Specification v2
1686 (1997) or v3 (2001, aka IEEE Std 1003.1-2001) are no longer supported. They
1687 may continue to work and non-intrusive patches to support them may be
1688 accepted. Matthias Andree
1689 * The default for --smtphost is now always "localhost" regardless of
1690 authentication types and protocols, so as to simplify configurations for
1691 workstations where the SMTP daemon only listens on the loopback interface.
1692 Sunil Shetye & Matthias Andree
1693 Amendment, 2006-01-04:
1694 * fetchmail's idfile (.fetchids) is no longer written directly, but the ids are
1695 written to a temporary file which is renamed into place after being written
1696 completely. This is to avoid writing incomplete idfiles when running out of
1697 space, which would cause excessive duplicate refetches of messages, this might
1698 make matters even worse. This means that fetchmail requires write permission
1699 on the directory holding the idfile. This will usually affect system-global
1700 daemons only, for instance, Debian. Found by Dan Jacobson. Matthias Andree.
1701 Escalated to "incompatible", 2006-01-13:
1702 * Try to obtain FQDN as our own host by default, rather than using "localhost".
1703 If hostname cannot be qualified, complain noisily and continue, unless
1704 Kerberos, ODMR or ETRN are used (these have always required an FQDN).
1705 Partial fix of Debian Bug#150137. Fixes Debian Bug#316454. Matthias Andree
1707 # CHANGES RELEVANT TO PACKAGERS AND USERS
1708 * fetchmailconf is now a shell wrapper that calls the byte-compiled
1709 fetchmailconf.py script, which is now installed in the regular python
1710 directory. Matthias Andree.
1711 * The --enable-inet6 configure option was removed. The code is mostly protocol
1712 agnostic, a fully IPv6 aware OS is expected to provide getaddrinfo(),
1713 getnameinfo() and the macro AF_INET6. Matthias Andree.
1714 * gettext (intl/) has been removed from the fetchmail package. Install GNU
1715 gettext 0.14 separately for NLS (i18n). Matthias Andree
1716 * Added Russian translation, courtesy of Pavel Maryanov of the
1717 Russian translation team. (MA)
1718 * Updated and re-enabled Czech translation, by Miloslav Trmac (MA).
1719 * Dropped da=Danish, el=Greek and tr=Turkish translations which have more than
1720 10% (61+) untranslated or fuzzy messages. Matthias Andree.
1722 # OTHER USER-VISIBLE CHANGES
1723 * Sunil Shetye's fix to force fetchsizelimit to 1 for APOP and RPOP. (ESR)
1724 * PopDel.py removed from contrib at author's request. (ESR)
1725 * Matthias Andree's fix for Sunil Shetye's fetch-split patch. (ESR)
1726 * Include James Stone's moldremover.py script. (ESR)
1727 * Enable .fetchmailrc permissions checking under Cygwin. (ESR)
1728 * Nalin Dahyabai's fix for POP3 strong authentication. (ESR)
1729 * Revised Nalin Dahyabai's fix for POP3 strong authentication (the
1730 original version would go into an infinite loop when CAPA failed;
1731 found by David Greaves.) (MA)
1732 * HOME_ETC patch for PLD Linux. (ESR)
1733 * Sunil Shetye's fix for SSL configuration. (ESR)
1734 * Simon Josefsson's patch for GSS library support. (ESR)
1735 * Added Andrey Lelikov's recipe for Hotmail and Lycos Webmail. (ESR)
1736 * Remove blank between MAIL FROM: and <, which causes Cyrus to complain.
1737 Patch by Phil Endecott. (RF)
1738 * Build fixes for HESIOD and resolv.h trouble on FreeBSD. (MA)
1739 * Fabrice Bellet's fix for Red Hat bug #113492, fetchmail hangs in IMAP
1740 mode after EXPUNGE when the server (Dovecot 0.99.10) doesn't update
1741 RECENT and EXISTS counts. (MA)
1742 * Holger Mauermann's bounce patch, to use a NULL envelope from, not
1743 write a Return-Path header (both to meet RFC-2821), changed From,
1744 added Subject header, rewording the human readable part. Fixes Debian
1746 * Merge Sunil Shetye's time.h handling fix. (MA)
1747 * Merge Gerd von Egidy's patch to avoid a segfault in multidrop/received
1748 mode when the Received: headers are malformatted. (MA)
1749 * MIME-encode bodies and Subject headers of warning messages, limiting
1750 the header to 7 bits. (MA)
1751 * Normalize most locale codesets to IANA codesets, based on
1752 norm_charmap.c by Markus Kuhn. (MA)
1753 * Remove sleep(3) after POP3 login, patch by Brian Candler. (MA)
1754 * Fix option parsing bug that trashes the showdots setting when more
1755 than one server is configured. Patch by Brian Candler. (MA)
1756 * Honor sslcertpath setting even if sslcertck is unset. Patch by Brian
1758 * SSL certificate checking fixes, don't display same error message twice
1759 in succession, make sure that Common Name and fingerprint checking are
1760 only done once. Print all validation warnings/errors even if not in
1761 verbose mode. Patch by Brian Candler. (MA)
1762 * Import Bjorn Reese and Daniel Stenberg's MIT-licensed Trio 1.10 from
1763 http://daniel.haxx.se/projects/trio/ for systems that do not support
1764 snprintf or vsnprintf. (MA)
1765 * Clean up the horrible #ifdef HAVE_[V]SNPRINTF that made the code
1766 unreadable. Use Trio where [v]snprintf is/are missing. (MA)
1767 * Default to Linux 2.2 /proc/net/dev format, and use uname(2) to determine the
1768 kernel version instead of calling uname(1). Thanks to Paul Slootman. (MA)
1769 * Be more careful when swapping UID lists or writing the .fetchids file,
1770 requested by Manfred Weihs. (MA)
1771 * Print a warning if multidrop configuration is attempted without
1772 envelope option. (MA)
1773 * Split information on fetchmail versions before 6.0.0 to a separate
1775 * Merge SuSE patches: (sent by Stanislav Brabec, merged by Matthias Andree)
1776 - fetchmail-6.2.5-declaration.patch (double sigint_handler decl/getpass.c)
1777 - fetchmail-6.2.5-implicit-declaration.patch (missing #include)
1778 - fetchmail-6.2.5-random-result.patch (uninitialized variable/opie.c)
1779 * Revised some bogus assertions about POP3 LAST and UIDL use in the
1780 manual page. UIDL isn't flaky as the man page suggested, but a
1781 reliability feature. In fact, IMAP4 code is flaky in that it relies on
1782 the upstream seen flags. (MA)
1783 * Miloslav Trmac's patch for fetchmailconf to support string-type values
1784 of the "port" variable, avoiding "port None" corruption in .fetchmailrc.
1785 To fix Redhat Bug #55623 (MA)
1786 * de.po fixes from Nico Golde (MA)
1787 * es.po fixes from Jesus Roncero, Debian bug #286044 (MA)
1788 * sink.c fix from Cesar Eduardo Barros, to avoid double @ in address
1789 when username contains an @ and the envelope sender is null, Debian
1791 * configure.ac cleanups by Miloslav Trmac (MA)
1792 * Miloslav Trmac's fix to reply_hack() type, for systems where
1793 sizeof(int) != sizeof(size_t). (MA)
1794 * Nalin Dahyabhai's fix for driver.c to not call the private Kerberos
1795 krb5_init_ets() function. Sent by Miloslav Trmac. (MA)
1796 * Nalin Dahyabhai's fix for sink.c/transact.c to reserve sufficient
1797 space for \r\n trailers in snprintf calls. Sent by Miloslav Trmac,
1798 possibly fixing Red Hat bug #114470. (MA).
1799 * Nalin Dahyabhai's patch to use the krb5-config script, if present.
1800 Sent by Miloslav Trmac. (MA)
1801 * Nalin Dahyabhai's fix to make rpa.c compile. Sent by Miloslav Trmac. (MA)
1802 * Trivial fetchmailconf.man to redirect to fetchmail.1.
1803 Reported by Miloslav Trmac. (MA)
1804 * Internationalization (i18n) updates by Miloslav Trmac. (MA)
1805 * Fix "couldn't find canonical DNS name of NN (MM)" for hosts that have
1806 only IPv6 addresses. Matthias Andree.
1807 * Revised INSTALL after question from Brian Candler, inet6-apps is no
1808 longer available: remove inet6-apps hints for IPv6, and add some
1809 apologetic message for IPsec. Note the code may be removed in a future
1810 version. Matthias Andree.
1811 * Brian Candler's FAQ update about SSL certificate verification. (MA)
1812 * Nico Golde's patch to support "proto RPOP" in the configuration file,
1813 reported by Dr. Andreas Krüger, Debian bug #242384 (MA)
1814 * Skip sending POP3 PASS command when USER command failed. Matthias Andree.
1815 * Run fetchmail.man through automatic spell checker. Matthias Andree.
1816 * Major fetchmail(1) manual page overhaul by R. Hannes Beinert, to
1817 clarify singledrop vs. multidrop operation. (MA)
1818 * Make tracepolls a server option, as documented. Fixes Debian bug
1819 #156094. Matthias Andree.
1820 * Fix some minor inaccuracies (RFC-1893 related, grammar/spelling) in
1822 * Rename ESR's design notes to esrs-design-notes.html and add a new
1823 design-notes.html document. The NOTES file will contain both of them.
1825 * Fix Debian bug #301964, fetchmail leaks sockets when SSL negotiation
1826 fails. Fix suggested by Goswin Brederlow. (MA)
1827 * Really fix Debian Bug#207919 (garbage in Received: lines when smtphost set),
1828 patch by Tobias Diedrich. The 6.2.5 NEWS claimed Gregan's patch had fixed
1829 #207919 but it had fixed #212484 instead and #207919 remained unfixed in
1830 6.2.5. The entry below has been corrected to read #212484 now. (MA)
1831 * When writing the PID file, write a FHS 2.3 compliant PID file.
1832 Fixes Debian bug #230615. Matthias Andree.
1833 * Make ODMR really silent, suppress "fetchmail: receiving message
1834 data". Fixes Debian Bug#296163. Matthias Andree.
1835 * Add From: header to warning emails. Debian Bug#244828. Matthias Andree.
1836 * Fix IMAP code to use password of arbitrary length from configuration
1837 file (although not when read interactively). Debian Bug#276424.
1839 * Document that fetchmail may automatically enable UIDL option.
1840 Debian Bug#304701. Matthias Andree.
1841 * Put *BOLD* text into the manual page near --mda to state unmistakably that
1842 the --mda %T and %F substitutions add single quotes, hoping to avoid bogus
1843 bug reports such as Debian Bug #224564. Matthias Andree
1844 * Rename lock_release to fm_lock_release, to avoid namespace collision on
1845 Darwin. NetBSD PR#28543 (pkg/28543). Matthias Andree.
1846 * The RFC-822 parser no longer strips the last character of bare addresses.
1848 * The IP address matching code was broken and
1849 1. didn't search exhaustively, but matched only the first IP address of the
1850 server's queryname against the IP addresses of the server name to match.
1851 2. didn't match IP aliases versus MX hosts. Matthias Andree
1852 * The "port" option, while still understood, is being replaced by the "service"
1853 option, which is now supported even without --enable-inet6. Matthias Andree.
1854 * The default distribution format is now bzip2. Matthias Andree.
1855 * fetchmailconf redirects fetchmail's input from /dev/null so it doesn't
1856 wait for the user to enter a password when the user doesn't even see
1857 the prompt. Reported by Michal Marek. Matthias Andree.
1858 * Write RFC-compliant BSMTP envelopes. Reported by Nico Golde. Matthias Andree.
1859 * Fix --with-gssapi compilation problem. Simon Josefsson. (MA)
1860 * Foster protocol-independence to support IPv6 better, for instance, providing
1861 IPv6 addresses in Received: headers. Matthias Andree.
1862 * Received: headers now enclose the for <...> destination address in angle
1863 brackets for consistency with Postfix. Matthias Andree.
1864 * Operating systems that do not support at least one of gethostbyname,
1865 gethostbyname_r, getipnodebyname are no longer supported. Matthias Andree.
1866 * Fixes to --with-hesiod option. Sunil Shetye. (MA)
1867 * Delete oversized messages with the new --limitflush option. Debian
1868 Bug#212240. Sunil Shetye. (MA)
1869 * Fix MacOS X compilation failures in sink.c (ru_*time has incomplete type).
1870 Berlios Bug #4725. Matthias Andree.
1871 * Fix "auth ntlm" to send AUTH NTLM (rather than AUTH MSN). Add "auth msn"
1872 officially. Reported by Yves Boisjoly. Matthias Andree
1873 * Expunge between IMAP folders when polling multiple folders.
1875 * Fix IMAP expunged message counting. Sunil Shetye. (MA)
1876 * Add full support for --service option. Matthias Andree
1877 * When getaddrinfo() fails resolving a service, log getaddrinfo() error. (MA)
1878 * Fix bogus "cannot resolve service * to port number" error. Simon Barner. (MA)
1879 * Failure to set up SSL connections now results in PS_SOCKET. Suggested by
1880 Thomas Wolff. Matthias Andree.
1881 * Kerberos IV detection fix for FreeBSD 4. Simon Barner. (MA)
1882 * Fix display and documentation of --envelope option. Matthias Andree
1883 * Make "envelope 'Delivered-To'" work with dropdelivered. Timothy Lee. (MA)
1884 * Add -DBIND_8_COMPAT to Darwin (MacOS X) compiles, to fix build problems on
1885 newer Darwin versions. Matthias Andree.
1886 * fetchmail should now automatically detect if OpenSSL requires -ldl.
1888 * Fix Solaris build with --disable-nls (blastwave.org). Matthias Andree.
1889 * Missed --port/--service/--ssl cleanups in the manual. Reminder from Thomas
1891 * Complain in POP3 if NTLM/MSN auth is requested but had not been enabled at
1892 compile time. This configuration mismatch now causes an error message and
1893 authentication failure. Found by Yves Boisjoly. Matthias Andree
1894 * fetchmailconf now allows expert users to choose the authorization type and
1895 also offers MSN and NTLM, suggested by Yves Boisjoly. Matthias Andree
1896 * fetchmailconf now (as of 1.49) writes its version to the comment of the
1897 saved run control file. Matthias Andree
1898 * Properly shut down SSL connections. Berlios Patch #647 by Arkadiusz
1900 * Global variable cleanup, to fix daemon mode reinitialization problems. Patch
1901 by Sunil Shetye. (MA)
1902 * fetchmailconf -h documents the fetchmailconf -h option. Matthias Andree
1903 * fetchmailconf -V now prints the fetchmailconf version. Matthias Andree
1904 * Add support for SubjectAltName (RFC-2595 or 2818), to avoid bogus certificate
1905 mismatch errors. Patch by Roland Stigge, Debian Bug#201113. (MA)
1906 * make fetchmail --silent --quit really silent, Debian Bug #229014 by Dr.
1907 Andreas Krüger. Matthias Andree
1908 * cleanup --quit handling again (so that --silent --quit just kills the
1909 existing daemon, rather than continue running), and document it more clearly.
1911 * Print an error message if multiple "defaults" records are found in the
1912 configuration file. Matthias Andree
1913 * Bury on_exit officially - the necessary code had been missing from 6.0.0,
1914 6.2.0, 6.2.5. Matthias Andree
1915 * Exit with error if the lock file cannot be read. Matthias Andree
1916 * Exit with error if the lock file cannot be created exclusively, this got
1917 broken in a 6.2.6-pre, 6.2.5.2 and older were fine. Matthias Andree
1918 * Do not break some other process's lockfile in "-q" mode, but wait for the
1919 other process's exit. Matthias Andree
1920 * Man page: --sslfingerprint points user to x509(1ssl) and gives an example
1921 how to use it. Debian Bug#213484, Eduard Bloch. (MA)
1922 * fetchmailconf now sets the service properly after autoprobe. Fixes Debian
1923 Bug#320645. Matthias Andree
1924 * Man page: Fix Debian Bug#241883, making global options more clear. Matt
1925 Swift, Matthias Andree.
1926 * When eating IMAP message trailer, don't see any line containing "OK" as the
1927 end of the trailer, but wait for the proper tagged OK line. To work around
1928 the qmail + Courier-IMAP problem in Debian Bug#338007. Matthias Andree
1929 * Fix Debian Bug#317761: when trying to send a bounce message, don't bail out
1930 if we cannot qualify our own hostname, so we aren't losing the bounce.
1931 Instead, pass the buck on to the SMTP server and use our own unqualified
1932 hostname. Matthias Andree
1933 * Revise some error messages so they are less confusing. Sunil Shetye.
1934 * Man page: update --smtphost documentation. Sunil Shetye, Matthias Andree.
1935 * Man page: clarify --loghost works only while detached. Matthias Andree
1936 * Man page: update --smtpaddress documentation. Sunil Shetye.
1937 * Fix several memory leaks and bugs in the SMTP/LMTP retry logic where
1938 fetchmail confused UNIX and Internet domain sockets. Sunil Shetye.
1939 * Man page (BUGS): document that passwords are length limited. Matthias Andree
1940 * Man page: Document that quoted strings that run across line boundaries
1941 contain the control characters (CR or LF). Document explicitly the backslash
1942 escape sequences and their differences from the escape sequences used in the
1943 C programming language. Matthias Andree
1944 * Fix segfault when run control file ends with a backslash inside an
1945 unterminated quoted string. Matthias Andree.
1946 * In quoted strings, support backslash as last character on a line to join the
1947 following line to the current. Matthias Andree.
1948 * Parsing untagged IMAP responses is more robust now. Matthias Andree.
1949 * Man page: Remove some procmail praises in --mda documentation, suggest
1950 maildrop instead, warn of procmail fallthrough behavior. Matthias Andree.
1951 * Man page: Revise AUTHORS and SEE ALSO sections. Matthias Andree.
1952 * Updated translations: Albanian [sq] (Besnik Bleta), Catalan [ca] (Ernest
1953 Adrogué Calveras), Czech [cs] (Miloslav Trmac), German [de] (MA),
1954 Spanish (Castilian) [es] (Javier Kohen), French [fr] (MA),
1955 Polish [pl] (Jakub Bogusz), Russian [ru] (Pavel Maryanov).
1956 * In oversized warning messages, print the account name, too. Fixes Debian
1957 Bug#213299. Sunil Shetye (MA).
1958 * Fix installation without Python. Sunil Shetye, reported by Peter Church. (MA)
1959 * Update Japanese translation. Fixes Debian Bug#329342, Takeshi Hamasaki. (MA)
1960 * Fix imap.c size safeguard that broke on x86_64 architecture. Matthias Andree
1961 * The FAQ is now available for duplex DIN A4 printing in PDF format.
1962 Don't bother to ask for a Letter version, I don't care. Matthias Andree
1963 * Man page: Use \- in the manual page where appropriate so that copy & paste
1964 works. I hope we got them all. Héctor García, Matthias Andree.
1967 * Switched to automake. Matthias Andree.
1968 * Got rid of alloca() in fetchmail proper. Matthias Andree
1969 * Got rid of ipv6-connect, inner_connect and thereabouts. Matthias Andree
1971 --------------------------------------------------------------------------------
1973 fetchmail-6.2.5 (Wed Oct 15 18:39:22 EDT 2003), 23079 lines:
1975 * Updated Spanish, Turkish, and German translation files.
1976 * Matthew Gregan's patch to handle garbage lengths from dbmail;
1977 closes Debian bug #212484.
1978 * Fix IMAP query so new-message count doesn't include deleted messages.
1979 * Man page typo fix, closes Debian bug #205892.
1980 * OpenSSL cleanup patches from levinedl@acm.org.
1981 * Benjamin Drieu's patch to fix Debian bug #212240, no oversized-message
1982 flushing if both "flush" and "limit" were specified.
1983 * Benjamin Drieu's patch for Debian bug #156592, incorrect handing of
1985 * Smash all NULs out of headers right after the socket read.
1986 * Dup-killer code now keys on an MD5 hash of the raw headers.
1987 * Sunil Shetye's patches to break up fetching of sizes and UIDLs.
1989 There are 599 people on fetchmail-friends and 748 on fetchmail-announce.
1991 fetchmail-6.2.4 (Wed Aug 13 04:27:35 EDT 2003), 22625 lines:
1993 * Updated German, Spanish, Catalan, and Turkish translations.
1994 * IDLE is now supported using NOOP commands even if the server doesn't support
1995 the IMAP IDLE extension. Patch by Chris Boyle.
1996 * Sunil Shetye's patch to do better password shrouding.
1997 * Sunil Shetye's bug-fix rollup patch.
1998 * Introduce a translation item for the word "seen".
1999 * Back out the hack to deal with lack of byte stuffing on some POP3 servers.
2000 * Thomas Steudten's patch to improve SMTP handling of 550 errors.
2002 There are 585 people on fetchmail-friends and 745 on fetchmail-announce.
2004 fetchmail-6.2.3 (Thu Jul 17 14:53:00 EDT 2003), 22490 lines:
2006 * French, German, Danish, Spanish, and Turkish translations updated.
2007 * Brian Sammon's patch to deal with malformed message lines containing NULs.
2008 * Fai's patch to ignore all but the first Return-Path (some spams have
2009 more than one of these).
2010 * Benjamin Drieu's patch to properly byte-stuff when talking to BSMTP.
2011 Fixes Debian bug #184469.
2012 * Benjamin Drieu's patch to enable auth=cram-md5.
2013 Fixes Debian bug #185232.
2014 * Sunil Shetye's configure.in patch to avoid spurious search order messages
2016 * Header-reading code now copes better with lines ending in \n only.
2017 * Elias Israel's patches for POP3 NTLM support and dealing with byte-
2018 stuffing failures at socket level.
2020 There are 580 people on fetchmail-friends and 750 on fetchmail-announce.
2022 fetchmail-6.2.2 (Fri Feb 28 21:34:26 EST 2003), 22345 lines:
2024 * Sunil Shetye's patch to improve behavior on empty messages.
2025 * Conform to RFC2595; reissue capability probes after successful
2026 STARTTLS negotiation.
2027 * Sunil's patch to make handling of failed STARTTLS more graceful.
2028 * Sunil's JF2 fix patch for .fetchmailrc security.
2029 * Christophe GIAUME <christophe@giaume.com> finished the implementation
2031 * Jason Tishler's fix patch for Cygwin.
2032 * Support ssh-style authentication in POP3
2033 * Fix for Debian bug #108977, clean up config file evaluation,
2036 There are 554 people on fetchmail-friends and 727 on fetchmail-announce.
2038 fetchmail-6.2.1 (Tue Jan 14 08:17:19 EST 2003), 22219 lines:
2040 * Updated German, Turkish, Spanish, and Danish translation files.
2041 * Integrated Sunil Shetye's patch to make mark_seen an explicit method.
2042 * Removed FAQ warning about GMX and associated fetchmailconf check,
2043 we have a report that its servers are conformant now.
2044 * Another Sunil patch to fix a minor bug in bouncemail generation.
2046 There are 536 people on fetchmail-friends and 716 on fetchmail-announce.
2048 fetchmail-6.2.0 (Fri Dec 13 00:10:07 EST 2002), 22235 lines:
2050 * Applied Steffen Esser's fix for a buffer-overflow bug in rfc822.c
2051 * Updated Danish, German, and Turkish translation files.
2052 * Sunil Shetye's SMTP timeout patch.
2054 There are 538 people on fetchmail-friends and 701 on fetchmail-announce.
2056 --------------------------------------------------------------------------------
2058 fetchmail-6.1.3 (Thu Nov 28 05:35:15 EST 2002), 22203 lines:
2060 * Updated Turkish, Danish, German, Spanish, Catalan po files.
2061 * Added Slovak support.
2062 * Configure.in update for autoconf 2.5 (Art Haas).
2063 * Be case-insensitive when looking for IMAP responses.
2064 * Fix logout-after-idle-delivery bug (Sunil Shetye).
2065 * Sunil Shetye's patch to bulletproof end-of-header detection.
2066 * Sunil's fix for the STARTTLS problem -- repoll if TLS nabdshake
2067 fails. The attempt to set up STARTTLS can be suppressed with 'sslproto ""'.
2069 There are 540 people on fetchmail-friends and 701 on fetchmail-announce.
2071 fetchmail-6.1.2 (Thu Oct 31 11:41:02 EST 2002), 22135 lines:
2073 * Jan Klaverstijn's verbosity-lowering patch.
2074 * Updated Turkish, German, Catalan, and Danish translation files.
2075 * Fix processing of POP3 messages with missing bodies.
2076 * Minor fixes by Sunil Shetye: fix generation of auth fail note, handle
2077 unexpected SIGALRM, plug memory leak, handle lines beginning with '\0',
2078 try to bulletproof error handling against read failures.
2080 There are 535 people on fetchmail-friends and 696 on fetchmail-announce.
2082 fetchmail-6.1.1 (Fri Oct 18 14:53:51 EDT 2002), 22087 lines:
2084 * OTP fix patches from Stanislav Brabec <utx@penguin.cz>
2085 * fix patch for writing antispam capability correctly in conf.c.
2086 * Fix patches for Debian bugs #162571, #156592.
2087 * Correction to manpage re -b and qmail.
2088 * Patch to disable use of STLS if auth passwd is specified.
2089 * Fix specfile generation to handle SSL correctly.
2090 * New Danish, Turkish, and Catalan translation files.
2091 * Improved ODMR debug messages.
2092 * IMAP efficiency hack; don't fetch sizes unless needed.
2093 * Detect and rewrite invalid return paths beginning with @.
2094 * Fix for subtle freeing bug that suppressed information in some bounce msgs.
2095 * Newline fix patches for internationalization files.
2096 * Fix reversed test guarding authentication-failure warnings.
2097 * Fix POP3 breakage starting at 5.9.14.
2099 There are 529 people on fetchmail-friends and 693 on fetchmail-announce.
2101 fetchmail-6.1.0 (Sun Sep 22 18:31:23 EDT 2002), 21999 lines:
2103 * Updated French translation.
2104 * Stefan Esser's fix for potential remote vulnerability in multidrop mode.
2105 This is an important security fix!
2107 There are 519 people on fetchmail-friends and 680 on fetchmail-announce.
2109 --------------------------------------------------------------------------------
2111 fetchmail-6.0.0 (Tue Sep 17 19:48:25 EDT 2002), 21972 lines:
2113 * Applied Matt Kraai's fix for minor Debian bug #144539.
2114 * Nerijus Baliunas's patch to support STARTTLS over IMAP.
2115 * More cleanups and minor bugfixes from Sunil Shetye.
2116 * Default antispam-response list is now empty.
2117 * Updated de and po translations.
2119 --------------------------------------------------------------------------------
2121 There are 520 people on fetchmail-friends and 683 on fetchmail-announce.
2123 --------------------------------------------------------------------------------
2125 vim:tw=80 com=bf\:* ts=8 sts=8 sw=8 ai: