- dovecot - dovecot.org

[Dovecot] dovecot patch: do scanf correctly
by Jann Horn 24 Nov '13

24 Nov '13

As far as I can tell, there are two places in which dovecot does not use scanf properly. This patch should fix that. -------------------------------------------------------------------------------- diff -r 8a3d0426f514 src/lib-dns/dns-lookup.c --- a/src/lib-dns/dns-lookup.c Sat Nov 23 23:22:09 2013 +0000 +++ b/src/lib-dns/dns-lookup.c Sun Nov 24 06:30:12 2013 +0100 @@ -102,7 +102,7 @@ } /* first line: <ret> <ip count> */ if (sscanf(line, "%d %u", &result->ret, - &result->ips_count) == 0) + &result->ips_count) != 2) return -1; if (result->ret != 0) { result->error = net_gethosterror(result->ret); diff -r 8a3d0426f514 src/lib-settings/settings.c --- a/src/lib-settings/settings.c Sat Nov 23 23:22:09 2013 +0000 +++ b/src/lib-settings/settings.c Sun Nov 24 06:30:12 2013 +0100 @@ -44,7 +44,7 @@ { int num; - if (!sscanf(value, "%i", &num) || num < 0) + if (sscanf(value, "%i", &num) != 1 || num < 0) return t_strconcat("Invalid number: ", value, NULL); *result = num; return NULL; -------------------------------------------------------------------------------- Specifically, both invocations disregard that sscanf("", "%d", &num) and similar make sscanf return EOF (-1 on my system) and the first invocation also does not account for the case that the second argument is not matched.

1 0

Re: [Dovecot] 2.2.9
by Rick Romero 23 Nov '13

23 Nov '13

Thanks for providing and maintaining a great piece of software. Rick

1 0

[Dovecot] Dict client unescaping sieve script
by Joshua Perry 22 Nov '13

22 Nov '13

I've created a dict service that listens on a unix socket and answers queries for sieve scripts (among other things). As I understand it (from the source code at http://hg.dovecot.org/dovecot-2.2/file/tip/src/lib-dict/dict-client.c) the dict client will unescape \001n, \001t, and \0011 to line feeds, tabs, and the \001 character respectively. In my service I am escaping those three characters in my response (if I don't escape them the line-oriented nature of the protocol causes a failure for multiline sieve scripts) but every time LDA attempts to process a sieve script I get an error in the logs (see below) showing sieve choking on \001 characters. Is there some configuration value I've missed or something? *dovecot log* Nov 8 23:04:54 www dovecot: lmtp(29940, josh(a)redacted.com) pxg7JxZufVL0dAAAPhZyyg: sieve: failed to compile script dict:proxy:/var/run/dovecot-auth.sock:sieve;name=main script (view user logfile /var/mail/vhosts/redacted.com/josh/.dovecot.sieve.log for more information) *.dovecot.sieve.log* sieve: info: started log at Nov 08 16:14:38. main script: line 1: error: unexpected character(s) starting with 0x01. main script: line 1: error: unexpected unknown characters found at (the presumed) end of file. main script: error: parse failed.

2 2

[Dovecot] dovecot + sieve redirect failed
by Tibor Korocz 22 Nov '13

22 Nov '13

Hi Guys, I have a strange problem and I couldn't find any solution, I hope somebody could help me. I'm using postfix+dovecot+sieve combination and I tried to set up redirect. .sieve: require ["fileinto", "regex", "date", "relational", "vacation"]; redirect "xxxx(a)gmail.com"; keep; But I get some error in the mail.log: sieve: msgid=<1562626302.336239.1385116787283>: failed to redirect message to <xxxxx(a)gmail.com> (refer to server log for more information) Nov 22 11:39:50 server dovecot: lda(xxx@xxxx): sieve: msgid=<1562626302.336239.1385116787283>: stored mail into mailbox 'INBOX' Nov 22 11:39:50 server dovecot: lda(xxx@xxxx): Error: sieve: execution of script /var/vmail/xxxx/.sieve failed, but implicit keep was successful (user logfile /var/vmail/xxxx/.sieve.log may reveal additional details) .sieve.log: error: msgid=<1562626302.336239.1385116787283>: failed to redirect message to <xxxx(a)gmail.com> (refer to server log for more information). I couldn't find any other details in the logs. Any idea what is the problem and what is the solution? Thanks, Tibi

2 1

[Dovecot] removing html via dovecot/sieve
by Bonz 22 Nov '13

22 Nov '13

is there a howto or similar for configuring dovecot as a user to strip mime. ideally i would like all html copies deleted, but graphics, etc dumped to ~/public_html and replaced with a url. i have a super crappy link so html, etc usually jams my spool. i can usually wget things however hence this desired approach. bonz.

2 1

[Dovecot] Personal Sieve Resieving Default Sieve?
by Jim McNamara 22 Nov '13

22 Nov '13

Hello all. In my dovecot install we have a default.sieve script which filters incoming mail based on the subdomain of our main domain. All it does is if the mail comes from A.domain.com move in to folder A in inbox, and if it comes from B.domain.com, move to folder B also in inbox. In my personal sieve script, I'd like to grab a particular email each day that the default.sieve saw coming from A.domain.com and moved it to folder A, and either copy or move it to another folder, C. My google-fu was weak while looking for second sieve passes undoing the first, so I thought it best to ask here. Thanks for your time.

2 1

[Dovecot] Pigeonhole make check fails with Dovecot 2.2.8
by Tamsy 22 Nov '13

22 Nov '13

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Is it just here? When compiling dovecot-2.2-pigeonhole-0.4.2 at the stage "make check" Pigeonhole is throwing the following message and stops: Test case: ./tests/test-header.svtest: 1: Test 'Basic functionality' SUCCEEDED 2: Test 'Matching empty key' SUCCEEDED 3: Test 'Ignoring whitespace' FAILED: header test does not strip leading whitespace 4: Test 'Absent or empty header' SUCCEEDED 5: Test 'Invalid header name' SUCCEEDED 6: Test 'Folded header' SUCCEEDED I have tried to compile Pigeonhole 0.4.2 and also the latest dovecot-2-2-pigeonhole-b7434a02fc51 from the Mercurial repository but the outcome at "make test" is always the same: 3: Test 'Ignoring whitespace' FAILED: header test does not strip leading whitespace Cheers, Tamsy -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (MingW32) iQEcBAEBAgAGBQJSjC2QAAoJEKYXVM1dyOfZ9xsIAKDC7goLP3N2rr90Lf0P6gam wQFmYY3HUMIKi3FKWcCy6DQBN6Qr33QWTk/b4OY1LDMOETolZTWIEwvlix+YdI8x oRdKQXWHR9k+jd9etWkNuuk2xUnbjz353IOMZ3Yc+spuv0JyUvYhef6pfq3MLTYQ R+YW4aoNp3XstAKcknnkduNwrfRXoZoFT/YOU3VQyDUwUa6u07JhuL+onZtb0mO+ HsdD4NNY28YqHY1/eoT3YNRbgtRxyOfzxh3f17ITbquM0ZgiPJucWX7HpIiwNJFu 3gqkceDLqqAL9DKUjDapSyR2IjxJlRSGcxQckxlkGXzcaTNNWqAn7mMxSr3hEaY= =EY7D -----END PGP SIGNATURE-----

5 12

[Dovecot] (no subject)
by Axel Reinhold 22 Nov '13

22 Nov '13

Subject: dovecot-2.2.8 does no more quit correctly after update to dovecot-2.2.8 monit is no more validating pop3 server due to dovecot mo more logging out correctly "+OK Logging out" is missing With 2.2.7: [root@pie ~]# telnet glen 110 Trying 188.138.31.225... Connected to glen. Escape character is '^]'. +OK Dovecot ready. <1da9.1.528e3113.qpIWrdBNJbF3qp/tA9R+sQ==(a)glen.ypack.de> quit +OK Logging out Connection closed by foreign host. with 2.2.8: [root@pie ~]# telnet glen 110 Trying 188.138.31.225... Connected to glen. Escape character is '^]'. +OK Dovecot ready. <1fc9.1.528e31b1.W4FN6a7gzBMYNOBoxP43OQ==(a)glen.ypack.de> quit Connection closed by foreign host. After quit the "+OK Logging out" is missing! New option? Or bug? Regards Axel

2 1

[Dovecot] dovecot-2.2.7: Fatal: master: service(imap): child 49545 killed with signal 11 (core dumped)
by Marcin Mirosław 21 Nov '13

21 Nov '13

Hi! After upgrade I'm noticing many coredummps, below is backtrace: $ echo "bt"|gdb .... #0 o_stream_zlib_flush (stream=0xc8be5b27640) at ostream-zlib.c:222 222 if (o_stream_zlib_send_flush(zstream) < 0) (gdb) #0 o_stream_zlib_flush (stream=0xc8be5b27640) at ostream-zlib.c:222 #1 0x00007a695e8f09cd in o_stream_flush (stream=stream@entry=0xc8be5b276d0) at ostream.c:147 #2 0x00007a695e83c06f in o_stream_zlib_close (stream=0xc8be5b27640, close_parent=true) at ostream-zlib.c:35 #3 0x00007a695e8f0375 in o_stream_close_full (stream=stream@entry=0xc8be5b276d0, close_parents=close_parents@entry=true) at ostream.c:49 #4 0x00007a695e8f06e2 in o_stream_close (stream=stream@entry=0xc8be5b276d0) at ostream.c:85 #5 0x00007a695e8f1738 in o_stream_copy_error_from_parent (_stream=_stream@entry=0xc8be5b27640) at ostream.c:434 #6 0x00007a695e83c5cc in o_stream_zlib_flush (stream=0xc8be5b27640) at ostream-zlib.c:227 #7 0x00007a695e8f09cd in o_stream_flush (stream=stream@entry=0xc8be5b276d0) at ostream.c:147 #8 0x00007a695e83c06f in o_stream_zlib_close (stream=0xc8be5b27640, close_parent=true) at ostream-zlib.c:35 #9 0x00007a695e8f0375 in o_stream_close_full (stream=stream@entry=0xc8be5b276d0, close_parents=close_parents@entry=true) at ostream.c:49 #10 0x00007a695e8f06e2 in o_stream_close (stream=stream@entry=0xc8be5b276d0) at ostream.c:85 #11 0x00007a695e8f1738 in o_stream_copy_error_from_parent (_stream=_stream@entry=0xc8be5b27640) at ostream.c:434 #12 0x00007a695e83c5cc in o_stream_zlib_flush (stream=0xc8be5b27640) at ostream-zlib.c:227 #13 0x00007a695e8f09cd in o_stream_flush (stream=stream@entry=0xc8be5b276d0) at ostream.c:147 #14 0x00007a695e83c06f in o_stream_zlib_close (stream=0xc8be5b27640, close_parent=true) at ostream-zlib.c:35 #15 0x00007a695e8f0375 in o_stream_close_full (stream=stream@entry=0xc8be5b276d0, close_parents=close_parents@entry=true) at ostream.c:49 #16 0x00007a695e8f06e2 in o_stream_close (stream=stream@entry=0xc8be5b276d0) at ostream.c:85 #17 0x00007a695e8f1738 in o_stream_copy_error_from_parent (_stream=_stream@entry=0xc8be5b27640) at ostream.c:434 #18 0x00007a695e83c5cc in o_stream_zlib_flush (stream=0xc8be5b27640) at ostream-zlib.c:227 #19 0x00007a695e8f09cd in o_stream_flush (stream=stream@entry=0xc8be5b276d0) at ostream.c:147 #20 0x00007a695e83c06f in o_stream_zlib_close (stream=0xc8be5b27640, close_parent=true) at ostream-zlib.c:35 #21 0x00007a695e8f0375 in o_stream_close_full (stream=stream@entry=0xc8be5b276d0, close_parents=close_parents@entry=true) at ostream.c:49 #22 0x00007a695e8f06e2 in o_stream_close (stream=stream@entry=0xc8be5b276d0) at ostream.c:85 #23 0x00007a695e8f1738 in o_stream_copy_error_from_parent (_stream=_stream@entry=0xc8be5b27640) at ostream.c:434 #24 0x00007a695e83c5cc in o_stream_zlib_flush (stream=0xc8be5b27640) at ostream-zlib.c:227 #25 0x00007a695e8f09cd in o_stream_flush (stream=stream@entry=0xc8be5b276d0) at ostream.c:147 #26 0x00007a695e83c06f in o_stream_zlib_close (stream=0xc8be5b27640, close_parent=true) at ostream-zlib.c:35 [...] #285878 0x00007a695e8f0adc in o_stream_default_cork (_stream=0xc8be5b27640, set=<optimized out>) at ostream.c:419 #285879 o_stream_default_cork (_stream=0xc8be5b27640, set=<optimized out>) at ostream.c:412 #285880 0x00000c8be506b72d in client_input (client=0xc8be5b07b00) at imap-client.c:945 #285881 0x00007a695e8e4867 in io_loop_call_io (io=0xc8be5b8f370) at ioloop.c:387 #285882 0x00007a695e8e5d1f in io_loop_handler_run (ioloop=ioloop@entry=0xc8be5ab50f0) at ioloop-epoll.c:220 #285883 0x00007a695e8e4178 in io_loop_run (ioloop=0xc8be5ab50f0) at ioloop.c:411 #285884 0x00007a695e87a847 in master_service_run (service=0xc8be5ab4f80, callback=callback@entry=0xc8be5077f20 <client_connected>) at master-service.c:566 #285885 0x00000c8be50785d4 in main (argc=1, argv=0xc8be5ab4d50) at main.c:400 And gdb is killing by OOM. I suspect this can be related to imapc. Regards, Marcin

1 2

[Dovecot] dsync-2.2.7 incorrectly synchronizes subscription status of renamed mailbox
by Karol Jurak 21 Nov '13

21 Nov '13

Hi, It seems that dsync-2.2.7 doesn't correctly synchronize subscription status of a renamed mailbox. The situation is as follows. There are two servers: A and B, and a test user test_mdbox. Mailboxes (mdbox) of this user on both servers are synchronized. Specifically on both of them there is a (subscribed) Test1 mailbox. Replication plugin is disabled. I rename Test1 to Test2 on A with Thunderbird. The output of 'doveadm mailbox list', contents of subscriptions file and the output of 'doveadm dump -t mailboxlog dovecot.mailbox.log' (below) confirm that the rename succeeded on A. #2976: rename 2b84f621c0fd4ba8bd514c5c43ab9a89 (2013-11-19 14:30:35) #3000: subscribe 2b84f621c0fd4ba8bd514c5c43ab9a89 (2013-11-19 14:30:35) #3024: unsubscribe 99ea7bf70f6e69ad71659995677b43f8 (2013-11-19 14:30:35) Subsequently I run 'doveadm sync -u test_mdbox remote:B'. Nothing changes on A. On B 'doveadm mailbox list' shows that Test1 was renamed to Test2, but subscriptions file still contains Test1 and not Test2. The output of 'doveadm dump -t mailboxlog dovecot.mailbox.log' confirms that only rename was performed: #2160: rename 2b84f621c0fd4ba8bd514c5c43ab9a89 (2013-11-19 14:57:22) Only another 'doveadm sync -u test_mdbox remote:B' fixes subscriptions on B. The following records are added to dovecot.mailbox.log: #2184: unsubscribe 99ea7bf70f6e69ad71659995677b43f8 (2013-11-19 15:00:32) #2208: subscribe 2b84f621c0fd4ba8bd514c5c43ab9a89 (2013-11-19 15:00:32) -- Karol Jurak

2 2